From ed81a111d1999a6b62f2cfea21c241e557b800da Mon Sep 17 00:00:00 2001 From: Marijn Schouten Date: Thu, 19 Jun 2025 15:56:32 +0000 Subject: wasm: rm static mut --- library/std/src/sys/alloc/wasm.rs | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-) diff --git a/library/std/src/sys/alloc/wasm.rs b/library/std/src/sys/alloc/wasm.rs index c8fab992a88..48e2fdd4ecc 100644 --- a/library/std/src/sys/alloc/wasm.rs +++ b/library/std/src/sys/alloc/wasm.rs @@ -16,12 +16,15 @@ //! The crate itself provides a global allocator which on wasm has no //! synchronization as there are no threads! -// FIXME(static_mut_refs): Do not allow `static_mut_refs` lint -#![allow(static_mut_refs)] +use core::cell::SyncUnsafeCell; use crate::alloc::{GlobalAlloc, Layout, System}; -static mut DLMALLOC: dlmalloc::Dlmalloc = dlmalloc::Dlmalloc::new(); +struct SyncDlmalloc(dlmalloc::Dlmalloc); +unsafe impl Sync for SyncDlmalloc {} + +static DLMALLOC: SyncUnsafeCell = + SyncUnsafeCell::new(SyncDlmalloc(dlmalloc::Dlmalloc::new())); #[stable(feature = "alloc_system_type", since = "1.28.0")] unsafe impl GlobalAlloc for System { @@ -30,7 +33,7 @@ unsafe impl GlobalAlloc for System { // SAFETY: DLMALLOC access is guaranteed to be safe because the lock gives us unique and non-reentrant access. // Calling malloc() is safe because preconditions on this function match the trait method preconditions. let _lock = lock::lock(); - unsafe { DLMALLOC.malloc(layout.size(), layout.align()) } + unsafe { (*DLMALLOC.get()).0.malloc(layout.size(), layout.align()) } } #[inline] @@ -38,7 +41,7 @@ unsafe impl GlobalAlloc for System { // SAFETY: DLMALLOC access is guaranteed to be safe because the lock gives us unique and non-reentrant access. // Calling calloc() is safe because preconditions on this function match the trait method preconditions. let _lock = lock::lock(); - unsafe { DLMALLOC.calloc(layout.size(), layout.align()) } + unsafe { (*DLMALLOC.get()).0.calloc(layout.size(), layout.align()) } } #[inline] @@ -46,7 +49,7 @@ unsafe impl GlobalAlloc for System { // SAFETY: DLMALLOC access is guaranteed to be safe because the lock gives us unique and non-reentrant access. // Calling free() is safe because preconditions on this function match the trait method preconditions. let _lock = lock::lock(); - unsafe { DLMALLOC.free(ptr, layout.size(), layout.align()) } + unsafe { (*DLMALLOC.get()).0.free(ptr, layout.size(), layout.align()) } } #[inline] @@ -54,7 +57,7 @@ unsafe impl GlobalAlloc for System { // SAFETY: DLMALLOC access is guaranteed to be safe because the lock gives us unique and non-reentrant access. // Calling realloc() is safe because preconditions on this function match the trait method preconditions. let _lock = lock::lock(); - unsafe { DLMALLOC.realloc(ptr, layout.size(), layout.align(), new_size) } + unsafe { (*DLMALLOC.get()).0.realloc(ptr, layout.size(), layout.align(), new_size) } } } -- cgit 1.4.1-3-g733a5 From a8d417d1a53dc22c42054327aca52a3b6a034150 Mon Sep 17 00:00:00 2001 From: y21 <30553356+y21@users.noreply.github.com> Date: Sat, 29 Mar 2025 22:46:36 +0100 Subject: rework `useless_vec` lint --- clippy_lints/src/vec.rs | 218 +++++++++++++++++++++++++++++------------------- 1 file changed, 132 insertions(+), 86 deletions(-) diff --git a/clippy_lints/src/vec.rs b/clippy_lints/src/vec.rs index 7b6a25123e8..1646f222021 100644 --- a/clippy_lints/src/vec.rs +++ b/clippy_lints/src/vec.rs @@ -1,4 +1,6 @@ use std::collections::BTreeMap; +use std::collections::btree_map::Entry; +use std::mem; use std::ops::ControlFlow; use clippy_config::Conf; @@ -20,15 +22,36 @@ use rustc_span::{DesugaringKind, Span}; pub struct UselessVec { too_large_for_stack: u64, msrv: Msrv, - span_to_lint_map: BTreeMap>, + /// Maps from a `vec![]` source callsite invocation span to the "state" (i.e., whether we can + /// emit a warning there or not). + /// + /// The purpose of this is to buffer lints up until `check_expr_post` so that we can cancel a + /// lint while visiting, because a `vec![]` invocation span can appear multiple times when + /// it is passed as a macro argument, once in a context that doesn't require a `Vec<_>` and + /// another time that does. Consider: + /// ``` + /// macro_rules! m { + /// ($v:expr) => { + /// let a = $v; + /// $v.push(3); + /// } + /// } + /// m!(vec![1, 2]); + /// ``` + /// The macro invocation expands to two `vec![1, 2]` invocations. If we eagerly suggest changing + /// the first `vec![1, 2]` (which is shared with the other expn) to an array which indeed would + /// work, we get a false positive warning on the `$v.push(3)` which really requires `$v` to + /// be a vector. + span_to_state: BTreeMap, allow_in_test: bool, } + impl UselessVec { pub fn new(conf: &'static Conf) -> Self { Self { too_large_for_stack: conf.too_large_for_stack, msrv: conf.msrv, - span_to_lint_map: BTreeMap::new(), + span_to_state: BTreeMap::new(), allow_in_test: conf.allow_useless_vec_in_tests, } } @@ -62,17 +85,28 @@ declare_clippy_lint! { impl_lint_pass!(UselessVec => [USELESS_VEC]); -impl<'tcx> LateLintPass<'tcx> for UselessVec { - fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>) { - let Some(vec_args) = higher::VecArgs::hir(cx, expr.peel_borrows()) else { - return; - }; - if self.allow_in_test && is_in_test(cx.tcx, expr.hir_id) { - return; - } - // the parent callsite of this `vec!` expression, or span to the borrowed one such as `&vec!` - let callsite = expr.span.parent_callsite().unwrap_or(expr.span); +/// The "state" of a `vec![]` invocation, indicating whether it can or cannot be changed. +enum VecState { + Change { + suggest_ty: SuggestedType, + vec_snippet: String, + expr_hir_id: HirId, + }, + NoChange, +} +enum VecToArray { + /// Expression does not need to be a `Vec<_>` and its type can be changed to an array (or + /// slice). + Possible, + /// Expression must be a `Vec<_>`. Type cannot change. + Impossible, +} + +impl UselessVec { + /// Checks if the surrounding environment requires this expression to actually be of type + /// `Vec<_>`, or if it can be changed to `&[]`/`[]` without causing type errors. + fn expr_usage_requires_vec(&mut self, cx: &LateContext<'_>, expr: &Expr<'_>) -> VecToArray { match cx.tcx.parent_hir_node(expr.hir_id) { // search for `let foo = vec![_]` expressions where all uses of `foo` // adjust to slices or call a method that exist on slices (e.g. len) @@ -100,110 +134,122 @@ impl<'tcx> LateLintPass<'tcx> for UselessVec { .is_continue(); if only_slice_uses { - self.check_vec_macro(cx, &vec_args, callsite, expr.hir_id, SuggestedType::Array); + VecToArray::Possible } else { - self.span_to_lint_map.insert(callsite, None); + VecToArray::Impossible } }, // if the local pattern has a specified type, do not lint. Node::LetStmt(LetStmt { ty: Some(_), .. }) if higher::VecArgs::hir(cx, expr).is_some() => { - self.span_to_lint_map.insert(callsite, None); + VecToArray::Impossible }, // search for `for _ in vec![...]` Node::Expr(Expr { span, .. }) if span.is_desugaring(DesugaringKind::ForLoop) && self.msrv.meets(cx, msrvs::ARRAY_INTO_ITERATOR) => { - let suggest_slice = suggest_type(expr); - self.check_vec_macro(cx, &vec_args, callsite, expr.hir_id, suggest_slice); + VecToArray::Possible }, // search for `&vec![_]` or `vec![_]` expressions where the adjusted type is `&[_]` _ => { - let suggest_slice = suggest_type(expr); - if adjusts_to_slice(cx, expr) { - self.check_vec_macro(cx, &vec_args, callsite, expr.hir_id, suggest_slice); + VecToArray::Possible } else { - self.span_to_lint_map.insert(callsite, None); + VecToArray::Impossible } }, } } +} + +impl<'tcx> LateLintPass<'tcx> for UselessVec { + fn check_expr(&mut self, cx: &LateContext<'tcx>, expr: &'tcx Expr<'_>) { + if let Some(vec_args) = higher::VecArgs::hir(cx, expr.peel_borrows()) + // The `vec![]` or `&vec![]` invocation span. + && let vec_span = expr.span.parent_callsite().unwrap_or(expr.span) + && !vec_span.from_expansion() + { + if self.allow_in_test && is_in_test(cx.tcx, expr.hir_id) { + return; + } + + match self.expr_usage_requires_vec(cx, expr) { + VecToArray::Possible => { + let suggest_ty = suggest_type(expr); + + // Size and `Copy` checks don't depend on the enclosing usage of the expression + // and don't need to be inserted into the state map. + let vec_snippet = match vec_args { + higher::VecArgs::Repeat(expr, len) => { + if is_copy(cx, cx.typeck_results().expr_ty(expr)) + && let Some(Constant::Int(length)) = ConstEvalCtxt::new(cx).eval(len) + && let Ok(length) = u64::try_from(length) + && size_of(cx, expr) + .checked_mul(length) + .is_some_and(|size| size <= self.too_large_for_stack) + { + suggest_ty.snippet(cx, Some(expr.span), Some(len.span)) + } else { + return; + } + }, + higher::VecArgs::Vec(args) => { + if let Ok(length) = u64::try_from(args.len()) + && size_of(cx, expr) + .checked_mul(length) + .is_some_and(|size| size <= self.too_large_for_stack) + { + suggest_ty.snippet( + cx, + args.first().zip(args.last()).map(|(first, last)| { + first.span.source_callsite().to(last.span.source_callsite()) + }), + None, + ) + } else { + return; + } + }, + }; + + if let Entry::Vacant(entry) = self.span_to_state.entry(vec_span) { + entry.insert(VecState::Change { + suggest_ty, + vec_snippet, + expr_hir_id: expr.hir_id, + }); + } + }, + VecToArray::Impossible => { + self.span_to_state.insert(vec_span, VecState::NoChange); + }, + } + } + } fn check_crate_post(&mut self, cx: &LateContext<'tcx>) { - for (span, lint_opt) in &self.span_to_lint_map { - if let Some((hir_id, suggest_slice, snippet, applicability)) = lint_opt { - let help_msg = format!("you can use {} directly", suggest_slice.desc()); - span_lint_hir_and_then(cx, USELESS_VEC, *hir_id, *span, "useless use of `vec!`", |diag| { - // If the `vec!` macro contains comment, better not make the suggestion machine - // applicable as it would remove them. - let applicability = if *applicability != Applicability::Unspecified - && let source_map = cx.tcx.sess.source_map() - && span_contains_comment(source_map, *span) - { + for (span, state) in mem::take(&mut self.span_to_state) { + if let VecState::Change { + suggest_ty, + vec_snippet, + expr_hir_id, + } = state + { + span_lint_hir_and_then(cx, USELESS_VEC, expr_hir_id, span, "useless use of `vec!`", |diag| { + let help_msg = format!("you can use {} directly", suggest_ty.desc()); + // If the `vec!` macro contains comment, better not make the suggestion machine applicable as it + // would remove them. + let applicability = if span_contains_comment(cx.tcx.sess.source_map(), span) { Applicability::Unspecified } else { - *applicability + Applicability::MachineApplicable }; - diag.span_suggestion(*span, help_msg, snippet, applicability); + diag.span_suggestion(span, help_msg, vec_snippet, applicability); }); } } } } -impl UselessVec { - fn check_vec_macro<'tcx>( - &mut self, - cx: &LateContext<'tcx>, - vec_args: &higher::VecArgs<'tcx>, - span: Span, - hir_id: HirId, - suggest_slice: SuggestedType, - ) { - if span.from_expansion() { - return; - } - - let snippet = match *vec_args { - higher::VecArgs::Repeat(elem, len) => { - if let Some(Constant::Int(len_constant)) = ConstEvalCtxt::new(cx).eval(len) { - // vec![ty; N] works when ty is Clone, [ty; N] requires it to be Copy also - if !is_copy(cx, cx.typeck_results().expr_ty(elem)) { - return; - } - - #[expect(clippy::cast_possible_truncation)] - if len_constant as u64 * size_of(cx, elem) > self.too_large_for_stack { - return; - } - - suggest_slice.snippet(cx, Some(elem.span), Some(len.span)) - } else { - return; - } - }, - higher::VecArgs::Vec(args) => { - let args_span = if let Some(last) = args.iter().last() { - if args.len() as u64 * size_of(cx, last) > self.too_large_for_stack { - return; - } - Some(args[0].span.source_callsite().to(last.span.source_callsite())) - } else { - None - }; - suggest_slice.snippet(cx, args_span, None) - }, - }; - - self.span_to_lint_map.entry(span).or_insert(Some(( - hir_id, - suggest_slice, - snippet, - Applicability::MachineApplicable, - ))); - } -} - #[derive(Copy, Clone)] pub(crate) enum SuggestedType { /// Suggest using a slice `&[..]` / `&mut [..]` -- cgit 1.4.1-3-g733a5 From bfd73ccae051b32686486d1849ea92750d259ad6 Mon Sep 17 00:00:00 2001 From: y21 <30553356+y21@users.noreply.github.com> Date: Thu, 3 Apr 2025 17:26:22 +0200 Subject: Add test case for array false positive --- tests/ui/vec.fixed | 9 +++++++++ tests/ui/vec.rs | 9 +++++++++ tests/ui/vec.stderr | 8 +++++++- 3 files changed, 25 insertions(+), 1 deletion(-) diff --git a/tests/ui/vec.fixed b/tests/ui/vec.fixed index f360a8afadf..55742459c92 100644 --- a/tests/ui/vec.fixed +++ b/tests/ui/vec.fixed @@ -242,3 +242,12 @@ fn issue_12101() { for a in &[1, 2] {} //~^ useless_vec } + +fn issue_14531() { + // The lint used to suggest using an array rather than a reference to a slice. + + fn requires_ref_slice(v: &[()]) {} + let v = &[]; + //~^ useless_vec + requires_ref_slice(v); +} diff --git a/tests/ui/vec.rs b/tests/ui/vec.rs index a779d33557c..fbf7131323c 100644 --- a/tests/ui/vec.rs +++ b/tests/ui/vec.rs @@ -242,3 +242,12 @@ fn issue_12101() { for a in &(vec![1, 2]) {} //~^ useless_vec } + +fn issue_14531() { + // The lint used to suggest using an array rather than a reference to a slice. + + fn requires_ref_slice(v: &[()]) {} + let v = &vec![]; + //~^ useless_vec + requires_ref_slice(v); +} diff --git a/tests/ui/vec.stderr b/tests/ui/vec.stderr index 806d6617200..d16c8a8944a 100644 --- a/tests/ui/vec.stderr +++ b/tests/ui/vec.stderr @@ -127,5 +127,11 @@ error: useless use of `vec!` LL | for a in &(vec![1, 2]) {} | ^^^^^^^^^^^^^ help: you can use a slice directly: `&[1, 2]` -error: aborting due to 21 previous errors +error: useless use of `vec!` + --> tests/ui/vec.rs:250:13 + | +LL | let v = &vec![]; + | ^^^^^^^ help: you can use a slice directly: `&[]` + +error: aborting due to 22 previous errors -- cgit 1.4.1-3-g733a5 From 47e8c315f3e600357f0147b867de798b1b96cd03 Mon Sep 17 00:00:00 2001 From: y21 <30553356+y21@users.noreply.github.com> Date: Sat, 21 Jun 2025 15:34:31 +0200 Subject: Make sure spans are from the root --- clippy_lints/src/vec.rs | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/clippy_lints/src/vec.rs b/clippy_lints/src/vec.rs index 1646f222021..52b30ddce12 100644 --- a/clippy_lints/src/vec.rs +++ b/clippy_lints/src/vec.rs @@ -25,7 +25,7 @@ pub struct UselessVec { /// Maps from a `vec![]` source callsite invocation span to the "state" (i.e., whether we can /// emit a warning there or not). /// - /// The purpose of this is to buffer lints up until `check_expr_post` so that we can cancel a + /// The purpose of this is to buffer lints up until `check_crate_post` so that we can cancel a /// lint while visiting, because a `vec![]` invocation span can appear multiple times when /// it is passed as a macro argument, once in a context that doesn't require a `Vec<_>` and /// another time that does. Consider: @@ -187,7 +187,11 @@ impl<'tcx> LateLintPass<'tcx> for UselessVec { .checked_mul(length) .is_some_and(|size| size <= self.too_large_for_stack) { - suggest_ty.snippet(cx, Some(expr.span), Some(len.span)) + suggest_ty.snippet( + cx, + Some(expr.span.source_callsite()), + Some(len.span.source_callsite()), + ) } else { return; } @@ -267,11 +271,17 @@ impl SuggestedType { } fn snippet(self, cx: &LateContext<'_>, args_span: Option, len_span: Option) -> String { + // Invariant of the lint as implemented: all spans are from the root context (and as a result, + // always trivially crate-local). + assert!(args_span.is_none_or(|s| !s.from_expansion())); + assert!(len_span.is_none_or(|s| !s.from_expansion())); + let maybe_args = args_span - .and_then(|sp| sp.get_source_text(cx)) + .map(|sp| sp.get_source_text(cx).expect("spans are always crate-local")) .map_or(String::new(), |x| x.to_owned()); let maybe_len = len_span - .and_then(|sp| sp.get_source_text(cx).map(|s| format!("; {s}"))) + .map(|sp| sp.get_source_text(cx).expect("spans are always crate-local")) + .map(|st| format!("; {st}")) .unwrap_or_default(); match self { -- cgit 1.4.1-3-g733a5 From 3ddd2ef8c518efe10c811a229f056e96864d83c1 Mon Sep 17 00:00:00 2001 From: Manish Goregaokar Date: Wed, 2 Jul 2025 13:15:07 -0700 Subject: Mention that casting to *const () is a way to roundtrip with from_raw_parts --- library/core/src/ptr/metadata.rs | 3 +++ 1 file changed, 3 insertions(+) diff --git a/library/core/src/ptr/metadata.rs b/library/core/src/ptr/metadata.rs index 0deac3621e8..dc3ec3fd199 100644 --- a/library/core/src/ptr/metadata.rs +++ b/library/core/src/ptr/metadata.rs @@ -106,6 +106,9 @@ pub const fn metadata(ptr: *const T) -> ::Metadat /// For slices, see the documentation of [`slice::from_raw_parts`] for safety requirements. /// For trait objects, the metadata must come from a pointer to the same underlying erased type. /// +/// If you are attempting to deconstruct a DST in a generic context to be reconstructed later, +/// a thin pointer can always be obtained by casting `*const T` to `*const ()`. +/// /// [`slice::from_raw_parts`]: crate::slice::from_raw_parts #[unstable(feature = "ptr_metadata", issue = "81513")] #[inline] -- cgit 1.4.1-3-g733a5 From a3208108b0b2b98004dac023aaeb794c1ec9149a Mon Sep 17 00:00:00 2001 From: Ricardo Fernández Serrata <76864299+Rudxain@users.noreply.github.com> Date: Fri, 4 Jul 2025 19:10:39 -0400 Subject: fix(lib-std-fs): handle `usize` overflow in `read` & `read_to_string` --- library/std/src/fs.rs | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/library/std/src/fs.rs b/library/std/src/fs.rs index 72ad7c244ee..9fbfd92eaf8 100644 --- a/library/std/src/fs.rs +++ b/library/std/src/fs.rs @@ -304,7 +304,7 @@ pub struct DirBuilder { pub fn read>(path: P) -> io::Result> { fn inner(path: &Path) -> io::Result> { let mut file = File::open(path)?; - let size = file.metadata().map(|m| m.len() as usize).ok(); + let size = file.metadata().map(|m| usize::try_from(m.len()).unwrap_or(usize::MAX)).ok(); let mut bytes = Vec::try_with_capacity(size.unwrap_or(0))?; io::default_read_to_end(&mut file, &mut bytes, size)?; Ok(bytes) @@ -346,7 +346,7 @@ pub fn read>(path: P) -> io::Result> { pub fn read_to_string>(path: P) -> io::Result { fn inner(path: &Path) -> io::Result { let mut file = File::open(path)?; - let size = file.metadata().map(|m| m.len() as usize).ok(); + let size = file.metadata().map(|m| usize::try_from(m.len()).unwrap_or(usize::MAX)).ok(); let mut string = String::new(); string.try_reserve_exact(size.unwrap_or(0))?; io::default_read_to_string(&mut file, &mut string, size)?; -- cgit 1.4.1-3-g733a5 From 905342525f88b757a4f7026b2b51a21621999fd1 Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Wed, 9 Jul 2025 21:27:27 -0400 Subject: Add a mailmap entry for gnzlbg The submodule->subtree changes add a lot of commits with the GitHub email. --- .mailmap | 1 + 1 file changed, 1 insertion(+) diff --git a/.mailmap b/.mailmap index 90533e81b39..68abb9003a0 100644 --- a/.mailmap +++ b/.mailmap @@ -255,6 +255,7 @@ Guillaume Gomez Guillaume Gomez ggomez Guillaume Gomez Guillaume Gomez Guillaume Gomez Guillaume Gomez +gnzlbg hamidreza kalbasi Hanna Kruppe Heather -- cgit 1.4.1-3-g733a5 From 2261154549996eb7737a895e5df5fa894dc03e42 Mon Sep 17 00:00:00 2001 From: kennytm Date: Thu, 10 Jul 2025 18:22:18 +0800 Subject: core: add Peekable::next_if_map --- library/core/src/iter/adapters/peekable.rs | 102 ++++++++++++++++++++++ library/coretests/tests/iter/adapters/peekable.rs | 86 ++++++++++++++++++ library/coretests/tests/lib.rs | 1 + 3 files changed, 189 insertions(+) diff --git a/library/core/src/iter/adapters/peekable.rs b/library/core/src/iter/adapters/peekable.rs index a6522659620..a55de75d56c 100644 --- a/library/core/src/iter/adapters/peekable.rs +++ b/library/core/src/iter/adapters/peekable.rs @@ -317,6 +317,108 @@ impl Peekable { { self.next_if(|next| next == expected) } + + /// Consumes the next value of this iterator and applies a function `f` on it, + /// returning the result if the closure returns `Ok`. + /// + /// Otherwise if the closure returns `Err` the value is put back for the next iteration. + /// + /// The content of the `Err` variant is typically the original value of the closure, + /// but this is not required. If a different value is returned, + /// the next `peek()` or `next()` call will result in this new value. + /// This is similar to modifying the output of `peek_mut()`. + /// + /// If the closure panics, the next value will always be consumed and dropped + /// even if the panic is caught, because the closure never returned an `Err` value to put back. + /// + /// # Examples + /// + /// Parse the leading decimal number from an iterator of characters. + /// ``` + /// #![feature(peekable_next_if_map)] + /// let mut iter = "125 GOTO 10".chars().peekable(); + /// let mut line_num = 0_u32; + /// while let Some(digit) = iter.next_if_map(|c| c.to_digit(10).ok_or(c)) { + /// line_num = line_num * 10 + digit; + /// } + /// assert_eq!(line_num, 125); + /// assert_eq!(iter.collect::(), " GOTO 10"); + /// ``` + /// + /// Matching custom types. + /// ``` + /// #![feature(peekable_next_if_map)] + /// + /// #[derive(Debug, PartialEq, Eq)] + /// enum Node { + /// Comment(String), + /// Red(String), + /// Green(String), + /// Blue(String), + /// } + /// + /// /// Combines all consecutive `Comment` nodes into a single one. + /// fn combine_comments(nodes: Vec) -> Vec { + /// let mut result = Vec::with_capacity(nodes.len()); + /// let mut iter = nodes.into_iter().peekable(); + /// let mut comment_text = None::; + /// loop { + /// // Typically the closure in .next_if_map() matches on the input, + /// // extracts the desired pattern into an `Ok`, + /// // and puts the rest into an `Err`. + /// while let Some(text) = iter.next_if_map(|node| match node { + /// Node::Comment(text) => Ok(text), + /// other => Err(other), + /// }) { + /// comment_text.get_or_insert_default().push_str(&text); + /// } + /// + /// if let Some(text) = comment_text.take() { + /// result.push(Node::Comment(text)); + /// } + /// if let Some(node) = iter.next() { + /// result.push(node); + /// } else { + /// break; + /// } + /// } + /// result + /// } + ///# assert_eq!( // hiding the test to avoid cluttering the documentation. + ///# combine_comments(vec![ + ///# Node::Comment("The".to_owned()), + ///# Node::Comment("Quick".to_owned()), + ///# Node::Comment("Brown".to_owned()), + ///# Node::Red("Fox".to_owned()), + ///# Node::Green("Jumped".to_owned()), + ///# Node::Comment("Over".to_owned()), + ///# Node::Blue("The".to_owned()), + ///# Node::Comment("Lazy".to_owned()), + ///# Node::Comment("Dog".to_owned()), + ///# ]), + ///# vec![ + ///# Node::Comment("TheQuickBrown".to_owned()), + ///# Node::Red("Fox".to_owned()), + ///# Node::Green("Jumped".to_owned()), + ///# Node::Comment("Over".to_owned()), + ///# Node::Blue("The".to_owned()), + ///# Node::Comment("LazyDog".to_owned()), + ///# ], + ///# ) + /// ``` + #[unstable(feature = "peekable_next_if_map", issue = "143702")] + pub fn next_if_map(&mut self, f: impl FnOnce(I::Item) -> Result) -> Option { + let unpeek = if let Some(item) = self.next() { + match f(item) { + Ok(result) => return Some(result), + Err(item) => Some(item), + } + } else { + None + }; + self.peeked = Some(unpeek); + None + } } #[unstable(feature = "trusted_len", issue = "37572")] diff --git a/library/coretests/tests/iter/adapters/peekable.rs b/library/coretests/tests/iter/adapters/peekable.rs index 7f4341b8902..f0549e8d6c2 100644 --- a/library/coretests/tests/iter/adapters/peekable.rs +++ b/library/coretests/tests/iter/adapters/peekable.rs @@ -271,3 +271,89 @@ fn test_peekable_non_fused() { assert_eq!(iter.peek(), None); assert_eq!(iter.next_back(), None); } + +#[test] +fn test_peekable_next_if_map_mutation() { + fn collatz((mut num, mut len): (u64, u32)) -> Result { + let jump = num.trailing_zeros(); + num >>= jump; + len += jump; + if num == 1 { Ok(len) } else { Err((3 * num + 1, len + 1)) } + } + + let mut iter = once((3, 0)).peekable(); + assert_eq!(iter.peek(), Some(&(3, 0))); + assert_eq!(iter.next_if_map(collatz), None); + assert_eq!(iter.peek(), Some(&(10, 1))); + assert_eq!(iter.next_if_map(collatz), None); + assert_eq!(iter.peek(), Some(&(16, 3))); + assert_eq!(iter.next_if_map(collatz), Some(7)); + assert_eq!(iter.peek(), None); + assert_eq!(iter.next_if_map(collatz), None); +} + +#[test] +#[cfg_attr(not(panic = "unwind"), ignore = "test requires unwinding support")] +fn test_peekable_next_if_map_panic() { + use core::cell::Cell; + use std::panic::{AssertUnwindSafe, catch_unwind}; + + struct BitsetOnDrop<'a> { + value: u32, + cell: &'a Cell, + } + impl<'a> Drop for BitsetOnDrop<'a> { + fn drop(&mut self) { + self.cell.update(|v| v | self.value); + } + } + + let cell = &Cell::new(0); + let mut it = [ + BitsetOnDrop { value: 1, cell }, + BitsetOnDrop { value: 2, cell }, + BitsetOnDrop { value: 4, cell }, + BitsetOnDrop { value: 8, cell }, + ] + .into_iter() + .peekable(); + + // sanity check, .peek() won't consume the value, .next() will transfer ownership. + let item = it.peek().unwrap(); + assert_eq!(item.value, 1); + assert_eq!(cell.get(), 0); + let item = it.next().unwrap(); + assert_eq!(item.value, 1); + assert_eq!(cell.get(), 0); + drop(item); + assert_eq!(cell.get(), 1); + + // next_if_map returning Ok should transfer the value out. + let item = it.next_if_map(Ok).unwrap(); + assert_eq!(item.value, 2); + assert_eq!(cell.get(), 1); + drop(item); + assert_eq!(cell.get(), 3); + + // next_if_map returning Err should not drop anything. + assert_eq!(it.next_if_map::<()>(Err), None); + assert_eq!(cell.get(), 3); + assert_eq!(it.peek().unwrap().value, 4); + assert_eq!(cell.get(), 3); + + // next_if_map panicking should consume and drop the item. + let result = catch_unwind({ + let mut it = AssertUnwindSafe(&mut it); + move || it.next_if_map::<()>(|_| panic!()) + }); + assert!(result.is_err()); + assert_eq!(cell.get(), 7); + assert_eq!(it.next().unwrap().value, 8); + assert_eq!(cell.get(), 15); + assert!(it.peek().is_none()); + + // next_if_map should *not* execute the closure if the iterator is exhausted. + assert!(it.next_if_map::<()>(|_| panic!()).is_none()); + assert!(it.peek().is_none()); + assert_eq!(cell.get(), 15); +} diff --git a/library/coretests/tests/lib.rs b/library/coretests/tests/lib.rs index fdef736c0c0..18e6e96c235 100644 --- a/library/coretests/tests/lib.rs +++ b/library/coretests/tests/lib.rs @@ -78,6 +78,7 @@ #![feature(next_index)] #![feature(numfmt)] #![feature(pattern)] +#![feature(peekable_next_if_map)] #![feature(pointer_is_aligned_to)] #![feature(portable_simd)] #![feature(ptr_metadata)] -- cgit 1.4.1-3-g733a5 From 84f834b87f1f0b2eb56e94652eb3a4b440402a93 Mon Sep 17 00:00:00 2001 From: Ralf Jung Date: Sat, 19 Jul 2025 23:23:40 +0200 Subject: atomicrmw on pointers: move integer-pointer cast hacks into backend --- src/builder.rs | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/src/builder.rs b/src/builder.rs index a4ec4bf8dea..032f5d0a622 100644 --- a/src/builder.rs +++ b/src/builder.rs @@ -1656,6 +1656,7 @@ impl<'a, 'gcc, 'tcx> BuilderMethods<'a, 'tcx> for Builder<'a, 'gcc, 'tcx> { dst: RValue<'gcc>, src: RValue<'gcc>, order: AtomicOrdering, + ret_ptr: bool, ) -> RValue<'gcc> { let size = get_maybe_pointer_size(src); let name = match op { @@ -1683,6 +1684,9 @@ impl<'a, 'gcc, 'tcx> BuilderMethods<'a, 'tcx> for Builder<'a, 'gcc, 'tcx> { let atomic_function = self.context.get_builtin_function(name); let order = self.context.new_rvalue_from_int(self.i32_type, order.to_gcc()); + // FIXME: If `ret_ptr` is true and `src` is an integer, we should really tell GCC + // that this is a pointer operation that needs to preserve provenance -- but like LLVM, + // GCC does not currently seems to support that. let void_ptr_type = self.context.new_type::<*mut ()>(); let volatile_void_ptr_type = void_ptr_type.make_volatile(); let dst = self.context.new_cast(self.location, dst, volatile_void_ptr_type); @@ -1690,7 +1694,8 @@ impl<'a, 'gcc, 'tcx> BuilderMethods<'a, 'tcx> for Builder<'a, 'gcc, 'tcx> { let new_src_type = atomic_function.get_param(1).to_rvalue().get_type(); let src = self.context.new_bitcast(self.location, src, new_src_type); let res = self.context.new_call(self.location, atomic_function, &[dst, src, order]); - self.context.new_cast(self.location, res, src.get_type()) + let res_type = if ret_ptr { void_ptr_type } else { src.get_type() }; + self.context.new_cast(self.location, res, res_type) } fn atomic_fence(&mut self, order: AtomicOrdering, scope: SynchronizationScope) { -- cgit 1.4.1-3-g733a5 From bd33a02ea6c01a7053451b0205edb01e4172a601 Mon Sep 17 00:00:00 2001 From: Samuel Tardieu Date: Sat, 7 Jun 2025 08:25:20 +0200 Subject: Get the block content from the proper context The first statement of the block might have been in a different context from the expression. Walk up to the right context to get bounds properly. Also, switch to `snippet_with_applicability()` since we know that we are in the right context already. --- clippy_lints/src/if_then_some_else_none.rs | 32 ++++++++++--------- tests/ui/if_then_some_else_none.fixed | 41 ++++++++++++++++++++++++ tests/ui/if_then_some_else_none.rs | 51 ++++++++++++++++++++++++++++++ tests/ui/if_then_some_else_none.stderr | 25 ++++++++++++++- 4 files changed, 133 insertions(+), 16 deletions(-) diff --git a/clippy_lints/src/if_then_some_else_none.rs b/clippy_lints/src/if_then_some_else_none.rs index 7158f9419c1..b50d91f1014 100644 --- a/clippy_lints/src/if_then_some_else_none.rs +++ b/clippy_lints/src/if_then_some_else_none.rs @@ -2,16 +2,16 @@ use clippy_config::Conf; use clippy_utils::diagnostics::span_lint_and_then; use clippy_utils::eager_or_lazy::switch_to_eager_eval; use clippy_utils::msrvs::{self, Msrv}; -use clippy_utils::source::snippet_with_context; +use clippy_utils::source::{snippet_with_applicability, snippet_with_context, walk_span_to_context}; use clippy_utils::sugg::Sugg; use clippy_utils::{ contains_return, expr_adjustment_requires_coercion, higher, is_else_clause, is_in_const_context, is_res_lang_ctor, - path_res, peel_blocks, + path_res, peel_blocks, sym, }; use rustc_errors::Applicability; use rustc_hir::LangItem::{OptionNone, OptionSome}; use rustc_hir::{Expr, ExprKind}; -use rustc_lint::{LateContext, LateLintPass, LintContext}; +use rustc_lint::{LateContext, LateLintPass}; use rustc_session::impl_lint_pass; declare_clippy_lint! { @@ -71,21 +71,21 @@ impl<'tcx> LateLintPass<'tcx> for IfThenSomeElseNone { && let ExprKind::Block(then_block, _) = then.kind && let Some(then_expr) = then_block.expr && let ExprKind::Call(then_call, [then_arg]) = then_expr.kind - && let ctxt = expr.span.ctxt() - && then_expr.span.ctxt() == ctxt + && !expr.span.from_expansion() + && !then_expr.span.from_expansion() && is_res_lang_ctor(cx, path_res(cx, then_call), OptionSome) && is_res_lang_ctor(cx, path_res(cx, peel_blocks(els)), OptionNone) && !is_else_clause(cx.tcx, expr) && !is_in_const_context(cx) - && !expr.span.in_external_macro(cx.sess().source_map()) && self.msrv.meets(cx, msrvs::BOOL_THEN) && !contains_return(then_block.stmts) { let method_name = if switch_to_eager_eval(cx, expr) && self.msrv.meets(cx, msrvs::BOOL_THEN_SOME) { - "then_some" + sym::then_some } else { - "then" + sym::then }; + let ctxt = expr.span.ctxt(); span_lint_and_then( cx, @@ -98,16 +98,18 @@ impl<'tcx> LateLintPass<'tcx> for IfThenSomeElseNone { } let mut app = Applicability::MachineApplicable; - let cond_snip = Sugg::hir_with_context(cx, cond, expr.span.ctxt(), "[condition]", &mut app) + let cond_snip = Sugg::hir_with_context(cx, cond, ctxt, "[condition]", &mut app) .maybe_paren() .to_string(); let arg_snip = snippet_with_context(cx, then_arg.span, ctxt, "[body]", &mut app).0; - let method_body = if let Some(first_stmt) = then_block.stmts.first() { - let (block_snippet, _) = - snippet_with_context(cx, first_stmt.span.until(then_arg.span), ctxt, "..", &mut app); - let closure = if method_name == "then" { "|| " } else { "" }; - format!("{closure} {{ {block_snippet}; {arg_snip} }}") - } else if method_name == "then" { + let method_body = if let Some(first_stmt) = then_block.stmts.first() + && let Some(first_stmt_span) = walk_span_to_context(first_stmt.span, ctxt) + { + let block_snippet = + snippet_with_applicability(cx, first_stmt_span.until(then_expr.span), "..", &mut app); + let closure = if method_name == sym::then { "|| " } else { "" }; + format!("{closure} {{ {} {arg_snip} }}", block_snippet.trim_end()) + } else if method_name == sym::then { (std::borrow::Cow::Borrowed("|| ") + arg_snip).into_owned() } else { arg_snip.into_owned() diff --git a/tests/ui/if_then_some_else_none.fixed b/tests/ui/if_then_some_else_none.fixed index d14a805b666..0fd130609ae 100644 --- a/tests/ui/if_then_some_else_none.fixed +++ b/tests/ui/if_then_some_else_none.fixed @@ -165,3 +165,44 @@ mod issue15257 { do_something((i % 2 == 0).then_some(closure_fn)); } } + +fn issue15005() { + struct Counter { + count: u32, + } + + impl Counter { + fn new() -> Counter { + Counter { count: 0 } + } + } + + impl Iterator for Counter { + type Item = u32; + + fn next(&mut self) -> Option { + //~v if_then_some_else_none + (self.count < 5).then(|| { self.count += 1; self.count }) + } + } +} + +fn statements_from_macro() { + macro_rules! mac { + () => { + println!("foo"); + println!("bar"); + }; + } + //~v if_then_some_else_none + let _ = true.then(|| { mac!(); 42 }); +} + +fn dont_lint_inside_macros() { + macro_rules! mac { + ($cond:expr, $res:expr) => { + if $cond { Some($res) } else { None } + }; + } + let _: Option = mac!(true, 42); +} diff --git a/tests/ui/if_then_some_else_none.rs b/tests/ui/if_then_some_else_none.rs index bb0072f3157..640828aa9bf 100644 --- a/tests/ui/if_then_some_else_none.rs +++ b/tests/ui/if_then_some_else_none.rs @@ -211,3 +211,54 @@ mod issue15257 { }); } } + +fn issue15005() { + struct Counter { + count: u32, + } + + impl Counter { + fn new() -> Counter { + Counter { count: 0 } + } + } + + impl Iterator for Counter { + type Item = u32; + + fn next(&mut self) -> Option { + //~v if_then_some_else_none + if self.count < 5 { + self.count += 1; + Some(self.count) + } else { + None + } + } + } +} + +fn statements_from_macro() { + macro_rules! mac { + () => { + println!("foo"); + println!("bar"); + }; + } + //~v if_then_some_else_none + let _ = if true { + mac!(); + Some(42) + } else { + None + }; +} + +fn dont_lint_inside_macros() { + macro_rules! mac { + ($cond:expr, $res:expr) => { + if $cond { Some($res) } else { None } + }; + } + let _: Option = mac!(true, 42); +} diff --git a/tests/ui/if_then_some_else_none.stderr b/tests/ui/if_then_some_else_none.stderr index c2e624a0a73..58651a05594 100644 --- a/tests/ui/if_then_some_else_none.stderr +++ b/tests/ui/if_then_some_else_none.stderr @@ -116,5 +116,28 @@ LL | | None LL | | }); | |_________^ help: try: `(i % 2 == 0).then_some(closure_fn)` -error: aborting due to 11 previous errors +error: this could be simplified with `bool::then` + --> tests/ui/if_then_some_else_none.rs:231:13 + | +LL | / if self.count < 5 { +LL | | self.count += 1; +LL | | Some(self.count) +LL | | } else { +LL | | None +LL | | } + | |_____________^ help: try: `(self.count < 5).then(|| { self.count += 1; self.count })` + +error: this could be simplified with `bool::then` + --> tests/ui/if_then_some_else_none.rs:249:13 + | +LL | let _ = if true { + | _____________^ +LL | | mac!(); +LL | | Some(42) +LL | | } else { +LL | | None +LL | | }; + | |_____^ help: try: `true.then(|| { mac!(); 42 })` + +error: aborting due to 13 previous errors -- cgit 1.4.1-3-g733a5 From 613080b5f1d58a2496d9f8dc8c314d5432fe3d6f Mon Sep 17 00:00:00 2001 From: Jeremy Smart Date: Thu, 31 Jul 2025 23:46:04 -0400 Subject: add Option::reduce --- library/core/src/option.rs | 36 ++++++++++++++++++++++++++++++++++++ library/coretests/tests/lib.rs | 1 + 2 files changed, 37 insertions(+) diff --git a/library/core/src/option.rs b/library/core/src/option.rs index 0e481519f76..f3ddc05bc1a 100644 --- a/library/core/src/option.rs +++ b/library/core/src/option.rs @@ -1961,6 +1961,42 @@ impl Option { _ => None, } } + + /// Reduces two options into one, using the provided function if both are `Some`. + /// + /// If `self` is `Some(s)` and `other` is `Some(o)`, this method returns `Some(f(s, o))`. + /// Otherwise, if only one of `self` and `other` is `Some`, that one is returned. + /// If both `self` and `other` are `None`, `None` is returned. + /// + /// # Examples + /// + /// ``` + /// #![feature(option_reduce)] + /// + /// let s12 = Some(12); + /// let s17 = Some(17); + /// let n = None; + /// let f = |a, b| a + b; + /// + /// assert_eq!(s12.reduce(s17, f), Some(29)); + /// assert_eq!(s12.reduce(n, f), Some(12)); + /// assert_eq!(n.reduce(s17, f), Some(17)); + /// assert_eq!(n.reduce(n, f), None); + /// ``` + #[unstable(feature = "option_reduce", issue = "144273")] + pub fn reduce(self, other: Option, f: F) -> Option + where + T: Into, + U: Into, + F: FnOnce(T, U) -> R, + { + match (self, other) { + (Some(a), Some(b)) => Some(f(a, b)), + (Some(a), _) => Some(a.into()), + (_, Some(b)) => Some(b.into()), + _ => None, + } + } } impl Option<(T, U)> { diff --git a/library/coretests/tests/lib.rs b/library/coretests/tests/lib.rs index 4cfac9ecc2a..3f817ab128b 100644 --- a/library/coretests/tests/lib.rs +++ b/library/coretests/tests/lib.rs @@ -77,6 +77,7 @@ #![feature(never_type)] #![feature(next_index)] #![feature(numfmt)] +#![feature(option_reduce)] #![feature(pattern)] #![feature(pointer_is_aligned_to)] #![feature(portable_simd)] -- cgit 1.4.1-3-g733a5 From c0c48984b05e244fc93144d0c1d48db318639b3b Mon Sep 17 00:00:00 2001 From: Camille GILLOT Date: Sat, 2 Aug 2025 17:28:17 +0000 Subject: Method probe: check doc aliases after edit distance. --- compiler/rustc_hir_typeck/src/method/probe.rs | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/compiler/rustc_hir_typeck/src/method/probe.rs b/compiler/rustc_hir_typeck/src/method/probe.rs index 1f3969bd93c..f5584e0596a 100644 --- a/compiler/rustc_hir_typeck/src/method/probe.rs +++ b/compiler/rustc_hir_typeck/src/method/probe.rs @@ -2373,17 +2373,14 @@ impl<'a, 'tcx> ProbeContext<'a, 'tcx> { if !self.is_relevant_kind_for_mode(x.kind) { return false; } - if self.matches_by_doc_alias(x.def_id) { - return true; - } - match edit_distance_with_substrings( + if let Some(d) = edit_distance_with_substrings( name.as_str(), x.name().as_str(), max_dist, ) { - Some(d) => d > 0, - None => false, + return d > 0; } + self.matches_by_doc_alias(x.def_id) }) .copied() .collect() -- cgit 1.4.1-3-g733a5 From a7162e416e73afd4fc4c72f171b8166bb9d35b8c Mon Sep 17 00:00:00 2001 From: Samuel Tardieu Date: Sat, 2 Aug 2025 23:42:57 +0200 Subject: Remove references to two unknown lints in config `option_map_unwrap_or` and `seek_rewind` are not current lints, and cannot be referenced as having the `msrv` configuration option. --- book/src/lint_configuration.md | 2 -- clippy_config/src/conf.rs | 2 -- 2 files changed, 4 deletions(-) diff --git a/book/src/lint_configuration.md b/book/src/lint_configuration.md index 7f16f3a9810..6d554170f1f 100644 --- a/book/src/lint_configuration.md +++ b/book/src/lint_configuration.md @@ -873,7 +873,6 @@ The minimum rust version that the project supports. Defaults to the `rust-versio * [`needless_borrow`](https://rust-lang.github.io/rust-clippy/master/index.html#needless_borrow) * [`non_std_lazy_statics`](https://rust-lang.github.io/rust-clippy/master/index.html#non_std_lazy_statics) * [`option_as_ref_deref`](https://rust-lang.github.io/rust-clippy/master/index.html#option_as_ref_deref) -* [`option_map_unwrap_or`](https://rust-lang.github.io/rust-clippy/master/index.html#option_map_unwrap_or) * [`ptr_as_ptr`](https://rust-lang.github.io/rust-clippy/master/index.html#ptr_as_ptr) * [`question_mark`](https://rust-lang.github.io/rust-clippy/master/index.html#question_mark) * [`redundant_field_names`](https://rust-lang.github.io/rust-clippy/master/index.html#redundant_field_names) @@ -881,7 +880,6 @@ The minimum rust version that the project supports. Defaults to the `rust-versio * [`repeat_vec_with_capacity`](https://rust-lang.github.io/rust-clippy/master/index.html#repeat_vec_with_capacity) * [`same_item_push`](https://rust-lang.github.io/rust-clippy/master/index.html#same_item_push) * [`seek_from_current`](https://rust-lang.github.io/rust-clippy/master/index.html#seek_from_current) -* [`seek_rewind`](https://rust-lang.github.io/rust-clippy/master/index.html#seek_rewind) * [`to_digit_is_some`](https://rust-lang.github.io/rust-clippy/master/index.html#to_digit_is_some) * [`transmute_ptr_to_ref`](https://rust-lang.github.io/rust-clippy/master/index.html#transmute_ptr_to_ref) * [`tuple_array_conversions`](https://rust-lang.github.io/rust-clippy/master/index.html#tuple_array_conversions) diff --git a/clippy_config/src/conf.rs b/clippy_config/src/conf.rs index 8167d75583e..a9987b4e2ec 100644 --- a/clippy_config/src/conf.rs +++ b/clippy_config/src/conf.rs @@ -775,7 +775,6 @@ define_Conf! { needless_borrow, non_std_lazy_statics, option_as_ref_deref, - option_map_unwrap_or, ptr_as_ptr, question_mark, redundant_field_names, @@ -783,7 +782,6 @@ define_Conf! { repeat_vec_with_capacity, same_item_push, seek_from_current, - seek_rewind, to_digit_is_some, transmute_ptr_to_ref, tuple_array_conversions, -- cgit 1.4.1-3-g733a5 From 07b04a2e04348283acf9565f3de2c500cce4df9a Mon Sep 17 00:00:00 2001 From: Samuel Tardieu Date: Sat, 2 Aug 2025 23:52:39 +0200 Subject: Check that all configuration options reference existing lints --- tests/config-consistency.rs | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) create mode 100644 tests/config-consistency.rs diff --git a/tests/config-consistency.rs b/tests/config-consistency.rs new file mode 100644 index 00000000000..9e7ca26c7d4 --- /dev/null +++ b/tests/config-consistency.rs @@ -0,0 +1,30 @@ +#![feature(rustc_private)] + +// This test checks that all lints defined in `clippy_config::conf` in `#[lints]` +// attributes exist as Clippy lints. +// +// This test is a no-op if run as part of the compiler test suite +// and will always succeed. + +use std::collections::HashSet; + +#[test] +fn config_consistency() { + if option_env!("RUSTC_TEST_SUITE").is_some() { + return; + } + + let lint_names: HashSet = clippy_lints::declared_lints::LINTS + .iter() + .map(|lint_info| lint_info.lint.name.strip_prefix("clippy::").unwrap().to_lowercase()) + .collect(); + for conf in clippy_config::get_configuration_metadata() { + for lint in conf.lints { + assert!( + lint_names.contains(*lint), + "Configuration option {} references lint `{lint}` which does not exist", + conf.name + ); + } + } +} -- cgit 1.4.1-3-g733a5 From bf644e9d14d4e088303041d4136daefaa8ae6f86 Mon Sep 17 00:00:00 2001 From: Camille GILLOT Date: Sat, 2 Aug 2025 19:03:46 +0000 Subject: Use codegen_fn_attrs to check if function is naked. --- compiler/rustc_hir_typeck/src/lib.rs | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/compiler/rustc_hir_typeck/src/lib.rs b/compiler/rustc_hir_typeck/src/lib.rs index aae870f7ee3..72fd8db300b 100644 --- a/compiler/rustc_hir_typeck/src/lib.rs +++ b/compiler/rustc_hir_typeck/src/lib.rs @@ -46,12 +46,12 @@ use rustc_data_structures::unord::UnordSet; use rustc_errors::codes::*; use rustc_errors::{Applicability, ErrorGuaranteed, pluralize, struct_span_code_err}; use rustc_hir as hir; -use rustc_hir::attrs::AttributeKind; use rustc_hir::def::{DefKind, Res}; -use rustc_hir::{HirId, HirIdMap, Node, find_attr}; +use rustc_hir::{HirId, HirIdMap, Node}; use rustc_hir_analysis::check::{check_abi, check_custom_abi}; use rustc_hir_analysis::hir_ty_lowering::HirTyLowerer; use rustc_infer::traits::{ObligationCauseCode, ObligationInspector, WellFormedLoc}; +use rustc_middle::middle::codegen_fn_attrs::CodegenFnAttrFlags; use rustc_middle::query::Providers; use rustc_middle::ty::{self, Ty, TyCtxt, TypeVisitableExt}; use rustc_middle::{bug, span_bug}; @@ -174,7 +174,7 @@ fn typeck_with_inspect<'tcx>( .map(|(idx, ty)| fcx.normalize(arg_span(idx), ty)), ); - if find_attr!(tcx.get_all_attrs(def_id), AttributeKind::Naked(..)) { + if tcx.codegen_fn_attrs(def_id).flags.contains(CodegenFnAttrFlags::NAKED) { naked_functions::typeck_naked_fn(tcx, def_id, body); } -- cgit 1.4.1-3-g733a5 From c4b91bd9f87391c1dda3784205c0dd10b54ec2e8 Mon Sep 17 00:00:00 2001 From: Guillaume Gomez Date: Mon, 4 Aug 2025 10:49:43 +0200 Subject: Merge commit '482e8540a1b757ed7bccc2041c5400f051fdb01e' into subtree-update_cg_gcc_2025-08-04 --- build_system/src/abi_test.rs | 2 +- build_system/src/fuzz.rs | 8 +-- .../0001-Add-stdarch-Cargo.toml-for-testing.patch | 23 +++--- rust-toolchain | 2 +- src/allocator.rs | 14 ++-- src/builder.rs | 26 +++++-- src/int.rs | 76 +++++++++++++++++--- src/intrinsic/archs.rs | 60 ++++++++++++++++ src/intrinsic/mod.rs | 9 ++- src/intrinsic/simd.rs | 81 ++++++++++++++++++++++ src/lib.rs | 5 +- tests/failing-lto-tests.txt | 2 +- tests/failing-run-make-tests.txt | 1 - tests/failing-ui-tests.txt | 8 ++- tools/cspell_dicts/rustc_codegen_gcc.txt | 3 + 15 files changed, 275 insertions(+), 45 deletions(-) diff --git a/build_system/src/abi_test.rs b/build_system/src/abi_test.rs index 3c1531be27a..a85886d87f3 100644 --- a/build_system/src/abi_test.rs +++ b/build_system/src/abi_test.rs @@ -31,7 +31,7 @@ pub fn run() -> Result<(), String> { Some("clones/abi-cafe".as_ref()), true, ) - .map_err(|err| (format!("Git clone failed with message: {err:?}!")))?; + .map_err(|err| format!("Git clone failed with message: {err:?}!"))?; // Configure abi-cafe to use the exact same rustc version we use - this is crucial. // Otherwise, the concept of ABI compatibility becomes meanignless. std::fs::copy("rust-toolchain", "clones/abi-cafe/rust-toolchain") diff --git a/build_system/src/fuzz.rs b/build_system/src/fuzz.rs index 453211366b3..9714ce29af9 100644 --- a/build_system/src/fuzz.rs +++ b/build_system/src/fuzz.rs @@ -43,18 +43,18 @@ pub fn run() -> Result<(), String> { "--start" => { start = str::parse(&args.next().ok_or_else(|| "Fuzz start not provided!".to_string())?) - .map_err(|err| (format!("Fuzz start not a number {err:?}!")))?; + .map_err(|err| format!("Fuzz start not a number {err:?}!"))?; } "--count" => { count = str::parse(&args.next().ok_or_else(|| "Fuzz count not provided!".to_string())?) - .map_err(|err| (format!("Fuzz count not a number {err:?}!")))?; + .map_err(|err| format!("Fuzz count not a number {err:?}!"))?; } "-j" | "--jobs" => { threads = str::parse( &args.next().ok_or_else(|| "Fuzz thread count not provided!".to_string())?, ) - .map_err(|err| (format!("Fuzz thread count not a number {err:?}!")))?; + .map_err(|err| format!("Fuzz thread count not a number {err:?}!"))?; } _ => return Err(format!("Unknown option {arg}")), } @@ -66,7 +66,7 @@ pub fn run() -> Result<(), String> { Some("clones/rustlantis".as_ref()), true, ) - .map_err(|err| (format!("Git clone failed with message: {err:?}!")))?; + .map_err(|err| format!("Git clone failed with message: {err:?}!"))?; // Ensure that we are on the newest rustlantis commit. let cmd: &[&dyn AsRef] = &[&"git", &"pull", &"origin"]; diff --git a/patches/0001-Add-stdarch-Cargo.toml-for-testing.patch b/patches/0001-Add-stdarch-Cargo.toml-for-testing.patch index 9cc377850b9..3a8c37a8b8d 100644 --- a/patches/0001-Add-stdarch-Cargo.toml-for-testing.patch +++ b/patches/0001-Add-stdarch-Cargo.toml-for-testing.patch @@ -1,29 +1,28 @@ -From b8f3eed3053c9333b5dfbeaeb2a6a65a4b3156df Mon Sep 17 00:00:00 2001 -From: Antoni Boucher -Date: Tue, 29 Aug 2023 13:06:34 -0400 +From 190e26c9274b3c93a9ee3516b395590e6bd9213b Mon Sep 17 00:00:00 2001 +From: None +Date: Sun, 3 Aug 2025 19:54:56 -0400 Subject: [PATCH] Patch 0001-Add-stdarch-Cargo.toml-for-testing.patch --- - library/stdarch/Cargo.toml | 23 +++++++++++++++++++++++ - 1 file changed, 23 insertions(+) + library/stdarch/Cargo.toml | 20 ++++++++++++++++++++ + 1 file changed, 20 insertions(+) create mode 100644 library/stdarch/Cargo.toml diff --git a/library/stdarch/Cargo.toml b/library/stdarch/Cargo.toml new file mode 100644 -index 0000000..4c63700 +index 0000000..bd6725c --- /dev/null +++ b/library/stdarch/Cargo.toml -@@ -0,0 +1,21 @@ +@@ -0,0 +1,20 @@ +[workspace] +resolver = "1" +members = [ -+ "crates/core_arch", -+ "crates/std_detect", -+ "crates/stdarch-gen-arm", ++ "crates/*", + #"examples/" +] +exclude = [ -+ "crates/wasm-assert-instr-tests" ++ "crates/wasm-assert-instr-tests", ++ "rust_programs", +] + +[profile.release] @@ -36,5 +35,5 @@ index 0000000..4c63700 +opt-level = 3 +incremental = true -- -2.42.0 +2.50.1 diff --git a/rust-toolchain b/rust-toolchain index 2fe8ec4647f..058e734be5c 100644 --- a/rust-toolchain +++ b/rust-toolchain @@ -1,3 +1,3 @@ [toolchain] -channel = "nightly-2025-07-04" +channel = "nightly-2025-08-03" components = ["rust-src", "rustc-dev", "llvm-tools-preview"] diff --git a/src/allocator.rs b/src/allocator.rs index 0d8dc93274f..2a95a7368aa 100644 --- a/src/allocator.rs +++ b/src/allocator.rs @@ -99,11 +99,15 @@ fn create_const_value_function( let func = context.new_function(None, FunctionType::Exported, output, &[], name, false); #[cfg(feature = "master")] - func.add_attribute(FnAttribute::Visibility(symbol_visibility_to_gcc( - tcx.sess.default_visibility(), - ))); - - func.add_attribute(FnAttribute::AlwaysInline); + { + func.add_attribute(FnAttribute::Visibility(symbol_visibility_to_gcc( + tcx.sess.default_visibility(), + ))); + + // FIXME(antoyo): cg_llvm sets AlwaysInline, but AlwaysInline is different in GCC and using + // it here will causes linking errors when using LTO. + func.add_attribute(FnAttribute::Inline); + } if tcx.sess.must_emit_unwind_tables() { // TODO(antoyo): emit unwind tables. diff --git a/src/builder.rs b/src/builder.rs index 4aee211e2ef..34ade3d025f 100644 --- a/src/builder.rs +++ b/src/builder.rs @@ -540,9 +540,15 @@ impl<'a, 'gcc, 'tcx> BuilderMethods<'a, 'tcx> for Builder<'a, 'gcc, 'tcx> { fn ret(&mut self, mut value: RValue<'gcc>) { let expected_return_type = self.current_func().get_return_type(); - if !expected_return_type.is_compatible_with(value.get_type()) { - // NOTE: due to opaque pointers now being used, we need to cast here. - value = self.context.new_cast(self.location, value, expected_return_type); + let value_type = value.get_type(); + if !expected_return_type.is_compatible_with(value_type) { + // NOTE: due to opaque pointers now being used, we need to (bit)cast here. + if self.is_native_int_type(value_type) && self.is_native_int_type(expected_return_type) + { + value = self.context.new_cast(self.location, value, expected_return_type); + } else { + value = self.context.new_bitcast(self.location, value, expected_return_type); + } } self.llbb().end_with_return(self.location, value); } @@ -1279,11 +1285,19 @@ impl<'a, 'gcc, 'tcx> BuilderMethods<'a, 'tcx> for Builder<'a, 'gcc, 'tcx> { fn intcast( &mut self, - value: RValue<'gcc>, + mut value: RValue<'gcc>, dest_typ: Type<'gcc>, - _is_signed: bool, + is_signed: bool, ) -> RValue<'gcc> { - // NOTE: is_signed is for value, not dest_typ. + let value_type = value.get_type(); + if is_signed && !value_type.is_signed(self.cx) { + let signed_type = value_type.to_signed(self.cx); + value = self.gcc_int_cast(value, signed_type); + } else if !is_signed && value_type.is_signed(self.cx) { + let unsigned_type = value_type.to_unsigned(self.cx); + value = self.gcc_int_cast(value, unsigned_type); + } + self.gcc_int_cast(value, dest_typ) } diff --git a/src/int.rs b/src/int.rs index 6f21ce9352b..9fb7f6bad68 100644 --- a/src/int.rs +++ b/src/int.rs @@ -4,12 +4,15 @@ // cSpell:words cmpti divti modti mulodi muloti udivti umodti -use gccjit::{BinaryOp, ComparisonOp, FunctionType, Location, RValue, ToRValue, Type, UnaryOp}; +use gccjit::{ + BinaryOp, CType, ComparisonOp, FunctionType, Location, RValue, ToRValue, Type, UnaryOp, +}; use rustc_abi::{CanonAbi, Endian, ExternAbi}; use rustc_codegen_ssa::common::{IntPredicate, TypeKind}; use rustc_codegen_ssa::traits::{BackendTypes, BaseTypeCodegenMethods, BuilderMethods, OverflowOp}; use rustc_middle::ty::{self, Ty}; use rustc_target::callconv::{ArgAbi, ArgAttributes, FnAbi, PassMode}; +use rustc_type_ir::{Interner, TyKind}; use crate::builder::{Builder, ToGccComp}; use crate::common::{SignType, TypeReflection}; @@ -167,9 +170,9 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { if a_type.is_vector() { // Vector types need to be bitcast. // TODO(antoyo): perhaps use __builtin_convertvector for vector casting. - b = self.context.new_bitcast(self.location, b, a.get_type()); + b = self.context.new_bitcast(self.location, b, a_type); } else { - b = self.context.new_cast(self.location, b, a.get_type()); + b = self.context.new_cast(self.location, b, a_type); } } self.context.new_binary_op(self.location, operation, a_type, a, b) @@ -216,13 +219,22 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { operation_name: &str, signed: bool, a: RValue<'gcc>, - b: RValue<'gcc>, + mut b: RValue<'gcc>, ) -> RValue<'gcc> { let a_type = a.get_type(); let b_type = b.get_type(); if (self.is_native_int_type_or_bool(a_type) && self.is_native_int_type_or_bool(b_type)) || (a_type.is_vector() && b_type.is_vector()) { + if !a_type.is_compatible_with(b_type) { + if a_type.is_vector() { + // Vector types need to be bitcast. + // TODO(antoyo): perhaps use __builtin_convertvector for vector casting. + b = self.context.new_bitcast(self.location, b, a_type); + } else { + b = self.context.new_cast(self.location, b, a_type); + } + } self.context.new_binary_op(self.location, operation, a_type, a, b) } else { debug_assert!(a_type.dyncast_array().is_some()); @@ -351,6 +363,11 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { // TODO(antoyo): is it correct to use rhs type instead of the parameter typ? .new_local(self.location, rhs.get_type(), "binopResult") .get_address(self.location); + let new_type = type_kind_to_gcc_type(new_kind); + let new_type = self.context.new_c_type(new_type); + let lhs = self.context.new_cast(self.location, lhs, new_type); + let rhs = self.context.new_cast(self.location, rhs, new_type); + let res = self.context.new_cast(self.location, res, new_type.make_pointer()); let overflow = self.overflow_call(intrinsic, &[lhs, rhs, res], None); (res.dereference(self.location).to_rvalue(), overflow) } @@ -477,11 +494,27 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { let lhs_low = self.context.new_cast(self.location, self.low(lhs), unsigned_type); let rhs_low = self.context.new_cast(self.location, self.low(rhs), unsigned_type); + let mut lhs_high = self.high(lhs); + let mut rhs_high = self.high(rhs); + + match op { + IntPredicate::IntUGT + | IntPredicate::IntUGE + | IntPredicate::IntULT + | IntPredicate::IntULE => { + lhs_high = self.context.new_cast(self.location, lhs_high, unsigned_type); + rhs_high = self.context.new_cast(self.location, rhs_high, unsigned_type); + } + // TODO(antoyo): we probably need to handle signed comparison for unsigned + // integers. + _ => (), + } + let condition = self.context.new_comparison( self.location, ComparisonOp::LessThan, - self.high(lhs), - self.high(rhs), + lhs_high, + rhs_high, ); self.llbb().end_with_conditional(self.location, condition, block1, block2); @@ -495,8 +528,8 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { let condition = self.context.new_comparison( self.location, ComparisonOp::GreaterThan, - self.high(lhs), - self.high(rhs), + lhs_high, + rhs_high, ); block2.end_with_conditional(self.location, condition, block3, block4); @@ -620,7 +653,7 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { } } - pub fn gcc_xor(&self, a: RValue<'gcc>, b: RValue<'gcc>) -> RValue<'gcc> { + pub fn gcc_xor(&self, a: RValue<'gcc>, mut b: RValue<'gcc>) -> RValue<'gcc> { let a_type = a.get_type(); let b_type = b.get_type(); if a_type.is_vector() && b_type.is_vector() { @@ -628,6 +661,9 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { a ^ b } else if self.is_native_int_type_or_bool(a_type) && self.is_native_int_type_or_bool(b_type) { + if !a_type.is_compatible_with(b_type) { + b = self.context.new_cast(self.location, b, a_type); + } a ^ b } else { self.concat_low_high_rvalues( @@ -1042,3 +1078,25 @@ impl<'gcc, 'tcx> CodegenCx<'gcc, 'tcx> { self.context.new_array_constructor(None, typ, &values) } } + +fn type_kind_to_gcc_type(kind: TyKind) -> CType { + use rustc_middle::ty::IntTy::*; + use rustc_middle::ty::UintTy::*; + use rustc_middle::ty::{Int, Uint}; + + match kind { + Int(I8) => CType::Int8t, + Int(I16) => CType::Int16t, + Int(I32) => CType::Int32t, + Int(I64) => CType::Int64t, + Int(I128) => CType::Int128t, + + Uint(U8) => CType::UInt8t, + Uint(U16) => CType::UInt16t, + Uint(U32) => CType::UInt32t, + Uint(U64) => CType::UInt64t, + Uint(U128) => CType::UInt128t, + + _ => unimplemented!("Kind: {:?}", kind), + } +} diff --git a/src/intrinsic/archs.rs b/src/intrinsic/archs.rs index 915ed875e32..d1b2a93243d 100644 --- a/src/intrinsic/archs.rs +++ b/src/intrinsic/archs.rs @@ -95,8 +95,11 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { "cubema" => "__builtin_amdgcn_cubema", "cubesc" => "__builtin_amdgcn_cubesc", "cubetc" => "__builtin_amdgcn_cubetc", + "cvt.f16.bf8" => "__builtin_amdgcn_cvt_f16_bf8", + "cvt.f16.fp8" => "__builtin_amdgcn_cvt_f16_fp8", "cvt.f32.bf8" => "__builtin_amdgcn_cvt_f32_bf8", "cvt.f32.fp8" => "__builtin_amdgcn_cvt_f32_fp8", + "cvt.f32.fp8.e5m3" => "__builtin_amdgcn_cvt_f32_fp8_e5m3", "cvt.off.f32.i4" => "__builtin_amdgcn_cvt_off_f32_i4", "cvt.pk.bf8.f32" => "__builtin_amdgcn_cvt_pk_bf8_f32", "cvt.pk.f16.bf8" => "__builtin_amdgcn_cvt_pk_f16_bf8", @@ -181,6 +184,12 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { "dot4.f32.fp8.bf8" => "__builtin_amdgcn_dot4_f32_fp8_bf8", "dot4.f32.fp8.fp8" => "__builtin_amdgcn_dot4_f32_fp8_fp8", "ds.add.gs.reg.rtn" => "__builtin_amdgcn_ds_add_gs_reg_rtn", + "ds.atomic.async.barrier.arrive.b64" => { + "__builtin_amdgcn_ds_atomic_async_barrier_arrive_b64" + } + "ds.atomic.barrier.arrive.rtn.b64" => { + "__builtin_amdgcn_ds_atomic_barrier_arrive_rtn_b64" + } "ds.bpermute" => "__builtin_amdgcn_ds_bpermute", "ds.bpermute.fi.b32" => "__builtin_amdgcn_ds_bpermute_fi_b32", "ds.gws.barrier" => "__builtin_amdgcn_ds_gws_barrier", @@ -198,8 +207,32 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { "fdot2.f16.f16" => "__builtin_amdgcn_fdot2_f16_f16", "fdot2.f32.bf16" => "__builtin_amdgcn_fdot2_f32_bf16", "fdot2c.f32.bf16" => "__builtin_amdgcn_fdot2c_f32_bf16", + "flat.prefetch" => "__builtin_amdgcn_flat_prefetch", "fmul.legacy" => "__builtin_amdgcn_fmul_legacy", + "global.load.async.to.lds.b128" => { + "__builtin_amdgcn_global_load_async_to_lds_b128" + } + "global.load.async.to.lds.b32" => { + "__builtin_amdgcn_global_load_async_to_lds_b32" + } + "global.load.async.to.lds.b64" => { + "__builtin_amdgcn_global_load_async_to_lds_b64" + } + "global.load.async.to.lds.b8" => "__builtin_amdgcn_global_load_async_to_lds_b8", "global.load.lds" => "__builtin_amdgcn_global_load_lds", + "global.prefetch" => "__builtin_amdgcn_global_prefetch", + "global.store.async.from.lds.b128" => { + "__builtin_amdgcn_global_store_async_from_lds_b128" + } + "global.store.async.from.lds.b32" => { + "__builtin_amdgcn_global_store_async_from_lds_b32" + } + "global.store.async.from.lds.b64" => { + "__builtin_amdgcn_global_store_async_from_lds_b64" + } + "global.store.async.from.lds.b8" => { + "__builtin_amdgcn_global_store_async_from_lds_b8" + } "groupstaticsize" => "__builtin_amdgcn_groupstaticsize", "iglp.opt" => "__builtin_amdgcn_iglp_opt", "implicit.buffer.ptr" => "__builtin_amdgcn_implicit_buffer_ptr", @@ -291,6 +324,7 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { "s.incperflevel" => "__builtin_amdgcn_s_incperflevel", "s.memrealtime" => "__builtin_amdgcn_s_memrealtime", "s.memtime" => "__builtin_amdgcn_s_memtime", + "s.monitor.sleep" => "__builtin_amdgcn_s_monitor_sleep", "s.sendmsg" => "__builtin_amdgcn_s_sendmsg", "s.sendmsghalt" => "__builtin_amdgcn_s_sendmsghalt", "s.setprio" => "__builtin_amdgcn_s_setprio", @@ -300,11 +334,15 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { "s.sleep.var" => "__builtin_amdgcn_s_sleep_var", "s.ttracedata" => "__builtin_amdgcn_s_ttracedata", "s.ttracedata.imm" => "__builtin_amdgcn_s_ttracedata_imm", + "s.wait.asynccnt" => "__builtin_amdgcn_s_wait_asynccnt", "s.wait.event.export.ready" => "__builtin_amdgcn_s_wait_event_export_ready", + "s.wait.tensorcnt" => "__builtin_amdgcn_s_wait_tensorcnt", "s.waitcnt" => "__builtin_amdgcn_s_waitcnt", "sad.hi.u8" => "__builtin_amdgcn_sad_hi_u8", "sad.u16" => "__builtin_amdgcn_sad_u16", "sad.u8" => "__builtin_amdgcn_sad_u8", + "sat.pk4.i4.i8" => "__builtin_amdgcn_sat_pk4_i4_i8", + "sat.pk4.u4.u8" => "__builtin_amdgcn_sat_pk4_u4_u8", "sched.barrier" => "__builtin_amdgcn_sched_barrier", "sched.group.barrier" => "__builtin_amdgcn_sched_group_barrier", "sdot2" => "__builtin_amdgcn_sdot2", @@ -346,8 +384,13 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { "smfmac.i32.16x16x64.i8" => "__builtin_amdgcn_smfmac_i32_16x16x64_i8", "smfmac.i32.32x32x32.i8" => "__builtin_amdgcn_smfmac_i32_32x32x32_i8", "smfmac.i32.32x32x64.i8" => "__builtin_amdgcn_smfmac_i32_32x32x64_i8", + "struct.ptr.buffer.load.lds" => "__builtin_amdgcn_struct_ptr_buffer_load_lds", "sudot4" => "__builtin_amdgcn_sudot4", "sudot8" => "__builtin_amdgcn_sudot8", + "tensor.load.to.lds" => "__builtin_amdgcn_tensor_load_to_lds", + "tensor.load.to.lds.d2" => "__builtin_amdgcn_tensor_load_to_lds_d2", + "tensor.store.from.lds" => "__builtin_amdgcn_tensor_store_from_lds", + "tensor.store.from.lds.d2" => "__builtin_amdgcn_tensor_store_from_lds_d2", "udot2" => "__builtin_amdgcn_udot2", "udot4" => "__builtin_amdgcn_udot4", "udot8" => "__builtin_amdgcn_udot8", @@ -6326,6 +6369,23 @@ fn map_arch_intrinsic(full_name: &str) -> &'static str { } s390(name, full_name) } + "spv" => { + #[allow(non_snake_case)] + fn spv(name: &str, full_name: &str) -> &'static str { + match name { + // spv + "num.subgroups" => "__builtin_spirv_num_subgroups", + "subgroup.id" => "__builtin_spirv_subgroup_id", + "subgroup.local.invocation.id" => { + "__builtin_spirv_subgroup_local_invocation_id" + } + "subgroup.max.size" => "__builtin_spirv_subgroup_max_size", + "subgroup.size" => "__builtin_spirv_subgroup_size", + _ => unimplemented!("***** unsupported LLVM intrinsic {full_name}"), + } + } + spv(name, full_name) + } "ve" => { #[allow(non_snake_case)] fn ve(name: &str, full_name: &str) -> &'static str { diff --git a/src/intrinsic/mod.rs b/src/intrinsic/mod.rs index 0753ac1aeb8..eb0a5336a1f 100644 --- a/src/intrinsic/mod.rs +++ b/src/intrinsic/mod.rs @@ -925,10 +925,17 @@ impl<'a, 'gcc, 'tcx> Builder<'a, 'gcc, 'tcx> { // TODO(antoyo): use width? let arg_type = arg.get_type(); let result_type = self.u32_type; + let arg = if arg_type.is_signed(self.cx) { + let new_type = arg_type.to_unsigned(self.cx); + self.gcc_int_cast(arg, new_type) + } else { + arg + }; + let arg_type = arg.get_type(); let count_leading_zeroes = // TODO(antoyo): write a new function Type::is_compatible_with(&Type) and use it here // instead of using is_uint(). - if arg_type.is_uint(self.cx) { + if arg_type.is_uchar(self.cx) || arg_type.is_ushort(self.cx) || arg_type.is_uint(self.cx) { "__builtin_clz" } else if arg_type.is_ulong(self.cx) { diff --git a/src/intrinsic/simd.rs b/src/intrinsic/simd.rs index 350915a277e..fdc15d580ef 100644 --- a/src/intrinsic/simd.rs +++ b/src/intrinsic/simd.rs @@ -206,6 +206,28 @@ pub fn generic_simd_intrinsic<'a, 'gcc, 'tcx>( ); } + #[cfg(feature = "master")] + if name == sym::simd_funnel_shl { + return Ok(simd_funnel_shift( + bx, + args[0].immediate(), + args[1].immediate(), + args[2].immediate(), + true, + )); + } + + #[cfg(feature = "master")] + if name == sym::simd_funnel_shr { + return Ok(simd_funnel_shift( + bx, + args[0].immediate(), + args[1].immediate(), + args[2].immediate(), + false, + )); + } + if name == sym::simd_bswap { return Ok(simd_bswap(bx, args[0].immediate())); } @@ -1434,3 +1456,62 @@ pub fn generic_simd_intrinsic<'a, 'gcc, 'tcx>( unimplemented!("simd {}", name); } + +#[cfg(feature = "master")] +fn simd_funnel_shift<'a, 'gcc, 'tcx>( + bx: &mut Builder<'a, 'gcc, 'tcx>, + a: RValue<'gcc>, + b: RValue<'gcc>, + shift: RValue<'gcc>, + shift_left: bool, +) -> RValue<'gcc> { + use crate::common::SignType; + + let a_type = a.get_type(); + let vector_type = a_type.unqualified().dyncast_vector().expect("vector type"); + let num_units = vector_type.get_num_units(); + let elem_type = vector_type.get_element_type(); + + let (new_int_type, int_shift_val, int_mask) = if elem_type.is_compatible_with(bx.u8_type) + || elem_type.is_compatible_with(bx.i8_type) + { + (bx.u16_type, 8, u8::MAX as u64) + } else if elem_type.is_compatible_with(bx.u16_type) || elem_type.is_compatible_with(bx.i16_type) + { + (bx.u32_type, 16, u16::MAX as u64) + } else if elem_type.is_compatible_with(bx.u32_type) || elem_type.is_compatible_with(bx.i32_type) + { + (bx.u64_type, 32, u32::MAX as u64) + } else if elem_type.is_compatible_with(bx.u64_type) || elem_type.is_compatible_with(bx.i64_type) + { + (bx.u128_type, 64, u64::MAX) + } else { + unimplemented!("funnel shift on {:?}", elem_type); + }; + + let int_mask = bx.context.new_rvalue_from_long(new_int_type, int_mask as i64); + let int_shift_val = bx.context.new_rvalue_from_int(new_int_type, int_shift_val); + let mut elements = vec![]; + let unsigned_type = elem_type.to_unsigned(bx); + for i in 0..num_units { + let index = bx.context.new_rvalue_from_int(bx.int_type, i as i32); + let a_val = bx.context.new_vector_access(None, a, index).to_rvalue(); + let a_val = bx.context.new_bitcast(None, a_val, unsigned_type); + // TODO: we probably need to use gcc_int_cast instead. + let a_val = bx.gcc_int_cast(a_val, new_int_type); + let b_val = bx.context.new_vector_access(None, b, index).to_rvalue(); + let b_val = bx.context.new_bitcast(None, b_val, unsigned_type); + let b_val = bx.gcc_int_cast(b_val, new_int_type); + let shift_val = bx.context.new_vector_access(None, shift, index).to_rvalue(); + let shift_val = bx.gcc_int_cast(shift_val, new_int_type); + let mut val = a_val << int_shift_val | b_val; + if shift_left { + val = (val << shift_val) >> int_shift_val; + } else { + val = (val >> shift_val) & int_mask; + } + let val = bx.gcc_int_cast(val, elem_type); + elements.push(val); + } + bx.context.new_rvalue_from_vector(None, a_type, &elements) +} diff --git a/src/lib.rs b/src/lib.rs index 613315f77a6..b11f11d38e3 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -50,6 +50,7 @@ extern crate rustc_session; extern crate rustc_span; extern crate rustc_symbol_mangling; extern crate rustc_target; +extern crate rustc_type_ir; // This prevents duplicating functions and statics that are already part of the host rustc process. #[allow(unused_extern_crates)] @@ -362,9 +363,9 @@ impl WriteBackendMethods for GccCodegenBackend { _exported_symbols_for_lto: &[String], each_linked_rlib_for_lto: &[PathBuf], modules: Vec>, - diff_fncs: Vec, + diff_functions: Vec, ) -> Result, FatalError> { - if !diff_fncs.is_empty() { + if !diff_functions.is_empty() { unimplemented!(); } diff --git a/tests/failing-lto-tests.txt b/tests/failing-lto-tests.txt index b9126fb73a7..bf0633f7320 100644 --- a/tests/failing-lto-tests.txt +++ b/tests/failing-lto-tests.txt @@ -28,6 +28,6 @@ tests/ui/macros/macro-comma-behavior-rpass.rs tests/ui/macros/rfc-2011-nicer-assert-messages/assert-with-custom-errors-does-not-create-unnecessary-code.rs tests/ui/macros/rfc-2011-nicer-assert-messages/feature-gate-generic_assert.rs tests/ui/macros/stringify.rs -tests/ui/reexport-test-harness-main.rs tests/ui/rfcs/rfc-1937-termination-trait/termination-trait-in-test.rs tests/ui/binding/fn-arg-incomplete-pattern-drop-order.rs +tests/ui/lto/debuginfo-lto-alloc.rs diff --git a/tests/failing-run-make-tests.txt b/tests/failing-run-make-tests.txt index 842533cd3c6..29032b321fa 100644 --- a/tests/failing-run-make-tests.txt +++ b/tests/failing-run-make-tests.txt @@ -6,7 +6,6 @@ tests/run-make/doctests-keep-binaries/ tests/run-make/doctests-runtool/ tests/run-make/emit-shared-files/ tests/run-make/exit-code/ -tests/run-make/issue-22131/ tests/run-make/issue-64153/ tests/run-make/llvm-ident/ tests/run-make/native-link-modifier-bundle/ diff --git a/tests/failing-ui-tests.txt b/tests/failing-ui-tests.txt index 6979c04d534..41fb4729c07 100644 --- a/tests/failing-ui-tests.txt +++ b/tests/failing-ui-tests.txt @@ -10,11 +10,10 @@ tests/ui/iterators/iter-sum-overflow-overflow-checks.rs tests/ui/mir/mir_drop_order.rs tests/ui/mir/mir_let_chains_drop_order.rs tests/ui/mir/mir_match_guard_let_chains_drop_order.rs -tests/ui/oom_unwind.rs +tests/ui/panics/oom-panic-unwind.rs tests/ui/panic-runtime/abort-link-to-unwinding-crates.rs tests/ui/panic-runtime/abort.rs tests/ui/panic-runtime/link-to-abort.rs -tests/ui/unwind-no-uwtable.rs tests/ui/parser/unclosed-delimiter-in-dep.rs tests/ui/consts/missing_span_in_backtrace.rs tests/ui/drop/dynamic-drop.rs @@ -82,3 +81,8 @@ tests/ui/coroutine/panic-drops.rs tests/ui/coroutine/panic-safe.rs tests/ui/process/nofile-limit.rs tests/ui/simd/intrinsic/generic-arithmetic-pass.rs +tests/ui/linking/no-gc-encapsulation-symbols.rs +tests/ui/panics/unwind-force-no-unwind-tables.rs +tests/ui/attributes/fn-align-dyn.rs +tests/ui/linkage-attr/raw-dylib/elf/glibc-x86_64.rs +tests/ui/explicit-tail-calls/recursion-etc.rs diff --git a/tools/cspell_dicts/rustc_codegen_gcc.txt b/tools/cspell_dicts/rustc_codegen_gcc.txt index 31023e50ffa..4fb018b3ecd 100644 --- a/tools/cspell_dicts/rustc_codegen_gcc.txt +++ b/tools/cspell_dicts/rustc_codegen_gcc.txt @@ -8,6 +8,7 @@ clzll cmse codegened csky +ctfe ctlz ctpop cttz @@ -25,6 +26,7 @@ fwrapv gimple hrtb immediates +interner liblto llbb llcx @@ -47,6 +49,7 @@ mavx mcmodel minimumf minnumf +miri monomorphization monomorphizations monomorphized -- cgit 1.4.1-3-g733a5 From 1c793602ae4183d4b114e62affb4c7035cb130ad Mon Sep 17 00:00:00 2001 From: Antoni Boucher Date: Mon, 4 Aug 2025 09:47:33 -0400 Subject: Reenable m68k CI --- .github/workflows/m68k.yml | 38 ++++++++++++++++---------------------- 1 file changed, 16 insertions(+), 22 deletions(-) diff --git a/.github/workflows/m68k.yml b/.github/workflows/m68k.yml index 759d0d59e26..176ee8628b0 100644 --- a/.github/workflows/m68k.yml +++ b/.github/workflows/m68k.yml @@ -91,11 +91,7 @@ jobs: ./y.sh prepare --only-libcore --cross ./y.sh build --sysroot --features compiler-builtins-no-f16-f128 --target-triple m68k-unknown-linux-gnu ./y.sh test --mini-tests --target-triple m68k-unknown-linux-gnu - # FIXME: since https://github.com/rust-lang/rust/pull/140809, we cannot run programs for architectures not - # supported by the object crate, since this adds a dependency on symbols.o for the panic runtime. - # And as such, a wrong order of the object files in the linker command now fails with an undefined reference - # to some symbols like __rustc::rust_panic. - #CG_GCC_TEST_TARGET=m68k-unknown-linux-gnu ./y.sh test --cargo-tests --target-triple m68k-unknown-linux-gnu + CG_GCC_TEST_TARGET=m68k-unknown-linux-gnu ./y.sh test --cargo-tests --target-triple m68k-unknown-linux-gnu ./y.sh clean all - name: Prepare dependencies @@ -104,23 +100,21 @@ jobs: git config --global user.name "User" ./y.sh prepare --cross - # FIXME: We cannot run programs for architectures not supported by the object crate. See comment above. - #- name: Run tests - #run: | - #./y.sh test --target-triple m68k-unknown-linux-gnu --release --clean --build-sysroot --sysroot-features compiler-builtins-no-f16-f128 ${{ matrix.commands }} - - # FIXME: We cannot run programs for architectures not supported by the object crate. See comment above. - #- name: Run Hello World! - #run: | - #./y.sh build --target-triple m68k-unknown-linux-gnu - - #vm_dir=$(pwd)/vm - #cd tests/hello-world - #CG_RUSTFLAGS="-Clinker=m68k-unknown-linux-gnu-gcc" ../../y.sh cargo build --target m68k-unknown-linux-gnu - #sudo cp target/m68k-unknown-linux-gnu/debug/hello_world $vm_dir/home/ - #sudo chroot $vm_dir qemu-m68k-static /home/hello_world > hello_world_stdout - #expected_output="40" - #test $(cat hello_world_stdout) == $expected_output || (echo "Output differs. Actual output: $(cat hello_world_stdout)"; exit 1) + - name: Run tests + run: | + ./y.sh test --target-triple m68k-unknown-linux-gnu --release --clean --build-sysroot --sysroot-features compiler-builtins-no-f16-f128 ${{ matrix.commands }} + + - name: Run Hello World! + run: | + ./y.sh build --target-triple m68k-unknown-linux-gnu + + vm_dir=$(pwd)/vm + cd tests/hello-world + CG_RUSTFLAGS="-Clinker=m68k-unknown-linux-gnu-gcc" ../../y.sh cargo build --target m68k-unknown-linux-gnu + sudo cp target/m68k-unknown-linux-gnu/debug/hello_world $vm_dir/home/ + sudo chroot $vm_dir qemu-m68k-static /home/hello_world > hello_world_stdout + expected_output="40" + test $(cat hello_world_stdout) == $expected_output || (echo "Output differs. Actual output: $(cat hello_world_stdout)"; exit 1) # Summary job for the merge queue. # ALL THE PREVIOUS JOBS NEED TO BE ADDED TO THE `needs` SECTION OF THIS JOB! -- cgit 1.4.1-3-g733a5 From aa51e18ae6c86e7c44f3d77ea40a2a9c4e967235 Mon Sep 17 00:00:00 2001 From: Antoni Boucher Date: Mon, 4 Aug 2025 10:06:57 -0400 Subject: Remove outdated comment --- src/intrinsic/simd.rs | 1 - 1 file changed, 1 deletion(-) diff --git a/src/intrinsic/simd.rs b/src/intrinsic/simd.rs index fdc15d580ef..41363d6313d 100644 --- a/src/intrinsic/simd.rs +++ b/src/intrinsic/simd.rs @@ -1497,7 +1497,6 @@ fn simd_funnel_shift<'a, 'gcc, 'tcx>( let index = bx.context.new_rvalue_from_int(bx.int_type, i as i32); let a_val = bx.context.new_vector_access(None, a, index).to_rvalue(); let a_val = bx.context.new_bitcast(None, a_val, unsigned_type); - // TODO: we probably need to use gcc_int_cast instead. let a_val = bx.gcc_int_cast(a_val, new_int_type); let b_val = bx.context.new_vector_access(None, b, index).to_rvalue(); let b_val = bx.context.new_bitcast(None, b_val, unsigned_type); -- cgit 1.4.1-3-g733a5 From fa6986ebe749bc59d569e3f75d0b5764f2601d78 Mon Sep 17 00:00:00 2001 From: Esteban Küber Date: Fri, 1 Aug 2025 21:19:32 +0000 Subject: Suggest bounds in more cases, accounting for type parameters referenced in predicate --- .../src/error_reporting/traits/suggestions.rs | 77 +++++++++++++++++++--- compiler/rustc_trait_selection/src/lib.rs | 1 + ...failing_goal_with_repeat_expr_anon_const.stderr | 4 -- .../bound/suggest-complex-bound-on-method.fixed | 23 +++++++ .../bound/suggest-complex-bound-on-method.rs | 23 +++++++ .../bound/suggest-complex-bound-on-method.stderr | 22 +++++++ 6 files changed, 136 insertions(+), 14 deletions(-) create mode 100644 tests/ui/traits/bound/suggest-complex-bound-on-method.fixed create mode 100644 tests/ui/traits/bound/suggest-complex-bound-on-method.rs create mode 100644 tests/ui/traits/bound/suggest-complex-bound-on-method.stderr diff --git a/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs b/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs index c182fd99b17..d6997c8525e 100644 --- a/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs +++ b/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs @@ -31,8 +31,8 @@ use rustc_middle::ty::print::{ }; use rustc_middle::ty::{ self, AdtKind, GenericArgs, InferTy, IsSuggestable, Ty, TyCtxt, TypeFoldable, TypeFolder, - TypeSuperFoldable, TypeVisitableExt, TypeckResults, Upcast, suggest_arbitrary_trait_bound, - suggest_constraining_type_param, + TypeSuperFoldable, TypeSuperVisitable, TypeVisitableExt, TypeVisitor, TypeckResults, Upcast, + suggest_arbitrary_trait_bound, suggest_constraining_type_param, }; use rustc_middle::{bug, span_bug}; use rustc_span::def_id::LocalDefId; @@ -262,6 +262,9 @@ impl<'a, 'tcx> TypeErrCtxt<'a, 'tcx> { _ => (false, None), }; + let mut finder = ParamFinder { .. }; + finder.visit_binder(&trait_pred); + // FIXME: Add check for trait bound that is already present, particularly `?Sized` so we // don't suggest `T: Sized + ?Sized`. loop { @@ -410,6 +413,26 @@ impl<'a, 'tcx> TypeErrCtxt<'a, 'tcx> { } } + hir::Node::TraitItem(hir::TraitItem { + generics, + kind: hir::TraitItemKind::Fn(..), + .. + }) + | hir::Node::ImplItem(hir::ImplItem { + generics, + trait_item_def_id: None, + kind: hir::ImplItemKind::Fn(..), + .. + }) if finder.can_suggest_bound(generics) => { + // Missing generic type parameter bound. + suggest_arbitrary_trait_bound( + self.tcx, + generics, + err, + trait_pred, + associated_ty, + ); + } hir::Node::Item(hir::Item { kind: hir::ItemKind::Struct(_, generics, _) @@ -422,7 +445,7 @@ impl<'a, 'tcx> TypeErrCtxt<'a, 'tcx> { | hir::ItemKind::Const(_, generics, _, _) | hir::ItemKind::TraitAlias(_, generics, _), .. - }) if !param_ty => { + }) if finder.can_suggest_bound(generics) => { // Missing generic type parameter bound. if suggest_arbitrary_trait_bound( self.tcx, @@ -5034,8 +5057,7 @@ impl<'a, 'tcx> TypeErrCtxt<'a, 'tcx> { // Suggesting `T: ?Sized` is only valid in an ADT if `T` is only used in a // borrow. `struct S<'a, T: ?Sized>(&'a T);` is valid, `struct S(T);` // is not. Look for invalid "bare" parameter uses, and suggest using indirection. - let mut visitor = - FindTypeParam { param: param.name.ident().name, invalid_spans: vec![], nested: false }; + let mut visitor = FindTypeParam { param: param.name.ident().name, .. }; visitor.visit_item(item); if visitor.invalid_spans.is_empty() { return false; @@ -5198,7 +5220,7 @@ fn hint_missing_borrow<'tcx>( /// Used to suggest replacing associated types with an explicit type in `where` clauses. #[derive(Debug)] pub struct SelfVisitor<'v> { - pub paths: Vec<&'v hir::Ty<'v>>, + pub paths: Vec<&'v hir::Ty<'v>> = Vec::new(), pub name: Option, } @@ -5568,7 +5590,7 @@ fn point_at_assoc_type_restriction( ); // Search for the associated type `Self::{name}`, get // its type and suggest replacing the bound with it. - let mut visitor = SelfVisitor { paths: vec![], name: Some(name) }; + let mut visitor = SelfVisitor { name: Some(name), .. }; visitor.visit_trait_ref(trait_ref); for path in visitor.paths { err.span_suggestion_verbose( @@ -5579,7 +5601,7 @@ fn point_at_assoc_type_restriction( ); } } else { - let mut visitor = SelfVisitor { paths: vec![], name: None }; + let mut visitor = SelfVisitor { name: None, .. }; visitor.visit_trait_ref(trait_ref); let span: MultiSpan = visitor.paths.iter().map(|p| p.span).collect::>().into(); @@ -5609,8 +5631,8 @@ fn get_deref_type_and_refs(mut ty: Ty<'_>) -> (Ty<'_>, Vec) { /// `param: ?Sized` would be a valid constraint. struct FindTypeParam { param: rustc_span::Symbol, - invalid_spans: Vec, - nested: bool, + invalid_spans: Vec = Vec::new(), + nested: bool = false, } impl<'v> Visitor<'v> for FindTypeParam { @@ -5648,3 +5670,38 @@ impl<'v> Visitor<'v> for FindTypeParam { } } } + +/// Look for type parameters in predicates. We use this to identify whether a bound is suitable in +/// on a given item. +struct ParamFinder { + params: Vec = Vec::new(), +} + +impl<'tcx> TypeVisitor> for ParamFinder { + fn visit_ty(&mut self, t: Ty<'tcx>) -> Self::Result { + match t.kind() { + ty::Param(p) => self.params.push(p.name), + _ => {} + } + t.super_visit_with(self) + } +} + +impl ParamFinder { + /// Whether the `hir::Generics` of the current item can suggest the evaluated bound because its + /// references to type parameters are present in the generics. + fn can_suggest_bound(&self, generics: &hir::Generics<'_>) -> bool { + if self.params.is_empty() { + // There are no references to type parameters at all, so suggesting the bound + // would be reasonable. + return true; + } + generics.params.iter().any(|p| match p.name { + hir::ParamName::Plain(p_name) => { + // All of the parameters in the bound can be referenced in the current item. + self.params.iter().any(|p| *p == p_name.name || *p == kw::SelfUpper) + } + _ => true, + }) + } +} diff --git a/compiler/rustc_trait_selection/src/lib.rs b/compiler/rustc_trait_selection/src/lib.rs index e2b22f7bab7..fc0cf8f140a 100644 --- a/compiler/rustc_trait_selection/src/lib.rs +++ b/compiler/rustc_trait_selection/src/lib.rs @@ -19,6 +19,7 @@ #![feature(assert_matches)] #![feature(associated_type_defaults)] #![feature(box_patterns)] +#![feature(default_field_values)] #![feature(if_let_guard)] #![feature(iter_intersperse)] #![feature(iterator_try_reduce)] diff --git a/tests/ui/const-generics/failing_goal_with_repeat_expr_anon_const.stderr b/tests/ui/const-generics/failing_goal_with_repeat_expr_anon_const.stderr index 12de4f1dc30..cf5b55a0ce2 100644 --- a/tests/ui/const-generics/failing_goal_with_repeat_expr_anon_const.stderr +++ b/tests/ui/const-generics/failing_goal_with_repeat_expr_anon_const.stderr @@ -21,10 +21,6 @@ LL | [0u8; std::mem::size_of::()] == Self::P; | ^^ no implementation for `[u8; std::mem::size_of::()] == ::A` | = help: the trait `PartialEq<::A>` is not implemented for `[u8; std::mem::size_of::()]` -help: consider introducing a `where` clause, but there might be an alternative better way to express this requirement - | -LL | pub trait T where [u8; std::mem::size_of::()]: PartialEq<::A> { - | +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ error: aborting due to 3 previous errors diff --git a/tests/ui/traits/bound/suggest-complex-bound-on-method.fixed b/tests/ui/traits/bound/suggest-complex-bound-on-method.fixed new file mode 100644 index 00000000000..719835af0a9 --- /dev/null +++ b/tests/ui/traits/bound/suggest-complex-bound-on-method.fixed @@ -0,0 +1,23 @@ +//@ run-rustfix +#![allow(dead_code)] +struct Application; +// https://github.com/rust-lang/rust/issues/144734 +trait Trait { + type Error: std::error::Error; + + fn run(&self) -> Result<(), Self::Error>; +} + +#[derive(Debug)] +enum ApplicationError { + Quit, +} + +impl Application { + fn thing(&self, t: T) -> Result<(), ApplicationError> where ApplicationError: From<::Error> { + t.run()?; //~ ERROR E0277 + Ok(()) + } +} + +fn main() {} diff --git a/tests/ui/traits/bound/suggest-complex-bound-on-method.rs b/tests/ui/traits/bound/suggest-complex-bound-on-method.rs new file mode 100644 index 00000000000..39cec571242 --- /dev/null +++ b/tests/ui/traits/bound/suggest-complex-bound-on-method.rs @@ -0,0 +1,23 @@ +//@ run-rustfix +#![allow(dead_code)] +struct Application; +// https://github.com/rust-lang/rust/issues/144734 +trait Trait { + type Error: std::error::Error; + + fn run(&self) -> Result<(), Self::Error>; +} + +#[derive(Debug)] +enum ApplicationError { + Quit, +} + +impl Application { + fn thing(&self, t: T) -> Result<(), ApplicationError> { + t.run()?; //~ ERROR E0277 + Ok(()) + } +} + +fn main() {} diff --git a/tests/ui/traits/bound/suggest-complex-bound-on-method.stderr b/tests/ui/traits/bound/suggest-complex-bound-on-method.stderr new file mode 100644 index 00000000000..ac45f049541 --- /dev/null +++ b/tests/ui/traits/bound/suggest-complex-bound-on-method.stderr @@ -0,0 +1,22 @@ +error[E0277]: `?` couldn't convert the error to `ApplicationError` + --> $DIR/suggest-complex-bound-on-method.rs:18:16 + | +LL | t.run()?; + | -----^ the trait `From<::Error>` is not implemented for `ApplicationError` + | | + | this can't be annotated with `?` because it has type `Result<_, ::Error>` + | +note: `ApplicationError` needs to implement `From<::Error>` + --> $DIR/suggest-complex-bound-on-method.rs:12:1 + | +LL | enum ApplicationError { + | ^^^^^^^^^^^^^^^^^^^^^ + = note: the question mark operation (`?`) implicitly performs a conversion on the error value using the `From` trait +help: consider introducing a `where` clause, but there might be an alternative better way to express this requirement + | +LL | fn thing(&self, t: T) -> Result<(), ApplicationError> where ApplicationError: From<::Error> { + | +++++++++++++++++++++++++++++++++++++++++++++++++ + +error: aborting due to 1 previous error + +For more information about this error, try `rustc --explain E0277`. -- cgit 1.4.1-3-g733a5 From 9916ce362f712b7aa91cd7576a5dcfc575ffd621 Mon Sep 17 00:00:00 2001 From: Folkert de Vries Date: Fri, 1 Aug 2025 15:31:16 +0200 Subject: fix `#[loop_match]` on diverging loop this generated invalid MIR before --- compiler/rustc_mir_build/src/builder/expr/into.rs | 2 +- tests/ui/loop-match/diverges.rs | 44 +++++++++++++++++++++++ 2 files changed, 45 insertions(+), 1 deletion(-) create mode 100644 tests/ui/loop-match/diverges.rs diff --git a/compiler/rustc_mir_build/src/builder/expr/into.rs b/compiler/rustc_mir_build/src/builder/expr/into.rs index 82b883a99a1..ac87f671699 100644 --- a/compiler/rustc_mir_build/src/builder/expr/into.rs +++ b/compiler/rustc_mir_build/src/builder/expr/into.rs @@ -345,7 +345,7 @@ impl<'a, 'tcx> Builder<'a, 'tcx> { expr_span, |this| { this.lower_match_arms( - destination, + state_place, scrutinee_place_builder, scrutinee_span, arms, diff --git a/tests/ui/loop-match/diverges.rs b/tests/ui/loop-match/diverges.rs new file mode 100644 index 00000000000..f1b3ffb2076 --- /dev/null +++ b/tests/ui/loop-match/diverges.rs @@ -0,0 +1,44 @@ +//@ build-pass +//@ compile-flags: -Zvalidate-mir +#![allow(incomplete_features)] +#![feature(loop_match)] +#![crate_type = "lib"] + +// Test that a #[loop_match] without an explicit break from the loop generates valid MIR. + +fn break_to_block_unit() -> u8 { + let mut state = 0; + #[loop_match] + loop { + state = 'blk: { + match state { + _ => 'b: { + break 'b 2; + } + } + } + } +} + +fn break_to_block_value() -> u8 { + let mut state = 0u8; + #[loop_match] + 'a: loop { + state = 'blk: { + match state { + _ => break 'blk state, + } + } + } +} + +fn infinite_a(mut state: u8) { + #[loop_match] + loop { + state = 'blk: { + match state { + a => a, + } + } + } +} -- cgit 1.4.1-3-g733a5 From dccc075d05f0c28fab5775f7efa5550f24382ec6 Mon Sep 17 00:00:00 2001 From: Camille GILLOT Date: Mon, 4 Aug 2025 23:47:40 +0000 Subject: Check for rustc_has_incoherent_inherent_impls in incoherent_impls query. --- compiler/rustc_hir_typeck/src/method/probe.rs | 21 +++------------------ compiler/rustc_middle/src/ty/trait_def.rs | 7 +++++++ 2 files changed, 10 insertions(+), 18 deletions(-) diff --git a/compiler/rustc_hir_typeck/src/method/probe.rs b/compiler/rustc_hir_typeck/src/method/probe.rs index f5584e0596a..51de661a2bb 100644 --- a/compiler/rustc_hir_typeck/src/method/probe.rs +++ b/compiler/rustc_hir_typeck/src/method/probe.rs @@ -777,31 +777,16 @@ impl<'a, 'tcx> ProbeContext<'a, 'tcx> { self.assemble_inherent_candidates_from_object(generalized_self_ty); self.assemble_inherent_impl_candidates_for_type(p.def_id(), receiver_steps); - if self.tcx.has_attr(p.def_id(), sym::rustc_has_incoherent_inherent_impls) { - self.assemble_inherent_candidates_for_incoherent_ty( - raw_self_ty, - receiver_steps, - ); - } + self.assemble_inherent_candidates_for_incoherent_ty(raw_self_ty, receiver_steps); } ty::Adt(def, _) => { let def_id = def.did(); self.assemble_inherent_impl_candidates_for_type(def_id, receiver_steps); - if self.tcx.has_attr(def_id, sym::rustc_has_incoherent_inherent_impls) { - self.assemble_inherent_candidates_for_incoherent_ty( - raw_self_ty, - receiver_steps, - ); - } + self.assemble_inherent_candidates_for_incoherent_ty(raw_self_ty, receiver_steps); } ty::Foreign(did) => { self.assemble_inherent_impl_candidates_for_type(did, receiver_steps); - if self.tcx.has_attr(did, sym::rustc_has_incoherent_inherent_impls) { - self.assemble_inherent_candidates_for_incoherent_ty( - raw_self_ty, - receiver_steps, - ); - } + self.assemble_inherent_candidates_for_incoherent_ty(raw_self_ty, receiver_steps); } ty::Param(_) => { self.assemble_inherent_candidates_from_param(raw_self_ty); diff --git a/compiler/rustc_middle/src/ty/trait_def.rs b/compiler/rustc_middle/src/ty/trait_def.rs index 59e2b2a034d..4e38d969192 100644 --- a/compiler/rustc_middle/src/ty/trait_def.rs +++ b/compiler/rustc_middle/src/ty/trait_def.rs @@ -6,6 +6,7 @@ use rustc_hir as hir; use rustc_hir::def::DefKind; use rustc_hir::def_id::{DefId, LOCAL_CRATE}; use rustc_macros::{Decodable, Encodable, HashStable}; +use rustc_span::symbol::sym; use tracing::debug; use crate::query::LocalCrate; @@ -239,6 +240,12 @@ pub(super) fn trait_impls_of_provider(tcx: TyCtxt<'_>, trait_id: DefId) -> Trait /// Query provider for `incoherent_impls`. pub(super) fn incoherent_impls_provider(tcx: TyCtxt<'_>, simp: SimplifiedType) -> &[DefId] { + if let Some(def_id) = simp.def() + && !tcx.has_attr(def_id, sym::rustc_has_incoherent_inherent_impls) + { + return &[]; + } + let mut impls = Vec::new(); for cnum in iter::once(LOCAL_CRATE).chain(tcx.crates(()).iter().copied()) { for &impl_def_id in tcx.crate_incoherent_impls((cnum, simp)) { -- cgit 1.4.1-3-g733a5 From 5073886181cb95bfe899a26404c57d2944f2a4d4 Mon Sep 17 00:00:00 2001 From: Camille GILLOT Date: Sat, 2 Aug 2025 18:21:16 +0000 Subject: Gate checking rustc attrs on the feature. --- compiler/rustc_hir_typeck/src/callee.rs | 6 ++++-- compiler/rustc_hir_typeck/src/fn_ctxt/mod.rs | 16 ++++++++++++---- compiler/rustc_hir_typeck/src/upvar.rs | 2 +- compiler/rustc_hir_typeck/src/writeback.rs | 3 ++- 4 files changed, 19 insertions(+), 8 deletions(-) diff --git a/compiler/rustc_hir_typeck/src/callee.rs b/compiler/rustc_hir_typeck/src/callee.rs index 48bb45de53e..e19dd83f81d 100644 --- a/compiler/rustc_hir_typeck/src/callee.rs +++ b/compiler/rustc_hir_typeck/src/callee.rs @@ -511,7 +511,9 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { // Untranslatable diagnostics are okay for rustc internals #[allow(rustc::untranslatable_diagnostic)] #[allow(rustc::diagnostic_outside_of_impl)] - if self.tcx.has_attr(def_id, sym::rustc_evaluate_where_clauses) { + if self.has_rustc_attrs + && self.tcx.has_attr(def_id, sym::rustc_evaluate_where_clauses) + { let predicates = self.tcx.predicates_of(def_id); let predicates = predicates.instantiate(self.tcx, args); for (predicate, predicate_span) in predicates { @@ -894,7 +896,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { } // If we have `rustc_do_not_const_check`, do not check `[const]` bounds. - if self.tcx.has_attr(self.body_id, sym::rustc_do_not_const_check) { + if self.has_rustc_attrs && self.tcx.has_attr(self.body_id, sym::rustc_do_not_const_check) { return; } diff --git a/compiler/rustc_hir_typeck/src/fn_ctxt/mod.rs b/compiler/rustc_hir_typeck/src/fn_ctxt/mod.rs index 0c6226ce71e..74f27e85cba 100644 --- a/compiler/rustc_hir_typeck/src/fn_ctxt/mod.rs +++ b/compiler/rustc_hir_typeck/src/fn_ctxt/mod.rs @@ -126,6 +126,10 @@ pub(crate) struct FnCtxt<'a, 'tcx> { /// These are stored here so we may collect them when canonicalizing user /// type ascriptions later. pub(super) trait_ascriptions: RefCell>>>, + + /// Whether the current crate enables the `rustc_attrs` feature. + /// This allows to skip processing attributes in many places. + pub(super) has_rustc_attrs: bool, } impl<'a, 'tcx> FnCtxt<'a, 'tcx> { @@ -154,6 +158,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { diverging_fallback_behavior, diverging_block_behavior, trait_ascriptions: Default::default(), + has_rustc_attrs: root_ctxt.tcx.features().rustc_attrs(), } } @@ -525,10 +530,13 @@ fn parse_never_type_options_attr( let mut fallback = None; let mut block = None; - let items = tcx - .get_attr(CRATE_DEF_ID, sym::rustc_never_type_options) - .map(|attr| attr.meta_item_list().unwrap()) - .unwrap_or_default(); + let items = if tcx.features().rustc_attrs() { + tcx.get_attr(CRATE_DEF_ID, sym::rustc_never_type_options) + .map(|attr| attr.meta_item_list().unwrap()) + } else { + None + }; + let items = items.unwrap_or_default(); for item in items { if item.has_name(sym::fallback) && fallback.is_none() { diff --git a/compiler/rustc_hir_typeck/src/upvar.rs b/compiler/rustc_hir_typeck/src/upvar.rs index df38c3a1214..8c0ec78e0af 100644 --- a/compiler/rustc_hir_typeck/src/upvar.rs +++ b/compiler/rustc_hir_typeck/src/upvar.rs @@ -1747,7 +1747,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { } fn should_log_capture_analysis(&self, closure_def_id: LocalDefId) -> bool { - self.tcx.has_attr(closure_def_id, sym::rustc_capture_analysis) + self.has_rustc_attrs && self.tcx.has_attr(closure_def_id, sym::rustc_capture_analysis) } fn log_capture_analysis_first_pass( diff --git a/compiler/rustc_hir_typeck/src/writeback.rs b/compiler/rustc_hir_typeck/src/writeback.rs index 093de950d63..574406ecc8a 100644 --- a/compiler/rustc_hir_typeck/src/writeback.rs +++ b/compiler/rustc_hir_typeck/src/writeback.rs @@ -45,7 +45,8 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> { // This attribute causes us to dump some writeback information // in the form of errors, which is used for unit tests. - let rustc_dump_user_args = self.tcx.has_attr(item_def_id, sym::rustc_dump_user_args); + let rustc_dump_user_args = + self.has_rustc_attrs && self.tcx.has_attr(item_def_id, sym::rustc_dump_user_args); let mut wbcx = WritebackCx::new(self, body, rustc_dump_user_args); for param in body.params { -- cgit 1.4.1-3-g733a5 From 6b7f8fbcc584f4dc46081b5c690aeca7def88019 Mon Sep 17 00:00:00 2001 From: Camille GILLOT Date: Tue, 5 Aug 2025 01:55:20 +0000 Subject: Promote crash test. --- tests/crashes/124352.rs | 4 ---- tests/ui/attributes/malformed-never-type-options.rs | 8 ++++++++ .../ui/attributes/malformed-never-type-options.stderr | 19 +++++++++++++++++++ 3 files changed, 27 insertions(+), 4 deletions(-) delete mode 100644 tests/crashes/124352.rs create mode 100644 tests/ui/attributes/malformed-never-type-options.rs create mode 100644 tests/ui/attributes/malformed-never-type-options.stderr diff --git a/tests/crashes/124352.rs b/tests/crashes/124352.rs deleted file mode 100644 index e9eb4419e6a..00000000000 --- a/tests/crashes/124352.rs +++ /dev/null @@ -1,4 +0,0 @@ -//@ known-bug: #124352 -#![rustc_never_type_options(: Unsize = "hi")] - -fn main() {} diff --git a/tests/ui/attributes/malformed-never-type-options.rs b/tests/ui/attributes/malformed-never-type-options.rs new file mode 100644 index 00000000000..0c384be0e22 --- /dev/null +++ b/tests/ui/attributes/malformed-never-type-options.rs @@ -0,0 +1,8 @@ +//! Regression test for #124352 +//! The `rustc_*` attribute is malformed, but ICEing without a `feature(rustc_attrs)` is still bad. + +#![rustc_never_type_options(: Unsize = "hi")] +//~^ ERROR expected unsuffixed literal, found `:` +//~| ERROR use of an internal attribute + +fn main() {} diff --git a/tests/ui/attributes/malformed-never-type-options.stderr b/tests/ui/attributes/malformed-never-type-options.stderr new file mode 100644 index 00000000000..0d2ff4881f2 --- /dev/null +++ b/tests/ui/attributes/malformed-never-type-options.stderr @@ -0,0 +1,19 @@ +error: expected unsuffixed literal, found `:` + --> $DIR/malformed-never-type-options.rs:4:29 + | +LL | #![rustc_never_type_options(: Unsize = "hi")] + | ^ + +error[E0658]: use of an internal attribute + --> $DIR/malformed-never-type-options.rs:4:1 + | +LL | #![rustc_never_type_options(: Unsize = "hi")] + | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + | + = help: add `#![feature(rustc_attrs)]` to the crate attributes to enable + = note: the `#[rustc_never_type_options]` attribute is an internal implementation detail that will never be stable + = note: `rustc_never_type_options` is used to experiment with never type fallback and work on never type stabilization + +error: aborting due to 2 previous errors + +For more information about this error, try `rustc --explain E0658`. -- cgit 1.4.1-3-g733a5 From 2136c500cdbdd08fe7c25927b5d3ce790d75f68b Mon Sep 17 00:00:00 2001 From: Sebastian Poeplau Date: Fri, 1 Aug 2025 12:39:50 +0200 Subject: Preserve the .debug_gdb_scripts section Make sure that compiler and linker don't optimize the section's contents away by adding the global holding the data to "llvm.used". The volatile load in the main shim is retained because "llvm.used", which translates to SHF_GNU_RETAIN on ELF targets, requires a reasonably recent linker; emitting the volatile load ensures compatibility with older linkers, at least when libstd is used. Pretty printers in dylib dependencies are now emitted by the main crate instead of the dylib; apart from matching how rlibs are handled, this approach has the advantage that `omit_gdb_pretty_printer_section` keeps working with dylib dependencies. --- src/debuginfo.rs | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/debuginfo.rs b/src/debuginfo.rs index 4c8585192a1..4c0b6439523 100644 --- a/src/debuginfo.rs +++ b/src/debuginfo.rs @@ -254,7 +254,8 @@ impl<'gcc, 'tcx> DebugInfoCodegenMethods<'tcx> for CodegenCx<'gcc, 'tcx> { // TODO(antoyo): implement. } - fn debuginfo_finalize(&self) { + fn debuginfo_finalize(&mut self) { + // TODO: emit section `.debug_gdb_scripts`. self.context.set_debug_info(true) } -- cgit 1.4.1-3-g733a5 From 82aac02c5267a427fdc5e8e1abbba98fc11c832c Mon Sep 17 00:00:00 2001 From: bjorn3 <17426603+bjorn3@users.noreply.github.com> Date: Wed, 6 Aug 2025 18:01:07 +0000 Subject: Revert "Preserve the .debug_gdb_scripts section" This reverts commit 868bdde25b030e0b71a29a5dbc04a891036e702e. --- src/debuginfo.rs | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/debuginfo.rs b/src/debuginfo.rs index 4c0b6439523..4c8585192a1 100644 --- a/src/debuginfo.rs +++ b/src/debuginfo.rs @@ -254,8 +254,7 @@ impl<'gcc, 'tcx> DebugInfoCodegenMethods<'tcx> for CodegenCx<'gcc, 'tcx> { // TODO(antoyo): implement. } - fn debuginfo_finalize(&mut self) { - // TODO: emit section `.debug_gdb_scripts`. + fn debuginfo_finalize(&self) { self.context.set_debug_info(true) } -- cgit 1.4.1-3-g733a5 From decdb950bf016391145b582cd0756d3e67481ad6 Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Thu, 7 Aug 2025 01:05:01 -0500 Subject: symcheck: Store the section name in `SymInfo` if available Currently `SymInfo` stores a `Section`, which is just an index: SymInfo { section: Section( SectionIndex( 539, ), ), ... }, Look up and store the section name instead if possible, with a fallback to the `Section` debug printing. This makes output more clear and will allow us to filter by section name. --- .../crates/symbol-check/src/main.rs | 30 ++++++++++++++-------- 1 file changed, 20 insertions(+), 10 deletions(-) diff --git a/library/compiler-builtins/crates/symbol-check/src/main.rs b/library/compiler-builtins/crates/symbol-check/src/main.rs index 1312a717970..beb568a0f84 100644 --- a/library/compiler-builtins/crates/symbol-check/src/main.rs +++ b/library/compiler-builtins/crates/symbol-check/src/main.rs @@ -9,7 +9,7 @@ use std::process::{Command, Stdio}; use object::read::archive::{ArchiveFile, ArchiveMember}; use object::{ - File as ObjFile, Object, ObjectSymbol, Symbol, SymbolKind, SymbolScope, SymbolSection, + File as ObjFile, Object, ObjectSection, ObjectSymbol, Symbol, SymbolKind, SymbolScope, }; use serde_json::Value; @@ -154,7 +154,7 @@ struct SymInfo { name: String, kind: SymbolKind, scope: SymbolScope, - section: SymbolSection, + section: String, is_undefined: bool, is_global: bool, is_local: bool, @@ -165,12 +165,22 @@ struct SymInfo { } impl SymInfo { - fn new(sym: &Symbol, member: &ArchiveMember) -> Self { + fn new(sym: &Symbol, obj: &ObjFile, member: &ArchiveMember) -> Self { + // Include the section name if possible. Fall back to the `Section` debug impl if not. + let section = sym.section(); + let section_name = sym + .section() + .index() + .and_then(|idx| obj.section_by_index(idx).ok()) + .and_then(|sec| sec.name().ok()) + .map(ToString::to_string) + .unwrap_or_else(|| format!("{section:?}")); + Self { name: sym.name().expect("missing name").to_owned(), kind: sym.kind(), scope: sym.scope(), - section: sym.section(), + section: section_name, is_undefined: sym.is_undefined(), is_global: sym.is_global(), is_local: sym.is_local(), @@ -192,13 +202,13 @@ fn verify_no_duplicates(archive: &Archive) { let mut dups = Vec::new(); let mut found_any = false; - archive.for_each_symbol(|symbol, member| { + archive.for_each_symbol(|symbol, obj, member| { // Only check defined globals if !symbol.is_global() || symbol.is_undefined() { return; } - let sym = SymInfo::new(&symbol, member); + let sym = SymInfo::new(&symbol, obj, member); // x86-32 includes multiple copies of thunk symbols if sym.name.starts_with("__x86.get_pc_thunk") { @@ -244,7 +254,7 @@ fn verify_core_symbols(archive: &Archive) { let mut undefined = Vec::new(); let mut has_symbols = false; - archive.for_each_symbol(|symbol, member| { + archive.for_each_symbol(|symbol, obj, member| { has_symbols = true; // Find only symbols from `core` @@ -252,7 +262,7 @@ fn verify_core_symbols(archive: &Archive) { return; } - let sym = SymInfo::new(&symbol, member); + let sym = SymInfo::new(&symbol, obj, member); if sym.is_undefined { undefined.push(sym); } else { @@ -304,9 +314,9 @@ impl Archive { } /// For a given archive, do something with each symbol. - fn for_each_symbol(&self, mut f: impl FnMut(Symbol, &ArchiveMember)) { + fn for_each_symbol(&self, mut f: impl FnMut(Symbol, &ObjFile, &ArchiveMember)) { self.for_each_object(|obj, member| { - obj.symbols().for_each(|sym| f(sym, member)); + obj.symbols().for_each(|sym| f(sym, &obj, member)); }); } } -- cgit 1.4.1-3-g733a5 From ab0def2739744ae31d5738263acd0cc0e00e1a16 Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Thu, 7 Aug 2025 03:27:16 -0500 Subject: symcheck: Ignore symbols in `.debug_gdb_scripts` Since [1], our object files may now contain a GDB script section. These symbols wind up with multiple instances in the archive but are weak, so we can safely ignore them in our duplicates check. This resolves the current CI failures. [1]: https://github.com/rust-lang/rust/pull/143679 --- library/compiler-builtins/crates/symbol-check/src/main.rs | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/library/compiler-builtins/crates/symbol-check/src/main.rs b/library/compiler-builtins/crates/symbol-check/src/main.rs index beb568a0f84..129c959f23a 100644 --- a/library/compiler-builtins/crates/symbol-check/src/main.rs +++ b/library/compiler-builtins/crates/symbol-check/src/main.rs @@ -215,6 +215,11 @@ fn verify_no_duplicates(archive: &Archive) { return; } + // GDB pretty printing symbols may show up more than once but are weak. + if sym.section == ".debug_gdb_scripts" && sym.is_weak { + return; + } + // Windows has symbols for literal numeric constants, string literals, and MinGW pseudo- // relocations. These are allowed to have repeated definitions. let win_allowed_dup_pfx = ["__real@", "__xmm@", "??_C@_", ".refptr"]; -- cgit 1.4.1-3-g733a5 From 098d06b0300e7ec3d4174904c2d6a9d4444a30a2 Mon Sep 17 00:00:00 2001 From: bjorn3 <17426603+bjorn3@users.noreply.github.com> Date: Thu, 7 Aug 2025 13:13:04 +0000 Subject: Prevent name collisions with internal implementation details The implementation of the linkage attribute inside extern blocks defines symbols starting with _rust_extern_with_linkage_. If someone tries to also define this symbol you will get a symbol conflict or even an ICE. By adding an unpredictable component to the symbol name, this becomes less of an issue. --- src/consts.rs | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/consts.rs b/src/consts.rs index c04c75e1b11..873f1f1951c 100644 --- a/src/consts.rs +++ b/src/consts.rs @@ -6,6 +6,7 @@ use rustc_codegen_ssa::traits::{ BaseTypeCodegenMethods, ConstCodegenMethods, StaticCodegenMethods, }; use rustc_hir::def::DefKind; +use rustc_hir::def_id::LOCAL_CRATE; use rustc_middle::middle::codegen_fn_attrs::{CodegenFnAttrFlags, CodegenFnAttrs}; use rustc_middle::mir::interpret::{ self, ConstAllocation, ErrorHandled, Scalar as InterpScalar, read_target_uint, @@ -384,8 +385,8 @@ fn check_and_apply_linkage<'gcc, 'tcx>( // linkage and there are no definitions), then // `extern_with_linkage_foo` will instead be initialized to // zero. - let mut real_name = "_rust_extern_with_linkage_".to_string(); - real_name.push_str(sym); + let real_name = + format!("_rust_extern_with_linkage_{:016x}_{sym}", cx.tcx.stable_crate_id(LOCAL_CRATE)); let global2 = cx.define_global(&real_name, gcc_type, is_tls, attrs.link_section); // TODO(antoyo): set linkage. let value = cx.const_ptrcast(global1.get_address(None), gcc_type); -- cgit 1.4.1-3-g733a5 From 36ec648e0a951425a7dcee0c742a60df19a37939 Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Thu, 7 Aug 2025 03:42:02 -0500 Subject: Remove instances of `allow(improper_ctypes)` i128/u128 haven't flagged `improper_ctypes` for a while, and this just made it to stable [1]. Remove the `allow`s as they are no longer needed. [1]: https://blog.rust-lang.org/2025/08/07/Rust-1.89.0/#i128-and-u128-in-extern-c-functions --- library/compiler-builtins/builtins-test/benches/float_conv.rs | 1 - library/compiler-builtins/compiler-builtins/src/lib.rs | 4 ---- 2 files changed, 5 deletions(-) diff --git a/library/compiler-builtins/builtins-test/benches/float_conv.rs b/library/compiler-builtins/builtins-test/benches/float_conv.rs index e0f488eb685..40c13d270ac 100644 --- a/library/compiler-builtins/builtins-test/benches/float_conv.rs +++ b/library/compiler-builtins/builtins-test/benches/float_conv.rs @@ -1,4 +1,3 @@ -#![allow(improper_ctypes)] #![cfg_attr(f128_enabled, feature(f128))] use builtins_test::float_bench; diff --git a/library/compiler-builtins/compiler-builtins/src/lib.rs b/library/compiler-builtins/compiler-builtins/src/lib.rs index ca75f44e02a..b111dc0bd18 100644 --- a/library/compiler-builtins/compiler-builtins/src/lib.rs +++ b/library/compiler-builtins/compiler-builtins/src/lib.rs @@ -18,10 +18,6 @@ #![no_std] #![allow(unused_features)] #![allow(internal_features)] -// We use `u128` in a whole bunch of places which we currently agree with the -// compiler on ABIs and such, so we should be "good enough" for now and changes -// to the `u128` ABI will be reflected here. -#![allow(improper_ctypes, improper_ctypes_definitions)] // `mem::swap` cannot be used because it may generate references to memcpy in unoptimized code. #![allow(clippy::manual_swap)] // Support compiling on both stage0 and stage1 which may differ in supported stable features. -- cgit 1.4.1-3-g733a5 From 9a81b2a6a0932547c8e949a34d61cf789dbee675 Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Thu, 7 Aug 2025 15:45:52 -0500 Subject: Start runnning tests for aarch64-pc-windows-msvc This target is currently build-only. Switch to the windows-11-arm runner, which allows us to start running tests. --- library/compiler-builtins/.github/workflows/main.yaml | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/library/compiler-builtins/.github/workflows/main.yaml b/library/compiler-builtins/.github/workflows/main.yaml index c54df2e90b7..3afadbfe894 100644 --- a/library/compiler-builtins/.github/workflows/main.yaml +++ b/library/compiler-builtins/.github/workflows/main.yaml @@ -51,8 +51,7 @@ jobs: - target: aarch64-unknown-linux-gnu os: ubuntu-24.04-arm - target: aarch64-pc-windows-msvc - os: windows-2025 - build_only: 1 + os: windows-11-arm - target: arm-unknown-linux-gnueabi os: ubuntu-24.04 - target: arm-unknown-linux-gnueabihf -- cgit 1.4.1-3-g733a5 From d47b5e4ab63df3774727e6dec542dade6ef22d9a Mon Sep 17 00:00:00 2001 From: minxuanz Date: Fri, 8 Aug 2025 11:19:04 +0800 Subject: Fix wrong cache line size of riscv64 --- library/std/src/sync/mpmc/utils.rs | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/library/std/src/sync/mpmc/utils.rs b/library/std/src/sync/mpmc/utils.rs index 0cbc61160f7..e3bcb149f64 100644 --- a/library/std/src/sync/mpmc/utils.rs +++ b/library/std/src/sync/mpmc/utils.rs @@ -23,14 +23,13 @@ use crate::ops::{Deref, DerefMut}; any(target_arch = "x86_64", target_arch = "aarch64", target_arch = "powerpc64",), repr(align(128)) )] -// arm, mips, mips64, and riscv64 have 32-byte cache line size. +// arm, mips and mips64 have 32-byte cache line size. // // Sources: // - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_arm.go#L7 // - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_mips.go#L7 // - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_mipsle.go#L7 // - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_mips64x.go#L9 -// - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_riscv64.go#L7 #[cfg_attr( any( target_arch = "arm", @@ -38,7 +37,6 @@ use crate::ops::{Deref, DerefMut}; target_arch = "mips32r6", target_arch = "mips64", target_arch = "mips64r6", - target_arch = "riscv64", ), repr(align(32)) )] @@ -47,11 +45,12 @@ use crate::ops::{Deref, DerefMut}; // Sources: // - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_s390x.go#L7 #[cfg_attr(target_arch = "s390x", repr(align(256)))] -// x86 and wasm have 64-byte cache line size. +// x86, wasm and riscv have 64-byte cache line size. // // Sources: // - https://github.com/golang/go/blob/dda2991c2ea0c5914714469c4defc2562a907230/src/internal/cpu/cpu_x86.go#L9 // - https://github.com/golang/go/blob/3dd58676054223962cd915bb0934d1f9f489d4d2/src/internal/cpu/cpu_wasm.go#L7 +// - https://github.com/golang/go/blob/5e31f78c8a4ed1b872ddc194f0cd1ae931b37d7e/src/internal/cpu/cpu_riscv64.go#L7 // // All others are assumed to have 64-byte cache line size. #[cfg_attr( @@ -64,7 +63,6 @@ use crate::ops::{Deref, DerefMut}; target_arch = "mips32r6", target_arch = "mips64", target_arch = "mips64r6", - target_arch = "riscv64", target_arch = "s390x", )), repr(align(64)) -- cgit 1.4.1-3-g733a5 From e22a140c4596a7a439b7a823f6fcf5a82b9c69e0 Mon Sep 17 00:00:00 2001 From: Guillaume Gomez Date: Fri, 8 Aug 2025 18:08:08 +0200 Subject: Update gccjit dependency and libgccjit commit --- Cargo.lock | 8 ++++---- Cargo.toml | 2 +- libgccjit.version | 2 +- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/Cargo.lock b/Cargo.lock index 7f35c1a80bd..a5b972baf98 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -56,18 +56,18 @@ dependencies = [ [[package]] name = "gccjit" -version = "2.7.0" +version = "2.9.0" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "ae99a89184220d967dd300139f2d2ae7d52c1a69d632b24aacc57c54625254ce" +checksum = "4a0e310ef75f396cd11b2443b353d55376656ca92c13cba36f92b7aff346ac1a" dependencies = [ "gccjit_sys", ] [[package]] name = "gccjit_sys" -version = "0.8.0" +version = "0.8.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "24edb7bfe2b7b27c6d09ed23eebfcab0b359c8fe978433f902943e6f127a0f1b" +checksum = "95ed7572b30cd32430294dde6fb70822d58e67c6846a548647e8739776a0125b" dependencies = [ "libc", ] diff --git a/Cargo.toml b/Cargo.toml index 193348d1ef6..6031933bd2d 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -24,7 +24,7 @@ default = ["master"] [dependencies] object = { version = "0.37.0", default-features = false, features = ["std", "read"] } tempfile = "3.20" -gccjit = "2.7" +gccjit = "2.8" #gccjit = { git = "https://github.com/rust-lang/gccjit.rs" } # Local copy. diff --git a/libgccjit.version b/libgccjit.version index f62154968d3..dc9a0012864 100644 --- a/libgccjit.version +++ b/libgccjit.version @@ -1 +1 @@ -04ce66d8c918de9273bd7101638ad8724edf5e21 +4e995bd73c4490edfe5080ec6014d63aa9abed5f -- cgit 1.4.1-3-g733a5 From 3fb62b990c0bdb46412944862a28ad9b92a73b82 Mon Sep 17 00:00:00 2001 From: The rustc-josh-sync Cronjob Bot Date: Sat, 9 Aug 2025 01:53:44 +0000 Subject: Prepare for merging from rust-lang/rust This updates the rust-version file to ffb9d94dcf4ade0d534842be3672d5e9f47e1333. --- library/compiler-builtins/rust-version | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/library/compiler-builtins/rust-version b/library/compiler-builtins/rust-version index a4db05a8796..3928504c857 100644 --- a/library/compiler-builtins/rust-version +++ b/library/compiler-builtins/rust-version @@ -1 +1 @@ -82310651b93a594a3fd69015e1562186a080d94c +ffb9d94dcf4ade0d534842be3672d5e9f47e1333 -- cgit 1.4.1-3-g733a5 From 21bdfa14cbd99aba173751a2c723fb1e65de401b Mon Sep 17 00:00:00 2001 From: Aurelia Molzer <5550310+HeroicKatora@users.noreply.github.com> Date: Sat, 9 Aug 2025 17:54:43 +0200 Subject: Ensure consistent drop for hint::select_unpredictable There are a few alternatives to the implementation. The principal problem is that the selected value must be owned (in the sense of having any drop flag of sorts) when the unselected value is dropped, such that panic unwind goes through the drop of both. This ownership must then be passed on in return when the drop went smoothly. The basic way of achieving this is by extracting the selected value first, at the cost of relying on the optimizer a little more for detecting the copy as constructing the return value despite having a place in the body. --- library/core/src/hint.rs | 6 +++++- library/coretests/tests/hint.rs | 37 +++++++++++++++++++++++++++++++++++++ 2 files changed, 42 insertions(+), 1 deletion(-) diff --git a/library/core/src/hint.rs b/library/core/src/hint.rs index c72eeb9a9c9..53c17d6c887 100644 --- a/library/core/src/hint.rs +++ b/library/core/src/hint.rs @@ -790,8 +790,12 @@ pub fn select_unpredictable(condition: bool, true_val: T, false_val: T) -> T // is returned. This is necessary because the intrinsic doesn't drop the // value that is not selected. unsafe { + // Extract the selected value first, ensure it is dropped as well if dropping the unselected + // value panics. + let ret = crate::intrinsics::select_unpredictable(condition, &true_val, &false_val) + .assume_init_read(); crate::intrinsics::select_unpredictable(!condition, &mut true_val, &mut false_val) .assume_init_drop(); - crate::intrinsics::select_unpredictable(condition, true_val, false_val).assume_init() + ret } } diff --git a/library/coretests/tests/hint.rs b/library/coretests/tests/hint.rs index 032bbc1dcc8..9ef5567681e 100644 --- a/library/coretests/tests/hint.rs +++ b/library/coretests/tests/hint.rs @@ -21,3 +21,40 @@ fn select_unpredictable_drop() { assert!(a_dropped.get()); assert!(b_dropped.get()); } + +#[test] +#[should_panic] +fn select_unpredictable_drop_on_panic() { + use core::cell::Cell; + + struct X<'a> { + cell: &'a Cell, + expect: u16, + write: u16, + } + + impl Drop for X<'_> { + fn drop(&mut self) { + let value = self.cell.get(); + self.cell.set(self.write); + assert_eq!(value, self.expect); + } + } + + let cell = Cell::new(0); + + // Trigger a double-panic if the selected cell was not dropped during panic. + let _armed = X { cell: &cell, expect: 0xdead, write: 0 }; + let selected = X { cell: &cell, write: 0xdead, expect: 1 }; + let unselected = X { cell: &cell, write: 1, expect: 0xff }; + + // The correct drop order is: + // + // 1. `unselected` drops, writes 1, and panics as 0 != 0xff + // 2. `selected` drops during unwind, writes 0xdead and does not panic as 1 == 1 + // 3. `armed` drops during unwind, writes 0 and does not panic as 0xdead == 0xdead + // + // If `selected` is not dropped, `armed` panics as 1 != 0xdead + let _unreachable = + core::hint::select_unpredictable(core::hint::black_box(true), selected, unselected); +} -- cgit 1.4.1-3-g733a5 From 7d7fe3ff05c7c14834252e3345389e4dbf842cc5 Mon Sep 17 00:00:00 2001 From: Aurelia Molzer <5550310+HeroicKatora@users.noreply.github.com> Date: Sat, 9 Aug 2025 21:40:12 +0200 Subject: Adjust for codegen of unpredictable return value It seems important for LLVM that we select on the values by-value instead of reading and have no intermediate store. So make sure the guards selects both potential drops but defers the return value to the second selection. Since the two options alias we use raw mutable pointers instead of mutable references as before. --- library/core/src/hint.rs | 36 +++++++++++++++++++++++++++++++----- 1 file changed, 31 insertions(+), 5 deletions(-) diff --git a/library/core/src/hint.rs b/library/core/src/hint.rs index 53c17d6c887..1bb1c569508 100644 --- a/library/core/src/hint.rs +++ b/library/core/src/hint.rs @@ -786,16 +786,42 @@ pub fn select_unpredictable(condition: bool, true_val: T, false_val: T) -> T // Change this to use ManuallyDrop instead. let mut true_val = MaybeUninit::new(true_val); let mut false_val = MaybeUninit::new(false_val); + + struct DropOnPanic { + inner: *mut MaybeUninit, + } + + impl Drop for DropOnPanic { + fn drop(&mut self) { + // SAFETY: Must be guaranteed on construction of local type `DropOnPanic`. + unsafe { (*self.inner).assume_init_drop() } + } + } + + let true_ptr = (&mut true_val) as *mut _; + let false_ptr = (&mut false_val) as *mut _; + // SAFETY: The value that is not selected is dropped, and the selected one // is returned. This is necessary because the intrinsic doesn't drop the // value that is not selected. unsafe { // Extract the selected value first, ensure it is dropped as well if dropping the unselected // value panics. - let ret = crate::intrinsics::select_unpredictable(condition, &true_val, &false_val) - .assume_init_read(); - crate::intrinsics::select_unpredictable(!condition, &mut true_val, &mut false_val) - .assume_init_drop(); - ret + let (guard, drop) = crate::intrinsics::select_unpredictable( + condition, + (true_ptr, false_ptr), + (false_ptr, true_ptr), + ); + + // SAFETY: both pointers are to valid `MaybeUninit`, in both variants they do not alias but + // the two arguments we have selected from did alias each other. + let guard = DropOnPanic { inner: guard }; + (*drop).assume_init_drop(); + crate::mem::forget(guard); + + // Note that it is important to use the values here. Reading from the pointer we got makes + // LLVM forget the !unpredictable annotation sometimes (in tests, integer sized values in + // particular seemed to confuse it, also observed in llvm/llvm-project #82340). + crate::intrinsics::select_unpredictable(condition, true_val, false_val).assume_init() } } -- cgit 1.4.1-3-g733a5 From 531d5aa593592b16f379f9d6af5d9ae666a163b3 Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Sat, 9 Aug 2025 15:42:48 -0500 Subject: symcheck: Skip `__ymm@` symbols on Windows Like `__real@`, and `__xmm@`, Windows can emit duplicate `__ymm@` symbols for constants. --- library/compiler-builtins/crates/symbol-check/src/main.rs | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/library/compiler-builtins/crates/symbol-check/src/main.rs b/library/compiler-builtins/crates/symbol-check/src/main.rs index 129c959f23a..4e94552331a 100644 --- a/library/compiler-builtins/crates/symbol-check/src/main.rs +++ b/library/compiler-builtins/crates/symbol-check/src/main.rs @@ -222,7 +222,7 @@ fn verify_no_duplicates(archive: &Archive) { // Windows has symbols for literal numeric constants, string literals, and MinGW pseudo- // relocations. These are allowed to have repeated definitions. - let win_allowed_dup_pfx = ["__real@", "__xmm@", "??_C@_", ".refptr"]; + let win_allowed_dup_pfx = ["__real@", "__xmm@", "__ymm@", "??_C@_", ".refptr"]; if win_allowed_dup_pfx .iter() .any(|pfx| sym.name.starts_with(pfx)) -- cgit 1.4.1-3-g733a5 From ba5def8a85bc983460f65f889510accbc7b743ae Mon Sep 17 00:00:00 2001 From: Trevor Gross Date: Sat, 9 Aug 2025 06:29:01 -0500 Subject: Add __addhf3, __subhf3, __mulhf3, __{eq,ge,gt,le,lt,ne,unord}hf2 LLVM does not currently emit these, but it is being discussed as an option on platforms where `f32` is not hardware supported. Glibc/libgcc also has the comparison functions [1] already. The generic implementations for addition, subtraction, and multiplication work for f16 without any complications, as do comparisons, so add them here. [1]: https://sourceware.org/git/?p=glibc.git;a=commit;h=6ec6c77867af4ddfec7323e0ac6ede89effca852 --- .../builtins-test/tests/addsub.rs | 24 ++++++++--------- .../compiler-builtins/builtins-test/tests/cmp.rs | 21 +++++++++++++++ .../compiler-builtins/builtins-test/tests/mul.rs | 8 +++++- .../compiler-builtins/src/float/add.rs | 5 ++++ .../compiler-builtins/src/float/cmp.rs | 31 ++++++++++++++++++++++ .../compiler-builtins/src/float/mul.rs | 5 ++++ .../compiler-builtins/src/float/sub.rs | 5 ++++ 7 files changed, 85 insertions(+), 14 deletions(-) diff --git a/library/compiler-builtins/builtins-test/tests/addsub.rs b/library/compiler-builtins/builtins-test/tests/addsub.rs index abe7dde645e..f3334bd0e2d 100644 --- a/library/compiler-builtins/builtins-test/tests/addsub.rs +++ b/library/compiler-builtins/builtins-test/tests/addsub.rs @@ -1,4 +1,5 @@ #![allow(unused_macros)] +#![cfg_attr(f16_enabled, feature(f16))] #![cfg_attr(f128_enabled, feature(f128))] use builtins_test::*; @@ -115,28 +116,25 @@ macro_rules! float_sum { mod float_addsub { use super::*; + #[cfg(f16_enabled)] + float_sum! { + f16, __addhf3, __subhf3, Half, all(); + } + float_sum! { f32, __addsf3, __subsf3, Single, all(); f64, __adddf3, __subdf3, Double, all(); } -} - -#[cfg(f128_enabled)] -#[cfg(not(x86_no_sse))] -#[cfg(not(any(target_arch = "powerpc", target_arch = "powerpc64")))] -mod float_addsub_f128 { - use super::*; + #[cfg(f128_enabled)] + #[cfg(not(x86_no_sse))] + #[cfg(not(any(target_arch = "powerpc", target_arch = "powerpc64")))] float_sum! { f128, __addtf3, __subtf3, Quad, not(feature = "no-sys-f128"); } -} - -#[cfg(f128_enabled)] -#[cfg(any(target_arch = "powerpc", target_arch = "powerpc64"))] -mod float_addsub_f128_ppc { - use super::*; + #[cfg(f128_enabled)] + #[cfg(any(target_arch = "powerpc", target_arch = "powerpc64"))] float_sum! { f128, __addkf3, __subkf3, Quad, not(feature = "no-sys-f128"); } diff --git a/library/compiler-builtins/builtins-test/tests/cmp.rs b/library/compiler-builtins/builtins-test/tests/cmp.rs index a904dc5f7de..4b01b6ca1c7 100644 --- a/library/compiler-builtins/builtins-test/tests/cmp.rs +++ b/library/compiler-builtins/builtins-test/tests/cmp.rs @@ -1,5 +1,6 @@ #![allow(unused_macros)] #![allow(unreachable_code)] +#![cfg_attr(f16_enabled, feature(f16))] #![cfg_attr(f128_enabled, feature(f128))] use builtins_test::*; @@ -51,6 +52,26 @@ mod float_comparisons { }; } + #[test] + #[cfg(f16_enabled)] + fn cmp_f16() { + use compiler_builtins::float::cmp::{ + __eqhf2, __gehf2, __gthf2, __lehf2, __lthf2, __nehf2, __unordhf2, + }; + + fuzz_float_2(N, |x: f16, y: f16| { + assert_eq!(__unordhf2(x, y) != 0, x.is_nan() || y.is_nan()); + cmp!(f16, x, y, Half, all(), + 1, __lthf2; + 1, __lehf2; + 1, __eqhf2; + -1, __gehf2; + -1, __gthf2; + 1, __nehf2; + ); + }); + } + #[test] fn cmp_f32() { use compiler_builtins::float::cmp::{ diff --git a/library/compiler-builtins/builtins-test/tests/mul.rs b/library/compiler-builtins/builtins-test/tests/mul.rs index 3072b45dca0..bbf1157db42 100644 --- a/library/compiler-builtins/builtins-test/tests/mul.rs +++ b/library/compiler-builtins/builtins-test/tests/mul.rs @@ -1,5 +1,6 @@ -#![allow(unused_macros)] +#![cfg_attr(f16_enabled, feature(f16))] #![cfg_attr(f128_enabled, feature(f128))] +#![allow(unused_macros)] use builtins_test::*; @@ -117,6 +118,11 @@ macro_rules! float_mul { mod float_mul { use super::*; + #[cfg(f16_enabled)] + float_mul! { + f16, __mulhf3, Half, all(); + } + // FIXME(#616): Stop ignoring arches that don't have native support once fix for builtins is in // nightly. float_mul! { diff --git a/library/compiler-builtins/compiler-builtins/src/float/add.rs b/library/compiler-builtins/compiler-builtins/src/float/add.rs index 0cc362f705b..8dbfb0e1020 100644 --- a/library/compiler-builtins/compiler-builtins/src/float/add.rs +++ b/library/compiler-builtins/compiler-builtins/src/float/add.rs @@ -191,6 +191,11 @@ where } intrinsics! { + #[cfg(f16_enabled)] + pub extern "C" fn __addhf3(a: f16, b: f16) -> f16 { + add(a, b) + } + #[aapcs_on_arm] #[arm_aeabi_alias = __aeabi_fadd] pub extern "C" fn __addsf3(a: f32, b: f32) -> f32 { diff --git a/library/compiler-builtins/compiler-builtins/src/float/cmp.rs b/library/compiler-builtins/compiler-builtins/src/float/cmp.rs index f1e54dc1c83..8ab39c2b591 100644 --- a/library/compiler-builtins/compiler-builtins/src/float/cmp.rs +++ b/library/compiler-builtins/compiler-builtins/src/float/cmp.rs @@ -115,6 +115,37 @@ fn unord(a: F, b: F) -> bool { a_abs > inf_rep || b_abs > inf_rep } +#[cfg(f16_enabled)] +intrinsics! { + pub extern "C" fn __lehf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + cmp(a, b).to_le_abi() + } + + pub extern "C" fn __gehf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + cmp(a, b).to_ge_abi() + } + + pub extern "C" fn __unordhf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + unord(a, b) as crate::float::cmp::CmpResult + } + + pub extern "C" fn __eqhf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + cmp(a, b).to_le_abi() + } + + pub extern "C" fn __lthf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + cmp(a, b).to_le_abi() + } + + pub extern "C" fn __nehf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + cmp(a, b).to_le_abi() + } + + pub extern "C" fn __gthf2(a: f16, b: f16) -> crate::float::cmp::CmpResult { + cmp(a, b).to_ge_abi() + } +} + intrinsics! { pub extern "C" fn __lesf2(a: f32, b: f32) -> crate::float::cmp::CmpResult { cmp(a, b).to_le_abi() diff --git a/library/compiler-builtins/compiler-builtins/src/float/mul.rs b/library/compiler-builtins/compiler-builtins/src/float/mul.rs index dbed3095cda..49a2414eb5c 100644 --- a/library/compiler-builtins/compiler-builtins/src/float/mul.rs +++ b/library/compiler-builtins/compiler-builtins/src/float/mul.rs @@ -180,6 +180,11 @@ where } intrinsics! { + #[cfg(f16_enabled)] + pub extern "C" fn __mulhf3(a: f16, b: f16) -> f16 { + mul(a, b) + } + #[aapcs_on_arm] #[arm_aeabi_alias = __aeabi_fmul] pub extern "C" fn __mulsf3(a: f32, b: f32) -> f32 { diff --git a/library/compiler-builtins/compiler-builtins/src/float/sub.rs b/library/compiler-builtins/compiler-builtins/src/float/sub.rs index a0fd9dff97f..48ef33b0b82 100644 --- a/library/compiler-builtins/compiler-builtins/src/float/sub.rs +++ b/library/compiler-builtins/compiler-builtins/src/float/sub.rs @@ -1,6 +1,11 @@ use crate::float::Float; intrinsics! { + #[cfg(f16_enabled)] + pub extern "C" fn __subhf3(a: f16, b: f16) -> f16 { + crate::float::add::__addhf3(a, f16::from_bits(b.to_bits() ^ f16::SIGN_MASK)) + } + #[arm_aeabi_alias = __aeabi_fsub] pub extern "C" fn __subsf3(a: f32, b: f32) -> f32 { crate::float::add::__addsf3(a, f32::from_bits(b.to_bits() ^ f32::SIGN_MASK)) -- cgit 1.4.1-3-g733a5 From 307ebca295428fc14718ebae874e816f06532145 Mon Sep 17 00:00:00 2001 From: dvermd <315743+dvermd@users.noreply.github.com> Date: Sun, 10 Aug 2025 08:23:11 +0200 Subject: run rustfmt on tests/run/*.rs --- build_system/src/fmt.rs | 30 ++++++++++++++++++++++++++++-- 1 file changed, 28 insertions(+), 2 deletions(-) diff --git a/build_system/src/fmt.rs b/build_system/src/fmt.rs index 7e6594f50f9..91535f217e3 100644 --- a/build_system/src/fmt.rs +++ b/build_system/src/fmt.rs @@ -1,7 +1,7 @@ use std::ffi::OsStr; use std::path::Path; -use crate::utils::run_command_with_output; +use crate::utils::{run_command_with_output, walk_dir}; fn show_usage() { println!( @@ -32,5 +32,31 @@ pub fn run() -> Result<(), String> { if check { &[&"cargo", &"fmt", &"--check"] } else { &[&"cargo", &"fmt"] }; run_command_with_output(cmd, Some(Path::new(".")))?; - run_command_with_output(cmd, Some(Path::new("build_system"))) + run_command_with_output(cmd, Some(Path::new("build_system")))?; + + run_rustfmt_recursively("tests/run", check) +} + +fn run_rustfmt_recursively

(dir: P, check: bool) -> Result<(), String> +where + P: AsRef, +{ + walk_dir( + dir, + &mut |dir| run_rustfmt_recursively(dir, check), + &mut |file_path| { + if file_path.extension().filter(|ext| ext == &OsStr::new("rs")).is_some() { + let rustfmt_cmd: &[&dyn AsRef] = if check { + &[&"rustfmt", &"--check", &file_path] + } else { + &[&"rustfmt", &file_path] + }; + + run_command_with_output(rustfmt_cmd, Some(Path::new("."))) + } else { + Ok(()) + } + }, + true, + ) } -- cgit 1.4.1-3-g733a5 From dde77db2ab46cb4319ced77a20b97d01ab13647c Mon Sep 17 00:00:00 2001 From: dvermd <315743+dvermd@users.noreply.github.com> Date: Sun, 10 Aug 2025 08:24:48 +0200 Subject: fix tests/run formatting --- tests/run/int.rs | 10 ++++------ tests/run/int_overflow.rs | 2 +- tests/run/structs.rs | 8 ++------ tests/run/volatile.rs | 8 ++------ tests/run/volatile2.rs | 12 ++++++++++-- 5 files changed, 19 insertions(+), 21 deletions(-) diff --git a/tests/run/int.rs b/tests/run/int.rs index e20ecc23679..78675acb544 100644 --- a/tests/run/int.rs +++ b/tests/run/int.rs @@ -7,12 +7,10 @@ fn main() { use std::hint::black_box; macro_rules! check { - ($ty:ty, $expr:expr) => { - { - const EXPECTED: $ty = $expr; - assert_eq!($expr, EXPECTED); - } - }; + ($ty:ty, $expr:expr) => {{ + const EXPECTED: $ty = $expr; + assert_eq!($expr, EXPECTED); + }}; } check!(u32, (2220326408_u32 + black_box(1)) >> (32 - 6)); diff --git a/tests/run/int_overflow.rs b/tests/run/int_overflow.rs index 78872159f62..78e1cac57e0 100644 --- a/tests/run/int_overflow.rs +++ b/tests/run/int_overflow.rs @@ -12,7 +12,7 @@ fn main() { let arg_count = std::env::args().count(); let int = isize::MAX; - let _int = int + arg_count as isize; // overflow + let _int = int + arg_count as isize; // overflow // If overflow checking is disabled, we should reach here. #[cfg(not(debug_assertions))] diff --git a/tests/run/structs.rs b/tests/run/structs.rs index da73cbed9ae..e08e67837be 100644 --- a/tests/run/structs.rs +++ b/tests/run/structs.rs @@ -27,12 +27,8 @@ fn one() -> isize { #[no_mangle] extern "C" fn main(argc: i32, _argv: *const *const u8) -> i32 { - let test = Test { - field: one(), - }; - let two = Two { - two: 2, - }; + let test = Test { field: one() }; + let two = Two { two: 2 }; unsafe { libc::printf(b"%ld\n\0" as *const u8 as *const i8, test.field); libc::printf(b"%ld\n\0" as *const u8 as *const i8, two.two); diff --git a/tests/run/volatile.rs b/tests/run/volatile.rs index 94a7bdc5c06..dc11fbfa600 100644 --- a/tests/run/volatile.rs +++ b/tests/run/volatile.rs @@ -12,15 +12,11 @@ struct Struct { func: unsafe fn(*const ()), } -fn func(_ptr: *const ()) { -} +fn func(_ptr: *const ()) {} fn main() { let mut x = MaybeUninit::<&Struct>::uninit(); - x.write(&Struct { - pointer: std::ptr::null(), - func, - }); + x.write(&Struct { pointer: std::ptr::null(), func }); let x = unsafe { x.assume_init() }; let value = unsafe { (x as *const Struct).read_volatile() }; println!("{:?}", value); diff --git a/tests/run/volatile2.rs b/tests/run/volatile2.rs index bdcb8259878..ada112687d3 100644 --- a/tests/run/volatile2.rs +++ b/tests/run/volatile2.rs @@ -7,7 +7,14 @@ mod libc { #[link(name = "c")] extern "C" { pub fn sigaction(signum: i32, act: *const sigaction, oldact: *mut sigaction) -> i32; - pub fn mmap(addr: *mut (), len: usize, prot: i32, flags: i32, fd: i32, offset: i64) -> *mut (); + pub fn mmap( + addr: *mut (), + len: usize, + prot: i32, + flags: i32, + fd: i32, + offset: i64, + ) -> *mut (); pub fn mprotect(addr: *mut (), len: usize, prot: i32) -> i32; } @@ -54,7 +61,8 @@ fn main() { libc::MAP_PRIVATE | libc::MAP_ANONYMOUS, -1, 0, - ).cast(); + ) + .cast(); if STORAGE == libc::MAP_FAILED { panic!("error: mmap failed"); } -- cgit 1.4.1-3-g733a5 From 3e76b58453461a7ac04db4914caff584345d8448 Mon Sep 17 00:00:00 2001 From: Folkert de Vries Date: Fri, 1 Aug 2025 16:15:21 +0200 Subject: add place mention for `#[loop_match]` scrutinee --- compiler/rustc_mir_build/src/builder/expr/into.rs | 6 +- .../rustc_mir_build/src/builder/matches/mod.rs | 2 +- ...ch_diverges.break_to_block_unit.built.after.mir | 63 +++++++ .../loop_match_diverges.infinite_a.built.after.mir | 51 ++++++ tests/mir-opt/building/loop_match_diverges.rs | 68 ++++++++ .../loop_match_diverges.simple.built.after.mir | 187 +++++++++++++++++++++ 6 files changed, 374 insertions(+), 3 deletions(-) create mode 100644 tests/mir-opt/building/loop_match_diverges.break_to_block_unit.built.after.mir create mode 100644 tests/mir-opt/building/loop_match_diverges.infinite_a.built.after.mir create mode 100644 tests/mir-opt/building/loop_match_diverges.rs create mode 100644 tests/mir-opt/building/loop_match_diverges.simple.built.after.mir diff --git a/compiler/rustc_mir_build/src/builder/expr/into.rs b/compiler/rustc_mir_build/src/builder/expr/into.rs index ac87f671699..78bcc47d9d0 100644 --- a/compiler/rustc_mir_build/src/builder/expr/into.rs +++ b/compiler/rustc_mir_build/src/builder/expr/into.rs @@ -295,9 +295,11 @@ impl<'a, 'tcx> Builder<'a, 'tcx> { this.diverge_from(loop_block); // Logic for `match`. - let scrutinee_place_builder = - unpack!(body_block = this.as_place_builder(body_block, scrutinee)); let scrutinee_span = this.thir.exprs[scrutinee].span; + let scrutinee_place_builder = unpack!( + body_block = this.lower_scrutinee(body_block, scrutinee, scrutinee_span) + ); + let match_start_span = match_span.shrink_to_lo().to(scrutinee_span); let mut patterns = Vec::with_capacity(arms.len()); diff --git a/compiler/rustc_mir_build/src/builder/matches/mod.rs b/compiler/rustc_mir_build/src/builder/matches/mod.rs index 2c29b862841..e863f94af7e 100644 --- a/compiler/rustc_mir_build/src/builder/matches/mod.rs +++ b/compiler/rustc_mir_build/src/builder/matches/mod.rs @@ -407,7 +407,7 @@ impl<'a, 'tcx> Builder<'a, 'tcx> { } /// Evaluate the scrutinee and add the PlaceMention for it. - fn lower_scrutinee( + pub(crate) fn lower_scrutinee( &mut self, mut block: BasicBlock, scrutinee_id: ExprId, diff --git a/tests/mir-opt/building/loop_match_diverges.break_to_block_unit.built.after.mir b/tests/mir-opt/building/loop_match_diverges.break_to_block_unit.built.after.mir new file mode 100644 index 00000000000..6d779e46146 --- /dev/null +++ b/tests/mir-opt/building/loop_match_diverges.break_to_block_unit.built.after.mir @@ -0,0 +1,63 @@ +// MIR for `break_to_block_unit` after built + +fn break_to_block_unit() -> u8 { + let mut _0: u8; + let mut _1: i32; + let mut _2: !; + scope 1 { + debug state => _1; + } + + bb0: { + StorageLive(_1); + _1 = const 0_i32; + FakeRead(ForLet(None), _1); + StorageLive(_2); + goto -> bb1; + } + + bb1: { + falseUnwind -> [real: bb2, unwind: bb10]; + } + + bb2: { + PlaceMention(_1); + _1 = const 2_i32; + goto -> bb5; + } + + bb3: { + FakeRead(ForMatchedPlace(None), _1); + unreachable; + } + + bb4: { + goto -> bb6; + } + + bb5: { + goto -> bb6; + } + + bb6: { + goto -> bb7; + } + + bb7: { + goto -> bb1; + } + + bb8: { + unreachable; + } + + bb9: { + StorageDead(_2); + StorageDead(_1); + return; + } + + bb10 (cleanup): { + resume; + } +} diff --git a/tests/mir-opt/building/loop_match_diverges.infinite_a.built.after.mir b/tests/mir-opt/building/loop_match_diverges.infinite_a.built.after.mir new file mode 100644 index 00000000000..e3766744790 --- /dev/null +++ b/tests/mir-opt/building/loop_match_diverges.infinite_a.built.after.mir @@ -0,0 +1,51 @@ +// MIR for `infinite_a` after built + +fn infinite_a(_1: u8) -> () { + debug state => _1; + let mut _0: (); + let mut _2: !; + let _3: u8; + scope 1 { + debug a => _3; + } + + bb0: { + StorageLive(_2); + goto -> bb1; + } + + bb1: { + falseUnwind -> [real: bb2, unwind: bb7]; + } + + bb2: { + PlaceMention(_1); + StorageLive(_3); + _3 = copy _1; + _1 = copy _3; + StorageDead(_3); + goto -> bb4; + } + + bb3: { + FakeRead(ForMatchedPlace(None), _1); + unreachable; + } + + bb4: { + goto -> bb1; + } + + bb5: { + unreachable; + } + + bb6: { + StorageDead(_2); + return; + } + + bb7 (cleanup): { + resume; + } +} diff --git a/tests/mir-opt/building/loop_match_diverges.rs b/tests/mir-opt/building/loop_match_diverges.rs new file mode 100644 index 00000000000..774e195c33c --- /dev/null +++ b/tests/mir-opt/building/loop_match_diverges.rs @@ -0,0 +1,68 @@ +// skip-filecheck +#![allow(incomplete_features)] +#![feature(loop_match)] +#![crate_type = "lib"] + +// Test that a #[loop_match] without an explicit break from the loop generates valid MIR. + +enum State { + A, + B, + C, +} + +// EMIT_MIR loop_match_diverges.simple.built.after.mir +fn simple(mut state: State) -> State { + #[loop_match] + 'a: loop { + state = 'blk: { + match state { + State::A => { + #[const_continue] + break 'blk State::B; + } + State::B => { + if true { + #[const_continue] + break 'blk State::C; + } else { + #[const_continue] + break 'blk State::A; + } + } + State::C => break 'a, + } + }; + } + + state +} + +// EMIT_MIR loop_match_diverges.break_to_block_unit.built.after.mir +#[unsafe(no_mangle)] +fn break_to_block_unit() -> u8 { + let mut state = 0; + #[loop_match] + loop { + state = 'blk: { + match state { + _ => 'b: { + break 'b 2; + } + } + } + } +} + +// EMIT_MIR loop_match_diverges.infinite_a.built.after.mir +#[unsafe(no_mangle)] +fn infinite_a(mut state: u8) { + #[loop_match] + loop { + state = 'blk: { + match state { + a => a, + } + } + } +} diff --git a/tests/mir-opt/building/loop_match_diverges.simple.built.after.mir b/tests/mir-opt/building/loop_match_diverges.simple.built.after.mir new file mode 100644 index 00000000000..26476ad7762 --- /dev/null +++ b/tests/mir-opt/building/loop_match_diverges.simple.built.after.mir @@ -0,0 +1,187 @@ +// MIR for `simple` after built + +fn simple(_1: State) -> State { + debug state => _1; + let mut _0: State; + let _2: (); + let mut _3: isize; + let mut _4: !; + let mut _5: isize; + let mut _6: bool; + let mut _7: !; + let mut _8: isize; + let mut _9: !; + let mut _10: isize; + let mut _11: !; + + bb0: { + StorageLive(_2); + goto -> bb1; + } + + bb1: { + falseUnwind -> [real: bb2, unwind: bb37]; + } + + bb2: { + PlaceMention(_1); + _3 = discriminant(_1); + switchInt(move _3) -> [0: bb4, 1: bb6, 2: bb8, otherwise: bb3]; + } + + bb3: { + FakeRead(ForMatchedPlace(None), _1); + unreachable; + } + + bb4: { + falseEdge -> [real: bb11, imaginary: bb6]; + } + + bb5: { + goto -> bb3; + } + + bb6: { + falseEdge -> [real: bb10, imaginary: bb8]; + } + + bb7: { + goto -> bb3; + } + + bb8: { + _2 = const (); + goto -> bb36; + } + + bb9: { + goto -> bb3; + } + + bb10: { + StorageLive(_6); + _6 = const true; + switchInt(move _6) -> [0: bb17, otherwise: bb16]; + } + + bb11: { + _1 = State::B; + _5 = discriminant(_1); + falseEdge -> [real: bb12, imaginary: bb13]; + } + + bb12: { + goto -> bb10; + } + + bb13: { + goto -> bb34; + } + + bb14: { + unreachable; + } + + bb15: { + goto -> bb32; + } + + bb16: { + _1 = State::C; + _8 = discriminant(_1); + falseEdge -> [real: bb18, imaginary: bb19]; + } + + bb17: { + goto -> bb23; + } + + bb18: { + goto -> bb20; + } + + bb19: { + goto -> bb33; + } + + bb20: { + StorageDead(_6); + goto -> bb8; + } + + bb21: { + unreachable; + } + + bb22: { + goto -> bb29; + } + + bb23: { + _1 = State::A; + _10 = discriminant(_1); + falseEdge -> [real: bb24, imaginary: bb25]; + } + + bb24: { + goto -> bb26; + } + + bb25: { + goto -> bb33; + } + + bb26: { + StorageDead(_6); + goto -> bb11; + } + + bb27: { + unreachable; + } + + bb28: { + goto -> bb29; + } + + bb29: { + StorageDead(_6); + goto -> bb32; + } + + bb30: { + unreachable; + } + + bb31: { + goto -> bb32; + } + + bb32: { + goto -> bb35; + } + + bb33: { + StorageDead(_6); + goto -> bb34; + } + + bb34: { + goto -> bb35; + } + + bb35: { + goto -> bb1; + } + + bb36: { + StorageDead(_2); + _0 = move _1; + return; + } + + bb37 (cleanup): { + resume; + } +} -- cgit 1.4.1-3-g733a5 From a4cf5e9eca00527082fd58356e4adc8775aeee03 Mon Sep 17 00:00:00 2001 From: Chen Zhongyao Date: Mon, 14 Jul 2025 13:35:53 +0000 Subject: Add new Tier-3 target: riscv64a23-unknown-linux-gnu --- compiler/rustc_codegen_llvm/src/llvm_util.rs | 6 +- compiler/rustc_target/src/spec/mod.rs | 1 + .../spec/targets/riscv64a23_unknown_linux_gnu.rs | 27 +++++ compiler/rustc_target/src/target_features.rs | 113 +++++++++++++++++++++ src/bootstrap/src/core/sanity.rs | 1 + tests/assembly-llvm/targets/targets-elf.rs | 3 + tests/ui/check-cfg/target_feature.stderr | 16 +++ 7 files changed, 166 insertions(+), 1 deletion(-) create mode 100644 compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs diff --git a/compiler/rustc_codegen_llvm/src/llvm_util.rs b/compiler/rustc_codegen_llvm/src/llvm_util.rs index 28d2100f478..a85a894e42c 100644 --- a/compiler/rustc_codegen_llvm/src/llvm_util.rs +++ b/compiler/rustc_codegen_llvm/src/llvm_util.rs @@ -278,7 +278,11 @@ pub(crate) fn to_llvm_features<'a>(sess: &Session, s: &'a str) -> Option None, + ("riscv32" | "riscv64", "zacas" | "rva23s64" | "rva23u64" | "sha" | "ssnpm" | "supm") + if get_version().0 < 20 => + { + None + } ( "s390x", "message-security-assist-extension12" diff --git a/compiler/rustc_target/src/spec/mod.rs b/compiler/rustc_target/src/spec/mod.rs index b9fbff8db05..d72d8bd2d08 100644 --- a/compiler/rustc_target/src/spec/mod.rs +++ b/compiler/rustc_target/src/spec/mod.rs @@ -2137,6 +2137,7 @@ supported_targets! { ("riscv64gc-unknown-none-elf", riscv64gc_unknown_none_elf), ("riscv64gc-unknown-linux-gnu", riscv64gc_unknown_linux_gnu), ("riscv64gc-unknown-linux-musl", riscv64gc_unknown_linux_musl), + ("riscv64a23-unknown-linux-gnu", riscv64a23_unknown_linux_gnu), ("sparc-unknown-none-elf", sparc_unknown_none_elf), diff --git a/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs b/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs new file mode 100644 index 00000000000..27a0e27c969 --- /dev/null +++ b/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs @@ -0,0 +1,27 @@ +use std::borrow::Cow; + +use crate::spec::{CodeModel, SplitDebuginfo, Target, TargetMetadata, TargetOptions, base}; + +pub(crate) fn target() -> Target { + Target { + llvm_target: "riscv64-unknown-linux-gnu".into(), + metadata: TargetMetadata { + description: Some("RISC-V Linux (kernel 4.20, glibc 2.29)".into()), + tier: Some(3), + host_tools: Some(true), + std: Some(true), + }, + pointer_width: 64, + data_layout: "e-m:e-p:64:64-i64:64-i128:128-n32:64-S128".into(), + arch: "riscv64".into(), + options: TargetOptions { + code_model: Some(CodeModel::Medium), + cpu: "generic-rv64".into(), + features: "+rva23u64,+rva23s64".into(), + llvm_abiname: "lp64d".into(), + max_atomic_width: Some(64), + supported_split_debuginfo: Cow::Borrowed(&[SplitDebuginfo::Off]), + ..base::linux_gnu::opts() + }, + } +} diff --git a/compiler/rustc_target/src/target_features.rs b/compiler/rustc_target/src/target_features.rs index 297d9ed84c5..434f6a840b6 100644 --- a/compiler/rustc_target/src/target_features.rs +++ b/compiler/rustc_target/src/target_features.rs @@ -597,6 +597,119 @@ static RISCV_FEATURES: &[(&str, Stability, ImpliedFeatures)] = &[ ), ("m", Stable, &[]), ("relax", Unstable(sym::riscv_target_feature), &[]), + ( + "rva23s64", + Unstable(sym::riscv_target_feature), + &[ + "m", + "a", + "f", + "d", + "c", + "b", + "v", + "zicsr", + "zicntr", + "zihpm", + "ziccif", + "ziccrse", + "ziccamoa", + "zicclsm", + "zic64b", + "za64rs", + "zihintpause", + "zba", + "zbb", + "zbs", + "zicbom", + "zicbop", + "zicboz", + "zfhmin", + "zkt", + "zvfhmin", + "zvbb", + "zvkt", + "zihintntl", + "zicond", + "zimop", + "zcmop", + "zcb", + "zfa", + "zawrs", + "svbare", + "svade", + "ssccptr", + "sstvecd", + "sstvala", + "sscounterenw", + "svpbmt", + "svinval", + "svnapot", + "sstc", + "sscofpmf", + "ssnpm", + "ssu64xl", + "sha", + "supm", + ], + ), + ( + "rva23u64", + Unstable(sym::riscv_target_feature), + &[ + "m", + "a", + "f", + "d", + "c", + "b", + "v", + "zicsr", + "zicntr", + "zihpm", + "ziccif", + "ziccrse", + "ziccamoa", + "zicclsm", + "zic64b", + "za64rs", + "zihintpause", + "zba", + "zbb", + "zbs", + "zicbom", + "zicbop", + "zicboz", + "zfhmin", + "zkt", + "zvfhmin", + "zvbb", + "zvkt", + "zihintntl", + "zicond", + "zimop", + "zcmop", + "zcb", + "zfa", + "zawrs", + "supm", + ], + ), + ("sha", Unstable(sym::riscv_target_feature), &[]), + ("ssccptr", Unstable(sym::riscv_target_feature), &[]), + ("sscofpmf", Unstable(sym::riscv_target_feature), &[]), + ("sscounterenw", Unstable(sym::riscv_target_feature), &[]), + ("ssnpm", Unstable(sym::riscv_target_feature), &[]), + ("sstc", Unstable(sym::riscv_target_feature), &[]), + ("sstvala", Unstable(sym::riscv_target_feature), &[]), + ("sstvecd", Unstable(sym::riscv_target_feature), &[]), + ("ssu64xl", Unstable(sym::riscv_target_feature), &[]), + ("supm", Unstable(sym::riscv_target_feature), &[]), + ("svade", Unstable(sym::riscv_target_feature), &[]), + ("svbare", Unstable(sym::riscv_target_feature), &[]), + ("svinval", Unstable(sym::riscv_target_feature), &[]), + ("svnapot", Unstable(sym::riscv_target_feature), &[]), + ("svpbmt", Unstable(sym::riscv_target_feature), &[]), ("unaligned-scalar-mem", Unstable(sym::riscv_target_feature), &[]), ("unaligned-vector-mem", Unstable(sym::riscv_target_feature), &[]), ("v", Unstable(sym::riscv_target_feature), &["zvl128b", "zve64d"]), diff --git a/src/bootstrap/src/core/sanity.rs b/src/bootstrap/src/core/sanity.rs index 3080e641b5b..a2d5d3f0e37 100644 --- a/src/bootstrap/src/core/sanity.rs +++ b/src/bootstrap/src/core/sanity.rs @@ -34,6 +34,7 @@ pub struct Finder { // Targets can be removed from this list once they are present in the stage0 compiler (usually by updating the beta compiler of the bootstrap). const STAGE0_MISSING_TARGETS: &[&str] = &[ "armv7a-vex-v5", + "riscv64a23-unknown-linux-gnu", // just a dummy comment so the list doesn't get onelined ]; diff --git a/tests/assembly-llvm/targets/targets-elf.rs b/tests/assembly-llvm/targets/targets-elf.rs index ee63dffe9ea..44f536b6856 100644 --- a/tests/assembly-llvm/targets/targets-elf.rs +++ b/tests/assembly-llvm/targets/targets-elf.rs @@ -475,6 +475,9 @@ //@ revisions: riscv64gc_unknown_linux_gnu //@ [riscv64gc_unknown_linux_gnu] compile-flags: --target riscv64gc-unknown-linux-gnu //@ [riscv64gc_unknown_linux_gnu] needs-llvm-components: riscv +//@ revisions: riscv64a23_unknown_linux_gnu +//@ [riscv64a23_unknown_linux_gnu] compile-flags: --target riscv64a23-unknown-linux-gnu +//@ [riscv64a23_unknown_linux_gnu] needs-llvm-components: riscv //@ revisions: riscv64gc_unknown_linux_musl //@ [riscv64gc_unknown_linux_musl] compile-flags: --target riscv64gc-unknown-linux-musl //@ [riscv64gc_unknown_linux_musl] needs-llvm-components: riscv diff --git a/tests/ui/check-cfg/target_feature.stderr b/tests/ui/check-cfg/target_feature.stderr index 44fc23b6390..a484fb0e5f0 100644 --- a/tests/ui/check-cfg/target_feature.stderr +++ b/tests/ui/check-cfg/target_feature.stderr @@ -241,6 +241,8 @@ LL | cfg!(target_feature = "_UNEXPECTED_VALUE"); `relax` `relaxed-simd` `rtm` +`rva23s64` +`rva23u64` `sb` `scq` `sha` @@ -292,16 +294,27 @@ LL | cfg!(target_feature = "_UNEXPECTED_VALUE"); `soft-float` `spe` `ssbs` +`ssccptr` +`sscofpmf` +`sscounterenw` `sse` `sse2` `sse3` `sse4.1` `sse4.2` `sse4a` +`ssnpm` `ssse3` +`sstc` +`sstvala` +`sstvecd` +`ssu64xl` `ssve-fp8dot2` `ssve-fp8dot4` `ssve-fp8fma` +`supm` +`svade` +`svbare` `sve` `sve-b16b16` `sve2` @@ -310,6 +323,9 @@ LL | cfg!(target_feature = "_UNEXPECTED_VALUE"); `sve2-sha3` `sve2-sm4` `sve2p1` +`svinval` +`svnapot` +`svpbmt` `tail-call` `tbm` `thumb-mode` -- cgit 1.4.1-3-g733a5 From 988cd14ec1e1d07d8fadc2e76d370e1da8de4f46 Mon Sep 17 00:00:00 2001 From: Chen Zhongyao Date: Mon, 21 Jul 2025 16:31:01 +0000 Subject: Add a disabled builder for riscv64 rva23 emulated tests This will run all tests for `riscv64a23-unknown-linux-gnu` in a QEMU instance. --- .../host-x86_64/disabled/riscv64a23-gnu/Dockerfile | 108 +++++++++++++++++++++ .../disabled/riscv64a23-gnu/linux.config | 51 ++++++++++ src/tools/build-manifest/src/main.rs | 2 + src/tools/remote-test-client/src/main.rs | 29 +++++- 4 files changed, 189 insertions(+), 1 deletion(-) create mode 100644 src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile create mode 100644 src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config diff --git a/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile b/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile new file mode 100644 index 00000000000..86b0d795687 --- /dev/null +++ b/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile @@ -0,0 +1,108 @@ +FROM ubuntu:24.04 + +ARG DEBIAN_FRONTEND=noninteractive +RUN apt-get update -y && apt-get install -y --no-install-recommends \ + bc \ + bzip2 \ + ca-certificates \ + cmake \ + cpio \ + curl \ + file \ + flex \ + bison \ + g++ \ + g++-riscv64-linux-gnu \ + git \ + libc6-dev \ + libc6-dev-riscv64-cross \ + libssl-dev \ + make \ + ninja-build \ + python3 \ + xz-utils \ + opensbi \ + u-boot-qemu \ + libslirp0 \ + build-essential \ + pkg-config \ + libglib2.0-dev \ + libpixman-1-dev \ + libsdl2-dev \ + libfdt-dev \ + python3 \ + python3-pip + +ENV ARCH=riscv \ + CROSS_COMPILE=riscv64-linux-gnu- + +WORKDIR /build + +# From https://github.com/michaeljclark/busybear-linux/blob/master/conf/linux.config +COPY host-x86_64/riscv64a23-gnu/linux.config /build + +# qemu v10.0.2 fully support +RUN curl https://gitlab.com/qemu-project/qemu/-/archive/v10.0.2/qemu-v10.0.2.tar.bz2 | tar xjf - && \ + cd qemu-v10.0.2 && \ + ./configure --target-list=riscv64-softmmu \ + --enable-sdl --enable-debug --enable-fdt --enable-slirp && \ + make -j && make install + +# use the opensbi fw from apt-get install +RUN cp /usr/lib/riscv64-linux-gnu/opensbi/generic/fw_jump.bin /tmp + +# Compile the kernel that we're going to be emulating with. This is +# basically just done to be compatible with the QEMU target that we're going +# to be using when running tests. +RUN curl https://cdn.kernel.org/pub/linux/kernel/v6.x/linux-6.6.97.tar.xz | tar xJf - && \ + cp linux.config linux-6.6.97/.config && \ + cd /build/linux-6.6.97 && \ + make olddefconfig && \ + make -j$(nproc) Image && \ + cp arch/riscv/boot/Image /tmp && \ + rm -rf linux-6.6.97 + +# Compile an instance of busybox as this provides a lightweight system and init +# binary which we will boot into. Only trick here is configuring busybox to +# build static binaries. +RUN curl https://www.busybox.net/downloads/busybox-1.37.0.tar.bz2 | tar xjf - && \ + cd busybox-1.37.0 && \ + make defconfig && \ + sed -i 's/# CONFIG_STATIC is not set/CONFIG_STATIC=y/' .config && \ + sed -i 's/^CONFIG_TC=y$/# CONFIG_TC is not set/' .config && \ + sed -i 's/CONFIG_SHA1_HWACCEL=y/# CONFIG_SHA1_HWACCEL is not set/' .config && \ + sed -i 's/CONFIG_SHA256_HWACCEL=y/# CONFIG_SHA256_HWACCEL is not set/' .config && \ + make -j$(nproc) && \ + make install && \ + mv _install /tmp/rootfs && \ + cd /build && \ + rm -rf busybox-1.37.0 + +# Download the ubuntu rootfs, which we'll use as a chroot for all our tests. +WORKDIR /tmp +RUN mkdir rootfs/ubuntu +RUN curl https://cdimage.ubuntu.com/ubuntu-base/releases/24.04/release/ubuntu-base-24.04.2-base-riscv64.tar.gz | \ + tar xzf - -C rootfs/ubuntu && \ + cd rootfs && mkdir proc sys dev etc etc/init.d + +# Copy over our init script, which starts up our test server and also a few other +# misc tasks +COPY scripts/qemu-bare-bones-rcS rootfs/etc/init.d/rcS +RUN chmod +x rootfs/etc/init.d/rcS + +# Helper to quickly fill the entropy pool in the kernel +COPY scripts/qemu-bare-bones-addentropy.c /tmp/addentropy.c +RUN riscv64-linux-gnu-gcc addentropy.c -o rootfs/addentropy -static + +COPY scripts/sccache.sh /scripts/ +RUN sh /scripts/sccache.sh + +# Avoid "fatal: detected dubious ownership in repository at '/checkout'" error +RUN git config --global --add safe.directory "*" + +ENV RUST_CONFIGURE_ARGS \ + --set target.riscv64a23-unknown-linux-gnu.linker=riscv64-linux-gnu-gcc \ + --set target.riscv64a23-unknown-linux-gnu.qemu-rootfs=/tmp/rootfs +ENV SCRIPT python3 ../x.py --stage 2 test --host='' --target riscv64a23-unknown-linux-gnu + +ENV NO_CHANGE_USER=1 diff --git a/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config b/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config new file mode 100644 index 00000000000..5142664742f --- /dev/null +++ b/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config @@ -0,0 +1,51 @@ +CONFIG_DEFAULT_HOSTNAME="busybear" +CONFIG_SYSVIPC=y +CONFIG_POSIX_MQUEUE=y +CONFIG_IKCONFIG=y +CONFIG_IKCONFIG_PROC=y +CONFIG_CGROUPS=y +CONFIG_CGROUP_SCHED=y +CONFIG_CFS_BANDWIDTH=y +CONFIG_CGROUP_BPF=y +CONFIG_NAMESPACES=y +CONFIG_USER_NS=y +CONFIG_CHECKPOINT_RESTORE=y +CONFIG_BLK_DEV_INITRD=y +CONFIG_EXPERT=y +CONFIG_BPF_SYSCALL=y +CONFIG_SMP=y +CONFIG_MODULES=y +CONFIG_NET=y +CONFIG_PACKET=y +CONFIG_PACKET_DIAG=y +CONFIG_UNIX=y +CONFIG_INET=y +CONFIG_NETLINK_DIAG=y +# CONFIG_WIRELESS is not set +CONFIG_PCI=y +CONFIG_DEVTMPFS=y +CONFIG_BLK_DEV_LOOP=y +CONFIG_VIRTIO_BLK=y +CONFIG_NETDEVICES=y +CONFIG_VIRTIO_NET=y +# CONFIG_ETHERNET is not set +# CONFIG_WLAN is not set +CONFIG_SERIAL_8250=y +CONFIG_SERIAL_8250_CONSOLE=y +CONFIG_SERIAL_OF_PLATFORM=y +CONFIG_HVC_RISCV_SBI=y +# CONFIG_HW_RANDOM is not set +# CONFIG_USB_SUPPORT is not set +CONFIG_VIRTIO_MMIO=y +CONFIG_SIFIVE_PLIC=y +CONFIG_RAS=y +CONFIG_EXT2_FS=y +CONFIG_EXT3_FS=y +CONFIG_EXT4_FS_POSIX_ACL=y +CONFIG_AUTOFS4_FS=y +CONFIG_MSDOS_FS=y +CONFIG_VFAT_FS=y +CONFIG_TMPFS=y +# CONFIG_CRYPTO_ECHAINIV is not set +# CONFIG_CRYPTO_HW is not set +CONFIG_PRINTK_TIME=y diff --git a/src/tools/build-manifest/src/main.rs b/src/tools/build-manifest/src/main.rs index 0520eff0fa2..4f26c170de0 100644 --- a/src/tools/build-manifest/src/main.rs +++ b/src/tools/build-manifest/src/main.rs @@ -40,6 +40,7 @@ static HOSTS: &[&str] = &[ "powerpc64le-unknown-linux-gnu", "powerpc64le-unknown-linux-musl", "riscv64gc-unknown-linux-gnu", + "riscv64a23-unknown-linux-gnu", "s390x-unknown-linux-gnu", "sparcv9-sun-solaris", "x86_64-apple-darwin", @@ -155,6 +156,7 @@ static TARGETS: &[&str] = &[ "riscv64gc-unknown-none-elf", "riscv64gc-unknown-linux-gnu", "riscv64gc-unknown-linux-musl", + "riscv64a23-unknown-linux-gnu", "s390x-unknown-linux-gnu", "sparc64-unknown-linux-gnu", "sparcv9-sun-solaris", diff --git a/src/tools/remote-test-client/src/main.rs b/src/tools/remote-test-client/src/main.rs index b9741431b50..e5b51722f92 100644 --- a/src/tools/remote-test-client/src/main.rs +++ b/src/tools/remote-test-client/src/main.rs @@ -111,7 +111,9 @@ fn prepare_rootfs(target: &str, rootfs: &Path, server: &Path, rootfs_img: &Path) "arm-unknown-linux-gnueabihf" | "aarch64-unknown-linux-gnu" => { prepare_rootfs_cpio(rootfs, rootfs_img) } - "riscv64gc-unknown-linux-gnu" => prepare_rootfs_ext4(rootfs, rootfs_img), + "riscv64a23-unknown-linux-gnu" | "riscv64gc-unknown-linux-gnu" => { + prepare_rootfs_ext4(rootfs, rootfs_img) + } _ => panic!("{} is not supported", target), } } @@ -234,6 +236,31 @@ fn start_qemu_emulator(target: &str, rootfs: &Path, server: &Path, tmpdir: &Path .arg(&format!("file={},format=raw,id=hd0", &rootfs_img.to_string_lossy())); t!(cmd.spawn()); } + "riscv64a23-unknown-linux-gnu" => { + let mut cmd = Command::new("qemu-system-riscv64"); + cmd.arg("-nographic") + .arg("-machine") + .arg("virt") + .arg("-cpu") + .arg("rva23s64") + .arg("-m") + .arg("1024") + .arg("-bios") + .arg("/tmp/fw_jump.bin") + .arg("-kernel") + .arg("/tmp/Image") + .arg("-append") + .arg("quiet console=ttyS0 root=/dev/vda rw") + .arg("-netdev") + .arg("user,id=net0,hostfwd=tcp::12345-:12345") + .arg("-device") + .arg("virtio-net-device,netdev=net0,mac=00:00:00:00:00:00") + .arg("-device") + .arg("virtio-blk-device,drive=hd0") + .arg("-drive") + .arg(&format!("file={},format=raw,id=hd0,if=none", &rootfs_img.to_string_lossy())); + t!(cmd.spawn()); + } _ => panic!("cannot start emulator for: {}", target), } } -- cgit 1.4.1-3-g733a5 From 31ae666f0f57a5fefd924a9945e5aca816c001f8 Mon Sep 17 00:00:00 2001 From: Zhongyao Chen Date: Tue, 29 Jul 2025 21:20:03 +0000 Subject: Add target page for riscv64a23-unknown-linux-gnu --- src/doc/rustc/src/SUMMARY.md | 1 + src/doc/rustc/src/platform-support.md | 1 + .../riscv64a23-unknown-linux-gnu.md | 66 ++++++++++++++++++++++ 3 files changed, 68 insertions(+) create mode 100644 src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md diff --git a/src/doc/rustc/src/SUMMARY.md b/src/doc/rustc/src/SUMMARY.md index 25f154f1180..79e229595c8 100644 --- a/src/doc/rustc/src/SUMMARY.md +++ b/src/doc/rustc/src/SUMMARY.md @@ -104,6 +104,7 @@ - [riscv32imac-unknown-xous-elf](platform-support/riscv32imac-unknown-xous-elf.md) - [riscv64gc-unknown-linux-gnu](platform-support/riscv64gc-unknown-linux-gnu.md) - [riscv64gc-unknown-linux-musl](platform-support/riscv64gc-unknown-linux-musl.md) + - [riscv64a23-unknown-linux-gnu](platform-support/riscv64a23-unknown-linux-gnu.md) - [s390x-unknown-linux-gnu](platform-support/s390x-unknown-linux-gnu.md) - [s390x-unknown-linux-musl](platform-support/s390x-unknown-linux-musl.md) - [sparc-unknown-none-elf](./platform-support/sparc-unknown-none-elf.md) diff --git a/src/doc/rustc/src/platform-support.md b/src/doc/rustc/src/platform-support.md index 8ebaa8dd874..cf55d4bd68f 100644 --- a/src/doc/rustc/src/platform-support.md +++ b/src/doc/rustc/src/platform-support.md @@ -389,6 +389,7 @@ target | std | host | notes [`riscv64gc-unknown-nuttx-elf`](platform-support/nuttx.md) | ✓ | | RISC-V 64bit with NuttX [`riscv64gc-unknown-openbsd`](platform-support/openbsd.md) | ✓ | ✓ | OpenBSD/riscv64 [`riscv64imac-unknown-nuttx-elf`](platform-support/nuttx.md) | ✓ | | RISC-V 64bit with NuttX +[`riscv64a23-unknown-linux-gnu`](platform-support/riscv64a23-unknown-linux-gnu.md) | ✓ | ✓ | RISC-V Linux (kernel 4.20+, glibc 2.29) [`s390x-unknown-linux-musl`](platform-support/s390x-unknown-linux-musl.md) | ✓ | | S390x Linux (kernel 3.2, musl 1.2.3) `sparc-unknown-linux-gnu` | ✓ | | 32-bit SPARC Linux [`sparc-unknown-none-elf`](./platform-support/sparc-unknown-none-elf.md) | * | | Bare 32-bit SPARC V7+ diff --git a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md new file mode 100644 index 00000000000..00790ba197f --- /dev/null +++ b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md @@ -0,0 +1,66 @@ +# `riscv64a23-unknown-linux-gnu` + +**Tier: 3** + +RISC-V targets using the ratified [RVA23 Profile](https://github.com/riscv/riscv-profiles/blob/main/rva23-profile.adoc). +This target will enable all mandary features of rva23u64 and rva23s64 by default. + + +## Target maintainers + +[@ZhongyaoChen](https://github.com/ZhongyaoChen) +[@CaiWeiran](https://github.com/CaiWeiran) + +## Requirements + +This target requires: + +* Linux Kernel version 4.20 or later +* glibc 2.17 or later + + +## Building the target + +The target is distributed through `rustup`, and otherwise require no +special configuration. + +If you need to build your own Rust for some reason though, the target can be build with: + +```bash +./x build --target riscv64a23-unknown-linux-gnu +``` + +## Building Rust programs + +Add the target: + +```bash +rustup target add riscv64a23-unknown-linux-gnu +``` + +Then cross compile crates with: + +```bash +cargo build --target riscv64a23-unknown-linux-gnu +``` + +## Cross-compilation toolchains and Testing + +On Ubuntu 24.04, we can install compilation dependencies with: + +```bash +apt install -y git python3 g++ g++-riscv64-linux-gnu +``` + +Then build target with: + +```bash +./x build --target=riscv64a23-unknown-linux-gnu +``` + +There are no special requirements for testing and running the targets. +For testing cross-builds on the host, you can use the docker image. It will automatically set up a RISC-V QEMU emulator and run all the test suite. + +```bash +DEPLOY=1 ./src/ci/docker/run.sh riscv64a23-gnu +``` -- cgit 1.4.1-3-g733a5 From 93a8c1d12cee1af04d84ffc59597dbd877de671f Mon Sep 17 00:00:00 2001 From: zhongyao Date: Fri, 8 Aug 2025 14:42:51 +0800 Subject: Update src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md Co-authored-by: zachs18 <8355914+zachs18@users.noreply.github.com> --- .../riscv64a23-unknown-linux-gnu.md | 36 ++++------------------ 1 file changed, 6 insertions(+), 30 deletions(-) diff --git a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md index 00790ba197f..6db4a672936 100644 --- a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md +++ b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md @@ -2,10 +2,9 @@ **Tier: 3** -RISC-V targets using the ratified [RVA23 Profile](https://github.com/riscv/riscv-profiles/blob/main/rva23-profile.adoc). +RISC-V target using the ratified [RVA23 Profile](https://github.com/riscv/riscv-profiles/blob/main/src/rva23-profile.adoc). This target will enable all mandary features of rva23u64 and rva23s64 by default. - ## Target maintainers [@ZhongyaoChen](https://github.com/ZhongyaoChen) @@ -18,13 +17,11 @@ This target requires: * Linux Kernel version 4.20 or later * glibc 2.17 or later - ## Building the target -The target is distributed through `rustup`, and otherwise require no -special configuration. +Tier-3 target is not distributed through `rustup`. -If you need to build your own Rust for some reason though, the target can be build with: +You need to build your own Rust, the target can be build with: ```bash ./x build --target riscv64a23-unknown-linux-gnu @@ -32,35 +29,14 @@ If you need to build your own Rust for some reason though, the target can be bui ## Building Rust programs -Add the target: +Add the toolchain: ```bash -rustup target add riscv64a23-unknown-linux-gnu +rustup toolchain link rva23-toolchain {path-to-rust}/build/host/stage2 ``` Then cross compile crates with: ```bash -cargo build --target riscv64a23-unknown-linux-gnu -``` - -## Cross-compilation toolchains and Testing - -On Ubuntu 24.04, we can install compilation dependencies with: - -```bash -apt install -y git python3 g++ g++-riscv64-linux-gnu -``` - -Then build target with: - -```bash -./x build --target=riscv64a23-unknown-linux-gnu -``` - -There are no special requirements for testing and running the targets. -For testing cross-builds on the host, you can use the docker image. It will automatically set up a RISC-V QEMU emulator and run all the test suite. - -```bash -DEPLOY=1 ./src/ci/docker/run.sh riscv64a23-gnu +RUSTFLAGS="-C linker=riscv64-linux-gnu-gcc" cargo +rva23-toolchain build --target=riscv64a23-unknown-linux-gnu ``` -- cgit 1.4.1-3-g733a5 From 87fd28998c92a014935d879ae0129b09baa56588 Mon Sep 17 00:00:00 2001 From: Zhongyao Chen Date: Sat, 9 Aug 2025 12:27:38 +0000 Subject: Revert "Add a disabled builder for riscv64 rva23 emulated tests" This reverts commit 289688ca494ddc45bcd0b0706dc173c19fbb7caa. --- .../host-x86_64/disabled/riscv64a23-gnu/Dockerfile | 108 --------------------- .../disabled/riscv64a23-gnu/linux.config | 51 ---------- src/tools/build-manifest/src/main.rs | 2 - src/tools/remote-test-client/src/main.rs | 29 +----- 4 files changed, 1 insertion(+), 189 deletions(-) delete mode 100644 src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile delete mode 100644 src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config diff --git a/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile b/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile deleted file mode 100644 index 86b0d795687..00000000000 --- a/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/Dockerfile +++ /dev/null @@ -1,108 +0,0 @@ -FROM ubuntu:24.04 - -ARG DEBIAN_FRONTEND=noninteractive -RUN apt-get update -y && apt-get install -y --no-install-recommends \ - bc \ - bzip2 \ - ca-certificates \ - cmake \ - cpio \ - curl \ - file \ - flex \ - bison \ - g++ \ - g++-riscv64-linux-gnu \ - git \ - libc6-dev \ - libc6-dev-riscv64-cross \ - libssl-dev \ - make \ - ninja-build \ - python3 \ - xz-utils \ - opensbi \ - u-boot-qemu \ - libslirp0 \ - build-essential \ - pkg-config \ - libglib2.0-dev \ - libpixman-1-dev \ - libsdl2-dev \ - libfdt-dev \ - python3 \ - python3-pip - -ENV ARCH=riscv \ - CROSS_COMPILE=riscv64-linux-gnu- - -WORKDIR /build - -# From https://github.com/michaeljclark/busybear-linux/blob/master/conf/linux.config -COPY host-x86_64/riscv64a23-gnu/linux.config /build - -# qemu v10.0.2 fully support -RUN curl https://gitlab.com/qemu-project/qemu/-/archive/v10.0.2/qemu-v10.0.2.tar.bz2 | tar xjf - && \ - cd qemu-v10.0.2 && \ - ./configure --target-list=riscv64-softmmu \ - --enable-sdl --enable-debug --enable-fdt --enable-slirp && \ - make -j && make install - -# use the opensbi fw from apt-get install -RUN cp /usr/lib/riscv64-linux-gnu/opensbi/generic/fw_jump.bin /tmp - -# Compile the kernel that we're going to be emulating with. This is -# basically just done to be compatible with the QEMU target that we're going -# to be using when running tests. -RUN curl https://cdn.kernel.org/pub/linux/kernel/v6.x/linux-6.6.97.tar.xz | tar xJf - && \ - cp linux.config linux-6.6.97/.config && \ - cd /build/linux-6.6.97 && \ - make olddefconfig && \ - make -j$(nproc) Image && \ - cp arch/riscv/boot/Image /tmp && \ - rm -rf linux-6.6.97 - -# Compile an instance of busybox as this provides a lightweight system and init -# binary which we will boot into. Only trick here is configuring busybox to -# build static binaries. -RUN curl https://www.busybox.net/downloads/busybox-1.37.0.tar.bz2 | tar xjf - && \ - cd busybox-1.37.0 && \ - make defconfig && \ - sed -i 's/# CONFIG_STATIC is not set/CONFIG_STATIC=y/' .config && \ - sed -i 's/^CONFIG_TC=y$/# CONFIG_TC is not set/' .config && \ - sed -i 's/CONFIG_SHA1_HWACCEL=y/# CONFIG_SHA1_HWACCEL is not set/' .config && \ - sed -i 's/CONFIG_SHA256_HWACCEL=y/# CONFIG_SHA256_HWACCEL is not set/' .config && \ - make -j$(nproc) && \ - make install && \ - mv _install /tmp/rootfs && \ - cd /build && \ - rm -rf busybox-1.37.0 - -# Download the ubuntu rootfs, which we'll use as a chroot for all our tests. -WORKDIR /tmp -RUN mkdir rootfs/ubuntu -RUN curl https://cdimage.ubuntu.com/ubuntu-base/releases/24.04/release/ubuntu-base-24.04.2-base-riscv64.tar.gz | \ - tar xzf - -C rootfs/ubuntu && \ - cd rootfs && mkdir proc sys dev etc etc/init.d - -# Copy over our init script, which starts up our test server and also a few other -# misc tasks -COPY scripts/qemu-bare-bones-rcS rootfs/etc/init.d/rcS -RUN chmod +x rootfs/etc/init.d/rcS - -# Helper to quickly fill the entropy pool in the kernel -COPY scripts/qemu-bare-bones-addentropy.c /tmp/addentropy.c -RUN riscv64-linux-gnu-gcc addentropy.c -o rootfs/addentropy -static - -COPY scripts/sccache.sh /scripts/ -RUN sh /scripts/sccache.sh - -# Avoid "fatal: detected dubious ownership in repository at '/checkout'" error -RUN git config --global --add safe.directory "*" - -ENV RUST_CONFIGURE_ARGS \ - --set target.riscv64a23-unknown-linux-gnu.linker=riscv64-linux-gnu-gcc \ - --set target.riscv64a23-unknown-linux-gnu.qemu-rootfs=/tmp/rootfs -ENV SCRIPT python3 ../x.py --stage 2 test --host='' --target riscv64a23-unknown-linux-gnu - -ENV NO_CHANGE_USER=1 diff --git a/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config b/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config deleted file mode 100644 index 5142664742f..00000000000 --- a/src/ci/docker/host-x86_64/disabled/riscv64a23-gnu/linux.config +++ /dev/null @@ -1,51 +0,0 @@ -CONFIG_DEFAULT_HOSTNAME="busybear" -CONFIG_SYSVIPC=y -CONFIG_POSIX_MQUEUE=y -CONFIG_IKCONFIG=y -CONFIG_IKCONFIG_PROC=y -CONFIG_CGROUPS=y -CONFIG_CGROUP_SCHED=y -CONFIG_CFS_BANDWIDTH=y -CONFIG_CGROUP_BPF=y -CONFIG_NAMESPACES=y -CONFIG_USER_NS=y -CONFIG_CHECKPOINT_RESTORE=y -CONFIG_BLK_DEV_INITRD=y -CONFIG_EXPERT=y -CONFIG_BPF_SYSCALL=y -CONFIG_SMP=y -CONFIG_MODULES=y -CONFIG_NET=y -CONFIG_PACKET=y -CONFIG_PACKET_DIAG=y -CONFIG_UNIX=y -CONFIG_INET=y -CONFIG_NETLINK_DIAG=y -# CONFIG_WIRELESS is not set -CONFIG_PCI=y -CONFIG_DEVTMPFS=y -CONFIG_BLK_DEV_LOOP=y -CONFIG_VIRTIO_BLK=y -CONFIG_NETDEVICES=y -CONFIG_VIRTIO_NET=y -# CONFIG_ETHERNET is not set -# CONFIG_WLAN is not set -CONFIG_SERIAL_8250=y -CONFIG_SERIAL_8250_CONSOLE=y -CONFIG_SERIAL_OF_PLATFORM=y -CONFIG_HVC_RISCV_SBI=y -# CONFIG_HW_RANDOM is not set -# CONFIG_USB_SUPPORT is not set -CONFIG_VIRTIO_MMIO=y -CONFIG_SIFIVE_PLIC=y -CONFIG_RAS=y -CONFIG_EXT2_FS=y -CONFIG_EXT3_FS=y -CONFIG_EXT4_FS_POSIX_ACL=y -CONFIG_AUTOFS4_FS=y -CONFIG_MSDOS_FS=y -CONFIG_VFAT_FS=y -CONFIG_TMPFS=y -# CONFIG_CRYPTO_ECHAINIV is not set -# CONFIG_CRYPTO_HW is not set -CONFIG_PRINTK_TIME=y diff --git a/src/tools/build-manifest/src/main.rs b/src/tools/build-manifest/src/main.rs index 4f26c170de0..0520eff0fa2 100644 --- a/src/tools/build-manifest/src/main.rs +++ b/src/tools/build-manifest/src/main.rs @@ -40,7 +40,6 @@ static HOSTS: &[&str] = &[ "powerpc64le-unknown-linux-gnu", "powerpc64le-unknown-linux-musl", "riscv64gc-unknown-linux-gnu", - "riscv64a23-unknown-linux-gnu", "s390x-unknown-linux-gnu", "sparcv9-sun-solaris", "x86_64-apple-darwin", @@ -156,7 +155,6 @@ static TARGETS: &[&str] = &[ "riscv64gc-unknown-none-elf", "riscv64gc-unknown-linux-gnu", "riscv64gc-unknown-linux-musl", - "riscv64a23-unknown-linux-gnu", "s390x-unknown-linux-gnu", "sparc64-unknown-linux-gnu", "sparcv9-sun-solaris", diff --git a/src/tools/remote-test-client/src/main.rs b/src/tools/remote-test-client/src/main.rs index e5b51722f92..b9741431b50 100644 --- a/src/tools/remote-test-client/src/main.rs +++ b/src/tools/remote-test-client/src/main.rs @@ -111,9 +111,7 @@ fn prepare_rootfs(target: &str, rootfs: &Path, server: &Path, rootfs_img: &Path) "arm-unknown-linux-gnueabihf" | "aarch64-unknown-linux-gnu" => { prepare_rootfs_cpio(rootfs, rootfs_img) } - "riscv64a23-unknown-linux-gnu" | "riscv64gc-unknown-linux-gnu" => { - prepare_rootfs_ext4(rootfs, rootfs_img) - } + "riscv64gc-unknown-linux-gnu" => prepare_rootfs_ext4(rootfs, rootfs_img), _ => panic!("{} is not supported", target), } } @@ -236,31 +234,6 @@ fn start_qemu_emulator(target: &str, rootfs: &Path, server: &Path, tmpdir: &Path .arg(&format!("file={},format=raw,id=hd0", &rootfs_img.to_string_lossy())); t!(cmd.spawn()); } - "riscv64a23-unknown-linux-gnu" => { - let mut cmd = Command::new("qemu-system-riscv64"); - cmd.arg("-nographic") - .arg("-machine") - .arg("virt") - .arg("-cpu") - .arg("rva23s64") - .arg("-m") - .arg("1024") - .arg("-bios") - .arg("/tmp/fw_jump.bin") - .arg("-kernel") - .arg("/tmp/Image") - .arg("-append") - .arg("quiet console=ttyS0 root=/dev/vda rw") - .arg("-netdev") - .arg("user,id=net0,hostfwd=tcp::12345-:12345") - .arg("-device") - .arg("virtio-net-device,netdev=net0,mac=00:00:00:00:00:00") - .arg("-device") - .arg("virtio-blk-device,drive=hd0") - .arg("-drive") - .arg(&format!("file={},format=raw,id=hd0,if=none", &rootfs_img.to_string_lossy())); - t!(cmd.spawn()); - } _ => panic!("cannot start emulator for: {}", target), } } -- cgit 1.4.1-3-g733a5 From bbf08d87eb92f618120b6517651588950780b3e4 Mon Sep 17 00:00:00 2001 From: dianne Date: Tue, 12 Aug 2025 23:55:29 -0700 Subject: add a test with incorrect `if let`-`super let` drop order --- tests/ui/drop/if-let-super-let.rs | 112 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 112 insertions(+) create mode 100644 tests/ui/drop/if-let-super-let.rs diff --git a/tests/ui/drop/if-let-super-let.rs b/tests/ui/drop/if-let-super-let.rs new file mode 100644 index 00000000000..0f966e793a3 --- /dev/null +++ b/tests/ui/drop/if-let-super-let.rs @@ -0,0 +1,112 @@ +//! Test for #145328: ensure the lifetime of a `super let` binding within an `if let` scrutinee is +//! at most the scope of the `if` condition's temporaries. Additionally, test `pin!` since it's +//! implemented in terms of `super let` and exposes this behavior. +//@ run-pass +//@ revisions: e2021 e2024 +//@ [e2021] edition: 2021 +//@ [e2024] edition: 2024 + +#![feature(if_let_guard)] +#![feature(super_let)] +#![expect(irrefutable_let_patterns)] + +use std::cell::RefCell; +use std::pin::pin; + +fn main() { + // The `super let` bindings here should have the same scope as `if let` temporaries. + // In Rust 2021, this means it lives past the end of the `if` expression. + // In Rust 2024, this means it lives to the end of the `if`'s success block. + assert_drop_order(0..=2, |o| { + #[cfg(e2021)] + ( + if let _ = { super let _x = o.log(2); } { o.push(0) }, + o.push(1), + ); + #[cfg(e2024)] + ( + if let _ = { super let _x = o.log(2); } { o.push(0) }, + o.push(1), + ); + }); + assert_drop_order(0..=2, |o| { + #[cfg(e2021)] + ( + if let true = { super let _x = o.log(2); false } {} else { o.push(0) }, + o.push(1), + ); + #[cfg(e2024)] + ( + if let true = { super let _x = o.log(2); false } {} else { o.push(0) }, + o.push(1), + ); + }); + + // `pin!` should behave likewise. + assert_drop_order(0..=2, |o| { + #[cfg(e2021)] (if let _ = pin!(o.log(2)) { o.push(0) }, o.push(1)); + #[cfg(e2024)] (if let _ = pin!(o.log(2)) { o.push(0) }, o.push(1)); + }); + assert_drop_order(0..=2, |o| { + #[cfg(e2021)] + ( + if let None = Some(pin!(o.log(2))) {} else { o.push(0) }, + o.push(1), + ); + #[cfg(e2024)] + ( + if let None = Some(pin!(o.log(2))) {} else { o.push(0) }, + o.push(1), + ); + }); + + // `super let` bindings' scope should also be consistent with `if let` temporaries in guards. + // Here, that means the `super let` binding in the second guard condition operand should be + // dropped before the first operand's temporary. This is consistent across Editions. + assert_drop_order(0..=1, |o| { + match () { + _ if let _ = o.log(0) + && let _ = { super let _x = o.log(1); } => {} + _ => unreachable!(), + } + }); + assert_drop_order(0..=1, |o| { + match () { + _ if let _ = o.log(0) + && let _ = pin!(o.log(1)) => {} + _ => unreachable!(), + } + }); +} + +// # Test scaffolding... + +struct DropOrder(RefCell>); +struct LogDrop<'o>(&'o DropOrder, u64); + +impl DropOrder { + fn log(&self, n: u64) -> LogDrop<'_> { + LogDrop(self, n) + } + fn push(&self, n: u64) { + self.0.borrow_mut().push(n); + } +} + +impl<'o> Drop for LogDrop<'o> { + fn drop(&mut self) { + self.0.push(self.1); + } +} + +#[track_caller] +fn assert_drop_order( + ex: impl IntoIterator, + f: impl Fn(&DropOrder), +) { + let order = DropOrder(RefCell::new(Vec::new())); + f(&order); + let order = order.0.into_inner(); + let expected: Vec = ex.into_iter().collect(); + assert_eq!(order, expected); +} -- cgit 1.4.1-3-g733a5 From 8fc3938d9961568dfec38b10506f33fd9bc7c147 Mon Sep 17 00:00:00 2001 From: dianne Date: Wed, 13 Aug 2025 01:02:13 -0700 Subject: fix scope of `super let` bindings within `if let` They now use the enclosing temporary scope as their scope, regardless of which `ScopeData` was used to mark it. --- compiler/rustc_hir_analysis/src/check/region.rs | 8 ++--- compiler/rustc_middle/src/middle/region.rs | 39 +++++++++++++++++++++++++ compiler/rustc_middle/src/ty/rvalue_scopes.rs | 37 ++--------------------- tests/ui/drop/if-let-super-let.rs | 22 +++++++------- 4 files changed, 54 insertions(+), 52 deletions(-) diff --git a/compiler/rustc_hir_analysis/src/check/region.rs b/compiler/rustc_hir_analysis/src/check/region.rs index f5770b7312d..2ba7ed46f92 100644 --- a/compiler/rustc_hir_analysis/src/check/region.rs +++ b/compiler/rustc_hir_analysis/src/check/region.rs @@ -490,12 +490,8 @@ fn resolve_local<'tcx>( // // Iterate up to the enclosing destruction scope to find the same scope that will also // be used for the result of the block itself. - while let Some(s) = visitor.cx.var_parent { - let parent = visitor.scope_tree.parent_map.get(&s).cloned(); - if let Some(Scope { data: ScopeData::Destruction, .. }) = parent { - break; - } - visitor.cx.var_parent = parent; + if let Some(inner_scope) = visitor.cx.var_parent { + (visitor.cx.var_parent, _) = visitor.scope_tree.default_temporary_scope(inner_scope) } } } diff --git a/compiler/rustc_middle/src/middle/region.rs b/compiler/rustc_middle/src/middle/region.rs index 857d041224f..5367e5edd49 100644 --- a/compiler/rustc_middle/src/middle/region.rs +++ b/compiler/rustc_middle/src/middle/region.rs @@ -299,4 +299,43 @@ impl ScopeTree { true } + + /// Returns the scope of non-lifetime-extended temporaries within a given scope, as well as + /// whether we've recorded a potential backwards-incompatible change to lint on. + /// Returns `None` when no enclosing temporary scope is found, such as for static items. + pub fn default_temporary_scope(&self, inner: Scope) -> (Option, Option) { + let mut id = inner; + let mut backwards_incompatible = None; + + while let Some(&p) = self.parent_map.get(&id) { + match p.data { + ScopeData::Destruction => { + debug!("temporary_scope({inner:?}) = {id:?} [enclosing]"); + return (Some(id), backwards_incompatible); + } + ScopeData::IfThenRescope | ScopeData::MatchGuard => { + debug!("temporary_scope({inner:?}) = {p:?} [enclosing]"); + return (Some(p), backwards_incompatible); + } + ScopeData::Node + | ScopeData::CallSite + | ScopeData::Arguments + | ScopeData::IfThen + | ScopeData::Remainder(_) => { + // If we haven't already passed through a backwards-incompatible node, + // then check if we are passing through one now and record it if so. + // This is for now only working for cases where a temporary lifetime is + // *shortened*. + if backwards_incompatible.is_none() { + backwards_incompatible = + self.backwards_incompatible_scope.get(&p.local_id).copied(); + } + id = p + } + } + } + + debug!("temporary_scope({inner:?}) = None"); + (None, backwards_incompatible) + } } diff --git a/compiler/rustc_middle/src/ty/rvalue_scopes.rs b/compiler/rustc_middle/src/ty/rvalue_scopes.rs index 7dfe2d28051..8b92e48ed1a 100644 --- a/compiler/rustc_middle/src/ty/rvalue_scopes.rs +++ b/compiler/rustc_middle/src/ty/rvalue_scopes.rs @@ -35,41 +35,8 @@ impl RvalueScopes { // if there's one. Static items, for instance, won't // have an enclosing scope, hence no scope will be // returned. - let mut id = Scope { local_id: expr_id, data: ScopeData::Node }; - let mut backwards_incompatible = None; - - while let Some(&p) = region_scope_tree.parent_map.get(&id) { - match p.data { - ScopeData::Destruction => { - debug!("temporary_scope({expr_id:?}) = {id:?} [enclosing]"); - return (Some(id), backwards_incompatible); - } - ScopeData::IfThenRescope | ScopeData::MatchGuard => { - debug!("temporary_scope({expr_id:?}) = {p:?} [enclosing]"); - return (Some(p), backwards_incompatible); - } - ScopeData::Node - | ScopeData::CallSite - | ScopeData::Arguments - | ScopeData::IfThen - | ScopeData::Remainder(_) => { - // If we haven't already passed through a backwards-incompatible node, - // then check if we are passing through one now and record it if so. - // This is for now only working for cases where a temporary lifetime is - // *shortened*. - if backwards_incompatible.is_none() { - backwards_incompatible = region_scope_tree - .backwards_incompatible_scope - .get(&p.local_id) - .copied(); - } - id = p - } - } - } - - debug!("temporary_scope({expr_id:?}) = None"); - (None, backwards_incompatible) + region_scope_tree + .default_temporary_scope(Scope { local_id: expr_id, data: ScopeData::Node }) } /// Make an association between a sub-expression and an extended lifetime diff --git a/tests/ui/drop/if-let-super-let.rs b/tests/ui/drop/if-let-super-let.rs index 0f966e793a3..c6543e6d3dc 100644 --- a/tests/ui/drop/if-let-super-let.rs +++ b/tests/ui/drop/if-let-super-let.rs @@ -25,8 +25,8 @@ fn main() { ); #[cfg(e2024)] ( - if let _ = { super let _x = o.log(2); } { o.push(0) }, - o.push(1), + if let _ = { super let _x = o.log(1); } { o.push(0) }, + o.push(2), ); }); assert_drop_order(0..=2, |o| { @@ -37,15 +37,15 @@ fn main() { ); #[cfg(e2024)] ( - if let true = { super let _x = o.log(2); false } {} else { o.push(0) }, - o.push(1), + if let true = { super let _x = o.log(0); false } {} else { o.push(1) }, + o.push(2), ); }); // `pin!` should behave likewise. assert_drop_order(0..=2, |o| { #[cfg(e2021)] (if let _ = pin!(o.log(2)) { o.push(0) }, o.push(1)); - #[cfg(e2024)] (if let _ = pin!(o.log(2)) { o.push(0) }, o.push(1)); + #[cfg(e2024)] (if let _ = pin!(o.log(1)) { o.push(0) }, o.push(2)); }); assert_drop_order(0..=2, |o| { #[cfg(e2021)] @@ -55,8 +55,8 @@ fn main() { ); #[cfg(e2024)] ( - if let None = Some(pin!(o.log(2))) {} else { o.push(0) }, - o.push(1), + if let None = Some(pin!(o.log(0))) {} else { o.push(1) }, + o.push(2), ); }); @@ -65,15 +65,15 @@ fn main() { // dropped before the first operand's temporary. This is consistent across Editions. assert_drop_order(0..=1, |o| { match () { - _ if let _ = o.log(0) - && let _ = { super let _x = o.log(1); } => {} + _ if let _ = o.log(1) + && let _ = { super let _x = o.log(0); } => {} _ => unreachable!(), } }); assert_drop_order(0..=1, |o| { match () { - _ if let _ = o.log(0) - && let _ = pin!(o.log(1)) => {} + _ if let _ = o.log(1) + && let _ = pin!(o.log(0)) => {} _ => unreachable!(), } }); -- cgit 1.4.1-3-g733a5 From 52063ae24860bb5cb3abc6f3dd83904bc18a3b0f Mon Sep 17 00:00:00 2001 From: Hanna Kruppe Date: Sun, 10 Aug 2025 11:21:43 +0200 Subject: stabilize path_add_extension --- library/std/src/path.rs | 8 ++------ library/std/tests/path.rs | 2 +- 2 files changed, 3 insertions(+), 7 deletions(-) diff --git a/library/std/src/path.rs b/library/std/src/path.rs index 3b52804d6be..2af6249f3b9 100644 --- a/library/std/src/path.rs +++ b/library/std/src/path.rs @@ -1575,8 +1575,6 @@ impl PathBuf { /// # Examples /// /// ``` - /// #![feature(path_add_extension)] - /// /// use std::path::{Path, PathBuf}; /// /// let mut p = PathBuf::from("/feel/the"); @@ -1596,7 +1594,7 @@ impl PathBuf { /// p.add_extension(""); /// assert_eq!(Path::new("/feel/the.formatted.dark"), p.as_path()); /// ``` - #[unstable(feature = "path_add_extension", issue = "127292")] + #[stable(feature = "path_add_extension", since = "CURRENT_RUSTC_VERSION")] pub fn add_extension>(&mut self, extension: S) -> bool { self._add_extension(extension.as_ref()) } @@ -2846,8 +2844,6 @@ impl Path { /// # Examples /// /// ``` - /// #![feature(path_add_extension)] - /// /// use std::path::{Path, PathBuf}; /// /// let path = Path::new("foo.rs"); @@ -2858,7 +2854,7 @@ impl Path { /// assert_eq!(path.with_added_extension("xz"), PathBuf::from("foo.tar.gz.xz")); /// assert_eq!(path.with_added_extension("").with_added_extension("txt"), PathBuf::from("foo.tar.gz.txt")); /// ``` - #[unstable(feature = "path_add_extension", issue = "127292")] + #[stable(feature = "path_add_extension", since = "CURRENT_RUSTC_VERSION")] pub fn with_added_extension>(&self, extension: S) -> PathBuf { let mut new_path = self.to_path_buf(); new_path.add_extension(extension); diff --git a/library/std/tests/path.rs b/library/std/tests/path.rs index e1576a0d423..3577f0d9c7b 100644 --- a/library/std/tests/path.rs +++ b/library/std/tests/path.rs @@ -1,4 +1,4 @@ -#![feature(clone_to_uninit, path_add_extension, maybe_uninit_slice, normalize_lexically)] +#![feature(clone_to_uninit, maybe_uninit_slice, normalize_lexically)] use std::clone::CloneToUninit; use std::ffi::OsStr; -- cgit 1.4.1-3-g733a5 From 27beb72e87a328a64eec5a1e3433133d072aebe3 Mon Sep 17 00:00:00 2001 From: Sasha Pourcelot Date: Tue, 12 Aug 2025 20:22:45 +0200 Subject: Port the `#[linkage]` attribute to the new attribute system --- src/base.rs | 2 +- src/consts.rs | 2 +- src/mono_item.rs | 3 ++- 3 files changed, 4 insertions(+), 3 deletions(-) diff --git a/src/base.rs b/src/base.rs index c105916bbb2..e9d72e457a0 100644 --- a/src/base.rs +++ b/src/base.rs @@ -8,8 +8,8 @@ use rustc_codegen_ssa::ModuleCodegen; use rustc_codegen_ssa::base::maybe_create_entry_wrapper; use rustc_codegen_ssa::mono_item::MonoItemExt; use rustc_codegen_ssa::traits::DebugInfoCodegenMethods; +use rustc_hir::attrs::Linkage; use rustc_middle::dep_graph; -use rustc_middle::mir::mono::Linkage; #[cfg(feature = "master")] use rustc_middle::mir::mono::Visibility; use rustc_middle::ty::TyCtxt; diff --git a/src/consts.rs b/src/consts.rs index 873f1f1951c..619277eba8b 100644 --- a/src/consts.rs +++ b/src/consts.rs @@ -5,13 +5,13 @@ use rustc_abi::{self as abi, Align, HasDataLayout, Primitive, Size, WrappingRang use rustc_codegen_ssa::traits::{ BaseTypeCodegenMethods, ConstCodegenMethods, StaticCodegenMethods, }; +use rustc_hir::attrs::Linkage; use rustc_hir::def::DefKind; use rustc_hir::def_id::LOCAL_CRATE; use rustc_middle::middle::codegen_fn_attrs::{CodegenFnAttrFlags, CodegenFnAttrs}; use rustc_middle::mir::interpret::{ self, ConstAllocation, ErrorHandled, Scalar as InterpScalar, read_target_uint, }; -use rustc_middle::mir::mono::Linkage; use rustc_middle::ty::layout::LayoutOf; use rustc_middle::ty::{self, Instance}; use rustc_middle::{bug, span_bug}; diff --git a/src/mono_item.rs b/src/mono_item.rs index ff188c437da..35d44d21bcb 100644 --- a/src/mono_item.rs +++ b/src/mono_item.rs @@ -1,11 +1,12 @@ #[cfg(feature = "master")] use gccjit::{FnAttribute, VarAttribute}; use rustc_codegen_ssa::traits::PreDefineCodegenMethods; +use rustc_hir::attrs::Linkage; use rustc_hir::def::DefKind; use rustc_hir::def_id::{DefId, LOCAL_CRATE}; use rustc_middle::bug; use rustc_middle::middle::codegen_fn_attrs::CodegenFnAttrFlags; -use rustc_middle::mir::mono::{Linkage, Visibility}; +use rustc_middle::mir::mono::Visibility; use rustc_middle::ty::layout::{FnAbiOf, HasTypingEnv, LayoutOf}; use rustc_middle::ty::{self, Instance, TypeVisitableExt}; -- cgit 1.4.1-3-g733a5 From 0170d6cecd1e8592774ccd94dd2a5bf30ff2495a Mon Sep 17 00:00:00 2001 From: Daniel Wagner-Hall Date: Thu, 14 Aug 2025 00:53:34 +0100 Subject: or_then_unwrap suggestion preserves macro calls Before this change, the suggestion for `Option.or(Some(vec![])).unwrap()` expanded the `vec!` macro which broke the code, both in terms of readability, and because the expansion references `$crate` which cannot be inlined. --- clippy_lints/src/methods/or_then_unwrap.rs | 2 +- tests/ui/or_then_unwrap.fixed | 6 ++++++ tests/ui/or_then_unwrap.rs | 6 ++++++ tests/ui/or_then_unwrap.stderr | 10 ++++++++-- 4 files changed, 21 insertions(+), 3 deletions(-) diff --git a/clippy_lints/src/methods/or_then_unwrap.rs b/clippy_lints/src/methods/or_then_unwrap.rs index 3e64e15dc86..1a760ea733d 100644 --- a/clippy_lints/src/methods/or_then_unwrap.rs +++ b/clippy_lints/src/methods/or_then_unwrap.rs @@ -62,7 +62,7 @@ fn get_content_if_ctor_matches(cx: &LateContext<'_>, expr: &Expr<'_>, item: Lang if let ExprKind::Call(some_expr, [arg]) = expr.kind && is_res_lang_ctor(cx, path_res(cx, some_expr), item) { - Some(arg.span) + Some(arg.span.source_callsite()) } else { None } diff --git a/tests/ui/or_then_unwrap.fixed b/tests/ui/or_then_unwrap.fixed index ba9beef57af..9660b82fe7d 100644 --- a/tests/ui/or_then_unwrap.fixed +++ b/tests/ui/or_then_unwrap.fixed @@ -28,6 +28,12 @@ fn main() { // //~^^ or_then_unwrap + // Call with macro should preserve the macro call rather than expand it + let option: Option> = None; + let _ = option.unwrap_or(vec!["fallback"]); // should trigger lint + // + //~^^ or_then_unwrap + // as part of a method chain let option: Option<&str> = None; let _ = option.map(|v| v).unwrap_or("fallback").to_string().chars(); // should trigger lint diff --git a/tests/ui/or_then_unwrap.rs b/tests/ui/or_then_unwrap.rs index fac90249a24..c3873352116 100644 --- a/tests/ui/or_then_unwrap.rs +++ b/tests/ui/or_then_unwrap.rs @@ -28,6 +28,12 @@ fn main() { // //~^^ or_then_unwrap + // Call with macro should preserve the macro call rather than expand it + let option: Option> = None; + let _ = option.or(Some(vec!["fallback"])).unwrap(); // should trigger lint + // + //~^^ or_then_unwrap + // as part of a method chain let option: Option<&str> = None; let _ = option.map(|v| v).or(Some("fallback")).unwrap().to_string().chars(); // should trigger lint diff --git a/tests/ui/or_then_unwrap.stderr b/tests/ui/or_then_unwrap.stderr index 1160498c605..3e66b15edbd 100644 --- a/tests/ui/or_then_unwrap.stderr +++ b/tests/ui/or_then_unwrap.stderr @@ -14,10 +14,16 @@ LL | let _ = result.or::<&str>(Ok("fallback")).unwrap(); // should trigger l | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: try: `unwrap_or("fallback")` error: found `.or(Some(…)).unwrap()` - --> tests/ui/or_then_unwrap.rs:33:31 + --> tests/ui/or_then_unwrap.rs:33:20 + | +LL | let _ = option.or(Some(vec!["fallback"])).unwrap(); // should trigger lint + | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: try: `unwrap_or(vec!["fallback"])` + +error: found `.or(Some(…)).unwrap()` + --> tests/ui/or_then_unwrap.rs:39:31 | LL | let _ = option.map(|v| v).or(Some("fallback")).unwrap().to_string().chars(); // should trigger lint | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: try: `unwrap_or("fallback")` -error: aborting due to 3 previous errors +error: aborting due to 4 previous errors -- cgit 1.4.1-3-g733a5 From 12b57942d646d27e8d0a615ba4b8a0dea6cf3b9f Mon Sep 17 00:00:00 2001 From: Zachary S Date: Tue, 12 Aug 2025 17:42:43 -0500 Subject: Remove TmpLayout in layout_of_enum --- compiler/rustc_abi/src/layout.rs | 39 +++++++++------------------------------ 1 file changed, 9 insertions(+), 30 deletions(-) diff --git a/compiler/rustc_abi/src/layout.rs b/compiler/rustc_abi/src/layout.rs index c2405553756..5004d0c8022 100644 --- a/compiler/rustc_abi/src/layout.rs +++ b/compiler/rustc_abi/src/layout.rs @@ -594,23 +594,13 @@ impl LayoutCalculator { discr_range_of_repr: impl Fn(i128, i128) -> (Integer, bool), discriminants: impl Iterator, ) -> LayoutCalculatorResult { - // Until we've decided whether to use the tagged or - // niche filling LayoutData, we don't want to intern the - // variant layouts, so we can't store them in the - // overall LayoutData. Store the overall LayoutData - // and the variant LayoutDatas here until then. - struct TmpLayout { - layout: LayoutData, - variants: IndexVec>, - } - let dl = self.cx.data_layout(); // bail if the enum has an incoherent repr that cannot be computed if repr.packed() { return Err(LayoutCalculatorError::ReprConflict); } - let calculate_niche_filling_layout = || -> Option> { + let calculate_niche_filling_layout = || -> Option> { if repr.inhibit_enum_layout_opt() { return None; } @@ -746,7 +736,7 @@ impl LayoutCalculator { niche_start, }, tag_field: FieldIdx::new(0), - variants: IndexVec::new(), + variants: variant_layouts, }, fields: FieldsShape::Arbitrary { offsets: [niche_offset].into(), @@ -762,7 +752,7 @@ impl LayoutCalculator { randomization_seed: combined_seed, }; - Some(TmpLayout { layout, variants: variant_layouts }) + Some(layout) }; let niche_filling_layout = calculate_niche_filling_layout(); @@ -1093,7 +1083,7 @@ impl LayoutCalculator { tag, tag_encoding: TagEncoding::Direct, tag_field: FieldIdx::new(0), - variants: IndexVec::new(), + variants: layout_variants, }, fields: FieldsShape::Arbitrary { offsets: [Size::ZERO].into(), @@ -1109,18 +1099,16 @@ impl LayoutCalculator { randomization_seed: combined_seed, }; - let tagged_layout = TmpLayout { layout: tagged_layout, variants: layout_variants }; - - let mut best_layout = match (tagged_layout, niche_filling_layout) { + let best_layout = match (tagged_layout, niche_filling_layout) { (tl, Some(nl)) => { // Pick the smaller layout; otherwise, // pick the layout with the larger niche; otherwise, // pick tagged as it has simpler codegen. use cmp::Ordering::*; - let niche_size = |tmp_l: &TmpLayout| { - tmp_l.layout.largest_niche.map_or(0, |n| n.available(dl)) + let niche_size = |l: &LayoutData| { + l.largest_niche.map_or(0, |n| n.available(dl)) }; - match (tl.layout.size.cmp(&nl.layout.size), niche_size(&tl).cmp(&niche_size(&nl))) { + match (tl.size.cmp(&nl.size), niche_size(&tl).cmp(&niche_size(&nl))) { (Greater, _) => nl, (Equal, Less) => nl, _ => tl, @@ -1129,16 +1117,7 @@ impl LayoutCalculator { (tl, None) => tl, }; - // Now we can intern the variant layouts and store them in the enum layout. - best_layout.layout.variants = match best_layout.layout.variants { - Variants::Multiple { tag, tag_encoding, tag_field, .. } => { - Variants::Multiple { tag, tag_encoding, tag_field, variants: best_layout.variants } - } - Variants::Single { .. } | Variants::Empty => { - panic!("encountered a single-variant or empty enum during multi-variant layout") - } - }; - Ok(best_layout.layout) + Ok(best_layout) } fn univariant_biased< -- cgit 1.4.1-3-g733a5 From 26183d574197b2605a559097e604f4906c4892fc Mon Sep 17 00:00:00 2001 From: Antoni Boucher Date: Thu, 14 Aug 2025 09:05:09 -0400 Subject: Update to nightly-2025-08-14 --- rust-toolchain | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/rust-toolchain b/rust-toolchain index 058e734be5c..059d469e8d4 100644 --- a/rust-toolchain +++ b/rust-toolchain @@ -1,3 +1,3 @@ [toolchain] -channel = "nightly-2025-08-03" +channel = "nightly-2025-08-14" components = ["rust-src", "rustc-dev", "llvm-tools-preview"] -- cgit 1.4.1-3-g733a5 From 3f476d309853e98fee3663089fe7842ad5d3733a Mon Sep 17 00:00:00 2001 From: Antoni Boucher Date: Thu, 14 Aug 2025 09:51:01 -0400 Subject: Rename/remove some UI and run-make tests --- tests/failing-run-make-tests.txt | 3 --- tests/failing-ui-tests.txt | 4 +++- 2 files changed, 3 insertions(+), 4 deletions(-) diff --git a/tests/failing-run-make-tests.txt b/tests/failing-run-make-tests.txt index 29032b321fa..c5e22970c66 100644 --- a/tests/failing-run-make-tests.txt +++ b/tests/failing-run-make-tests.txt @@ -6,7 +6,6 @@ tests/run-make/doctests-keep-binaries/ tests/run-make/doctests-runtool/ tests/run-make/emit-shared-files/ tests/run-make/exit-code/ -tests/run-make/issue-64153/ tests/run-make/llvm-ident/ tests/run-make/native-link-modifier-bundle/ tests/run-make/remap-path-prefix-dwarf/ @@ -34,8 +33,6 @@ tests/run-make/c-link-to-rust-staticlib/ tests/run-make/foreign-double-unwind/ tests/run-make/foreign-exceptions/ tests/run-make/glibc-staticlib-args/ -tests/run-make/issue-36710/ -tests/run-make/issue-68794-textrel-on-minimal-lib/ tests/run-make/lto-smoke-c/ tests/run-make/return-non-c-like-enum/ diff --git a/tests/failing-ui-tests.txt b/tests/failing-ui-tests.txt index 41fb4729c07..e2615bce190 100644 --- a/tests/failing-ui-tests.txt +++ b/tests/failing-ui-tests.txt @@ -20,7 +20,7 @@ tests/ui/drop/dynamic-drop.rs tests/ui/rfcs/rfc-2091-track-caller/std-panic-locations.rs tests/ui/simd/issue-17170.rs tests/ui/simd/issue-39720.rs -tests/ui/issues/issue-14875.rs +tests/ui/drop/panic-during-drop-14875.rs tests/ui/issues/issue-29948.rs tests/ui/process/println-with-broken-pipe.rs tests/ui/lto/thin-lto-inlines2.rs @@ -86,3 +86,5 @@ tests/ui/panics/unwind-force-no-unwind-tables.rs tests/ui/attributes/fn-align-dyn.rs tests/ui/linkage-attr/raw-dylib/elf/glibc-x86_64.rs tests/ui/explicit-tail-calls/recursion-etc.rs +tests/ui/explicit-tail-calls/indexer.rs +tests/ui/explicit-tail-calls/drop-order.rs -- cgit 1.4.1-3-g733a5 From eff47b545a4065f87cc14943bc638096cf929325 Mon Sep 17 00:00:00 2001 From: Antoni Boucher Date: Thu, 14 Aug 2025 11:46:34 -0400 Subject: Don't set feature compiler-builtins-no-f16-f128 --- .github/workflows/m68k.yml | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/.github/workflows/m68k.yml b/.github/workflows/m68k.yml index 176ee8628b0..e49c62d6c93 100644 --- a/.github/workflows/m68k.yml +++ b/.github/workflows/m68k.yml @@ -82,14 +82,14 @@ jobs: - name: Build sample project with target defined as JSON spec run: | ./y.sh prepare --only-libcore --cross - ./y.sh build --sysroot --features compiler-builtins-no-f16-f128 --target-triple m68k-unknown-linux-gnu --target ${{ github.workspace }}/target_specs/m68k-unknown-linux-gnu.json + ./y.sh build --sysroot --target-triple m68k-unknown-linux-gnu --target ${{ github.workspace }}/target_specs/m68k-unknown-linux-gnu.json CG_RUSTFLAGS="-Clinker=m68k-unknown-linux-gnu-gcc" ./y.sh cargo build --manifest-path=./tests/hello-world/Cargo.toml --target ${{ github.workspace }}/target_specs/m68k-unknown-linux-gnu.json ./y.sh clean all - name: Build run: | ./y.sh prepare --only-libcore --cross - ./y.sh build --sysroot --features compiler-builtins-no-f16-f128 --target-triple m68k-unknown-linux-gnu + ./y.sh build --sysroot --target-triple m68k-unknown-linux-gnu ./y.sh test --mini-tests --target-triple m68k-unknown-linux-gnu CG_GCC_TEST_TARGET=m68k-unknown-linux-gnu ./y.sh test --cargo-tests --target-triple m68k-unknown-linux-gnu ./y.sh clean all @@ -102,7 +102,7 @@ jobs: - name: Run tests run: | - ./y.sh test --target-triple m68k-unknown-linux-gnu --release --clean --build-sysroot --sysroot-features compiler-builtins-no-f16-f128 ${{ matrix.commands }} + ./y.sh test --target-triple m68k-unknown-linux-gnu --release --clean --build-sysroot ${{ matrix.commands }} - name: Run Hello World! run: | -- cgit 1.4.1-3-g733a5 From e6eb52121365920fd112d9a3ed52c82d682b3944 Mon Sep 17 00:00:00 2001 From: Marcelo Domínguez Date: Thu, 14 Aug 2025 15:27:57 +0000 Subject: Complete functionality and general cleanup --- src/lib.rs | 6 ------ 1 file changed, 6 deletions(-) diff --git a/src/lib.rs b/src/lib.rs index b11f11d38e3..4025aba82da 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -93,7 +93,6 @@ use gccjit::{CType, Context, OptimizationLevel}; #[cfg(feature = "master")] use gccjit::{TargetInfo, Version}; use rustc_ast::expand::allocator::AllocatorKind; -use rustc_ast::expand::autodiff_attrs::AutoDiffItem; use rustc_codegen_ssa::back::lto::{SerializedModule, ThinModule}; use rustc_codegen_ssa::back::write::{ CodegenContext, FatLtoInput, ModuleConfig, TargetMachineFactoryFn, @@ -363,12 +362,7 @@ impl WriteBackendMethods for GccCodegenBackend { _exported_symbols_for_lto: &[String], each_linked_rlib_for_lto: &[PathBuf], modules: Vec>, - diff_functions: Vec, ) -> Result, FatalError> { - if !diff_functions.is_empty() { - unimplemented!(); - } - back::lto::run_fat(cgcx, each_linked_rlib_for_lto, modules) } -- cgit 1.4.1-3-g733a5 From c99224536152cff14639b64a3fa4f7a215fd037c Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sat, 9 Aug 2025 23:52:11 +0100 Subject: refactor: Include table sizes in comment at top of `unicode_data.rs` To make changes in table size obvious from git diffs --- library/core/src/unicode/unicode_data.rs | 10 ++++++++++ src/tools/unicode-table-generator/src/main.rs | 20 +++++++++----------- 2 files changed, 19 insertions(+), 11 deletions(-) diff --git a/library/core/src/unicode/unicode_data.rs b/library/core/src/unicode/unicode_data.rs index b57234bbee9..6059f7d6450 100644 --- a/library/core/src/unicode/unicode_data.rs +++ b/library/core/src/unicode/unicode_data.rs @@ -1,4 +1,14 @@ ///! This file is generated by `./x run src/tools/unicode-table-generator`; do not edit manually! +// Alphabetic : 1727 bytes, 142759 codepoints in 757 ranges (U+000041 - U+0323B0) using skiplist +// Case_Ignorable : 1053 bytes, 2749 codepoints in 452 ranges (U+000027 - U+0E01F0) using skiplist +// Cased : 407 bytes, 4578 codepoints in 159 ranges (U+000041 - U+01F18A) using skiplist +// Cc : 9 bytes, 65 codepoints in 2 ranges (U+000000 - U+0000A0) using skiplist +// Grapheme_Extend : 887 bytes, 2193 codepoints in 375 ranges (U+000300 - U+0E01F0) using skiplist +// Lowercase : 935 bytes, 2569 codepoints in 675 ranges (U+000061 - U+01E944) using bitset +// N : 457 bytes, 1911 codepoints in 144 ranges (U+000030 - U+01FBFA) using skiplist +// Uppercase : 799 bytes, 1978 codepoints in 656 ranges (U+000041 - U+01F18A) using bitset +// White_Space : 256 bytes, 25 codepoints in 10 ranges (U+000009 - U+003001) using cascading +// Total : 6530 bytes #[inline(always)] const fn bitset_search< diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index 6cdb82a87bd..c1017142097 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -239,6 +239,11 @@ fn main() { std::fs::write(&path, generate_tests(&write_location, ranges_by_property)).unwrap(); } + let mut table_file = String::new(); + table_file.push_str( + "///! This file is generated by `./x run src/tools/unicode-table-generator`; do not edit manually!\n", + ); + let mut total_bytes = 0; let mut modules = Vec::new(); for (property, ranges) in ranges_by_property { @@ -252,8 +257,8 @@ fn main() { } modules.push((property.to_lowercase().to_string(), emitter.file)); - println!( - "{:15}: {} bytes, {} codepoints in {} ranges ({} - {}) using {}", + table_file.push_str(&format!( + "// {:16}: {:5} bytes, {:6} codepoints in {:3} ranges (U+{:06X} - U+{:06X}) using {}\n", property, emitter.bytes_used, datapoints, @@ -261,15 +266,10 @@ fn main() { ranges.first().unwrap().start, ranges.last().unwrap().end, emitter.desc, - ); + )); total_bytes += emitter.bytes_used; } - - let mut table_file = String::new(); - - table_file.push_str( - "///! This file is generated by `./x run src/tools/unicode-table-generator`; do not edit manually!\n", - ); + table_file.push_str(&format!("// {:16}: {:5} bytes\n", "Total", total_bytes)); // Include the range search function table_file.push('\n'); @@ -296,8 +296,6 @@ fn main() { } std::fs::write(&write_location, format!("{}\n", table_file.trim_end())).unwrap(); - - println!("Total table sizes: {total_bytes} bytes"); } fn version() -> String { -- cgit 1.4.1-3-g733a5 From 69e1974bb0bfbcc679d29950b1e4540cd0b9b3ee Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 00:18:04 +0100 Subject: refactor: Include size of case conversion tables Include the sizes of the `to_lowercase` and `to_uppercase` tables in the total size calculations. --- library/core/src/unicode/unicode_data.rs | 12 ++++--- .../unicode-table-generator/src/case_mapping.rs | 41 +++++++++++++++------- src/tools/unicode-table-generator/src/main.rs | 7 +++- 3 files changed, 42 insertions(+), 18 deletions(-) diff --git a/library/core/src/unicode/unicode_data.rs b/library/core/src/unicode/unicode_data.rs index 6059f7d6450..787efcc0914 100644 --- a/library/core/src/unicode/unicode_data.rs +++ b/library/core/src/unicode/unicode_data.rs @@ -8,7 +8,9 @@ // N : 457 bytes, 1911 codepoints in 144 ranges (U+000030 - U+01FBFA) using skiplist // Uppercase : 799 bytes, 1978 codepoints in 656 ranges (U+000041 - U+01F18A) using bitset // White_Space : 256 bytes, 25 codepoints in 10 ranges (U+000009 - U+003001) using cascading -// Total : 6530 bytes +// to_lower : 11484 bytes +// to_upper : 13432 bytes +// Total : 31446 bytes #[inline(always)] const fn bitset_search< @@ -782,7 +784,7 @@ pub mod conversions { } } - static LOWERCASE_TABLE: &[(char, u32)] = &[ + static LOWERCASE_TABLE: &[(char, u32); 1434] = &[ ('\u{c0}', 224), ('\u{c1}', 225), ('\u{c2}', 226), ('\u{c3}', 227), ('\u{c4}', 228), ('\u{c5}', 229), ('\u{c6}', 230), ('\u{c7}', 231), ('\u{c8}', 232), ('\u{c9}', 233), ('\u{ca}', 234), ('\u{cb}', 235), ('\u{cc}', 236), ('\u{cd}', 237), ('\u{ce}', 238), @@ -1132,11 +1134,11 @@ pub mod conversions { ('\u{1e921}', 125251), ]; - static LOWERCASE_TABLE_MULTI: &[[char; 3]] = &[ + static LOWERCASE_TABLE_MULTI: &[[char; 3]; 1] = &[ ['i', '\u{307}', '\u{0}'], ]; - static UPPERCASE_TABLE: &[(char, u32)] = &[ + static UPPERCASE_TABLE: &[(char, u32); 1526] = &[ ('\u{b5}', 924), ('\u{df}', 4194304), ('\u{e0}', 192), ('\u{e1}', 193), ('\u{e2}', 194), ('\u{e3}', 195), ('\u{e4}', 196), ('\u{e5}', 197), ('\u{e6}', 198), ('\u{e7}', 199), ('\u{e8}', 200), ('\u{e9}', 201), ('\u{ea}', 202), ('\u{eb}', 203), ('\u{ec}', 204), @@ -1509,7 +1511,7 @@ pub mod conversions { ('\u{1e941}', 125215), ('\u{1e942}', 125216), ('\u{1e943}', 125217), ]; - static UPPERCASE_TABLE_MULTI: &[[char; 3]] = &[ + static UPPERCASE_TABLE_MULTI: &[[char; 3]; 102] = &[ ['S', 'S', '\u{0}'], ['\u{2bc}', 'N', '\u{0}'], ['J', '\u{30c}', '\u{0}'], ['\u{399}', '\u{308}', '\u{301}'], ['\u{3a5}', '\u{308}', '\u{301}'], ['\u{535}', '\u{552}', '\u{0}'], ['H', '\u{331}', '\u{0}'], ['T', '\u{308}', '\u{0}'], diff --git a/src/tools/unicode-table-generator/src/case_mapping.rs b/src/tools/unicode-table-generator/src/case_mapping.rs index 9c6454492e7..a8527ea9a42 100644 --- a/src/tools/unicode-table-generator/src/case_mapping.rs +++ b/src/tools/unicode-table-generator/src/case_mapping.rs @@ -6,20 +6,22 @@ use crate::{UnicodeData, fmt_list}; const INDEX_MASK: u32 = 1 << 22; -pub(crate) fn generate_case_mapping(data: &UnicodeData) -> String { +pub(crate) fn generate_case_mapping(data: &UnicodeData) -> (String, [usize; 2]) { let mut file = String::new(); write!(file, "const INDEX_MASK: u32 = 0x{INDEX_MASK:x};").unwrap(); file.push_str("\n\n"); file.push_str(HEADER.trim_start()); file.push('\n'); - file.push_str(&generate_tables("LOWER", &data.to_lower)); + let (lower_tables, lower_size) = generate_tables("LOWER", &data.to_lower); + file.push_str(&lower_tables); file.push_str("\n\n"); - file.push_str(&generate_tables("UPPER", &data.to_upper)); - file + let (upper_tables, upper_size) = generate_tables("UPPER", &data.to_upper); + file.push_str(&upper_tables); + (file, [lower_size, upper_size]) } -fn generate_tables(case: &str, data: &BTreeMap) -> String { +fn generate_tables(case: &str, data: &BTreeMap) -> (String, usize) { let mut mappings = Vec::with_capacity(data.len()); let mut multis = Vec::new(); @@ -46,16 +48,31 @@ fn generate_tables(case: &str, data: &BTreeMap) -> String } let mut tables = String::new(); - - write!(tables, "static {}CASE_TABLE: &[(char, u32)] = &[{}];", case, fmt_list(mappings)) - .unwrap(); + let mut size = 0; + + size += size_of_val(mappings.as_slice()); + write!( + tables, + "static {}CASE_TABLE: &[(char, u32); {}] = &[{}];", + case, + mappings.len(), + fmt_list(mappings), + ) + .unwrap(); tables.push_str("\n\n"); - write!(tables, "static {}CASE_TABLE_MULTI: &[[char; 3]] = &[{}];", case, fmt_list(multis)) - .unwrap(); - - tables + size += size_of_val(multis.as_slice()); + write!( + tables, + "static {}CASE_TABLE_MULTI: &[[char; 3]; {}] = &[{}];", + case, + multis.len(), + fmt_list(multis), + ) + .unwrap(); + + (tables, size) } struct CharEscape(char); diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index c1017142097..f755ad048e4 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -269,6 +269,11 @@ fn main() { )); total_bytes += emitter.bytes_used; } + let (conversions, sizes) = case_mapping::generate_case_mapping(&unicode_data); + for (name, size) in ["to_lower", "to_upper"].iter().zip(sizes) { + table_file.push_str(&format!("// {:16}: {:5} bytes\n", name, size)); + total_bytes += size; + } table_file.push_str(&format!("// {:16}: {:5} bytes\n", "Total", total_bytes)); // Include the range search function @@ -280,7 +285,7 @@ fn main() { table_file.push('\n'); - modules.push((String::from("conversions"), case_mapping::generate_case_mapping(&unicode_data))); + modules.push((String::from("conversions"), conversions)); for (name, contents) in modules { table_file.push_str("#[rustfmt::skip]\n"); -- cgit 1.4.1-3-g733a5 From 5d54ac5276eade9d9424fc4fbfa6f77bcd5d4940 Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 01:10:15 +0100 Subject: refactor: rewrite `ranges_from_set` The `merge_ranges` function was very complicated and hard to understand. Forunately, we can use `slice::chunk_by` to achieve the same thing. --- src/tools/unicode-table-generator/src/main.rs | 83 ++++++--------------------- 1 file changed, 17 insertions(+), 66 deletions(-) diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index f755ad048e4..bf0511a2c77 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -187,33 +187,19 @@ fn load_data() -> UnicodeData { } } - let mut properties: HashMap<&'static str, Vec>> = properties + let mut properties: Vec<(&'static str, Vec>)> = properties .into_iter() - .map(|(k, v)| { - ( - k, - v.into_iter() - .flat_map(|codepoints| match codepoints { - Codepoints::Single(c) => c - .scalar() - .map(|ch| ch as u32..ch as u32 + 1) - .into_iter() - .collect::>(), - Codepoints::Range(c) => c - .into_iter() - .flat_map(|c| c.scalar().map(|ch| ch as u32..ch as u32 + 1)) - .collect::>(), - }) - .collect::>>(), - ) + .map(|(prop, codepoints)| { + let codepoints = codepoints + .into_iter() + .flatten() + .flat_map(|cp| cp.scalar()) + .map(u32::from) + .collect::>(); + (prop, ranges_from_set(&codepoints)) }) .collect(); - for ranges in properties.values_mut() { - merge_ranges(ranges); - } - - let mut properties = properties.into_iter().collect::>(); properties.sort_by_key(|p| p.0); UnicodeData { ranges: properties, to_lower, to_upper } } @@ -402,48 +388,13 @@ fn generate_asserts(s: &mut String, property: &str, points: &[u32], truthy: bool } } +/// Group the elements of `set` into contigous ranges fn ranges_from_set(set: &[u32]) -> Vec> { - let mut ranges = set.iter().map(|e| (*e)..(*e + 1)).collect::>>(); - merge_ranges(&mut ranges); - ranges -} - -fn merge_ranges(ranges: &mut Vec>) { - loop { - let mut new_ranges = Vec::new(); - let mut idx_iter = 0..(ranges.len() - 1); - let mut should_insert_last = true; - while let Some(idx) = idx_iter.next() { - let cur = ranges[idx].clone(); - let next = ranges[idx + 1].clone(); - if cur.end == next.start { - if idx_iter.next().is_none() { - // We're merging the last element - should_insert_last = false; - } - new_ranges.push(cur.start..next.end); - } else { - // We're *not* merging the last element - should_insert_last = true; - new_ranges.push(cur); - } - } - if should_insert_last { - new_ranges.push(ranges.last().unwrap().clone()); - } - if new_ranges.len() == ranges.len() { - *ranges = new_ranges; - break; - } else { - *ranges = new_ranges; - } - } - - let mut last_end = None; - for range in ranges { - if let Some(last) = last_end { - assert!(range.start > last, "{range:?}"); - } - last_end = Some(range.end); - } + set.chunk_by(|a, b| a + 1 == *b) + .map(|chunk| { + let start = *chunk.first().unwrap(); + let end = *chunk.last().unwrap(); + start..(end + 1) + }) + .collect() } -- cgit 1.4.1-3-g733a5 From 30d1bc7ba869c0f86bc6d2e1d9ed1ad3b58f7865 Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 01:46:00 +0100 Subject: refactor: `generate_tests` Rewrite `generate_tests` to be more idiomatic. --- src/tools/unicode-table-generator/src/main.rs | 97 +++++++++++++-------------- 1 file changed, 45 insertions(+), 52 deletions(-) diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index bf0511a2c77..c9530fec48a 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -72,6 +72,8 @@ //! or not. use std::collections::{BTreeMap, HashMap}; +use std::fmt; +use std::fmt::Write; use std::ops::Range; use ucd_parse::Codepoints; @@ -222,7 +224,7 @@ fn main() { let ranges_by_property = &unicode_data.ranges; if let Some(path) = test_path { - std::fs::write(&path, generate_tests(&write_location, ranges_by_property)).unwrap(); + std::fs::write(&path, generate_tests(ranges_by_property).unwrap()).unwrap(); } let mut table_file = String::new(); @@ -326,66 +328,57 @@ fn fmt_list(values: impl IntoIterator) -> String { out } -fn generate_tests(data_path: &str, ranges: &[(&str, Vec>)]) -> String { +fn generate_tests(ranges: &[(&str, Vec>)]) -> Result { let mut s = String::new(); - s.push_str("#![allow(incomplete_features, unused)]\n"); - s.push_str("#![feature(const_generics)]\n\n"); - s.push_str("\n#[allow(unused)]\nuse std::hint;\n"); - s.push_str(&format!("#[path = \"{data_path}\"]\n")); - s.push_str("mod unicode_data;\n\n"); - - s.push_str("\nfn main() {\n"); - + writeln!(s, "#![feature(core_intrinsics)]")?; + writeln!(s, "#![allow(internal_features, dead_code)]")?; + writeln!(s, "// ignore-tidy-filelength")?; + writeln!(s, "use std::intrinsics;")?; + writeln!(s, "mod unicode_data;")?; + writeln!(s, "fn main() {{")?; for (property, ranges) in ranges { - s.push_str(&format!(r#" println!("Testing {property}");"#)); - s.push('\n'); - s.push_str(&format!(" {}_true();\n", property.to_lowercase())); - s.push_str(&format!(" {}_false();\n", property.to_lowercase())); - let mut is_true = Vec::new(); - let mut is_false = Vec::new(); - for ch_num in 0..(std::char::MAX as u32) { - if std::char::from_u32(ch_num).is_none() { - continue; - } - if ranges.iter().any(|r| r.contains(&ch_num)) { - is_true.push(ch_num); - } else { - is_false.push(ch_num); - } - } - - s.push_str(&format!(" fn {}_true() {{\n", property.to_lowercase())); - generate_asserts(&mut s, property, &is_true, true); - s.push_str(" }\n\n"); - s.push_str(&format!(" fn {}_false() {{\n", property.to_lowercase())); - generate_asserts(&mut s, property, &is_false, false); - s.push_str(" }\n\n"); + let prop = property.to_lowercase(); + writeln!(s, r#" println!("Testing {prop}");"#)?; + writeln!(s, " {prop}_true();")?; + writeln!(s, " {prop}_false();")?; + let (is_true, is_false): (Vec<_>, Vec<_>) = (char::MIN..=char::MAX) + .filter(|c| !c.is_ascii()) + .map(u32::from) + .partition(|c| ranges.iter().any(|r| r.contains(c))); + + writeln!(s, " fn {prop}_true() {{")?; + generate_asserts(&mut s, &prop, &is_true, true)?; + writeln!(s, " }}")?; + + writeln!(s, " fn {prop}_false() {{")?; + generate_asserts(&mut s, &prop, &is_false, false)?; + writeln!(s, " }}")?; } - s.push('}'); - s + writeln!(s, "}}")?; + Ok(s) } -fn generate_asserts(s: &mut String, property: &str, points: &[u32], truthy: bool) { +fn generate_asserts( + s: &mut String, + prop: &str, + points: &[u32], + truthy: bool, +) -> Result<(), fmt::Error> { + let truthy = if truthy { "" } else { "!" }; for range in ranges_from_set(points) { - if range.end == range.start + 1 { - s.push_str(&format!( - " assert!({}unicode_data::{}::lookup({:?}), \"{}\");\n", - if truthy { "" } else { "!" }, - property.to_lowercase(), - std::char::from_u32(range.start).unwrap(), - range.start, - )); - } else { - s.push_str(&format!(" for chn in {range:?}u32 {{\n")); - s.push_str(&format!( - " assert!({}unicode_data::{}::lookup(std::char::from_u32(chn).unwrap()), \"{{:?}}\", chn);\n", - if truthy { "" } else { "!" }, - property.to_lowercase(), - )); - s.push_str(" }\n"); + let start = char::from_u32(range.start).unwrap(); + let end = char::from_u32(range.end - 1).unwrap(); + match range.len() { + 1 => writeln!(s, " assert!({truthy}unicode_data::{prop}::lookup({start:?}));")?, + _ => { + writeln!(s, " for c in {start:?}..={end:?} {{")?; + writeln!(s, " assert!({truthy}unicode_data::{prop}::lookup(c));")?; + writeln!(s, " }}")?; + } } } + Ok(()) } /// Group the elements of `set` into contigous ranges -- cgit 1.4.1-3-g733a5 From c3ce0796544152460554d8b8db4e56528fe362db Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Sun, 10 Aug 2025 02:13:03 +0100 Subject: refactor: Add tests for case conversions --- .../unicode-table-generator/src/case_mapping.rs | 4 +- src/tools/unicode-table-generator/src/main.rs | 48 ++++++++++++++++++---- 2 files changed, 41 insertions(+), 11 deletions(-) diff --git a/src/tools/unicode-table-generator/src/case_mapping.rs b/src/tools/unicode-table-generator/src/case_mapping.rs index a8527ea9a42..49aef3ec33e 100644 --- a/src/tools/unicode-table-generator/src/case_mapping.rs +++ b/src/tools/unicode-table-generator/src/case_mapping.rs @@ -21,11 +21,11 @@ pub(crate) fn generate_case_mapping(data: &UnicodeData) -> (String, [usize; 2]) (file, [lower_size, upper_size]) } -fn generate_tables(case: &str, data: &BTreeMap) -> (String, usize) { +fn generate_tables(case: &str, data: &BTreeMap) -> (String, usize) { let mut mappings = Vec::with_capacity(data.len()); let mut multis = Vec::new(); - for (&key, &(a, b, c)) in data.iter() { + for (&key, &[a, b, c]) in data.iter() { let key = char::from_u32(key).unwrap(); if key.is_ascii() { diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index c9530fec48a..1d70eebdbc6 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -100,11 +100,11 @@ static PROPERTIES: &[&str] = &[ struct UnicodeData { ranges: Vec<(&'static str, Vec>)>, - to_upper: BTreeMap, - to_lower: BTreeMap, + to_upper: BTreeMap, + to_lower: BTreeMap, } -fn to_mapping(origin: u32, codepoints: Vec) -> Option<(u32, u32, u32)> { +fn to_mapping(origin: u32, codepoints: Vec) -> Option<[u32; 3]> { let mut a = None; let mut b = None; let mut c = None; @@ -125,7 +125,7 @@ fn to_mapping(origin: u32, codepoints: Vec) -> Option<(u32 } } - Some((a.unwrap(), b.unwrap_or(0), c.unwrap_or(0))) + Some([a.unwrap(), b.unwrap_or(0), c.unwrap_or(0)]) } static UNICODE_DIRECTORY: &str = "unicode-downloads"; @@ -165,12 +165,12 @@ fn load_data() -> UnicodeData { if let Some(mapped) = row.simple_lowercase_mapping && mapped != row.codepoint { - to_lower.insert(row.codepoint.value(), (mapped.value(), 0, 0)); + to_lower.insert(row.codepoint.value(), [mapped.value(), 0, 0]); } if let Some(mapped) = row.simple_uppercase_mapping && mapped != row.codepoint { - to_upper.insert(row.codepoint.value(), (mapped.value(), 0, 0)); + to_upper.insert(row.codepoint.value(), [mapped.value(), 0, 0]); } } @@ -224,7 +224,7 @@ fn main() { let ranges_by_property = &unicode_data.ranges; if let Some(path) = test_path { - std::fs::write(&path, generate_tests(ranges_by_property).unwrap()).unwrap(); + std::fs::write(&path, generate_tests(&unicode_data).unwrap()).unwrap(); } let mut table_file = String::new(); @@ -328,7 +328,7 @@ fn fmt_list(values: impl IntoIterator) -> String { out } -fn generate_tests(ranges: &[(&str, Vec>)]) -> Result { +fn generate_tests(data: &UnicodeData) -> Result { let mut s = String::new(); writeln!(s, "#![feature(core_intrinsics)]")?; writeln!(s, "#![allow(internal_features, dead_code)]")?; @@ -336,7 +336,7 @@ fn generate_tests(ranges: &[(&str, Vec>)]) -> Result>)]) -> Result = (char::MIN..=char::MAX) + .filter(|c| !c.is_ascii()) + .map(u32::from) + .filter(|c| !conversion.contains_key(c)) + .collect(); + let unmapped_ranges = ranges_from_set(&unmapped); + for range in unmapped_ranges { + let start = char::from_u32(range.start).unwrap(); + let end = char::from_u32(range.end - 1).unwrap(); + writeln!(s, " for c in {start:?}..={end:?} {{")?; + writeln!( + s, + r#" assert_eq!(unicode_data::conversions::{name}(c), [c, '\0', '\0']);"# + )?; + + writeln!(s, " }}")?; + } + } + writeln!(s, "}}")?; Ok(s) } -- cgit 1.4.1-3-g733a5 From 489940ee1d24a0a2bc388b465f171da9ab0cb16b Mon Sep 17 00:00:00 2001 From: Kivooeo Date: Fri, 15 Aug 2025 16:42:21 +0000 Subject: stabilize array repeat --- library/core/src/array/mod.rs | 4 +--- tests/codegen-llvm/array-repeat.rs | 1 - tests/codegen-llvm/iter-repeat-n-trivial-drop.rs | 1 - 3 files changed, 1 insertion(+), 5 deletions(-) diff --git a/library/core/src/array/mod.rs b/library/core/src/array/mod.rs index b3a498570f9..e994fa00a23 100644 --- a/library/core/src/array/mod.rs +++ b/library/core/src/array/mod.rs @@ -41,8 +41,6 @@ pub use iter::IntoIter; /// /// Creating multiple copies of a `String`: /// ```rust -/// #![feature(array_repeat)] -/// /// use std::array; /// /// let string = "Hello there!".to_string(); @@ -50,7 +48,7 @@ pub use iter::IntoIter; /// assert_eq!(strings, ["Hello there!", "Hello there!"]); /// ``` #[inline] -#[unstable(feature = "array_repeat", issue = "126695")] +#[stable(feature = "array_repeat", since = "CURRENT_RUSTC_VERSION")] pub fn repeat(val: T) -> [T; N] { from_trusted_iterator(repeat_n(val, N)) } diff --git a/tests/codegen-llvm/array-repeat.rs b/tests/codegen-llvm/array-repeat.rs index 4c755df9390..1c45341d764 100644 --- a/tests/codegen-llvm/array-repeat.rs +++ b/tests/codegen-llvm/array-repeat.rs @@ -1,7 +1,6 @@ //@ compile-flags: -Copt-level=3 #![crate_type = "lib"] -#![feature(array_repeat)] use std::array::repeat; diff --git a/tests/codegen-llvm/iter-repeat-n-trivial-drop.rs b/tests/codegen-llvm/iter-repeat-n-trivial-drop.rs index 28173530324..6f340978428 100644 --- a/tests/codegen-llvm/iter-repeat-n-trivial-drop.rs +++ b/tests/codegen-llvm/iter-repeat-n-trivial-drop.rs @@ -4,7 +4,6 @@ #![crate_type = "lib"] #![feature(iter_repeat_n)] -#![feature(array_repeat)] #[derive(Clone)] pub struct NotCopy(u16); -- cgit 1.4.1-3-g733a5 From b951b5dca116803b89380aae55ce9053a4674f31 Mon Sep 17 00:00:00 2001 From: Kivooeo Date: Fri, 15 Aug 2025 16:56:11 +0000 Subject: stabilize strict provenance atomic ptr --- library/core/src/sync/atomic.rs | 21 +++++++-------------- library/coretests/tests/lib.rs | 1 - library/std/src/lib.rs | 1 - src/tools/miri/tests/pass/atomic.rs | 1 - tests/codegen-llvm/atomicptr.rs | 1 - 5 files changed, 7 insertions(+), 18 deletions(-) diff --git a/library/core/src/sync/atomic.rs b/library/core/src/sync/atomic.rs index 44a6895f90a..7bd68bcd0bc 100644 --- a/library/core/src/sync/atomic.rs +++ b/library/core/src/sync/atomic.rs @@ -2199,7 +2199,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let atom = AtomicPtr::::new(core::ptr::null_mut()); @@ -2209,7 +2208,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_ptr_add(&self, val: usize, order: Ordering) -> *mut T { self.fetch_byte_add(val.wrapping_mul(size_of::()), order) @@ -2240,7 +2239,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let array = [1i32, 2i32]; @@ -2254,7 +2252,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_ptr_sub(&self, val: usize, order: Ordering) -> *mut T { self.fetch_byte_sub(val.wrapping_mul(size_of::()), order) @@ -2279,7 +2277,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let atom = AtomicPtr::::new(core::ptr::null_mut()); @@ -2289,7 +2286,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_byte_add(&self, val: usize, order: Ordering) -> *mut T { // SAFETY: data races are prevented by atomic intrinsics. @@ -2315,7 +2312,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let mut arr = [0i64, 1]; @@ -2325,7 +2321,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_byte_sub(&self, val: usize, order: Ordering) -> *mut T { // SAFETY: data races are prevented by atomic intrinsics. @@ -2361,7 +2357,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let pointer = &mut 3i64 as *mut i64; @@ -2376,7 +2371,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_or(&self, val: usize, order: Ordering) -> *mut T { // SAFETY: data races are prevented by atomic intrinsics. @@ -2412,7 +2407,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let pointer = &mut 3i64 as *mut i64; @@ -2426,7 +2420,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_and(&self, val: usize, order: Ordering) -> *mut T { // SAFETY: data races are prevented by atomic intrinsics. @@ -2462,7 +2456,6 @@ impl AtomicPtr { /// # Examples /// /// ``` - /// #![feature(strict_provenance_atomic_ptr)] /// use core::sync::atomic::{AtomicPtr, Ordering}; /// /// let pointer = &mut 3i64 as *mut i64; @@ -2474,7 +2467,7 @@ impl AtomicPtr { /// ``` #[inline] #[cfg(target_has_atomic = "ptr")] - #[unstable(feature = "strict_provenance_atomic_ptr", issue = "99108")] + #[stable(feature = "strict_provenance_atomic_ptr", since = "CURRENT_RUSTC_VERSION")] #[cfg_attr(miri, track_caller)] // even without panics, this helps for Miri backtraces pub fn fetch_xor(&self, val: usize, order: Ordering) -> *mut T { // SAFETY: data races are prevented by atomic intrinsics. diff --git a/library/coretests/tests/lib.rs b/library/coretests/tests/lib.rs index 0c4d49f3c99..92774e12681 100644 --- a/library/coretests/tests/lib.rs +++ b/library/coretests/tests/lib.rs @@ -95,7 +95,6 @@ #![feature(std_internals)] #![feature(step_trait)] #![feature(str_internals)] -#![feature(strict_provenance_atomic_ptr)] #![feature(strict_provenance_lints)] #![feature(test)] #![feature(trusted_len)] diff --git a/library/std/src/lib.rs b/library/std/src/lib.rs index f111fcb4a47..77c7878ce08 100644 --- a/library/std/src/lib.rs +++ b/library/std/src/lib.rs @@ -370,7 +370,6 @@ #![feature(slice_range)] #![feature(std_internals)] #![feature(str_internals)] -#![feature(strict_provenance_atomic_ptr)] #![feature(sync_unsafe_cell)] #![feature(temporary_niche_types)] #![feature(ub_checks)] diff --git a/src/tools/miri/tests/pass/atomic.rs b/src/tools/miri/tests/pass/atomic.rs index 3de34e570c7..d8ac5114f27 100644 --- a/src/tools/miri/tests/pass/atomic.rs +++ b/src/tools/miri/tests/pass/atomic.rs @@ -2,7 +2,6 @@ //@[tree]compile-flags: -Zmiri-tree-borrows //@compile-flags: -Zmiri-strict-provenance -#![feature(strict_provenance_atomic_ptr)] // FIXME(static_mut_refs): Do not allow `static_mut_refs` lint #![allow(static_mut_refs)] diff --git a/tests/codegen-llvm/atomicptr.rs b/tests/codegen-llvm/atomicptr.rs index ce6c4aa0d2b..9d5e618fe76 100644 --- a/tests/codegen-llvm/atomicptr.rs +++ b/tests/codegen-llvm/atomicptr.rs @@ -6,7 +6,6 @@ //@ compile-flags: -Copt-level=3 -Cno-prepopulate-passes #![crate_type = "lib"] -#![feature(strict_provenance_atomic_ptr)] use std::ptr::without_provenance_mut; use std::sync::atomic::AtomicPtr; -- cgit 1.4.1-3-g733a5 From 1bb9b151c9f9b5116254827f04add845aff33408 Mon Sep 17 00:00:00 2001 From: Karl Meakin Date: Mon, 11 Aug 2025 00:57:29 +0000 Subject: refactor: Hard-code `char::is_control` According to https://www.unicode.org/policies/stability_policy.html#Property_Value, the set of codepoints in `Cc` will never change. So we can hard-code the patterns to match against instead of using a table. --- library/core/src/char/methods.rs | 6 +++++- library/core/src/unicode/mod.rs | 1 - library/core/src/unicode/unicode_data.rs | 25 ------------------------- src/tools/unicode-table-generator/src/main.rs | 1 - 4 files changed, 5 insertions(+), 28 deletions(-) diff --git a/library/core/src/char/methods.rs b/library/core/src/char/methods.rs index 7ee0962721f..61ac7f8a339 100644 --- a/library/core/src/char/methods.rs +++ b/library/core/src/char/methods.rs @@ -950,7 +950,11 @@ impl char { #[stable(feature = "rust1", since = "1.0.0")] #[inline] pub fn is_control(self) -> bool { - unicode::Cc(self) + // According to + // https://www.unicode.org/policies/stability_policy.html#Property_Value, + // the set of codepoints in `Cc` will never change. + // So we can just hard-code the patterns to match against instead of using a table. + matches!(self, '\0'..='\x1f' | '\x7f'..='\u{9f}') } /// Returns `true` if this `char` has the `Grapheme_Extend` property. diff --git a/library/core/src/unicode/mod.rs b/library/core/src/unicode/mod.rs index 49dbdeb1a6d..e1cb69c3c4f 100644 --- a/library/core/src/unicode/mod.rs +++ b/library/core/src/unicode/mod.rs @@ -9,7 +9,6 @@ pub use unicode_data::conversions; #[rustfmt::skip] pub(crate) use unicode_data::alphabetic::lookup as Alphabetic; -pub(crate) use unicode_data::cc::lookup as Cc; pub(crate) use unicode_data::grapheme_extend::lookup as Grapheme_Extend; pub(crate) use unicode_data::lowercase::lookup as Lowercase; pub(crate) use unicode_data::n::lookup as N; diff --git a/library/core/src/unicode/unicode_data.rs b/library/core/src/unicode/unicode_data.rs index b57234bbee9..55f64f1e96e 100644 --- a/library/core/src/unicode/unicode_data.rs +++ b/library/core/src/unicode/unicode_data.rs @@ -358,31 +358,6 @@ pub mod cased { } } -#[rustfmt::skip] -pub mod cc { - use super::ShortOffsetRunHeader; - - static SHORT_OFFSET_RUNS: [ShortOffsetRunHeader; 1] = [ - ShortOffsetRunHeader::new(0, 1114272), - ]; - static OFFSETS: [u8; 5] = [ - 0, 32, 95, 33, 0, - ]; - pub fn lookup(c: char) -> bool { - const { - assert!(SHORT_OFFSET_RUNS.last().unwrap().0 > char::MAX as u32); - let mut i = 0; - while i < SHORT_OFFSET_RUNS.len() { - assert!(SHORT_OFFSET_RUNS[i].start_index() < OFFSETS.len()); - i += 1; - } - } - // SAFETY: We just ensured the last element of `SHORT_OFFSET_RUNS` is greater than `std::char::MAX` - // and the start indices of all elements in `SHORT_OFFSET_RUNS` are smaller than `OFFSETS.len()`. - unsafe { super::skip_search(c, &SHORT_OFFSET_RUNS, &OFFSETS) } - } -} - #[rustfmt::skip] pub mod grapheme_extend { use super::ShortOffsetRunHeader; diff --git a/src/tools/unicode-table-generator/src/main.rs b/src/tools/unicode-table-generator/src/main.rs index 6cdb82a87bd..38e5e8bbdb9 100644 --- a/src/tools/unicode-table-generator/src/main.rs +++ b/src/tools/unicode-table-generator/src/main.rs @@ -92,7 +92,6 @@ static PROPERTIES: &[&str] = &[ "Case_Ignorable", "Grapheme_Extend", "White_Space", - "Cc", "N", ]; -- cgit 1.4.1-3-g733a5 From 41405456daf54af9182e73277c417a6572d14200 Mon Sep 17 00:00:00 2001 From: Zihan Date: Sat, 16 Aug 2025 01:23:37 -0400 Subject: `missing_inline_in_public_items`: fix lint emission source HirId use trait item's HirId when emitting lint at `check_item` level changelog: [`missing_inline_in_public_items`]: fix trait item lint emission Signed-off-by: Zihan --- clippy_lints/src/missing_inline.rs | 38 ++++++++++++++++++-------------------- tests/ui/missing_inline.rs | 7 +++++++ 2 files changed, 25 insertions(+), 20 deletions(-) diff --git a/clippy_lints/src/missing_inline.rs b/clippy_lints/src/missing_inline.rs index d02952eb487..28555a61090 100644 --- a/clippy_lints/src/missing_inline.rs +++ b/clippy_lints/src/missing_inline.rs @@ -1,4 +1,4 @@ -use clippy_utils::diagnostics::span_lint; +use clippy_utils::diagnostics::{span_lint, span_lint_hir}; use rustc_hir::attrs::AttributeKind; use rustc_hir::def_id::DefId; use rustc_hir::{self as hir, Attribute, find_attr}; @@ -64,14 +64,20 @@ declare_clippy_lint! { "detects missing `#[inline]` attribute for public callables (functions, trait methods, methods...)" } -fn check_missing_inline_attrs(cx: &LateContext<'_>, attrs: &[Attribute], sp: Span, desc: &'static str) { +fn check_missing_inline_attrs( + cx: &LateContext<'_>, + attrs: &[Attribute], + sp: Span, + desc: &'static str, + hir_id: Option, +) { if !find_attr!(attrs, AttributeKind::Inline(..)) { - span_lint( - cx, - MISSING_INLINE_IN_PUBLIC_ITEMS, - sp, - format!("missing `#[inline]` for {desc}"), - ); + let msg = format!("missing `#[inline]` for {desc}"); + if let Some(hir_id) = hir_id { + span_lint_hir(cx, MISSING_INLINE_IN_PUBLIC_ITEMS, hir_id, sp, msg); + } else { + span_lint(cx, MISSING_INLINE_IN_PUBLIC_ITEMS, sp, msg); + } } } @@ -103,17 +109,9 @@ impl<'tcx> LateLintPass<'tcx> for MissingInline { let desc = "a function"; let attrs = cx.tcx.hir_attrs(it.hir_id()); - check_missing_inline_attrs(cx, attrs, it.span, desc); + check_missing_inline_attrs(cx, attrs, it.span, desc, None); }, - hir::ItemKind::Trait( - ref _constness, - ref _is_auto, - ref _unsafe, - _ident, - _generics, - _bounds, - trait_items, - ) => { + hir::ItemKind::Trait(.., trait_items) => { // note: we need to check if the trait is exported so we can't use // `LateLintPass::check_trait_item` here. for &tit in trait_items { @@ -127,7 +125,7 @@ impl<'tcx> LateLintPass<'tcx> for MissingInline { let desc = "a default trait method"; let item = cx.tcx.hir_trait_item(tit); let attrs = cx.tcx.hir_attrs(item.hir_id()); - check_missing_inline_attrs(cx, attrs, item.span, desc); + check_missing_inline_attrs(cx, attrs, item.span, desc, Some(tit.hir_id())); } }, } @@ -182,7 +180,7 @@ impl<'tcx> LateLintPass<'tcx> for MissingInline { } let attrs = cx.tcx.hir_attrs(impl_item.hir_id()); - check_missing_inline_attrs(cx, attrs, impl_item.span, desc); + check_missing_inline_attrs(cx, attrs, impl_item.span, desc, None); } } diff --git a/tests/ui/missing_inline.rs b/tests/ui/missing_inline.rs index 223c7447975..8e937d60951 100644 --- a/tests/ui/missing_inline.rs +++ b/tests/ui/missing_inline.rs @@ -97,3 +97,10 @@ pub mod issue15301 { println!("Just called a Rust function from Rust!"); } } + +pub mod issue15491 { + pub trait Foo { + #[allow(clippy::missing_inline_in_public_items)] + fn foo(&self) {} + } +} -- cgit 1.4.1-3-g733a5 From 57feb7fc018cd91cc89b7777184cbafc548367b0 Mon Sep 17 00:00:00 2001 From: Mu001999 Date: Sat, 16 Aug 2025 15:01:17 +0800 Subject: old testcase output --- ...-parentheses-to-call-closure-issue-145404.fixed | 13 +++++++++++++ ...use-parentheses-to-call-closure-issue-145404.rs | 13 +++++++++++++ ...parentheses-to-call-closure-issue-145404.stderr | 22 ++++++++++++++++++++++ 3 files changed, 48 insertions(+) create mode 100644 tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed create mode 100644 tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.rs create mode 100644 tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr diff --git a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed new file mode 100644 index 00000000000..4d29b9b0b1a --- /dev/null +++ b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed @@ -0,0 +1,13 @@ +//@ run-rustfix + +use std::fmt::Display; + +struct S; + +impl S { + fn call(&self, _: impl Display) {} +} + +fn main() { + S.call(|| "hello"()); //~ ERROR [E0277] +} diff --git a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.rs b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.rs new file mode 100644 index 00000000000..848629a6700 --- /dev/null +++ b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.rs @@ -0,0 +1,13 @@ +//@ run-rustfix + +use std::fmt::Display; + +struct S; + +impl S { + fn call(&self, _: impl Display) {} +} + +fn main() { + S.call(|| "hello"); //~ ERROR [E0277] +} diff --git a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr new file mode 100644 index 00000000000..ee51dbca804 --- /dev/null +++ b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr @@ -0,0 +1,22 @@ +error[E0277]: `{closure@$DIR/use-parentheses-to-call-closure-issue-145404.rs:12:12: 12:14}` doesn't implement `std::fmt::Display` + --> $DIR/use-parentheses-to-call-closure-issue-145404.rs:12:12 + | +LL | S.call(|| "hello"); + | ---- ^^^^^^^^^^ unsatisfied trait bound + | | + | required by a bound introduced by this call + | + = help: the trait `std::fmt::Display` is not implemented for closure `{closure@$DIR/use-parentheses-to-call-closure-issue-145404.rs:12:12: 12:14}` +note: required by a bound in `S::call` + --> $DIR/use-parentheses-to-call-closure-issue-145404.rs:8:28 + | +LL | fn call(&self, _: impl Display) {} + | ^^^^^^^ required by this bound in `S::call` +help: use parentheses to call this closure + | +LL | S.call(|| "hello"()); + | ++ + +error: aborting due to 1 previous error + +For more information about this error, try `rustc --explain E0277`. -- cgit 1.4.1-3-g733a5 From 5e8c2c3aec19c50efc22854c09564844a133bbaf Mon Sep 17 00:00:00 2001 From: Mu001999 Date: Sat, 16 Aug 2025 15:07:38 +0800 Subject: Add parentheses for closure when suggesting calling closure --- .../src/error_reporting/traits/suggestions.rs | 16 ++++++++++------ .../adt_const_params/const_param_ty_bad.stderr | 5 +++-- .../use-parentheses-to-call-closure-issue-145404.fixed | 2 +- .../use-parentheses-to-call-closure-issue-145404.stderr | 5 +++-- 4 files changed, 17 insertions(+), 11 deletions(-) diff --git a/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs b/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs index bc8c8a44405..3bcbeddf95c 100644 --- a/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs +++ b/compiler/rustc_trait_selection/src/error_reporting/traits/suggestions.rs @@ -820,16 +820,20 @@ impl<'a, 'tcx> TypeErrCtxt<'a, 'tcx> { if matches!(obligation.cause.code(), ObligationCauseCode::FunctionArg { .. }) && obligation.cause.span.can_be_used_for_suggestions() { + let (span, sugg) = if let Some(snippet) = + self.tcx.sess.source_map().span_to_snippet(obligation.cause.span).ok() + && snippet.starts_with("|") + { + (obligation.cause.span, format!("({snippet})({args})")) + } else { + (obligation.cause.span.shrink_to_hi(), format!("({args})")) + }; + // When the obligation error has been ensured to have been caused by // an argument, the `obligation.cause.span` points at the expression // of the argument, so we can provide a suggestion. Otherwise, we give // a more general note. - err.span_suggestion_verbose( - obligation.cause.span.shrink_to_hi(), - msg, - format!("({args})"), - Applicability::HasPlaceholders, - ); + err.span_suggestion_verbose(span, msg, sugg, Applicability::HasPlaceholders); } else if let DefIdOrName::DefId(def_id) = def_id_or_name { let name = match self.tcx.hir_get_if_local(def_id) { Some(hir::Node::Expr(hir::Expr { diff --git a/tests/ui/const-generics/adt_const_params/const_param_ty_bad.stderr b/tests/ui/const-generics/adt_const_params/const_param_ty_bad.stderr index 4c8a5e46751..c05584ef909 100644 --- a/tests/ui/const-generics/adt_const_params/const_param_ty_bad.stderr +++ b/tests/ui/const-generics/adt_const_params/const_param_ty_bad.stderr @@ -32,8 +32,9 @@ LL | fn check(_: impl std::marker::UnsizedConstParamTy) {} | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ required by this bound in `check` help: use parentheses to call this closure | -LL | check(|| {}()); - | ++ +LL - check(|| {}); +LL + check((|| {})()); + | error[E0277]: `fn()` can't be used as a const parameter type --> $DIR/const_param_ty_bad.rs:9:11 diff --git a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed index 4d29b9b0b1a..2835743fca2 100644 --- a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed +++ b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.fixed @@ -9,5 +9,5 @@ impl S { } fn main() { - S.call(|| "hello"()); //~ ERROR [E0277] + S.call((|| "hello")()); //~ ERROR [E0277] } diff --git a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr index ee51dbca804..cb6df5af7fb 100644 --- a/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr +++ b/tests/ui/suggestions/use-parentheses-to-call-closure-issue-145404.stderr @@ -14,8 +14,9 @@ LL | fn call(&self, _: impl Display) {} | ^^^^^^^ required by this bound in `S::call` help: use parentheses to call this closure | -LL | S.call(|| "hello"()); - | ++ +LL - S.call(|| "hello"); +LL + S.call((|| "hello")()); + | error: aborting due to 1 previous error -- cgit 1.4.1-3-g733a5 From 224fd13a9c583db4054fc241bd55f38f0b309dbf Mon Sep 17 00:00:00 2001 From: Zhongyao Chen Date: Mon, 18 Aug 2025 10:33:53 +0800 Subject: change riscv64a23 target reqirements comments: linux kernel 6.8.0, glibc 2.39 --- .../rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs | 2 +- src/doc/rustc/src/platform-support.md | 2 +- src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md | 5 ++--- 3 files changed, 4 insertions(+), 5 deletions(-) diff --git a/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs b/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs index 27a0e27c969..394f8b0685d 100644 --- a/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs +++ b/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs @@ -6,7 +6,7 @@ pub(crate) fn target() -> Target { Target { llvm_target: "riscv64-unknown-linux-gnu".into(), metadata: TargetMetadata { - description: Some("RISC-V Linux (kernel 4.20, glibc 2.29)".into()), + description: Some("RISC-V Linux (kernel 6.8.0, glibc 2.39)".into()), tier: Some(3), host_tools: Some(true), std: Some(true), diff --git a/src/doc/rustc/src/platform-support.md b/src/doc/rustc/src/platform-support.md index cf55d4bd68f..891932a1b6a 100644 --- a/src/doc/rustc/src/platform-support.md +++ b/src/doc/rustc/src/platform-support.md @@ -389,7 +389,7 @@ target | std | host | notes [`riscv64gc-unknown-nuttx-elf`](platform-support/nuttx.md) | ✓ | | RISC-V 64bit with NuttX [`riscv64gc-unknown-openbsd`](platform-support/openbsd.md) | ✓ | ✓ | OpenBSD/riscv64 [`riscv64imac-unknown-nuttx-elf`](platform-support/nuttx.md) | ✓ | | RISC-V 64bit with NuttX -[`riscv64a23-unknown-linux-gnu`](platform-support/riscv64a23-unknown-linux-gnu.md) | ✓ | ✓ | RISC-V Linux (kernel 4.20+, glibc 2.29) +[`riscv64a23-unknown-linux-gnu`](platform-support/riscv64a23-unknown-linux-gnu.md) | ✓ | ✓ | RISC-V Linux (kernel 6.8.0+, glibc 2.39) [`s390x-unknown-linux-musl`](platform-support/s390x-unknown-linux-musl.md) | ✓ | | S390x Linux (kernel 3.2, musl 1.2.3) `sparc-unknown-linux-gnu` | ✓ | | 32-bit SPARC Linux [`sparc-unknown-none-elf`](./platform-support/sparc-unknown-none-elf.md) | * | | Bare 32-bit SPARC V7+ diff --git a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md index 6db4a672936..a13796e0cf8 100644 --- a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md +++ b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md @@ -12,10 +12,9 @@ This target will enable all mandary features of rva23u64 and rva23s64 by default ## Requirements -This target requires: +This target can be sucessfully build on the following platform: ubuntu 24.04 (Linux Kernel version 6.8.0, glibc 2.39). -* Linux Kernel version 4.20 or later -* glibc 2.17 or later +Other platforms may work, but are not tested. Please contanct if you encounter any issues. ## Building the target -- cgit 1.4.1-3-g733a5 From ce07b5d3fc3d3bac072527ba29900f2238c11aba Mon Sep 17 00:00:00 2001 From: Zhongyao Chen Date: Tue, 19 Aug 2025 08:58:48 +0800 Subject: remove rva23s64 from riscv64a23-unknown-linux-gnu target --- compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs | 2 +- src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs b/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs index 394f8b0685d..60f2e7da042 100644 --- a/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs +++ b/compiler/rustc_target/src/spec/targets/riscv64a23_unknown_linux_gnu.rs @@ -17,7 +17,7 @@ pub(crate) fn target() -> Target { options: TargetOptions { code_model: Some(CodeModel::Medium), cpu: "generic-rv64".into(), - features: "+rva23u64,+rva23s64".into(), + features: "+rva23u64".into(), llvm_abiname: "lp64d".into(), max_atomic_width: Some(64), supported_split_debuginfo: Cow::Borrowed(&[SplitDebuginfo::Off]), diff --git a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md index a13796e0cf8..2cbaaa86654 100644 --- a/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md +++ b/src/doc/rustc/src/platform-support/riscv64a23-unknown-linux-gnu.md @@ -3,7 +3,7 @@ **Tier: 3** RISC-V target using the ratified [RVA23 Profile](https://github.com/riscv/riscv-profiles/blob/main/src/rva23-profile.adoc). -This target will enable all mandary features of rva23u64 and rva23s64 by default. +This target will enable all mandary features of rva23u64 by default. ## Target maintainers -- cgit 1.4.1-3-g733a5 From 45ea228c42cfb6712be36a0b4b64b43b8651a54f Mon Sep 17 00:00:00 2001 From: Zhongyao Chen Date: Tue, 19 Aug 2025 10:33:54 +0800 Subject: completely remove rva23s64 --- compiler/rustc_codegen_llvm/src/llvm_util.rs | 6 +-- compiler/rustc_target/src/target_features.rs | 70 ---------------------------- tests/ui/check-cfg/target_feature.stderr | 14 ------ 3 files changed, 1 insertion(+), 89 deletions(-) diff --git a/compiler/rustc_codegen_llvm/src/llvm_util.rs b/compiler/rustc_codegen_llvm/src/llvm_util.rs index a85a894e42c..d81c4c329b3 100644 --- a/compiler/rustc_codegen_llvm/src/llvm_util.rs +++ b/compiler/rustc_codegen_llvm/src/llvm_util.rs @@ -278,11 +278,7 @@ pub(crate) fn to_llvm_features<'a>(sess: &Session, s: &'a str) -> Option - { - None - } + ("riscv32" | "riscv64", "zacas" | "rva23u64" | "supm") if get_version().0 < 20 => None, ( "s390x", "message-security-assist-extension12" diff --git a/compiler/rustc_target/src/target_features.rs b/compiler/rustc_target/src/target_features.rs index 434f6a840b6..0729c44bc91 100644 --- a/compiler/rustc_target/src/target_features.rs +++ b/compiler/rustc_target/src/target_features.rs @@ -597,62 +597,6 @@ static RISCV_FEATURES: &[(&str, Stability, ImpliedFeatures)] = &[ ), ("m", Stable, &[]), ("relax", Unstable(sym::riscv_target_feature), &[]), - ( - "rva23s64", - Unstable(sym::riscv_target_feature), - &[ - "m", - "a", - "f", - "d", - "c", - "b", - "v", - "zicsr", - "zicntr", - "zihpm", - "ziccif", - "ziccrse", - "ziccamoa", - "zicclsm", - "zic64b", - "za64rs", - "zihintpause", - "zba", - "zbb", - "zbs", - "zicbom", - "zicbop", - "zicboz", - "zfhmin", - "zkt", - "zvfhmin", - "zvbb", - "zvkt", - "zihintntl", - "zicond", - "zimop", - "zcmop", - "zcb", - "zfa", - "zawrs", - "svbare", - "svade", - "ssccptr", - "sstvecd", - "sstvala", - "sscounterenw", - "svpbmt", - "svinval", - "svnapot", - "sstc", - "sscofpmf", - "ssnpm", - "ssu64xl", - "sha", - "supm", - ], - ), ( "rva23u64", Unstable(sym::riscv_target_feature), @@ -695,21 +639,7 @@ static RISCV_FEATURES: &[(&str, Stability, ImpliedFeatures)] = &[ "supm", ], ), - ("sha", Unstable(sym::riscv_target_feature), &[]), - ("ssccptr", Unstable(sym::riscv_target_feature), &[]), - ("sscofpmf", Unstable(sym::riscv_target_feature), &[]), - ("sscounterenw", Unstable(sym::riscv_target_feature), &[]), - ("ssnpm", Unstable(sym::riscv_target_feature), &[]), - ("sstc", Unstable(sym::riscv_target_feature), &[]), - ("sstvala", Unstable(sym::riscv_target_feature), &[]), - ("sstvecd", Unstable(sym::riscv_target_feature), &[]), - ("ssu64xl", Unstable(sym::riscv_target_feature), &[]), ("supm", Unstable(sym::riscv_target_feature), &[]), - ("svade", Unstable(sym::riscv_target_feature), &[]), - ("svbare", Unstable(sym::riscv_target_feature), &[]), - ("svinval", Unstable(sym::riscv_target_feature), &[]), - ("svnapot", Unstable(sym::riscv_target_feature), &[]), - ("svpbmt", Unstable(sym::riscv_target_feature), &[]), ("unaligned-scalar-mem", Unstable(sym::riscv_target_feature), &[]), ("unaligned-vector-mem", Unstable(sym::riscv_target_feature), &[]), ("v", Unstable(sym::riscv_target_feature), &["zvl128b", "zve64d"]), diff --git a/tests/ui/check-cfg/target_feature.stderr b/tests/ui/check-cfg/target_feature.stderr index a484fb0e5f0..218755a01f7 100644 --- a/tests/ui/check-cfg/target_feature.stderr +++ b/tests/ui/check-cfg/target_feature.stderr @@ -241,7 +241,6 @@ LL | cfg!(target_feature = "_UNEXPECTED_VALUE"); `relax` `relaxed-simd` `rtm` -`rva23s64` `rva23u64` `sb` `scq` @@ -294,27 +293,17 @@ LL | cfg!(target_feature = "_UNEXPECTED_VALUE"); `soft-float` `spe` `ssbs` -`ssccptr` -`sscofpmf` -`sscounterenw` `sse` `sse2` `sse3` `sse4.1` `sse4.2` `sse4a` -`ssnpm` `ssse3` -`sstc` -`sstvala` -`sstvecd` -`ssu64xl` `ssve-fp8dot2` `ssve-fp8dot4` `ssve-fp8fma` `supm` -`svade` -`svbare` `sve` `sve-b16b16` `sve2` @@ -323,9 +312,6 @@ LL | cfg!(target_feature = "_UNEXPECTED_VALUE"); `sve2-sha3` `sve2-sm4` `sve2p1` -`svinval` -`svnapot` -`svpbmt` `tail-call` `tbm` `thumb-mode` -- cgit 1.4.1-3-g733a5 From 443860d1ae4310e12188e05278fc34fd01ace1c6 Mon Sep 17 00:00:00 2001 From: Michael Howell Date: Mon, 18 Aug 2025 22:15:03 -0700 Subject: ptr_as_ptr: use correct span context for `pointer::cast` sugg --- clippy_lints/src/casts/ptr_as_ptr.rs | 2 +- tests/ui/ptr_as_ptr.fixed | 6 ++++++ tests/ui/ptr_as_ptr.rs | 6 ++++++ tests/ui/ptr_as_ptr.stderr | 8 +++++++- 4 files changed, 20 insertions(+), 2 deletions(-) diff --git a/clippy_lints/src/casts/ptr_as_ptr.rs b/clippy_lints/src/casts/ptr_as_ptr.rs index 89075409098..30e4adefd1f 100644 --- a/clippy_lints/src/casts/ptr_as_ptr.rs +++ b/clippy_lints/src/casts/ptr_as_ptr.rs @@ -78,7 +78,7 @@ pub(super) fn check<'tcx>(cx: &LateContext<'tcx>, expr: &Expr<'tcx>, msrv: Msrv) let method = snippet_with_applicability(cx, qpath_span_without_turbofish(method), "..", &mut app); ("try call directly", format!("{method}{turbofish}()")) } else { - let cast_expr_sugg = Sugg::hir_with_applicability(cx, cast_expr, "_", &mut app); + let cast_expr_sugg = Sugg::hir_with_context(cx, cast_expr, expr.span.ctxt(), "_", &mut app); ( "try `pointer::cast`, a safer alternative", diff --git a/tests/ui/ptr_as_ptr.fixed b/tests/ui/ptr_as_ptr.fixed index 78e1ceb480a..4457878b6fa 100644 --- a/tests/ui/ptr_as_ptr.fixed +++ b/tests/ui/ptr_as_ptr.fixed @@ -229,3 +229,9 @@ fn issue15283() { //~^ ptr_as_ptr } } + +fn issue15232() { + let mut b = Box::new(0i32); + let _ptr = std::ptr::addr_of_mut!(*b).cast::<()>(); + //~^ ptr_as_ptr +} diff --git a/tests/ui/ptr_as_ptr.rs b/tests/ui/ptr_as_ptr.rs index 70732cf0a6c..9124fc912d2 100644 --- a/tests/ui/ptr_as_ptr.rs +++ b/tests/ui/ptr_as_ptr.rs @@ -229,3 +229,9 @@ fn issue15283() { //~^ ptr_as_ptr } } + +fn issue15232() { + let mut b = Box::new(0i32); + let _ptr = std::ptr::addr_of_mut!(*b) as *mut (); + //~^ ptr_as_ptr +} diff --git a/tests/ui/ptr_as_ptr.stderr b/tests/ui/ptr_as_ptr.stderr index c0a2a4b6d20..af21c1e35f5 100644 --- a/tests/ui/ptr_as_ptr.stderr +++ b/tests/ui/ptr_as_ptr.stderr @@ -199,5 +199,11 @@ error: `as` casting between raw pointers without changing their constness LL | let _: fn() = std::mem::transmute(std::ptr::null::<()>() as *const u8); | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: try call directly: `std::ptr::null::()` -error: aborting due to 33 previous errors +error: `as` casting between raw pointers without changing their constness + --> tests/ui/ptr_as_ptr.rs:235:16 + | +LL | let _ptr = std::ptr::addr_of_mut!(*b) as *mut (); + | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: try `pointer::cast`, a safer alternative: `std::ptr::addr_of_mut!(*b).cast::<()>()` + +error: aborting due to 34 previous errors -- cgit 1.4.1-3-g733a5 From 6a24ae96eaf175b048e70dfb46dfb860ee7147d9 Mon Sep 17 00:00:00 2001 From: Tobias Bucher Date: Mon, 18 Aug 2025 11:50:27 +0200 Subject: Simplify macro generating `ToString` implementations for `&…&str` MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use deref coercion to let the compiler remove any amount of references. --- library/alloc/src/string.rs | 55 +++++++++++++++++---------------------------- 1 file changed, 21 insertions(+), 34 deletions(-) diff --git a/library/alloc/src/string.rs b/library/alloc/src/string.rs index 9eacbf00e42..b1f2cb598c8 100644 --- a/library/alloc/src/string.rs +++ b/library/alloc/src/string.rs @@ -2949,52 +2949,39 @@ impl SpecToString for i8 { } } -// Generic/generated code can sometimes have multiple, nested references -// for strings, including `&&&str`s that would never be written -// by hand. This macro generates twelve layers of nested `&`-impl -// for primitive strings. -#[cfg(not(no_global_oom_handling))] -macro_rules! to_string_str_wrap_in_ref { - {x $($x:ident)*} => { - &to_string_str_wrap_in_ref! { $($x)* } - }; - {} => { str }; -} -#[cfg(not(no_global_oom_handling))] -macro_rules! to_string_expr_wrap_in_deref { - {$self:expr ; x $($x:ident)*} => { - *(to_string_expr_wrap_in_deref! { $self ; $($x)* }) - }; - {$self:expr ;} => { $self }; -} #[cfg(not(no_global_oom_handling))] macro_rules! to_string_str { - {$($($x:ident)*),+} => { + {$($type:ty,)*} => { $( - impl SpecToString for to_string_str_wrap_in_ref!($($x)*) { + impl SpecToString for $type { #[inline] fn spec_to_string(&self) -> String { - String::from(to_string_expr_wrap_in_deref!(self ; $($x)*)) + let s: &str = self; + String::from(s) } } - )+ + )* }; } #[cfg(not(no_global_oom_handling))] to_string_str! { - x x x x x x x x x x x x, - x x x x x x x x x x x, - x x x x x x x x x x, - x x x x x x x x x, - x x x x x x x x, - x x x x x x x, - x x x x x x, - x x x x x, - x x x x, - x x x, - x x, - x, + // Generic/generated code can sometimes have multiple, nested references + // for strings, including `&&&str`s that would never be written + // by hand. + &&&&&&&&&&&&str, + &&&&&&&&&&&str, + &&&&&&&&&&str, + &&&&&&&&&str, + &&&&&&&&str, + &&&&&&&str, + &&&&&&str, + &&&&&str, + &&&&str, + &&&str, + &&str, + &str, + str, } #[cfg(not(no_global_oom_handling))] -- cgit 1.4.1-3-g733a5 From cb5a4d138a298853225d7561c243ce6e23064775 Mon Sep 17 00:00:00 2001 From: Tobias Bucher Date: Mon, 18 Aug 2025 11:51:31 +0200 Subject: Use `ToString` specialization macro also for `Cow` and `String` --- library/alloc/src/string.rs | 18 ++---------------- 1 file changed, 2 insertions(+), 16 deletions(-) diff --git a/library/alloc/src/string.rs b/library/alloc/src/string.rs index b1f2cb598c8..ec4d934a5a3 100644 --- a/library/alloc/src/string.rs +++ b/library/alloc/src/string.rs @@ -2966,6 +2966,8 @@ macro_rules! to_string_str { #[cfg(not(no_global_oom_handling))] to_string_str! { + Cow<'_, str>, + String, // Generic/generated code can sometimes have multiple, nested references // for strings, including `&&&str`s that would never be written // by hand. @@ -2984,22 +2986,6 @@ to_string_str! { str, } -#[cfg(not(no_global_oom_handling))] -impl SpecToString for Cow<'_, str> { - #[inline] - fn spec_to_string(&self) -> String { - self[..].to_owned() - } -} - -#[cfg(not(no_global_oom_handling))] -impl SpecToString for String { - #[inline] - fn spec_to_string(&self) -> String { - self.to_owned() - } -} - #[cfg(not(no_global_oom_handling))] impl SpecToString for fmt::Arguments<'_> { #[inline] -- cgit 1.4.1-3-g733a5 From a21d7d7529a0ef8c937c9ed541f990ae0125b725 Mon Sep 17 00:00:00 2001 From: Lorenz Leutgeb Date: Tue, 19 Aug 2025 18:59:25 +0200 Subject: Fix doc of `std::os::windows::io::BorrowedSocket::borrow_raw` A small oversight in 0cb69dec57f I noticed while reading. --- library/std/src/os/windows/io/socket.rs | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/library/std/src/os/windows/io/socket.rs b/library/std/src/os/windows/io/socket.rs index 1c228914de9..28e972925e6 100644 --- a/library/std/src/os/windows/io/socket.rs +++ b/library/std/src/os/windows/io/socket.rs @@ -54,7 +54,7 @@ impl BorrowedSocket<'_> { /// /// # Safety /// - /// The resource pointed to by `raw` must remain open for the duration of + /// The resource pointed to by `socket` must remain open for the duration of /// the returned `BorrowedSocket`, and it must not have the value /// `INVALID_SOCKET`. #[inline] -- cgit 1.4.1-3-g733a5 From ca7a0aff74668ec62f4eb84a6b7655a2dfb821e2 Mon Sep 17 00:00:00 2001 From: Stefan Schindler Date: Wed, 20 Aug 2025 01:03:18 +0200 Subject: Fix some typos --- library/compiler-builtins/compiler-builtins/src/float/add.rs | 2 +- library/compiler-builtins/compiler-builtins/src/probestack.rs | 2 +- library/compiler-builtins/etc/update-api-list.py | 2 +- library/compiler-builtins/libm/src/math/rem_pio2_large.rs | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/library/compiler-builtins/compiler-builtins/src/float/add.rs b/library/compiler-builtins/compiler-builtins/src/float/add.rs index 8dbfb0e1020..acdcd2ebe31 100644 --- a/library/compiler-builtins/compiler-builtins/src/float/add.rs +++ b/library/compiler-builtins/compiler-builtins/src/float/add.rs @@ -130,7 +130,7 @@ where return F::from_bits(MinInt::ZERO); } - // If partial cancellation occured, we need to left-shift the result + // If partial cancellation occurred, we need to left-shift the result // and adjust the exponent: if a_significand < implicit_bit << 3 { let shift = a_significand.leading_zeros() as i32 diff --git a/library/compiler-builtins/compiler-builtins/src/probestack.rs b/library/compiler-builtins/compiler-builtins/src/probestack.rs index 9a18216da99..72975485a77 100644 --- a/library/compiler-builtins/compiler-builtins/src/probestack.rs +++ b/library/compiler-builtins/compiler-builtins/src/probestack.rs @@ -73,7 +73,7 @@ pub unsafe extern "custom" fn __rust_probestack() { // page needed. // // Note that we're also testing against `8(%rsp)` to account for the 8 - // bytes pushed on the stack orginally with our return address. Using + // bytes pushed on the stack originally with our return address. Using // `8(%rsp)` simulates us testing the stack pointer in the caller's // context. diff --git a/library/compiler-builtins/etc/update-api-list.py b/library/compiler-builtins/etc/update-api-list.py index 28ff22f4cbb..76c75cbf4dc 100755 --- a/library/compiler-builtins/etc/update-api-list.py +++ b/library/compiler-builtins/etc/update-api-list.py @@ -34,7 +34,7 @@ def eprint(*args, **kwargs): @dataclass class Crate: - """Representation of public interfaces and function defintion locations in + """Representation of public interfaces and function definition locations in `libm`. """ diff --git a/library/compiler-builtins/libm/src/math/rem_pio2_large.rs b/library/compiler-builtins/libm/src/math/rem_pio2_large.rs index f1fdf3673a8..bb2c532916b 100644 --- a/library/compiler-builtins/libm/src/math/rem_pio2_large.rs +++ b/library/compiler-builtins/libm/src/math/rem_pio2_large.rs @@ -146,7 +146,7 @@ const PIO2: [f64; 8] = [ // x[i] = floor(z) // z = (z-x[i])*2**24 // -// y[] ouput result in an array of double precision numbers. +// y[] output result in an array of double precision numbers. // The dimension of y[] is: // 24-bit precision 1 // 53-bit precision 2 -- cgit 1.4.1-3-g733a5 From 7d09ce84ee0b9b63e44ec90241ff9e6584068164 Mon Sep 17 00:00:00 2001 From: "dependabot[bot]" <49699333+dependabot[bot]@users.noreply.github.com> Date: Wed, 20 Aug 2025 09:50:54 +0000 Subject: Bump slab from 0.4.10 to 0.4.11 in /tests/deps Bumps [slab](https://github.com/tokio-rs/slab) from 0.4.10 to 0.4.11. - [Release notes](https://github.com/tokio-rs/slab/releases) - [Changelog](https://github.com/tokio-rs/slab/blob/master/CHANGELOG.md) - [Commits](https://github.com/tokio-rs/slab/compare/v0.4.10...v0.4.11) --- updated-dependencies: - dependency-name: slab dependency-version: 0.4.11 dependency-type: indirect ... Signed-off-by: dependabot[bot] --- src/tools/miri/tests/deps/Cargo.lock | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/tools/miri/tests/deps/Cargo.lock b/src/tools/miri/tests/deps/Cargo.lock index 4b783ebdc4e..65ca4215c60 100644 --- a/src/tools/miri/tests/deps/Cargo.lock +++ b/src/tools/miri/tests/deps/Cargo.lock @@ -296,9 +296,9 @@ dependencies = [ [[package]] name = "slab" -version = "0.4.10" +version = "0.4.11" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "04dc19736151f35336d325007ac991178d504a119863a2fcb3758cdb5e52c50d" +checksum = "7a2ae44ef20feb57a68b23d846850f861394c2e02dc425a50098ae8c90267589" [[package]] name = "socket2" -- cgit 1.4.1-3-g733a5 From e49d0008f77ff963185801cf9dff4a72958f95fb Mon Sep 17 00:00:00 2001 From: Scott McMurray Date: Sat, 26 Jul 2025 01:09:45 -0700 Subject: Partial-stabilize the basics from `bigint_helper_methods` --- library/core/src/num/uint_macros.rs | 66 +++++++++++++++++++++++-------------- 1 file changed, 42 insertions(+), 24 deletions(-) diff --git a/library/core/src/num/uint_macros.rs b/library/core/src/num/uint_macros.rs index 10d9498d15e..c556d951034 100644 --- a/library/core/src/num/uint_macros.rs +++ b/library/core/src/num/uint_macros.rs @@ -2447,7 +2447,7 @@ macro_rules! uint_impl { } /// Calculates `self` + `rhs` + `carry` and returns a tuple containing - /// the sum and the output carry. + /// the sum and the output carry (in that order). /// /// Performs "ternary addition" of two integer operands and a carry-in /// bit, and returns an output integer and a carry-out bit. This allows @@ -2465,8 +2465,6 @@ macro_rules! uint_impl { /// # Examples /// /// ``` - /// #![feature(bigint_helper_methods)] - /// #[doc = concat!("// 3 MAX (a = 3 × 2^", stringify!($BITS), " + 2^", stringify!($BITS), " - 1)")] #[doc = concat!("// + 5 7 (b = 5 × 2^", stringify!($BITS), " + 7)")] /// // --------- @@ -2483,7 +2481,7 @@ macro_rules! uint_impl { /// /// assert_eq!((sum1, sum0), (9, 6)); /// ``` - #[unstable(feature = "bigint_helper_methods", issue = "85532")] + #[stable(feature = "unsigned_bigint_helpers", since = "CURRENT_RUSTC_VERSION")] #[rustc_const_unstable(feature = "bigint_helper_methods", issue = "85532")] #[must_use = "this returns the result of the operation, \ without modifying the original"] @@ -2559,8 +2557,6 @@ macro_rules! uint_impl { /// # Examples /// /// ``` - /// #![feature(bigint_helper_methods)] - /// #[doc = concat!("// 9 6 (a = 9 × 2^", stringify!($BITS), " + 6)")] #[doc = concat!("// - 5 7 (b = 5 × 2^", stringify!($BITS), " + 7)")] /// // --------- @@ -2577,7 +2573,7 @@ macro_rules! uint_impl { /// #[doc = concat!("assert_eq!((diff1, diff0), (3, ", stringify!($SelfT), "::MAX));")] /// ``` - #[unstable(feature = "bigint_helper_methods", issue = "85532")] + #[stable(feature = "unsigned_bigint_helpers", since = "CURRENT_RUSTC_VERSION")] #[rustc_const_unstable(feature = "bigint_helper_methods", issue = "85532")] #[must_use = "this returns the result of the operation, \ without modifying the original"] @@ -2651,10 +2647,12 @@ macro_rules! uint_impl { /// indicating whether an arithmetic overflow would occur. If an /// overflow would have occurred then the wrapped value is returned. /// + /// If you want the *value* of the overflow, rather than just *whether* + /// an overflow occurred, see [`Self::carrying_mul`]. + /// /// # Examples /// - /// Please note that this example is shared among integer types, which is why why `u32` - /// is used. + /// Please note that this example is shared among integer types, which is why `u32` is used. /// /// ``` /// assert_eq!(5u32.overflowing_mul(2), (10, false)); @@ -2670,16 +2668,38 @@ macro_rules! uint_impl { (a as Self, b) } - /// Calculates the complete product `self * rhs` without the possibility to overflow. + /// Calculates the complete double-width product `self * rhs`. /// /// This returns the low-order (wrapping) bits and the high-order (overflow) bits - /// of the result as two separate values, in that order. + /// of the result as two separate values, in that order. As such, + /// `a.widening_mul(b).0` produces the same result as `a.wrapping_mul(b)`. + /// + /// If you also need to add a value and carry to the wide result, then you want + /// [`Self::carrying_mul_add`] instead. /// /// If you also need to add a carry to the wide result, then you want /// [`Self::carrying_mul`] instead. /// + /// If you just want to know *whether* the multiplication overflowed, then you + /// want [`Self::overflowing_mul`] instead. + /// /// # Examples /// + /// ``` + /// #![feature(bigint_helper_methods)] + #[doc = concat!("assert_eq!(5_", stringify!($SelfT), ".widening_mul(7), (35, 0));")] + #[doc = concat!("assert_eq!(", stringify!($SelfT), "::MAX.widening_mul(", stringify!($SelfT), "::MAX), (1, ", stringify!($SelfT), "::MAX - 1));")] + /// ``` + /// + /// Compared to other `*_mul` methods: + /// ``` + /// #![feature(bigint_helper_methods)] + #[doc = concat!("assert_eq!(", stringify!($SelfT), "::widening_mul(1 << ", stringify!($BITS_MINUS_ONE), ", 6), (0, 3));")] + #[doc = concat!("assert_eq!(", stringify!($SelfT), "::overflowing_mul(1 << ", stringify!($BITS_MINUS_ONE), ", 6), (0, true));")] + #[doc = concat!("assert_eq!(", stringify!($SelfT), "::wrapping_mul(1 << ", stringify!($BITS_MINUS_ONE), ", 6), 0);")] + #[doc = concat!("assert_eq!(", stringify!($SelfT), "::checked_mul(1 << ", stringify!($BITS_MINUS_ONE), ", 6), None);")] + /// ``` + /// /// Please note that this example is shared among integer types, which is why `u32` is used. /// /// ``` @@ -2706,14 +2726,13 @@ macro_rules! uint_impl { /// additional amount of overflow. This allows for chaining together multiple /// multiplications to create "big integers" which represent larger values. /// - /// If you don't need the `carry`, then you can use [`Self::widening_mul`] instead. + /// If you also need to add a value, then use [`Self::carrying_mul_add`]. /// /// # Examples /// /// Please note that this example is shared among integer types, which is why `u32` is used. /// /// ``` - /// #![feature(bigint_helper_methods)] /// assert_eq!(5u32.carrying_mul(2, 0), (10, 0)); /// assert_eq!(5u32.carrying_mul(2, 10), (20, 0)); /// assert_eq!(1_000_000_000u32.carrying_mul(10, 0), (1410065408, 2)); @@ -2771,7 +2790,7 @@ macro_rules! uint_impl { /// 789_u16.wrapping_mul(456).wrapping_add(123), /// ); /// ``` - #[unstable(feature = "bigint_helper_methods", issue = "85532")] + #[stable(feature = "unsigned_bigint_helpers", since = "CURRENT_RUSTC_VERSION")] #[rustc_const_unstable(feature = "bigint_helper_methods", issue = "85532")] #[must_use = "this returns the result of the operation, \ without modifying the original"] @@ -2780,18 +2799,20 @@ macro_rules! uint_impl { Self::carrying_mul_add(self, rhs, carry, 0) } - /// Calculates the "full multiplication" `self * rhs + carry1 + carry2` - /// without the possibility to overflow. + /// Calculates the "full multiplication" `self * rhs + carry1 + carry2`. /// /// This returns the low-order (wrapping) bits and the high-order (overflow) bits /// of the result as two separate values, in that order. /// + /// This cannot overflow, as the double-width result has exactly enough + /// space for the largest possible result. This is equivalent to how, in + /// decimal, 9 × 9 + 9 + 9 = 81 + 18 = 99 = 9×10⁰ + 9×10¹ = 10² - 1. + /// /// Performs "long multiplication" which takes in an extra amount to add, and may return an /// additional amount of overflow. This allows for chaining together multiple /// multiplications to create "big integers" which represent larger values. /// - /// If you don't need either `carry`, then you can use [`Self::widening_mul`] instead, - /// and if you only need one `carry`, then you can use [`Self::carrying_mul`] instead. + /// If you don't need the `add` part, then you can use [`Self::carrying_mul`] instead. /// /// # Examples /// @@ -2799,7 +2820,6 @@ macro_rules! uint_impl { /// which explains why `u32` is used here. /// /// ``` - /// #![feature(bigint_helper_methods)] /// assert_eq!(5u32.carrying_mul_add(2, 0, 0), (10, 0)); /// assert_eq!(5u32.carrying_mul_add(2, 10, 10), (30, 0)); /// assert_eq!(1_000_000_000u32.carrying_mul_add(10, 0, 0), (1410065408, 2)); @@ -2816,8 +2836,6 @@ macro_rules! uint_impl { /// using `u8` for simplicity of the demonstration. /// /// ``` - /// #![feature(bigint_helper_methods)] - /// /// fn quadratic_mul(a: [u8; N], b: [u8; N]) -> [u8; N] { /// let mut out = [0; N]; /// for j in 0..N { @@ -2832,13 +2850,13 @@ macro_rules! uint_impl { /// // -1 * -1 == 1 /// assert_eq!(quadratic_mul([0xFF; 3], [0xFF; 3]), [1, 0, 0]); /// - /// assert_eq!(u32::wrapping_mul(0x9e3779b9, 0x7f4a7c15), 0xCFFC982D); + /// assert_eq!(u32::wrapping_mul(0x9e3779b9, 0x7f4a7c15), 0xcffc982d); /// assert_eq!( /// quadratic_mul(u32::to_le_bytes(0x9e3779b9), u32::to_le_bytes(0x7f4a7c15)), - /// u32::to_le_bytes(0xCFFC982D) + /// u32::to_le_bytes(0xcffc982d) /// ); /// ``` - #[unstable(feature = "bigint_helper_methods", issue = "85532")] + #[stable(feature = "unsigned_bigint_helpers", since = "CURRENT_RUSTC_VERSION")] #[rustc_const_unstable(feature = "bigint_helper_methods", issue = "85532")] #[must_use = "this returns the result of the operation, \ without modifying the original"] -- cgit 1.4.1-3-g733a5 From d216ca0506d6a70ca70fc29974ed9155beaabf7a Mon Sep 17 00:00:00 2001 From: Esteban Küber Date: Mon, 18 Aug 2025 16:52:02 +0000 Subject: Detect missing `if let` or `let-else` During `let` binding parse error and encountering a block, detect if there is a likely missing `if` or `else`: ``` error: expected one of `.`, `;`, `?`, `else`, or an operator, found `{` --> $DIR/missing-if-let-or-let-else.rs:14:25 | LL | let Some(x) = foo() { | ^ expected one of `.`, `;`, `?`, `else`, or an operator | help: you might have meant to use `if let` | LL | if let Some(x) = foo() { | ++ help: alternatively, you might have meant to use `let else` | LL | let Some(x) = foo() else { | ++++ ``` --- compiler/rustc_parse/src/lib.rs | 1 + compiler/rustc_parse/src/parser/stmt.rs | 94 ++++++++++++++++++++++ tests/ui/uninhabited/missing-if-let-or-let-else.rs | 24 ++++++ .../uninhabited/missing-if-let-or-let-else.stderr | 39 +++++++++ 4 files changed, 158 insertions(+) create mode 100644 tests/ui/uninhabited/missing-if-let-or-let-else.rs create mode 100644 tests/ui/uninhabited/missing-if-let-or-let-else.stderr diff --git a/compiler/rustc_parse/src/lib.rs b/compiler/rustc_parse/src/lib.rs index 2050c5f9608..0dd5d89491b 100644 --- a/compiler/rustc_parse/src/lib.rs +++ b/compiler/rustc_parse/src/lib.rs @@ -6,6 +6,7 @@ #![feature(assert_matches)] #![feature(box_patterns)] #![feature(debug_closure_helpers)] +#![feature(default_field_values)] #![feature(if_let_guard)] #![feature(iter_intersperse)] #![recursion_limit = "256"] diff --git a/compiler/rustc_parse/src/parser/stmt.rs b/compiler/rustc_parse/src/parser/stmt.rs index b4943ff7de6..732c653e4bc 100644 --- a/compiler/rustc_parse/src/parser/stmt.rs +++ b/compiler/rustc_parse/src/parser/stmt.rs @@ -6,6 +6,7 @@ use ast::Label; use rustc_ast as ast; use rustc_ast::token::{self, Delimiter, InvisibleOrigin, MetaVarKind, TokenKind}; use rustc_ast::util::classify::{self, TrailingBrace}; +use rustc_ast::visit::{Visitor, walk_expr}; use rustc_ast::{ AttrStyle, AttrVec, Block, BlockCheckMode, DUMMY_NODE_ID, Expr, ExprKind, HasAttrs, Local, LocalKind, MacCall, MacCallStmt, MacStmtStyle, Recovered, Stmt, StmtKind, @@ -783,6 +784,71 @@ impl<'a> Parser<'a> { Ok(self.mk_block(stmts, s, lo.to(self.prev_token.span))) } + fn recover_missing_let_else(&mut self, err: &mut Diag<'_>, pat: &ast::Pat, stmt_span: Span) { + if self.token.kind != token::OpenBrace { + return; + } + match pat.kind { + ast::PatKind::Ident(..) | ast::PatKind::Missing | ast::PatKind::Wild => { + // Not if let or let else + return; + } + _ => {} + } + let snapshot = self.create_snapshot_for_diagnostic(); + let block_span = self.token.span; + let (if_let, let_else) = match self.parse_block() { + Ok(block) => { + let mut idents = vec![]; + pat.walk(&mut |pat: &ast::Pat| { + if let ast::PatKind::Ident(_, ident, _) = pat.kind { + idents.push(ident); + } + true + }); + // Collect all bindings in pattern and see if they appear in the block. Likely meant + // to write `if let`. See if the block has a return. Likely meant to write + // `let else`. + let mut visitor = IdentFinder { idents, .. }; + visitor.visit_block(&block); + + (visitor.references_ident, visitor.has_return) + } + Err(e) => { + e.cancel(); + self.restore_snapshot(snapshot); + (false, false) + } + }; + + let mut alternatively = ""; + if if_let || !let_else { + alternatively = "alternatively, "; + err.span_suggestion_verbose( + stmt_span.shrink_to_lo(), + "you might have meant to use `if let`", + "if ".to_string(), + if if_let { + Applicability::MachineApplicable + } else { + Applicability::MaybeIncorrect + }, + ); + } + if let_else || !if_let { + err.span_suggestion_verbose( + block_span.shrink_to_lo(), + format!("{alternatively}you might have meant to use `let else`"), + "else ".to_string(), + if let_else { + Applicability::MachineApplicable + } else { + Applicability::MaybeIncorrect + }, + ); + } + } + fn recover_missing_dot(&mut self, err: &mut Diag<'_>) { let Some((ident, _)) = self.token.ident() else { return; @@ -977,6 +1043,7 @@ impl<'a> Parser<'a> { self.check_mistyped_turbofish_with_multiple_type_params(e, expr).map_err( |mut e| { self.recover_missing_dot(&mut e); + self.recover_missing_let_else(&mut e, &local.pat, stmt.span); e }, )?; @@ -1065,3 +1132,30 @@ impl<'a> Parser<'a> { self.mk_block(thin_vec![self.mk_stmt_err(span, guar)], BlockCheckMode::Default, span) } } + +struct IdentFinder { + idents: Vec, + /// If a block references one of the bindings introduced by the let pattern, we likely meant to + /// use `if let`. + /// This is pre-expansion, so if we encounter `let Some(x) = foo() { println!("{x}") }` we won't + /// find it. + references_ident: bool = false, + /// If a block has a `return`, then we know with high certainty that the + has_return: bool = false, +} + +impl<'a> Visitor<'a> for IdentFinder { + fn visit_ident(&mut self, ident: &Ident) { + for i in &self.idents { + if ident.name == i.name { + self.references_ident = true; + } + } + } + fn visit_expr(&mut self, node: &'a Expr) { + if let ExprKind::Ret(..) = node.kind { + self.has_return = true; + } + walk_expr(self, node); + } +} diff --git a/tests/ui/uninhabited/missing-if-let-or-let-else.rs b/tests/ui/uninhabited/missing-if-let-or-let-else.rs new file mode 100644 index 00000000000..51fedb79756 --- /dev/null +++ b/tests/ui/uninhabited/missing-if-let-or-let-else.rs @@ -0,0 +1,24 @@ +fn a() { + let Some(x) = foo() { //~ ERROR expected one of + //~^ HELP you might have meant to use `if let` + let y = x; + } +} +fn b() { + let Some(x) = foo() { //~ ERROR expected one of + //~^ HELP you might have meant to use `let else` + return; + } +} +fn c() { + let Some(x) = foo() { //~ ERROR expected one of + //~^ HELP you might have meant to use `if let` + //~| HELP alternatively, you might have meant to use `let else` + // The parser check happens pre-macro-expansion, so we don't know for sure. + println!("{x}"); + } +} +fn foo() -> Option { + Some(42) +} +fn main() {} diff --git a/tests/ui/uninhabited/missing-if-let-or-let-else.stderr b/tests/ui/uninhabited/missing-if-let-or-let-else.stderr new file mode 100644 index 00000000000..4b78a0fa16e --- /dev/null +++ b/tests/ui/uninhabited/missing-if-let-or-let-else.stderr @@ -0,0 +1,39 @@ +error: expected one of `.`, `;`, `?`, `else`, or an operator, found `{` + --> $DIR/missing-if-let-or-let-else.rs:2:25 + | +LL | let Some(x) = foo() { + | ^ expected one of `.`, `;`, `?`, `else`, or an operator + | +help: you might have meant to use `if let` + | +LL | if let Some(x) = foo() { + | ++ + +error: expected one of `.`, `;`, `?`, `else`, or an operator, found `{` + --> $DIR/missing-if-let-or-let-else.rs:8:25 + | +LL | let Some(x) = foo() { + | ^ expected one of `.`, `;`, `?`, `else`, or an operator + | +help: you might have meant to use `let else` + | +LL | let Some(x) = foo() else { + | ++++ + +error: expected one of `.`, `;`, `?`, `else`, or an operator, found `{` + --> $DIR/missing-if-let-or-let-else.rs:14:25 + | +LL | let Some(x) = foo() { + | ^ expected one of `.`, `;`, `?`, `else`, or an operator + | +help: you might have meant to use `if let` + | +LL | if let Some(x) = foo() { + | ++ +help: alternatively, you might have meant to use `let else` + | +LL | let Some(x) = foo() else { + | ++++ + +error: aborting due to 3 previous errors + -- cgit 1.4.1-3-g733a5 From b51a8ed2439e21f2980160bdf6af043e42d55434 Mon Sep 17 00:00:00 2001 From: Ed Page Date: Fri, 8 Aug 2025 16:31:53 -0500 Subject: test(rustfmt): Verify frontmatter is preserved This is to prove that the frontmatter is preserved. The choices in tests is intended for showing the different parts of the proposed Style Guide for frontmatters. --- src/tools/rustfmt/tests/source/frontmatter_compact.rs | 8 ++++++++ src/tools/rustfmt/tests/source/frontmatter_escaped.rs | 13 +++++++++++++ src/tools/rustfmt/tests/source/frontmatter_spaced.rs | 16 ++++++++++++++++ src/tools/rustfmt/tests/target/frontmatter_compact.rs | 8 ++++++++ src/tools/rustfmt/tests/target/frontmatter_escaped.rs | 13 +++++++++++++ src/tools/rustfmt/tests/target/frontmatter_spaced.rs | 16 ++++++++++++++++ 6 files changed, 74 insertions(+) create mode 100644 src/tools/rustfmt/tests/source/frontmatter_compact.rs create mode 100644 src/tools/rustfmt/tests/source/frontmatter_escaped.rs create mode 100644 src/tools/rustfmt/tests/source/frontmatter_spaced.rs create mode 100644 src/tools/rustfmt/tests/target/frontmatter_compact.rs create mode 100644 src/tools/rustfmt/tests/target/frontmatter_escaped.rs create mode 100644 src/tools/rustfmt/tests/target/frontmatter_spaced.rs diff --git a/src/tools/rustfmt/tests/source/frontmatter_compact.rs b/src/tools/rustfmt/tests/source/frontmatter_compact.rs new file mode 100644 index 00000000000..21d4c6f4b61 --- /dev/null +++ b/src/tools/rustfmt/tests/source/frontmatter_compact.rs @@ -0,0 +1,8 @@ +#!/usr/bin/env cargo +---identifier +[dependencies] +regex = "1" +--- +#![feature(frontmatter)] + +fn main() {} diff --git a/src/tools/rustfmt/tests/source/frontmatter_escaped.rs b/src/tools/rustfmt/tests/source/frontmatter_escaped.rs new file mode 100644 index 00000000000..0d026377566 --- /dev/null +++ b/src/tools/rustfmt/tests/source/frontmatter_escaped.rs @@ -0,0 +1,13 @@ +#!/usr/bin/env cargo +------------ +package.description = """ +Header +----- + +Body +""" +------------ + +#![feature(frontmatter)] + +fn main() {} diff --git a/src/tools/rustfmt/tests/source/frontmatter_spaced.rs b/src/tools/rustfmt/tests/source/frontmatter_spaced.rs new file mode 100644 index 00000000000..ee0bb81705c --- /dev/null +++ b/src/tools/rustfmt/tests/source/frontmatter_spaced.rs @@ -0,0 +1,16 @@ +#!/usr/bin/env cargo + + +--- identifier +[dependencies] +regex = "1" + +--- + + + + + +#![feature(frontmatter)] + +fn main() {} diff --git a/src/tools/rustfmt/tests/target/frontmatter_compact.rs b/src/tools/rustfmt/tests/target/frontmatter_compact.rs new file mode 100644 index 00000000000..21d4c6f4b61 --- /dev/null +++ b/src/tools/rustfmt/tests/target/frontmatter_compact.rs @@ -0,0 +1,8 @@ +#!/usr/bin/env cargo +---identifier +[dependencies] +regex = "1" +--- +#![feature(frontmatter)] + +fn main() {} diff --git a/src/tools/rustfmt/tests/target/frontmatter_escaped.rs b/src/tools/rustfmt/tests/target/frontmatter_escaped.rs new file mode 100644 index 00000000000..0d026377566 --- /dev/null +++ b/src/tools/rustfmt/tests/target/frontmatter_escaped.rs @@ -0,0 +1,13 @@ +#!/usr/bin/env cargo +------------ +package.description = """ +Header +----- + +Body +""" +------------ + +#![feature(frontmatter)] + +fn main() {} diff --git a/src/tools/rustfmt/tests/target/frontmatter_spaced.rs b/src/tools/rustfmt/tests/target/frontmatter_spaced.rs new file mode 100644 index 00000000000..ee0bb81705c --- /dev/null +++ b/src/tools/rustfmt/tests/target/frontmatter_spaced.rs @@ -0,0 +1,16 @@ +#!/usr/bin/env cargo + + +--- identifier +[dependencies] +regex = "1" + +--- + + + + + +#![feature(frontmatter)] + +fn main() {} -- cgit 1.4.1-3-g733a5 From 5a2fceefd312ec027bdeaa89ebefbe4c33d94de1 Mon Sep 17 00:00:00 2001 From: ltdk Date: Wed, 13 Aug 2025 01:16:09 -0400 Subject: Copy WTF-8 code into core/alloc (for better diffs) --- library/alloc/src/wtf8/mod.rs | 1083 ++++++++++++++++++++++++++++++ library/alloc/src/wtf8/tests.rs | 751 +++++++++++++++++++++ library/core/src/wtf8.rs | 1083 ++++++++++++++++++++++++++++++ library/std/src/sys_common/wtf8.rs | 1083 ------------------------------ library/std/src/sys_common/wtf8/tests.rs | 751 --------------------- 5 files changed, 2917 insertions(+), 1834 deletions(-) create mode 100644 library/alloc/src/wtf8/mod.rs create mode 100644 library/alloc/src/wtf8/tests.rs create mode 100644 library/core/src/wtf8.rs delete mode 100644 library/std/src/sys_common/wtf8.rs delete mode 100644 library/std/src/sys_common/wtf8/tests.rs diff --git a/library/alloc/src/wtf8/mod.rs b/library/alloc/src/wtf8/mod.rs new file mode 100644 index 00000000000..50bde88b5a4 --- /dev/null +++ b/library/alloc/src/wtf8/mod.rs @@ -0,0 +1,1083 @@ +//! Implementation of [the WTF-8 encoding](https://simonsapin.github.io/wtf-8/). +//! +//! This library uses Rust’s type system to maintain +//! [well-formedness](https://simonsapin.github.io/wtf-8/#well-formed), +//! like the `String` and `&str` types do for UTF-8. +//! +//! Since [WTF-8 must not be used +//! for interchange](https://simonsapin.github.io/wtf-8/#intended-audience), +//! this library deliberately does not provide access to the underlying bytes +//! of WTF-8 strings, +//! nor can it decode WTF-8 from arbitrary bytes. +//! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points. + +// this module is imported from @SimonSapin's repo and has tons of dead code on +// unix (it's mostly used on windows), so don't worry about dead code here. +#![allow(dead_code)] + +#[cfg(test)] +mod tests; + +use core::char::{MAX_LEN_UTF8, MAX_LEN_UTF16, encode_utf8_raw, encode_utf16_raw}; +use core::clone::CloneToUninit; +use core::str::next_code_point; + +use crate::borrow::Cow; +use crate::collections::TryReserveError; +use crate::hash::{Hash, Hasher}; +use crate::iter::FusedIterator; +use crate::rc::Rc; +use crate::sync::Arc; +use crate::sys_common::AsInner; +use crate::{fmt, mem, ops, slice, str}; + +const UTF8_REPLACEMENT_CHARACTER: &str = "\u{FFFD}"; + +/// A Unicode code point: from U+0000 to U+10FFFF. +/// +/// Compares with the `char` type, +/// which represents a Unicode scalar value: +/// a code point that is not a surrogate (U+D800 to U+DFFF). +#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)] +pub struct CodePoint { + value: u32, +} + +/// Format the code point as `U+` followed by four to six hexadecimal digits. +/// Example: `U+1F4A9` +impl fmt::Debug for CodePoint { + #[inline] + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(formatter, "U+{:04X}", self.value) + } +} + +impl CodePoint { + /// Unsafely creates a new `CodePoint` without checking the value. + /// + /// Only use when `value` is known to be less than or equal to 0x10FFFF. + #[inline] + pub unsafe fn from_u32_unchecked(value: u32) -> CodePoint { + CodePoint { value } + } + + /// Creates a new `CodePoint` if the value is a valid code point. + /// + /// Returns `None` if `value` is above 0x10FFFF. + #[inline] + pub fn from_u32(value: u32) -> Option { + match value { + 0..=0x10FFFF => Some(CodePoint { value }), + _ => None, + } + } + + /// Creates a new `CodePoint` from a `char`. + /// + /// Since all Unicode scalar values are code points, this always succeeds. + #[inline] + pub fn from_char(value: char) -> CodePoint { + CodePoint { value: value as u32 } + } + + /// Returns the numeric value of the code point. + #[inline] + pub fn to_u32(&self) -> u32 { + self.value + } + + /// Returns the numeric value of the code point if it is a leading surrogate. + #[inline] + pub fn to_lead_surrogate(&self) -> Option { + match self.value { + lead @ 0xD800..=0xDBFF => Some(lead as u16), + _ => None, + } + } + + /// Returns the numeric value of the code point if it is a trailing surrogate. + #[inline] + pub fn to_trail_surrogate(&self) -> Option { + match self.value { + trail @ 0xDC00..=0xDFFF => Some(trail as u16), + _ => None, + } + } + + /// Optionally returns a Unicode scalar value for the code point. + /// + /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF). + #[inline] + pub fn to_char(&self) -> Option { + match self.value { + 0xD800..=0xDFFF => None, + _ => Some(unsafe { char::from_u32_unchecked(self.value) }), + } + } + + /// Returns a Unicode scalar value for the code point. + /// + /// Returns `'\u{FFFD}'` (the replacement character “�”) + /// if the code point is a surrogate (from U+D800 to U+DFFF). + #[inline] + pub fn to_char_lossy(&self) -> char { + self.to_char().unwrap_or('\u{FFFD}') + } +} + +/// An owned, growable string of well-formed WTF-8 data. +/// +/// Similar to `String`, but can additionally contain surrogate code points +/// if they’re not in a surrogate pair. +#[derive(Eq, PartialEq, Ord, PartialOrd, Clone)] +pub struct Wtf8Buf { + bytes: Vec, + + /// Do we know that `bytes` holds a valid UTF-8 encoding? We can easily + /// know this if we're constructed from a `String` or `&str`. + /// + /// It is possible for `bytes` to have valid UTF-8 without this being + /// set, such as when we're concatenating `&Wtf8`'s and surrogates become + /// paired, as we don't bother to rescan the entire string. + is_known_utf8: bool, +} + +impl ops::Deref for Wtf8Buf { + type Target = Wtf8; + + fn deref(&self) -> &Wtf8 { + self.as_slice() + } +} + +impl ops::DerefMut for Wtf8Buf { + fn deref_mut(&mut self) -> &mut Wtf8 { + self.as_mut_slice() + } +} + +/// Formats the string in double quotes, with characters escaped according to +/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, +/// where each `x` is a hexadecimal digit. +/// +/// For example, the code units [U+0061, U+D800, U+000A] are formatted as +/// `"a\u{D800}\n"`. +impl fmt::Debug for Wtf8Buf { + #[inline] + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, formatter) + } +} + +/// Formats the string with unpaired surrogates substituted with the replacement +/// character, U+FFFD. +impl fmt::Display for Wtf8Buf { + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + if let Some(s) = self.as_known_utf8() { + fmt::Display::fmt(s, formatter) + } else { + fmt::Display::fmt(&**self, formatter) + } + } +} + +impl Wtf8Buf { + /// Creates a new, empty WTF-8 string. + #[inline] + pub fn new() -> Wtf8Buf { + Wtf8Buf { bytes: Vec::new(), is_known_utf8: true } + } + + /// Creates a new, empty WTF-8 string with pre-allocated capacity for `capacity` bytes. + #[inline] + pub fn with_capacity(capacity: usize) -> Wtf8Buf { + Wtf8Buf { bytes: Vec::with_capacity(capacity), is_known_utf8: true } + } + + /// Creates a WTF-8 string from a WTF-8 byte vec. + /// + /// Since the byte vec is not checked for valid WTF-8, this function is + /// marked unsafe. + #[inline] + pub unsafe fn from_bytes_unchecked(value: Vec) -> Wtf8Buf { + Wtf8Buf { bytes: value, is_known_utf8: false } + } + + /// Creates a WTF-8 string from a UTF-8 `String`. + /// + /// This takes ownership of the `String` and does not copy. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub const fn from_string(string: String) -> Wtf8Buf { + Wtf8Buf { bytes: string.into_bytes(), is_known_utf8: true } + } + + /// Creates a WTF-8 string from a UTF-8 `&str` slice. + /// + /// This copies the content of the slice. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_str(s: &str) -> Wtf8Buf { + Wtf8Buf { bytes: s.as_bytes().to_vec(), is_known_utf8: true } + } + + pub fn clear(&mut self) { + self.bytes.clear(); + self.is_known_utf8 = true; + } + + /// Creates a WTF-8 string from a potentially ill-formed UTF-16 slice of 16-bit code units. + /// + /// This is lossless: calling `.encode_wide()` on the resulting string + /// will always return the original code units. + pub fn from_wide(v: &[u16]) -> Wtf8Buf { + let mut string = Wtf8Buf::with_capacity(v.len()); + for item in char::decode_utf16(v.iter().cloned()) { + match item { + Ok(ch) => string.push_char(ch), + Err(surrogate) => { + let surrogate = surrogate.unpaired_surrogate(); + // Surrogates are known to be in the code point range. + let code_point = unsafe { CodePoint::from_u32_unchecked(surrogate as u32) }; + // The string will now contain an unpaired surrogate. + string.is_known_utf8 = false; + // Skip the WTF-8 concatenation check, + // surrogate pairs are already decoded by decode_utf16 + string.push_code_point_unchecked(code_point); + } + } + } + string + } + + /// Appends the given `char` to the end of this string. + /// This does **not** include the WTF-8 concatenation check or `is_known_utf8` check. + /// Copied from String::push. + fn push_code_point_unchecked(&mut self, code_point: CodePoint) { + let mut bytes = [0; MAX_LEN_UTF8]; + let bytes = encode_utf8_raw(code_point.value, &mut bytes); + self.bytes.extend_from_slice(bytes) + } + + #[inline] + pub fn as_slice(&self) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(&self.bytes) } + } + + #[inline] + pub fn as_mut_slice(&mut self) -> &mut Wtf8 { + // Safety: `Wtf8` doesn't expose any way to mutate the bytes that would + // cause them to change from well-formed UTF-8 to ill-formed UTF-8, + // which would break the assumptions of the `is_known_utf8` field. + unsafe { Wtf8::from_mut_bytes_unchecked(&mut self.bytes) } + } + + /// Converts the string to UTF-8 without validation, if it was created from + /// valid UTF-8. + #[inline] + fn as_known_utf8(&self) -> Option<&str> { + if self.is_known_utf8 { + // SAFETY: The buffer is known to be valid UTF-8. + Some(unsafe { str::from_utf8_unchecked(self.as_bytes()) }) + } else { + None + } + } + + /// Reserves capacity for at least `additional` more bytes to be inserted + /// in the given `Wtf8Buf`. + /// The collection may reserve more space to avoid frequent reallocations. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + #[inline] + pub fn reserve(&mut self, additional: usize) { + self.bytes.reserve(additional) + } + + /// Tries to reserve capacity for at least `additional` more bytes to be + /// inserted in the given `Wtf8Buf`. The `Wtf8Buf` may reserve more space to + /// avoid frequent reallocations. After calling `try_reserve`, capacity will + /// be greater than or equal to `self.len() + additional`. Does nothing if + /// capacity is already sufficient. This method preserves the contents even + /// if an error occurs. + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + #[inline] + pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.bytes.try_reserve(additional) + } + + #[inline] + pub fn reserve_exact(&mut self, additional: usize) { + self.bytes.reserve_exact(additional) + } + + /// Tries to reserve the minimum capacity for exactly `additional` more + /// bytes to be inserted in the given `Wtf8Buf`. After calling + /// `try_reserve_exact`, capacity will be greater than or equal to + /// `self.len() + additional` if it returns `Ok(())`. + /// Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the `Wtf8Buf` more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`try_reserve`] if future insertions are expected. + /// + /// [`try_reserve`]: Wtf8Buf::try_reserve + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + #[inline] + pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.bytes.try_reserve_exact(additional) + } + + #[inline] + pub fn shrink_to_fit(&mut self) { + self.bytes.shrink_to_fit() + } + + #[inline] + pub fn shrink_to(&mut self, min_capacity: usize) { + self.bytes.shrink_to(min_capacity) + } + + #[inline] + pub fn leak<'a>(self) -> &'a mut Wtf8 { + unsafe { Wtf8::from_mut_bytes_unchecked(self.bytes.leak()) } + } + + /// Returns the number of bytes that this string buffer can hold without reallocating. + #[inline] + pub fn capacity(&self) -> usize { + self.bytes.capacity() + } + + /// Append a UTF-8 slice at the end of the string. + #[inline] + pub fn push_str(&mut self, other: &str) { + self.bytes.extend_from_slice(other.as_bytes()) + } + + /// Append a WTF-8 slice at the end of the string. + /// + /// This replaces newly paired surrogates at the boundary + /// with a supplementary code point, + /// like concatenating ill-formed UTF-16 strings effectively would. + #[inline] + pub fn push_wtf8(&mut self, other: &Wtf8) { + match ((&*self).final_lead_surrogate(), other.initial_trail_surrogate()) { + // Replace newly paired surrogates by a supplementary code point. + (Some(lead), Some(trail)) => { + let len_without_lead_surrogate = self.len() - 3; + self.bytes.truncate(len_without_lead_surrogate); + let other_without_trail_surrogate = &other.bytes[3..]; + // 4 bytes for the supplementary code point + self.bytes.reserve(4 + other_without_trail_surrogate.len()); + self.push_char(decode_surrogate_pair(lead, trail)); + self.bytes.extend_from_slice(other_without_trail_surrogate); + } + _ => { + // If we'll be pushing a string containing a surrogate, we may + // no longer have UTF-8. + if self.is_known_utf8 && other.next_surrogate(0).is_some() { + self.is_known_utf8 = false; + } + + self.bytes.extend_from_slice(&other.bytes); + } + } + } + + /// Append a Unicode scalar value at the end of the string. + #[inline] + pub fn push_char(&mut self, c: char) { + self.push_code_point_unchecked(CodePoint::from_char(c)) + } + + /// Append a code point at the end of the string. + /// + /// This replaces newly paired surrogates at the boundary + /// with a supplementary code point, + /// like concatenating ill-formed UTF-16 strings effectively would. + #[inline] + pub fn push(&mut self, code_point: CodePoint) { + if let Some(trail) = code_point.to_trail_surrogate() { + if let Some(lead) = (&*self).final_lead_surrogate() { + let len_without_lead_surrogate = self.len() - 3; + self.bytes.truncate(len_without_lead_surrogate); + self.push_char(decode_surrogate_pair(lead, trail)); + return; + } + + // We're pushing a trailing surrogate. + self.is_known_utf8 = false; + } else if code_point.to_lead_surrogate().is_some() { + // We're pushing a leading surrogate. + self.is_known_utf8 = false; + } + + // No newly paired surrogates at the boundary. + self.push_code_point_unchecked(code_point) + } + + /// Shortens a string to the specified length. + /// + /// # Panics + /// + /// Panics if `new_len` > current length, + /// or if `new_len` is not a code point boundary. + #[inline] + pub fn truncate(&mut self, new_len: usize) { + assert!(is_code_point_boundary(self, new_len)); + self.bytes.truncate(new_len) + } + + /// Consumes the WTF-8 string and tries to convert it to a vec of bytes. + #[inline] + pub fn into_bytes(self) -> Vec { + self.bytes + } + + /// Consumes the WTF-8 string and tries to convert it to UTF-8. + /// + /// This does not copy the data. + /// + /// If the contents are not well-formed UTF-8 + /// (that is, if the string contains surrogates), + /// the original WTF-8 string is returned instead. + pub fn into_string(self) -> Result { + if self.is_known_utf8 || self.next_surrogate(0).is_none() { + Ok(unsafe { String::from_utf8_unchecked(self.bytes) }) + } else { + Err(self) + } + } + + /// Consumes the WTF-8 string and converts it lossily to UTF-8. + /// + /// This does not copy the data (but may overwrite parts of it in place). + /// + /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”) + pub fn into_string_lossy(mut self) -> String { + if !self.is_known_utf8 { + let mut pos = 0; + while let Some((surrogate_pos, _)) = self.next_surrogate(pos) { + pos = surrogate_pos + 3; + // Surrogates and the replacement character are all 3 bytes, so + // they can substituted in-place. + self.bytes[surrogate_pos..pos] + .copy_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + } + } + unsafe { String::from_utf8_unchecked(self.bytes) } + } + + /// Converts this `Wtf8Buf` into a boxed `Wtf8`. + #[inline] + pub fn into_box(self) -> Box { + // SAFETY: relies on `Wtf8` being `repr(transparent)`. + unsafe { mem::transmute(self.bytes.into_boxed_slice()) } + } + + /// Converts a `Box` into a `Wtf8Buf`. + pub fn from_box(boxed: Box) -> Wtf8Buf { + let bytes: Box<[u8]> = unsafe { mem::transmute(boxed) }; + Wtf8Buf { bytes: bytes.into_vec(), is_known_utf8: false } + } + + /// Provides plumbing to core `Vec::extend_from_slice`. + /// More well behaving alternative to allowing outer types + /// full mutable access to the core `Vec`. + #[inline] + pub(crate) fn extend_from_slice(&mut self, other: &[u8]) { + self.bytes.extend_from_slice(other); + self.is_known_utf8 = false; + } +} + +/// Creates a new WTF-8 string from an iterator of code points. +/// +/// This replaces surrogate code point pairs with supplementary code points, +/// like concatenating ill-formed UTF-16 strings effectively would. +impl FromIterator for Wtf8Buf { + fn from_iter>(iter: T) -> Wtf8Buf { + let mut string = Wtf8Buf::new(); + string.extend(iter); + string + } +} + +/// Append code points from an iterator to the string. +/// +/// This replaces surrogate code point pairs with supplementary code points, +/// like concatenating ill-formed UTF-16 strings effectively would. +impl Extend for Wtf8Buf { + fn extend>(&mut self, iter: T) { + let iterator = iter.into_iter(); + let (low, _high) = iterator.size_hint(); + // Lower bound of one byte per code point (ASCII only) + self.bytes.reserve(low); + iterator.for_each(move |code_point| self.push(code_point)); + } + + #[inline] + fn extend_one(&mut self, code_point: CodePoint) { + self.push(code_point); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + // Lower bound of one byte per code point (ASCII only) + self.bytes.reserve(additional); + } +} + +/// A borrowed slice of well-formed WTF-8 data. +/// +/// Similar to `&str`, but can additionally contain surrogate code points +/// if they’re not in a surrogate pair. +#[derive(Eq, Ord, PartialEq, PartialOrd)] +#[repr(transparent)] +pub struct Wtf8 { + bytes: [u8], +} + +impl AsInner<[u8]> for Wtf8 { + #[inline] + fn as_inner(&self) -> &[u8] { + &self.bytes + } +} + +/// Formats the string in double quotes, with characters escaped according to +/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, +/// where each `x` is a hexadecimal digit. +impl fmt::Debug for Wtf8 { + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + fn write_str_escaped(f: &mut fmt::Formatter<'_>, s: &str) -> fmt::Result { + use crate::fmt::Write; + for c in s.chars().flat_map(|c| c.escape_debug()) { + f.write_char(c)? + } + Ok(()) + } + + formatter.write_str("\"")?; + let mut pos = 0; + while let Some((surrogate_pos, surrogate)) = self.next_surrogate(pos) { + write_str_escaped(formatter, unsafe { + str::from_utf8_unchecked(&self.bytes[pos..surrogate_pos]) + })?; + write!(formatter, "\\u{{{:x}}}", surrogate)?; + pos = surrogate_pos + 3; + } + write_str_escaped(formatter, unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) })?; + formatter.write_str("\"") + } +} + +/// Formats the string with unpaired surrogates substituted with the replacement +/// character, U+FFFD. +impl fmt::Display for Wtf8 { + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + let wtf8_bytes = &self.bytes; + let mut pos = 0; + loop { + match self.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + formatter.write_str(unsafe { + str::from_utf8_unchecked(&wtf8_bytes[pos..surrogate_pos]) + })?; + formatter.write_str(UTF8_REPLACEMENT_CHARACTER)?; + pos = surrogate_pos + 3; + } + None => { + let s = unsafe { str::from_utf8_unchecked(&wtf8_bytes[pos..]) }; + if pos == 0 { return s.fmt(formatter) } else { return formatter.write_str(s) } + } + } + } + } +} + +impl Wtf8 { + /// Creates a WTF-8 slice from a UTF-8 `&str` slice. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_str(value: &str) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) } + } + + /// Creates a WTF-8 slice from a WTF-8 byte slice. + /// + /// Since the byte slice is not checked for valid WTF-8, this functions is + /// marked unsafe. + #[inline] + pub unsafe fn from_bytes_unchecked(value: &[u8]) -> &Wtf8 { + // SAFETY: start with &[u8], end with fancy &[u8] + unsafe { &*(value as *const [u8] as *const Wtf8) } + } + + /// Creates a mutable WTF-8 slice from a mutable WTF-8 byte slice. + /// + /// Since the byte slice is not checked for valid WTF-8, this functions is + /// marked unsafe. + #[inline] + unsafe fn from_mut_bytes_unchecked(value: &mut [u8]) -> &mut Wtf8 { + // SAFETY: start with &mut [u8], end with fancy &mut [u8] + unsafe { &mut *(value as *mut [u8] as *mut Wtf8) } + } + + /// Returns the length, in WTF-8 bytes. + #[inline] + pub fn len(&self) -> usize { + self.bytes.len() + } + + #[inline] + pub fn is_empty(&self) -> bool { + self.bytes.is_empty() + } + + /// Returns the code point at `position` if it is in the ASCII range, + /// or `b'\xFF'` otherwise. + /// + /// # Panics + /// + /// Panics if `position` is beyond the end of the string. + #[inline] + pub fn ascii_byte_at(&self, position: usize) -> u8 { + match self.bytes[position] { + ascii_byte @ 0x00..=0x7F => ascii_byte, + _ => 0xFF, + } + } + + /// Returns an iterator for the string’s code points. + #[inline] + pub fn code_points(&self) -> Wtf8CodePoints<'_> { + Wtf8CodePoints { bytes: self.bytes.iter() } + } + + /// Access raw bytes of WTF-8 data + #[inline] + pub fn as_bytes(&self) -> &[u8] { + &self.bytes + } + + /// Tries to convert the string to UTF-8 and return a `&str` slice. + /// + /// Returns `None` if the string contains surrogates. + /// + /// This does not copy the data. + #[inline] + pub fn as_str(&self) -> Result<&str, str::Utf8Error> { + str::from_utf8(&self.bytes) + } + + /// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. + pub fn to_owned(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_vec(), is_known_utf8: false } + } + + /// Lossily converts the string to UTF-8. + /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. + /// + /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). + /// + /// This only copies the data if necessary (if it contains any surrogate). + pub fn to_string_lossy(&self) -> Cow<'_, str> { + let Some((surrogate_pos, _)) = self.next_surrogate(0) else { + return Cow::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }); + }; + let wtf8_bytes = &self.bytes; + let mut utf8_bytes = Vec::with_capacity(self.len()); + utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); + utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + let mut pos = surrogate_pos + 3; + loop { + match self.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); + utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + pos = surrogate_pos + 3; + } + None => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); + return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); + } + } + } + } + + /// Converts the WTF-8 string to potentially ill-formed UTF-16 + /// and return an iterator of 16-bit code units. + /// + /// This is lossless: + /// calling `Wtf8Buf::from_ill_formed_utf16` on the resulting code units + /// would always return the original WTF-8 string. + #[inline] + pub fn encode_wide(&self) -> EncodeWide<'_> { + EncodeWide { code_points: self.code_points(), extra: 0 } + } + + #[inline] + fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { + let mut iter = self.bytes[pos..].iter(); + loop { + let b = *iter.next()?; + if b < 0x80 { + pos += 1; + } else if b < 0xE0 { + iter.next(); + pos += 2; + } else if b == 0xED { + match (iter.next(), iter.next()) { + (Some(&b2), Some(&b3)) if b2 >= 0xA0 => { + return Some((pos, decode_surrogate(b2, b3))); + } + _ => pos += 3, + } + } else if b < 0xF0 { + iter.next(); + iter.next(); + pos += 3; + } else { + iter.next(); + iter.next(); + iter.next(); + pos += 4; + } + } + } + + #[inline] + fn final_lead_surrogate(&self) -> Option { + match self.bytes { + [.., 0xED, b2 @ 0xA0..=0xAF, b3] => Some(decode_surrogate(b2, b3)), + _ => None, + } + } + + #[inline] + fn initial_trail_surrogate(&self) -> Option { + match self.bytes { + [0xED, b2 @ 0xB0..=0xBF, b3, ..] => Some(decode_surrogate(b2, b3)), + _ => None, + } + } + + pub fn clone_into(&self, buf: &mut Wtf8Buf) { + buf.is_known_utf8 = false; + self.bytes.clone_into(&mut buf.bytes); + } + + /// Boxes this `Wtf8`. + #[inline] + pub fn into_box(&self) -> Box { + let boxed: Box<[u8]> = self.bytes.into(); + unsafe { mem::transmute(boxed) } + } + + /// Creates a boxed, empty `Wtf8`. + pub fn empty_box() -> Box { + let boxed: Box<[u8]> = Default::default(); + unsafe { mem::transmute(boxed) } + } + + #[inline] + pub fn into_arc(&self) -> Arc { + let arc: Arc<[u8]> = Arc::from(&self.bytes); + unsafe { Arc::from_raw(Arc::into_raw(arc) as *const Wtf8) } + } + + #[inline] + pub fn into_rc(&self) -> Rc { + let rc: Rc<[u8]> = Rc::from(&self.bytes); + unsafe { Rc::from_raw(Rc::into_raw(rc) as *const Wtf8) } + } + + #[inline] + pub fn make_ascii_lowercase(&mut self) { + self.bytes.make_ascii_lowercase() + } + + #[inline] + pub fn make_ascii_uppercase(&mut self) { + self.bytes.make_ascii_uppercase() + } + + #[inline] + pub fn to_ascii_lowercase(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_ascii_lowercase(), is_known_utf8: false } + } + + #[inline] + pub fn to_ascii_uppercase(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_ascii_uppercase(), is_known_utf8: false } + } + + #[inline] + pub fn is_ascii(&self) -> bool { + self.bytes.is_ascii() + } + + #[inline] + pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { + self.bytes.eq_ignore_ascii_case(&other.bytes) + } +} + +/// Returns a slice of the given string for the byte range \[`begin`..`end`). +/// +/// # Panics +/// +/// Panics when `begin` and `end` do not point to code point boundaries, +/// or point beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::Range) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if range.start <= range.end + && is_code_point_boundary(self, range.start) + && is_code_point_boundary(self, range.end) + { + unsafe { slice_unchecked(self, range.start, range.end) } + } else { + slice_error_fail(self, range.start, range.end) + } + } +} + +/// Returns a slice of the given string from byte `begin` to its end. +/// +/// # Panics +/// +/// Panics when `begin` is not at a code point boundary, +/// or is beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::RangeFrom) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if is_code_point_boundary(self, range.start) { + unsafe { slice_unchecked(self, range.start, self.len()) } + } else { + slice_error_fail(self, range.start, self.len()) + } + } +} + +/// Returns a slice of the given string from its beginning to byte `end`. +/// +/// # Panics +/// +/// Panics when `end` is not at a code point boundary, +/// or is beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::RangeTo) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if is_code_point_boundary(self, range.end) { + unsafe { slice_unchecked(self, 0, range.end) } + } else { + slice_error_fail(self, 0, range.end) + } + } +} + +impl ops::Index for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, _range: ops::RangeFull) -> &Wtf8 { + self + } +} + +#[inline] +fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { + // The first byte is assumed to be 0xED + 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F +} + +#[inline] +fn decode_surrogate_pair(lead: u16, trail: u16) -> char { + let code_point = 0x10000 + ((((lead - 0xD800) as u32) << 10) | (trail - 0xDC00) as u32); + unsafe { char::from_u32_unchecked(code_point) } +} + +/// Copied from str::is_char_boundary +#[inline] +pub fn is_code_point_boundary(slice: &Wtf8, index: usize) -> bool { + if index == 0 { + return true; + } + match slice.bytes.get(index) { + None => index == slice.len(), + Some(&b) => (b as i8) >= -0x40, + } +} + +/// Verify that `index` is at the edge of either a valid UTF-8 codepoint +/// (i.e. a codepoint that's not a surrogate) or of the whole string. +/// +/// These are the cases currently permitted by `OsStr::slice_encoded_bytes`. +/// Splitting between surrogates is valid as far as WTF-8 is concerned, but +/// we do not permit it in the public API because WTF-8 is considered an +/// implementation detail. +#[track_caller] +#[inline] +pub fn check_utf8_boundary(slice: &Wtf8, index: usize) { + if index == 0 { + return; + } + match slice.bytes.get(index) { + Some(0xED) => (), // Might be a surrogate + Some(&b) if (b as i8) >= -0x40 => return, + Some(_) => panic!("byte index {index} is not a codepoint boundary"), + None if index == slice.len() => return, + None => panic!("byte index {index} is out of bounds"), + } + if slice.bytes[index + 1] >= 0xA0 { + // There's a surrogate after index. Now check before index. + if index >= 3 && slice.bytes[index - 3] == 0xED && slice.bytes[index - 2] >= 0xA0 { + panic!("byte index {index} lies between surrogate codepoints"); + } + } +} + +/// Copied from core::str::raw::slice_unchecked +#[inline] +pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { + // SAFETY: memory layout of a &[u8] and &Wtf8 are the same + unsafe { + let len = end - begin; + let start = s.as_bytes().as_ptr().add(begin); + Wtf8::from_bytes_unchecked(slice::from_raw_parts(start, len)) + } +} + +/// Copied from core::str::raw::slice_error_fail +#[inline(never)] +pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { + assert!(begin <= end); + panic!("index {begin} and/or {end} in `{s:?}` do not lie on character boundary"); +} + +/// Iterator for the code points of a WTF-8 string. +/// +/// Created with the method `.code_points()`. +#[derive(Clone)] +pub struct Wtf8CodePoints<'a> { + bytes: slice::Iter<'a, u8>, +} + +impl Iterator for Wtf8CodePoints<'_> { + type Item = CodePoint; + + #[inline] + fn next(&mut self) -> Option { + // SAFETY: `self.bytes` has been created from a WTF-8 string + unsafe { next_code_point(&mut self.bytes).map(|c| CodePoint { value: c }) } + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + let len = self.bytes.len(); + (len.saturating_add(3) / 4, Some(len)) + } +} + +/// Generates a wide character sequence for potentially ill-formed UTF-16. +#[stable(feature = "rust1", since = "1.0.0")] +#[derive(Clone)] +pub struct EncodeWide<'a> { + code_points: Wtf8CodePoints<'a>, + extra: u16, +} + +// Copied from libunicode/u_str.rs +#[stable(feature = "rust1", since = "1.0.0")] +impl Iterator for EncodeWide<'_> { + type Item = u16; + + #[inline] + fn next(&mut self) -> Option { + if self.extra != 0 { + let tmp = self.extra; + self.extra = 0; + return Some(tmp); + } + + let mut buf = [0; MAX_LEN_UTF16]; + self.code_points.next().map(|code_point| { + let n = encode_utf16_raw(code_point.value, &mut buf).len(); + if n == 2 { + self.extra = buf[1]; + } + buf[0] + }) + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + let (low, high) = self.code_points.size_hint(); + let ext = (self.extra != 0) as usize; + // every code point gets either one u16 or two u16, + // so this iterator is between 1 or 2 times as + // long as the underlying iterator. + (low + ext, high.and_then(|n| n.checked_mul(2)).and_then(|n| n.checked_add(ext))) + } +} + +#[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")] +impl FusedIterator for EncodeWide<'_> {} + +impl Hash for CodePoint { + #[inline] + fn hash(&self, state: &mut H) { + self.value.hash(state) + } +} + +impl Hash for Wtf8Buf { + #[inline] + fn hash(&self, state: &mut H) { + state.write(&self.bytes); + 0xfeu8.hash(state) + } +} + +impl Hash for Wtf8 { + #[inline] + fn hash(&self, state: &mut H) { + state.write(&self.bytes); + 0xfeu8.hash(state) + } +} + +#[unstable(feature = "clone_to_uninit", issue = "126799")] +unsafe impl CloneToUninit for Wtf8 { + #[inline] + #[cfg_attr(debug_assertions, track_caller)] + unsafe fn clone_to_uninit(&self, dst: *mut u8) { + // SAFETY: we're just a transparent wrapper around [u8] + unsafe { self.bytes.clone_to_uninit(dst) } + } +} diff --git a/library/alloc/src/wtf8/tests.rs b/library/alloc/src/wtf8/tests.rs new file mode 100644 index 00000000000..b57c99a8452 --- /dev/null +++ b/library/alloc/src/wtf8/tests.rs @@ -0,0 +1,751 @@ +use super::*; + +#[test] +fn code_point_from_u32() { + assert!(CodePoint::from_u32(0).is_some()); + assert!(CodePoint::from_u32(0xD800).is_some()); + assert!(CodePoint::from_u32(0x10FFFF).is_some()); + assert!(CodePoint::from_u32(0x110000).is_none()); +} + +#[test] +fn code_point_to_u32() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0).to_u32(), 0); + assert_eq!(c(0xD800).to_u32(), 0xD800); + assert_eq!(c(0x10FFFF).to_u32(), 0x10FFFF); +} + +#[test] +fn code_point_to_lead_surrogate() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0).to_lead_surrogate(), None); + assert_eq!(c(0xE9).to_lead_surrogate(), None); + assert_eq!(c(0xD800).to_lead_surrogate(), Some(0xD800)); + assert_eq!(c(0xDBFF).to_lead_surrogate(), Some(0xDBFF)); + assert_eq!(c(0xDC00).to_lead_surrogate(), None); + assert_eq!(c(0xDFFF).to_lead_surrogate(), None); + assert_eq!(c(0x1F4A9).to_lead_surrogate(), None); + assert_eq!(c(0x10FFFF).to_lead_surrogate(), None); +} + +#[test] +fn code_point_to_trail_surrogate() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0).to_trail_surrogate(), None); + assert_eq!(c(0xE9).to_trail_surrogate(), None); + assert_eq!(c(0xD800).to_trail_surrogate(), None); + assert_eq!(c(0xDBFF).to_trail_surrogate(), None); + assert_eq!(c(0xDC00).to_trail_surrogate(), Some(0xDC00)); + assert_eq!(c(0xDFFF).to_trail_surrogate(), Some(0xDFFF)); + assert_eq!(c(0x1F4A9).to_trail_surrogate(), None); + assert_eq!(c(0x10FFFF).to_trail_surrogate(), None); +} + +#[test] +fn code_point_from_char() { + assert_eq!(CodePoint::from_char('a').to_u32(), 0x61); + assert_eq!(CodePoint::from_char('💩').to_u32(), 0x1F4A9); +} + +#[test] +fn code_point_to_string() { + assert_eq!(format!("{:?}", CodePoint::from_char('a')), "U+0061"); + assert_eq!(format!("{:?}", CodePoint::from_char('💩')), "U+1F4A9"); +} + +#[test] +fn code_point_to_char() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0x61).to_char(), Some('a')); + assert_eq!(c(0x1F4A9).to_char(), Some('💩')); + assert_eq!(c(0xD800).to_char(), None); +} + +#[test] +fn code_point_to_char_lossy() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + assert_eq!(c(0x61).to_char_lossy(), 'a'); + assert_eq!(c(0x1F4A9).to_char_lossy(), '💩'); + assert_eq!(c(0xD800).to_char_lossy(), '\u{FFFD}'); +} + +#[test] +fn wtf8buf_new() { + assert_eq!(Wtf8Buf::new().bytes, b""); +} + +#[test] +fn wtf8buf_from_str() { + assert_eq!(Wtf8Buf::from_str("").bytes, b""); + assert_eq!(Wtf8Buf::from_str("aé 💩").bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); +} + +#[test] +fn wtf8buf_from_string() { + assert_eq!(Wtf8Buf::from_string(String::from("")).bytes, b""); + assert_eq!(Wtf8Buf::from_string(String::from("aé 💩")).bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); +} + +#[test] +fn wtf8buf_from_wide() { + let buf = Wtf8Buf::from_wide(&[]); + assert_eq!(buf.bytes, b""); + assert!(buf.is_known_utf8); + + let buf = Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xDCA9]); + assert_eq!(buf.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert!(buf.is_known_utf8); + + let buf = Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]); + assert_eq!(buf.bytes, b"a\xC3\xA9 \xED\xA0\xBD\xF0\x9F\x92\xA9"); + assert!(!buf.is_known_utf8); + + let buf = Wtf8Buf::from_wide(&[0xD800]); + assert_eq!(buf.bytes, b"\xED\xA0\x80"); + assert!(!buf.is_known_utf8); + + let buf = Wtf8Buf::from_wide(&[0xDBFF]); + assert_eq!(buf.bytes, b"\xED\xAF\xBF"); + assert!(!buf.is_known_utf8); + + let buf = Wtf8Buf::from_wide(&[0xDC00]); + assert_eq!(buf.bytes, b"\xED\xB0\x80"); + assert!(!buf.is_known_utf8); + + let buf = Wtf8Buf::from_wide(&[0xDFFF]); + assert_eq!(buf.bytes, b"\xED\xBF\xBF"); + assert!(!buf.is_known_utf8); +} + +#[test] +fn wtf8buf_push_str() { + let mut string = Wtf8Buf::new(); + assert_eq!(string.bytes, b""); + assert!(string.is_known_utf8); + + string.push_str("aé 💩"); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert!(string.is_known_utf8); +} + +#[test] +fn wtf8buf_push_char() { + let mut string = Wtf8Buf::from_str("aé "); + assert_eq!(string.bytes, b"a\xC3\xA9 "); + assert!(string.is_known_utf8); + + string.push_char('💩'); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert!(string.is_known_utf8); +} + +#[test] +fn wtf8buf_push() { + let mut string = Wtf8Buf::from_str("aé "); + assert_eq!(string.bytes, b"a\xC3\xA9 "); + assert!(string.is_known_utf8); + + string.push(CodePoint::from_char('💩')); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert!(string.is_known_utf8); + + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + + let mut string = Wtf8Buf::new(); + string.push(c(0xD83D)); // lead + assert!(!string.is_known_utf8); + string.push(c(0xDCA9)); // trail + assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! + + let mut string = Wtf8Buf::new(); + string.push(c(0xD83D)); // lead + assert!(!string.is_known_utf8); + string.push(c(0x20)); // not surrogate + string.push(c(0xDCA9)); // trail + assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xD800)); // lead + assert!(!string.is_known_utf8); + string.push(c(0xDBFF)); // lead + assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xD800)); // lead + assert!(!string.is_known_utf8); + string.push(c(0xE000)); // not surrogate + assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xD7FF)); // not surrogate + assert!(string.is_known_utf8); + string.push(c(0xDC00)); // trail + assert!(!string.is_known_utf8); + assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + + let mut string = Wtf8Buf::new(); + string.push(c(0x61)); // not surrogate, < 3 bytes + assert!(string.is_known_utf8); + string.push(c(0xDC00)); // trail + assert!(!string.is_known_utf8); + assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); + + let mut string = Wtf8Buf::new(); + string.push(c(0xDC00)); // trail + assert!(!string.is_known_utf8); + assert_eq!(string.bytes, b"\xED\xB0\x80"); +} + +#[test] +fn wtf8buf_push_wtf8() { + let mut string = Wtf8Buf::from_str("aé"); + assert_eq!(string.bytes, b"a\xC3\xA9"); + string.push_wtf8(Wtf8::from_str(" 💩")); + assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert!(string.is_known_utf8); + + fn w(v: &[u8]) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(v) } + } + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\xBD")); // lead + string.push_wtf8(w(b"\xED\xB2\xA9")); // trail + assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\xBD")); // lead + string.push_wtf8(w(b" ")); // not surrogate + string.push_wtf8(w(b"\xED\xB2\xA9")); // trail + assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); + assert!(!string.is_known_utf8); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\x80")); // lead + string.push_wtf8(w(b"\xED\xAF\xBF")); // lead + assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + assert!(!string.is_known_utf8); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xA0\x80")); // lead + string.push_wtf8(w(b"\xEE\x80\x80")); // not surrogate + assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); + assert!(!string.is_known_utf8); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\x9F\xBF")); // not surrogate + string.push_wtf8(w(b"\xED\xB0\x80")); // trail + assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + assert!(!string.is_known_utf8); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"a")); // not surrogate, < 3 bytes + string.push_wtf8(w(b"\xED\xB0\x80")); // trail + assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); + assert!(!string.is_known_utf8); + + let mut string = Wtf8Buf::new(); + string.push_wtf8(w(b"\xED\xB0\x80")); // trail + assert_eq!(string.bytes, b"\xED\xB0\x80"); + assert!(!string.is_known_utf8); +} + +#[test] +fn wtf8buf_truncate() { + let mut string = Wtf8Buf::from_str("aé"); + assert!(string.is_known_utf8); + + string.truncate(3); + assert_eq!(string.bytes, b"a\xC3\xA9"); + assert!(string.is_known_utf8); + + string.truncate(1); + assert_eq!(string.bytes, b"a"); + assert!(string.is_known_utf8); + + string.truncate(0); + assert_eq!(string.bytes, b""); + assert!(string.is_known_utf8); +} + +#[test] +fn wtf8buf_truncate_around_non_bmp() { + let mut string = Wtf8Buf::from_str("💩"); + assert!(string.is_known_utf8); + + string.truncate(4); + assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); + assert!(string.is_known_utf8); + + string.truncate(0); + assert_eq!(string.bytes, b""); + assert!(string.is_known_utf8); +} + +#[test] +#[should_panic] +fn wtf8buf_truncate_fail_code_point_boundary() { + let mut string = Wtf8Buf::from_str("aé"); + string.truncate(2); +} + +#[test] +#[should_panic] +fn wtf8buf_truncate_fail_longer() { + let mut string = Wtf8Buf::from_str("aé"); + string.truncate(4); +} + +#[test] +#[should_panic] +fn wtf8buf_truncate_splitting_non_bmp3() { + let mut string = Wtf8Buf::from_str("💩"); + assert!(string.is_known_utf8); + string.truncate(3); +} + +#[test] +#[should_panic] +fn wtf8buf_truncate_splitting_non_bmp2() { + let mut string = Wtf8Buf::from_str("💩"); + assert!(string.is_known_utf8); + string.truncate(2); +} + +#[test] +#[should_panic] +fn wtf8buf_truncate_splitting_non_bmp1() { + let mut string = Wtf8Buf::from_str("💩"); + assert!(string.is_known_utf8); + string.truncate(1); +} + +#[test] +fn wtf8buf_into_string() { + let mut string = Wtf8Buf::from_str("aé 💩"); + assert!(string.is_known_utf8); + assert_eq!(string.clone().into_string(), Ok(String::from("aé 💩"))); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert!(!string.is_known_utf8); + assert_eq!(string.clone().into_string(), Err(string)); +} + +#[test] +fn wtf8buf_into_string_lossy() { + let mut string = Wtf8Buf::from_str("aé 💩"); + assert_eq!(string.clone().into_string_lossy(), String::from("aé 💩")); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!(string.clone().into_string_lossy(), String::from("aé 💩�")); +} + +#[test] +fn wtf8buf_from_iterator() { + fn f(values: &[u32]) -> Wtf8Buf { + values.iter().map(|&c| CodePoint::from_u32(c).unwrap()).collect::() + } + assert_eq!( + f(&[0x61, 0xE9, 0x20, 0x1F4A9]), + Wtf8Buf { bytes: b"a\xC3\xA9 \xF0\x9F\x92\xA9".to_vec(), is_known_utf8: true } + ); + + assert_eq!(f(&[0xD83D, 0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!( + f(&[0xD83D, 0x20, 0xDCA9]), + Wtf8Buf { bytes: b"\xED\xA0\xBD \xED\xB2\xA9".to_vec(), is_known_utf8: false } + ); + assert_eq!( + f(&[0xD800, 0xDBFF]), + Wtf8Buf { bytes: b"\xED\xA0\x80\xED\xAF\xBF".to_vec(), is_known_utf8: false } + ); + assert_eq!( + f(&[0xD800, 0xE000]), + Wtf8Buf { bytes: b"\xED\xA0\x80\xEE\x80\x80".to_vec(), is_known_utf8: false } + ); + assert_eq!( + f(&[0xD7FF, 0xDC00]), + Wtf8Buf { bytes: b"\xED\x9F\xBF\xED\xB0\x80".to_vec(), is_known_utf8: false } + ); + assert_eq!( + f(&[0x61, 0xDC00]), + Wtf8Buf { bytes: b"\x61\xED\xB0\x80".to_vec(), is_known_utf8: false } + ); + assert_eq!(f(&[0xDC00]), Wtf8Buf { bytes: b"\xED\xB0\x80".to_vec(), is_known_utf8: false }); +} + +#[test] +fn wtf8buf_extend() { + fn e(initial: &[u32], extended: &[u32]) -> Wtf8Buf { + fn c(value: &u32) -> CodePoint { + CodePoint::from_u32(*value).unwrap() + } + let mut string = initial.iter().map(c).collect::(); + string.extend(extended.iter().map(c)); + string + } + + assert_eq!( + e(&[0x61, 0xE9], &[0x20, 0x1F4A9]), + Wtf8Buf { bytes: b"a\xC3\xA9 \xF0\x9F\x92\xA9".to_vec(), is_known_utf8: true } + ); + + assert_eq!(e(&[0xD83D], &[0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!( + e(&[0xD83D, 0x20], &[0xDCA9]), + Wtf8Buf { bytes: b"\xED\xA0\xBD \xED\xB2\xA9".to_vec(), is_known_utf8: false } + ); + assert_eq!( + e(&[0xD800], &[0xDBFF]), + Wtf8Buf { bytes: b"\xED\xA0\x80\xED\xAF\xBF".to_vec(), is_known_utf8: false } + ); + assert_eq!( + e(&[0xD800], &[0xE000]), + Wtf8Buf { bytes: b"\xED\xA0\x80\xEE\x80\x80".to_vec(), is_known_utf8: false } + ); + assert_eq!( + e(&[0xD7FF], &[0xDC00]), + Wtf8Buf { bytes: b"\xED\x9F\xBF\xED\xB0\x80".to_vec(), is_known_utf8: false } + ); + assert_eq!( + e(&[0x61], &[0xDC00]), + Wtf8Buf { bytes: b"\x61\xED\xB0\x80".to_vec(), is_known_utf8: false } + ); + assert_eq!( + e(&[], &[0xDC00]), + Wtf8Buf { bytes: b"\xED\xB0\x80".to_vec(), is_known_utf8: false } + ); +} + +#[test] +fn wtf8buf_show() { + let mut string = Wtf8Buf::from_str("a\té \u{7f}💩\r"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!(format!("{string:?}"), "\"a\\té \\u{7f}\u{1f4a9}\\r\\u{d800}\""); +} + +#[test] +fn wtf8buf_as_slice() { + assert_eq!(Wtf8Buf::from_str("aé").as_slice(), Wtf8::from_str("aé")); +} + +#[test] +fn wtf8buf_show_str() { + let text = "a\té 💩\r"; + let string = Wtf8Buf::from_str(text); + assert_eq!(format!("{text:?}"), format!("{string:?}")); +} + +#[test] +fn wtf8_from_str() { + assert_eq!(&Wtf8::from_str("").bytes, b""); + assert_eq!(&Wtf8::from_str("aé 💩").bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); +} + +#[test] +fn wtf8_len() { + assert_eq!(Wtf8::from_str("").len(), 0); + assert_eq!(Wtf8::from_str("aé 💩").len(), 8); +} + +#[test] +fn wtf8_slice() { + assert_eq!(&Wtf8::from_str("aé 💩")[1..4].bytes, b"\xC3\xA9 "); +} + +#[test] +#[should_panic] +fn wtf8_slice_not_code_point_boundary() { + let _ = &Wtf8::from_str("aé 💩")[2..4]; +} + +#[test] +fn wtf8_slice_from() { + assert_eq!(&Wtf8::from_str("aé 💩")[1..].bytes, b"\xC3\xA9 \xF0\x9F\x92\xA9"); +} + +#[test] +#[should_panic] +fn wtf8_slice_from_not_code_point_boundary() { + let _ = &Wtf8::from_str("aé 💩")[2..]; +} + +#[test] +fn wtf8_slice_to() { + assert_eq!(&Wtf8::from_str("aé 💩")[..4].bytes, b"a\xC3\xA9 "); +} + +#[test] +#[should_panic] +fn wtf8_slice_to_not_code_point_boundary() { + let _ = &Wtf8::from_str("aé 💩")[5..]; +} + +#[test] +fn wtf8_ascii_byte_at() { + let slice = Wtf8::from_str("aé 💩"); + assert_eq!(slice.ascii_byte_at(0), b'a'); + assert_eq!(slice.ascii_byte_at(1), b'\xFF'); + assert_eq!(slice.ascii_byte_at(2), b'\xFF'); + assert_eq!(slice.ascii_byte_at(3), b' '); + assert_eq!(slice.ascii_byte_at(4), b'\xFF'); +} + +#[test] +fn wtf8_code_points() { + fn c(value: u32) -> CodePoint { + CodePoint::from_u32(value).unwrap() + } + fn cp(string: &Wtf8Buf) -> Vec> { + string.code_points().map(|c| c.to_char()).collect::>() + } + let mut string = Wtf8Buf::from_str("é "); + assert_eq!(cp(&string), [Some('é'), Some(' ')]); + string.push(c(0xD83D)); + assert_eq!(cp(&string), [Some('é'), Some(' '), None]); + string.push(c(0xDCA9)); + assert_eq!(cp(&string), [Some('é'), Some(' '), Some('💩')]); +} + +#[test] +fn wtf8_as_str() { + assert_eq!(Wtf8::from_str("").as_str(), Ok("")); + assert_eq!(Wtf8::from_str("aé 💩").as_str(), Ok("aé 💩")); + let mut string = Wtf8Buf::new(); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert!(string.as_str().is_err()); +} + +#[test] +fn wtf8_to_string_lossy() { + assert_eq!(Wtf8::from_str("").to_string_lossy(), Cow::Borrowed("")); + assert_eq!(Wtf8::from_str("aé 💩").to_string_lossy(), Cow::Borrowed("aé 💩")); + let mut string = Wtf8Buf::from_str("aé 💩"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + let expected: Cow<'_, str> = Cow::Owned(String::from("aé 💩�")); + assert_eq!(string.to_string_lossy(), expected); +} + +#[test] +fn wtf8_display() { + fn d(b: &[u8]) -> String { + (&unsafe { Wtf8::from_bytes_unchecked(b) }).to_string() + } + + assert_eq!("", d("".as_bytes())); + assert_eq!("aé 💩", d("aé 💩".as_bytes())); + + let mut string = Wtf8Buf::from_str("aé 💩"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + assert_eq!("aé 💩�", d(string.as_inner())); +} + +#[test] +fn wtf8_encode_wide() { + let mut string = Wtf8Buf::from_str("aé "); + string.push(CodePoint::from_u32(0xD83D).unwrap()); + string.push_char('💩'); + assert_eq!( + string.encode_wide().collect::>(), + vec![0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9] + ); +} + +#[test] +fn wtf8_encode_wide_size_hint() { + let string = Wtf8Buf::from_str("\u{12345}"); + let mut iter = string.encode_wide(); + assert_eq!((1, Some(8)), iter.size_hint()); + iter.next().unwrap(); + assert_eq!((1, Some(1)), iter.size_hint()); + iter.next().unwrap(); + assert_eq!((0, Some(0)), iter.size_hint()); + assert!(iter.next().is_none()); +} + +#[test] +fn wtf8_clone_into() { + let mut string = Wtf8Buf::new(); + Wtf8::from_str("green").clone_into(&mut string); + assert_eq!(string.bytes, b"green"); + + let mut string = Wtf8Buf::from_str("green"); + Wtf8::from_str("").clone_into(&mut string); + assert_eq!(string.bytes, b""); + + let mut string = Wtf8Buf::from_str("red"); + Wtf8::from_str("green").clone_into(&mut string); + assert_eq!(string.bytes, b"green"); + + let mut string = Wtf8Buf::from_str("green"); + Wtf8::from_str("red").clone_into(&mut string); + assert_eq!(string.bytes, b"red"); + + let mut string = Wtf8Buf::from_str("green"); + assert!(string.is_known_utf8); + unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").clone_into(&mut string) }; + assert_eq!(string.bytes, b"\xED\xA0\x80"); + assert!(!string.is_known_utf8); +} + +#[test] +fn wtf8_to_ascii_lowercase() { + let lowercase = Wtf8::from_str("").to_ascii_lowercase(); + assert_eq!(lowercase.bytes, b""); + + let lowercase = Wtf8::from_str("GrEeN gRaPeS! 🍇").to_ascii_lowercase(); + assert_eq!(lowercase.bytes, b"green grapes! \xf0\x9f\x8d\x87"); + + let lowercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_ascii_lowercase() }; + assert_eq!(lowercase.bytes, b"\xED\xA0\x80"); + assert!(!lowercase.is_known_utf8); +} + +#[test] +fn wtf8_to_ascii_uppercase() { + let uppercase = Wtf8::from_str("").to_ascii_uppercase(); + assert_eq!(uppercase.bytes, b""); + + let uppercase = Wtf8::from_str("GrEeN gRaPeS! 🍇").to_ascii_uppercase(); + assert_eq!(uppercase.bytes, b"GREEN GRAPES! \xf0\x9f\x8d\x87"); + + let uppercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_ascii_uppercase() }; + assert_eq!(uppercase.bytes, b"\xED\xA0\x80"); + assert!(!uppercase.is_known_utf8); +} + +#[test] +fn wtf8_make_ascii_lowercase() { + let mut lowercase = Wtf8Buf::from_str(""); + lowercase.make_ascii_lowercase(); + assert_eq!(lowercase.bytes, b""); + + let mut lowercase = Wtf8Buf::from_str("GrEeN gRaPeS! 🍇"); + lowercase.make_ascii_lowercase(); + assert_eq!(lowercase.bytes, b"green grapes! \xf0\x9f\x8d\x87"); + + let mut lowercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + lowercase.make_ascii_lowercase(); + assert_eq!(lowercase.bytes, b"\xED\xA0\x80"); + assert!(!lowercase.is_known_utf8); +} + +#[test] +fn wtf8_make_ascii_uppercase() { + let mut uppercase = Wtf8Buf::from_str(""); + uppercase.make_ascii_uppercase(); + assert_eq!(uppercase.bytes, b""); + + let mut uppercase = Wtf8Buf::from_str("GrEeN gRaPeS! 🍇"); + uppercase.make_ascii_uppercase(); + assert_eq!(uppercase.bytes, b"GREEN GRAPES! \xf0\x9f\x8d\x87"); + + let mut uppercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + uppercase.make_ascii_uppercase(); + assert_eq!(uppercase.bytes, b"\xED\xA0\x80"); + assert!(!uppercase.is_known_utf8); +} + +#[test] +fn wtf8_to_owned() { + let string = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + assert_eq!(string.bytes, b"\xED\xA0\x80"); + assert!(!string.is_known_utf8); +} + +#[test] +fn wtf8_valid_utf8_boundaries() { + let mut string = Wtf8Buf::from_str("aé 💩"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + string.push(CodePoint::from_u32(0xD800).unwrap()); + check_utf8_boundary(&string, 0); + check_utf8_boundary(&string, 1); + check_utf8_boundary(&string, 3); + check_utf8_boundary(&string, 4); + check_utf8_boundary(&string, 8); + check_utf8_boundary(&string, 14); + assert_eq!(string.len(), 14); + + string.push_char('a'); + check_utf8_boundary(&string, 14); + check_utf8_boundary(&string, 15); + + let mut string = Wtf8Buf::from_str("a"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + check_utf8_boundary(&string, 1); + + let mut string = Wtf8Buf::from_str("\u{D7FF}"); + string.push(CodePoint::from_u32(0xD800).unwrap()); + check_utf8_boundary(&string, 3); + + let mut string = Wtf8Buf::new(); + string.push(CodePoint::from_u32(0xD800).unwrap()); + string.push_char('\u{D7FF}'); + check_utf8_boundary(&string, 3); +} + +#[test] +#[should_panic(expected = "byte index 4 is out of bounds")] +fn wtf8_utf8_boundary_out_of_bounds() { + let string = Wtf8::from_str("aé"); + check_utf8_boundary(&string, 4); +} + +#[test] +#[should_panic(expected = "byte index 1 is not a codepoint boundary")] +fn wtf8_utf8_boundary_inside_codepoint() { + let string = Wtf8::from_str("é"); + check_utf8_boundary(&string, 1); +} + +#[test] +#[should_panic(expected = "byte index 1 is not a codepoint boundary")] +fn wtf8_utf8_boundary_inside_surrogate() { + let mut string = Wtf8Buf::new(); + string.push(CodePoint::from_u32(0xD800).unwrap()); + check_utf8_boundary(&string, 1); +} + +#[test] +#[should_panic(expected = "byte index 3 lies between surrogate codepoints")] +fn wtf8_utf8_boundary_between_surrogates() { + let mut string = Wtf8Buf::new(); + string.push(CodePoint::from_u32(0xD800).unwrap()); + string.push(CodePoint::from_u32(0xD800).unwrap()); + check_utf8_boundary(&string, 3); +} + +#[test] +fn wobbled_wtf8_plus_bytes_isnt_utf8() { + let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + assert!(!string.is_known_utf8); + string.extend_from_slice(b"some utf-8"); + assert!(!string.is_known_utf8); +} + +#[test] +fn wobbled_wtf8_plus_str_isnt_utf8() { + let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + assert!(!string.is_known_utf8); + string.push_str("some utf-8"); + assert!(!string.is_known_utf8); +} + +#[test] +fn unwobbly_wtf8_plus_utf8_is_utf8() { + let mut string: Wtf8Buf = Wtf8Buf::from_str("hello world"); + assert!(string.is_known_utf8); + string.push_str("some utf-8"); + assert!(string.is_known_utf8); +} diff --git a/library/core/src/wtf8.rs b/library/core/src/wtf8.rs new file mode 100644 index 00000000000..50bde88b5a4 --- /dev/null +++ b/library/core/src/wtf8.rs @@ -0,0 +1,1083 @@ +//! Implementation of [the WTF-8 encoding](https://simonsapin.github.io/wtf-8/). +//! +//! This library uses Rust’s type system to maintain +//! [well-formedness](https://simonsapin.github.io/wtf-8/#well-formed), +//! like the `String` and `&str` types do for UTF-8. +//! +//! Since [WTF-8 must not be used +//! for interchange](https://simonsapin.github.io/wtf-8/#intended-audience), +//! this library deliberately does not provide access to the underlying bytes +//! of WTF-8 strings, +//! nor can it decode WTF-8 from arbitrary bytes. +//! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points. + +// this module is imported from @SimonSapin's repo and has tons of dead code on +// unix (it's mostly used on windows), so don't worry about dead code here. +#![allow(dead_code)] + +#[cfg(test)] +mod tests; + +use core::char::{MAX_LEN_UTF8, MAX_LEN_UTF16, encode_utf8_raw, encode_utf16_raw}; +use core::clone::CloneToUninit; +use core::str::next_code_point; + +use crate::borrow::Cow; +use crate::collections::TryReserveError; +use crate::hash::{Hash, Hasher}; +use crate::iter::FusedIterator; +use crate::rc::Rc; +use crate::sync::Arc; +use crate::sys_common::AsInner; +use crate::{fmt, mem, ops, slice, str}; + +const UTF8_REPLACEMENT_CHARACTER: &str = "\u{FFFD}"; + +/// A Unicode code point: from U+0000 to U+10FFFF. +/// +/// Compares with the `char` type, +/// which represents a Unicode scalar value: +/// a code point that is not a surrogate (U+D800 to U+DFFF). +#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)] +pub struct CodePoint { + value: u32, +} + +/// Format the code point as `U+` followed by four to six hexadecimal digits. +/// Example: `U+1F4A9` +impl fmt::Debug for CodePoint { + #[inline] + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + write!(formatter, "U+{:04X}", self.value) + } +} + +impl CodePoint { + /// Unsafely creates a new `CodePoint` without checking the value. + /// + /// Only use when `value` is known to be less than or equal to 0x10FFFF. + #[inline] + pub unsafe fn from_u32_unchecked(value: u32) -> CodePoint { + CodePoint { value } + } + + /// Creates a new `CodePoint` if the value is a valid code point. + /// + /// Returns `None` if `value` is above 0x10FFFF. + #[inline] + pub fn from_u32(value: u32) -> Option { + match value { + 0..=0x10FFFF => Some(CodePoint { value }), + _ => None, + } + } + + /// Creates a new `CodePoint` from a `char`. + /// + /// Since all Unicode scalar values are code points, this always succeeds. + #[inline] + pub fn from_char(value: char) -> CodePoint { + CodePoint { value: value as u32 } + } + + /// Returns the numeric value of the code point. + #[inline] + pub fn to_u32(&self) -> u32 { + self.value + } + + /// Returns the numeric value of the code point if it is a leading surrogate. + #[inline] + pub fn to_lead_surrogate(&self) -> Option { + match self.value { + lead @ 0xD800..=0xDBFF => Some(lead as u16), + _ => None, + } + } + + /// Returns the numeric value of the code point if it is a trailing surrogate. + #[inline] + pub fn to_trail_surrogate(&self) -> Option { + match self.value { + trail @ 0xDC00..=0xDFFF => Some(trail as u16), + _ => None, + } + } + + /// Optionally returns a Unicode scalar value for the code point. + /// + /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF). + #[inline] + pub fn to_char(&self) -> Option { + match self.value { + 0xD800..=0xDFFF => None, + _ => Some(unsafe { char::from_u32_unchecked(self.value) }), + } + } + + /// Returns a Unicode scalar value for the code point. + /// + /// Returns `'\u{FFFD}'` (the replacement character “�”) + /// if the code point is a surrogate (from U+D800 to U+DFFF). + #[inline] + pub fn to_char_lossy(&self) -> char { + self.to_char().unwrap_or('\u{FFFD}') + } +} + +/// An owned, growable string of well-formed WTF-8 data. +/// +/// Similar to `String`, but can additionally contain surrogate code points +/// if they’re not in a surrogate pair. +#[derive(Eq, PartialEq, Ord, PartialOrd, Clone)] +pub struct Wtf8Buf { + bytes: Vec, + + /// Do we know that `bytes` holds a valid UTF-8 encoding? We can easily + /// know this if we're constructed from a `String` or `&str`. + /// + /// It is possible for `bytes` to have valid UTF-8 without this being + /// set, such as when we're concatenating `&Wtf8`'s and surrogates become + /// paired, as we don't bother to rescan the entire string. + is_known_utf8: bool, +} + +impl ops::Deref for Wtf8Buf { + type Target = Wtf8; + + fn deref(&self) -> &Wtf8 { + self.as_slice() + } +} + +impl ops::DerefMut for Wtf8Buf { + fn deref_mut(&mut self) -> &mut Wtf8 { + self.as_mut_slice() + } +} + +/// Formats the string in double quotes, with characters escaped according to +/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, +/// where each `x` is a hexadecimal digit. +/// +/// For example, the code units [U+0061, U+D800, U+000A] are formatted as +/// `"a\u{D800}\n"`. +impl fmt::Debug for Wtf8Buf { + #[inline] + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&**self, formatter) + } +} + +/// Formats the string with unpaired surrogates substituted with the replacement +/// character, U+FFFD. +impl fmt::Display for Wtf8Buf { + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + if let Some(s) = self.as_known_utf8() { + fmt::Display::fmt(s, formatter) + } else { + fmt::Display::fmt(&**self, formatter) + } + } +} + +impl Wtf8Buf { + /// Creates a new, empty WTF-8 string. + #[inline] + pub fn new() -> Wtf8Buf { + Wtf8Buf { bytes: Vec::new(), is_known_utf8: true } + } + + /// Creates a new, empty WTF-8 string with pre-allocated capacity for `capacity` bytes. + #[inline] + pub fn with_capacity(capacity: usize) -> Wtf8Buf { + Wtf8Buf { bytes: Vec::with_capacity(capacity), is_known_utf8: true } + } + + /// Creates a WTF-8 string from a WTF-8 byte vec. + /// + /// Since the byte vec is not checked for valid WTF-8, this function is + /// marked unsafe. + #[inline] + pub unsafe fn from_bytes_unchecked(value: Vec) -> Wtf8Buf { + Wtf8Buf { bytes: value, is_known_utf8: false } + } + + /// Creates a WTF-8 string from a UTF-8 `String`. + /// + /// This takes ownership of the `String` and does not copy. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub const fn from_string(string: String) -> Wtf8Buf { + Wtf8Buf { bytes: string.into_bytes(), is_known_utf8: true } + } + + /// Creates a WTF-8 string from a UTF-8 `&str` slice. + /// + /// This copies the content of the slice. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_str(s: &str) -> Wtf8Buf { + Wtf8Buf { bytes: s.as_bytes().to_vec(), is_known_utf8: true } + } + + pub fn clear(&mut self) { + self.bytes.clear(); + self.is_known_utf8 = true; + } + + /// Creates a WTF-8 string from a potentially ill-formed UTF-16 slice of 16-bit code units. + /// + /// This is lossless: calling `.encode_wide()` on the resulting string + /// will always return the original code units. + pub fn from_wide(v: &[u16]) -> Wtf8Buf { + let mut string = Wtf8Buf::with_capacity(v.len()); + for item in char::decode_utf16(v.iter().cloned()) { + match item { + Ok(ch) => string.push_char(ch), + Err(surrogate) => { + let surrogate = surrogate.unpaired_surrogate(); + // Surrogates are known to be in the code point range. + let code_point = unsafe { CodePoint::from_u32_unchecked(surrogate as u32) }; + // The string will now contain an unpaired surrogate. + string.is_known_utf8 = false; + // Skip the WTF-8 concatenation check, + // surrogate pairs are already decoded by decode_utf16 + string.push_code_point_unchecked(code_point); + } + } + } + string + } + + /// Appends the given `char` to the end of this string. + /// This does **not** include the WTF-8 concatenation check or `is_known_utf8` check. + /// Copied from String::push. + fn push_code_point_unchecked(&mut self, code_point: CodePoint) { + let mut bytes = [0; MAX_LEN_UTF8]; + let bytes = encode_utf8_raw(code_point.value, &mut bytes); + self.bytes.extend_from_slice(bytes) + } + + #[inline] + pub fn as_slice(&self) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(&self.bytes) } + } + + #[inline] + pub fn as_mut_slice(&mut self) -> &mut Wtf8 { + // Safety: `Wtf8` doesn't expose any way to mutate the bytes that would + // cause them to change from well-formed UTF-8 to ill-formed UTF-8, + // which would break the assumptions of the `is_known_utf8` field. + unsafe { Wtf8::from_mut_bytes_unchecked(&mut self.bytes) } + } + + /// Converts the string to UTF-8 without validation, if it was created from + /// valid UTF-8. + #[inline] + fn as_known_utf8(&self) -> Option<&str> { + if self.is_known_utf8 { + // SAFETY: The buffer is known to be valid UTF-8. + Some(unsafe { str::from_utf8_unchecked(self.as_bytes()) }) + } else { + None + } + } + + /// Reserves capacity for at least `additional` more bytes to be inserted + /// in the given `Wtf8Buf`. + /// The collection may reserve more space to avoid frequent reallocations. + /// + /// # Panics + /// + /// Panics if the new capacity exceeds `isize::MAX` bytes. + #[inline] + pub fn reserve(&mut self, additional: usize) { + self.bytes.reserve(additional) + } + + /// Tries to reserve capacity for at least `additional` more bytes to be + /// inserted in the given `Wtf8Buf`. The `Wtf8Buf` may reserve more space to + /// avoid frequent reallocations. After calling `try_reserve`, capacity will + /// be greater than or equal to `self.len() + additional`. Does nothing if + /// capacity is already sufficient. This method preserves the contents even + /// if an error occurs. + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + #[inline] + pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.bytes.try_reserve(additional) + } + + #[inline] + pub fn reserve_exact(&mut self, additional: usize) { + self.bytes.reserve_exact(additional) + } + + /// Tries to reserve the minimum capacity for exactly `additional` more + /// bytes to be inserted in the given `Wtf8Buf`. After calling + /// `try_reserve_exact`, capacity will be greater than or equal to + /// `self.len() + additional` if it returns `Ok(())`. + /// Does nothing if the capacity is already sufficient. + /// + /// Note that the allocator may give the `Wtf8Buf` more space than it + /// requests. Therefore, capacity can not be relied upon to be precisely + /// minimal. Prefer [`try_reserve`] if future insertions are expected. + /// + /// [`try_reserve`]: Wtf8Buf::try_reserve + /// + /// # Errors + /// + /// If the capacity overflows, or the allocator reports a failure, then an error + /// is returned. + #[inline] + pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { + self.bytes.try_reserve_exact(additional) + } + + #[inline] + pub fn shrink_to_fit(&mut self) { + self.bytes.shrink_to_fit() + } + + #[inline] + pub fn shrink_to(&mut self, min_capacity: usize) { + self.bytes.shrink_to(min_capacity) + } + + #[inline] + pub fn leak<'a>(self) -> &'a mut Wtf8 { + unsafe { Wtf8::from_mut_bytes_unchecked(self.bytes.leak()) } + } + + /// Returns the number of bytes that this string buffer can hold without reallocating. + #[inline] + pub fn capacity(&self) -> usize { + self.bytes.capacity() + } + + /// Append a UTF-8 slice at the end of the string. + #[inline] + pub fn push_str(&mut self, other: &str) { + self.bytes.extend_from_slice(other.as_bytes()) + } + + /// Append a WTF-8 slice at the end of the string. + /// + /// This replaces newly paired surrogates at the boundary + /// with a supplementary code point, + /// like concatenating ill-formed UTF-16 strings effectively would. + #[inline] + pub fn push_wtf8(&mut self, other: &Wtf8) { + match ((&*self).final_lead_surrogate(), other.initial_trail_surrogate()) { + // Replace newly paired surrogates by a supplementary code point. + (Some(lead), Some(trail)) => { + let len_without_lead_surrogate = self.len() - 3; + self.bytes.truncate(len_without_lead_surrogate); + let other_without_trail_surrogate = &other.bytes[3..]; + // 4 bytes for the supplementary code point + self.bytes.reserve(4 + other_without_trail_surrogate.len()); + self.push_char(decode_surrogate_pair(lead, trail)); + self.bytes.extend_from_slice(other_without_trail_surrogate); + } + _ => { + // If we'll be pushing a string containing a surrogate, we may + // no longer have UTF-8. + if self.is_known_utf8 && other.next_surrogate(0).is_some() { + self.is_known_utf8 = false; + } + + self.bytes.extend_from_slice(&other.bytes); + } + } + } + + /// Append a Unicode scalar value at the end of the string. + #[inline] + pub fn push_char(&mut self, c: char) { + self.push_code_point_unchecked(CodePoint::from_char(c)) + } + + /// Append a code point at the end of the string. + /// + /// This replaces newly paired surrogates at the boundary + /// with a supplementary code point, + /// like concatenating ill-formed UTF-16 strings effectively would. + #[inline] + pub fn push(&mut self, code_point: CodePoint) { + if let Some(trail) = code_point.to_trail_surrogate() { + if let Some(lead) = (&*self).final_lead_surrogate() { + let len_without_lead_surrogate = self.len() - 3; + self.bytes.truncate(len_without_lead_surrogate); + self.push_char(decode_surrogate_pair(lead, trail)); + return; + } + + // We're pushing a trailing surrogate. + self.is_known_utf8 = false; + } else if code_point.to_lead_surrogate().is_some() { + // We're pushing a leading surrogate. + self.is_known_utf8 = false; + } + + // No newly paired surrogates at the boundary. + self.push_code_point_unchecked(code_point) + } + + /// Shortens a string to the specified length. + /// + /// # Panics + /// + /// Panics if `new_len` > current length, + /// or if `new_len` is not a code point boundary. + #[inline] + pub fn truncate(&mut self, new_len: usize) { + assert!(is_code_point_boundary(self, new_len)); + self.bytes.truncate(new_len) + } + + /// Consumes the WTF-8 string and tries to convert it to a vec of bytes. + #[inline] + pub fn into_bytes(self) -> Vec { + self.bytes + } + + /// Consumes the WTF-8 string and tries to convert it to UTF-8. + /// + /// This does not copy the data. + /// + /// If the contents are not well-formed UTF-8 + /// (that is, if the string contains surrogates), + /// the original WTF-8 string is returned instead. + pub fn into_string(self) -> Result { + if self.is_known_utf8 || self.next_surrogate(0).is_none() { + Ok(unsafe { String::from_utf8_unchecked(self.bytes) }) + } else { + Err(self) + } + } + + /// Consumes the WTF-8 string and converts it lossily to UTF-8. + /// + /// This does not copy the data (but may overwrite parts of it in place). + /// + /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”) + pub fn into_string_lossy(mut self) -> String { + if !self.is_known_utf8 { + let mut pos = 0; + while let Some((surrogate_pos, _)) = self.next_surrogate(pos) { + pos = surrogate_pos + 3; + // Surrogates and the replacement character are all 3 bytes, so + // they can substituted in-place. + self.bytes[surrogate_pos..pos] + .copy_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + } + } + unsafe { String::from_utf8_unchecked(self.bytes) } + } + + /// Converts this `Wtf8Buf` into a boxed `Wtf8`. + #[inline] + pub fn into_box(self) -> Box { + // SAFETY: relies on `Wtf8` being `repr(transparent)`. + unsafe { mem::transmute(self.bytes.into_boxed_slice()) } + } + + /// Converts a `Box` into a `Wtf8Buf`. + pub fn from_box(boxed: Box) -> Wtf8Buf { + let bytes: Box<[u8]> = unsafe { mem::transmute(boxed) }; + Wtf8Buf { bytes: bytes.into_vec(), is_known_utf8: false } + } + + /// Provides plumbing to core `Vec::extend_from_slice`. + /// More well behaving alternative to allowing outer types + /// full mutable access to the core `Vec`. + #[inline] + pub(crate) fn extend_from_slice(&mut self, other: &[u8]) { + self.bytes.extend_from_slice(other); + self.is_known_utf8 = false; + } +} + +/// Creates a new WTF-8 string from an iterator of code points. +/// +/// This replaces surrogate code point pairs with supplementary code points, +/// like concatenating ill-formed UTF-16 strings effectively would. +impl FromIterator for Wtf8Buf { + fn from_iter>(iter: T) -> Wtf8Buf { + let mut string = Wtf8Buf::new(); + string.extend(iter); + string + } +} + +/// Append code points from an iterator to the string. +/// +/// This replaces surrogate code point pairs with supplementary code points, +/// like concatenating ill-formed UTF-16 strings effectively would. +impl Extend for Wtf8Buf { + fn extend>(&mut self, iter: T) { + let iterator = iter.into_iter(); + let (low, _high) = iterator.size_hint(); + // Lower bound of one byte per code point (ASCII only) + self.bytes.reserve(low); + iterator.for_each(move |code_point| self.push(code_point)); + } + + #[inline] + fn extend_one(&mut self, code_point: CodePoint) { + self.push(code_point); + } + + #[inline] + fn extend_reserve(&mut self, additional: usize) { + // Lower bound of one byte per code point (ASCII only) + self.bytes.reserve(additional); + } +} + +/// A borrowed slice of well-formed WTF-8 data. +/// +/// Similar to `&str`, but can additionally contain surrogate code points +/// if they’re not in a surrogate pair. +#[derive(Eq, Ord, PartialEq, PartialOrd)] +#[repr(transparent)] +pub struct Wtf8 { + bytes: [u8], +} + +impl AsInner<[u8]> for Wtf8 { + #[inline] + fn as_inner(&self) -> &[u8] { + &self.bytes + } +} + +/// Formats the string in double quotes, with characters escaped according to +/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, +/// where each `x` is a hexadecimal digit. +impl fmt::Debug for Wtf8 { + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + fn write_str_escaped(f: &mut fmt::Formatter<'_>, s: &str) -> fmt::Result { + use crate::fmt::Write; + for c in s.chars().flat_map(|c| c.escape_debug()) { + f.write_char(c)? + } + Ok(()) + } + + formatter.write_str("\"")?; + let mut pos = 0; + while let Some((surrogate_pos, surrogate)) = self.next_surrogate(pos) { + write_str_escaped(formatter, unsafe { + str::from_utf8_unchecked(&self.bytes[pos..surrogate_pos]) + })?; + write!(formatter, "\\u{{{:x}}}", surrogate)?; + pos = surrogate_pos + 3; + } + write_str_escaped(formatter, unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) })?; + formatter.write_str("\"") + } +} + +/// Formats the string with unpaired surrogates substituted with the replacement +/// character, U+FFFD. +impl fmt::Display for Wtf8 { + fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { + let wtf8_bytes = &self.bytes; + let mut pos = 0; + loop { + match self.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + formatter.write_str(unsafe { + str::from_utf8_unchecked(&wtf8_bytes[pos..surrogate_pos]) + })?; + formatter.write_str(UTF8_REPLACEMENT_CHARACTER)?; + pos = surrogate_pos + 3; + } + None => { + let s = unsafe { str::from_utf8_unchecked(&wtf8_bytes[pos..]) }; + if pos == 0 { return s.fmt(formatter) } else { return formatter.write_str(s) } + } + } + } + } +} + +impl Wtf8 { + /// Creates a WTF-8 slice from a UTF-8 `&str` slice. + /// + /// Since WTF-8 is a superset of UTF-8, this always succeeds. + #[inline] + pub fn from_str(value: &str) -> &Wtf8 { + unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) } + } + + /// Creates a WTF-8 slice from a WTF-8 byte slice. + /// + /// Since the byte slice is not checked for valid WTF-8, this functions is + /// marked unsafe. + #[inline] + pub unsafe fn from_bytes_unchecked(value: &[u8]) -> &Wtf8 { + // SAFETY: start with &[u8], end with fancy &[u8] + unsafe { &*(value as *const [u8] as *const Wtf8) } + } + + /// Creates a mutable WTF-8 slice from a mutable WTF-8 byte slice. + /// + /// Since the byte slice is not checked for valid WTF-8, this functions is + /// marked unsafe. + #[inline] + unsafe fn from_mut_bytes_unchecked(value: &mut [u8]) -> &mut Wtf8 { + // SAFETY: start with &mut [u8], end with fancy &mut [u8] + unsafe { &mut *(value as *mut [u8] as *mut Wtf8) } + } + + /// Returns the length, in WTF-8 bytes. + #[inline] + pub fn len(&self) -> usize { + self.bytes.len() + } + + #[inline] + pub fn is_empty(&self) -> bool { + self.bytes.is_empty() + } + + /// Returns the code point at `position` if it is in the ASCII range, + /// or `b'\xFF'` otherwise. + /// + /// # Panics + /// + /// Panics if `position` is beyond the end of the string. + #[inline] + pub fn ascii_byte_at(&self, position: usize) -> u8 { + match self.bytes[position] { + ascii_byte @ 0x00..=0x7F => ascii_byte, + _ => 0xFF, + } + } + + /// Returns an iterator for the string’s code points. + #[inline] + pub fn code_points(&self) -> Wtf8CodePoints<'_> { + Wtf8CodePoints { bytes: self.bytes.iter() } + } + + /// Access raw bytes of WTF-8 data + #[inline] + pub fn as_bytes(&self) -> &[u8] { + &self.bytes + } + + /// Tries to convert the string to UTF-8 and return a `&str` slice. + /// + /// Returns `None` if the string contains surrogates. + /// + /// This does not copy the data. + #[inline] + pub fn as_str(&self) -> Result<&str, str::Utf8Error> { + str::from_utf8(&self.bytes) + } + + /// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. + pub fn to_owned(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_vec(), is_known_utf8: false } + } + + /// Lossily converts the string to UTF-8. + /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. + /// + /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). + /// + /// This only copies the data if necessary (if it contains any surrogate). + pub fn to_string_lossy(&self) -> Cow<'_, str> { + let Some((surrogate_pos, _)) = self.next_surrogate(0) else { + return Cow::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }); + }; + let wtf8_bytes = &self.bytes; + let mut utf8_bytes = Vec::with_capacity(self.len()); + utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); + utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + let mut pos = surrogate_pos + 3; + loop { + match self.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); + utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + pos = surrogate_pos + 3; + } + None => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); + return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); + } + } + } + } + + /// Converts the WTF-8 string to potentially ill-formed UTF-16 + /// and return an iterator of 16-bit code units. + /// + /// This is lossless: + /// calling `Wtf8Buf::from_ill_formed_utf16` on the resulting code units + /// would always return the original WTF-8 string. + #[inline] + pub fn encode_wide(&self) -> EncodeWide<'_> { + EncodeWide { code_points: self.code_points(), extra: 0 } + } + + #[inline] + fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { + let mut iter = self.bytes[pos..].iter(); + loop { + let b = *iter.next()?; + if b < 0x80 { + pos += 1; + } else if b < 0xE0 { + iter.next(); + pos += 2; + } else if b == 0xED { + match (iter.next(), iter.next()) { + (Some(&b2), Some(&b3)) if b2 >= 0xA0 => { + return Some((pos, decode_surrogate(b2, b3))); + } + _ => pos += 3, + } + } else if b < 0xF0 { + iter.next(); + iter.next(); + pos += 3; + } else { + iter.next(); + iter.next(); + iter.next(); + pos += 4; + } + } + } + + #[inline] + fn final_lead_surrogate(&self) -> Option { + match self.bytes { + [.., 0xED, b2 @ 0xA0..=0xAF, b3] => Some(decode_surrogate(b2, b3)), + _ => None, + } + } + + #[inline] + fn initial_trail_surrogate(&self) -> Option { + match self.bytes { + [0xED, b2 @ 0xB0..=0xBF, b3, ..] => Some(decode_surrogate(b2, b3)), + _ => None, + } + } + + pub fn clone_into(&self, buf: &mut Wtf8Buf) { + buf.is_known_utf8 = false; + self.bytes.clone_into(&mut buf.bytes); + } + + /// Boxes this `Wtf8`. + #[inline] + pub fn into_box(&self) -> Box { + let boxed: Box<[u8]> = self.bytes.into(); + unsafe { mem::transmute(boxed) } + } + + /// Creates a boxed, empty `Wtf8`. + pub fn empty_box() -> Box { + let boxed: Box<[u8]> = Default::default(); + unsafe { mem::transmute(boxed) } + } + + #[inline] + pub fn into_arc(&self) -> Arc { + let arc: Arc<[u8]> = Arc::from(&self.bytes); + unsafe { Arc::from_raw(Arc::into_raw(arc) as *const Wtf8) } + } + + #[inline] + pub fn into_rc(&self) -> Rc { + let rc: Rc<[u8]> = Rc::from(&self.bytes); + unsafe { Rc::from_raw(Rc::into_raw(rc) as *const Wtf8) } + } + + #[inline] + pub fn make_ascii_lowercase(&mut self) { + self.bytes.make_ascii_lowercase() + } + + #[inline] + pub fn make_ascii_uppercase(&mut self) { + self.bytes.make_ascii_uppercase() + } + + #[inline] + pub fn to_ascii_lowercase(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_ascii_lowercase(), is_known_utf8: false } + } + + #[inline] + pub fn to_ascii_uppercase(&self) -> Wtf8Buf { + Wtf8Buf { bytes: self.bytes.to_ascii_uppercase(), is_known_utf8: false } + } + + #[inline] + pub fn is_ascii(&self) -> bool { + self.bytes.is_ascii() + } + + #[inline] + pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { + self.bytes.eq_ignore_ascii_case(&other.bytes) + } +} + +/// Returns a slice of the given string for the byte range \[`begin`..`end`). +/// +/// # Panics +/// +/// Panics when `begin` and `end` do not point to code point boundaries, +/// or point beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::Range) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if range.start <= range.end + && is_code_point_boundary(self, range.start) + && is_code_point_boundary(self, range.end) + { + unsafe { slice_unchecked(self, range.start, range.end) } + } else { + slice_error_fail(self, range.start, range.end) + } + } +} + +/// Returns a slice of the given string from byte `begin` to its end. +/// +/// # Panics +/// +/// Panics when `begin` is not at a code point boundary, +/// or is beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::RangeFrom) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if is_code_point_boundary(self, range.start) { + unsafe { slice_unchecked(self, range.start, self.len()) } + } else { + slice_error_fail(self, range.start, self.len()) + } + } +} + +/// Returns a slice of the given string from its beginning to byte `end`. +/// +/// # Panics +/// +/// Panics when `end` is not at a code point boundary, +/// or is beyond the end of the string. +impl ops::Index> for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, range: ops::RangeTo) -> &Wtf8 { + // is_code_point_boundary checks that the index is in [0, .len()] + if is_code_point_boundary(self, range.end) { + unsafe { slice_unchecked(self, 0, range.end) } + } else { + slice_error_fail(self, 0, range.end) + } + } +} + +impl ops::Index for Wtf8 { + type Output = Wtf8; + + #[inline] + fn index(&self, _range: ops::RangeFull) -> &Wtf8 { + self + } +} + +#[inline] +fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { + // The first byte is assumed to be 0xED + 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F +} + +#[inline] +fn decode_surrogate_pair(lead: u16, trail: u16) -> char { + let code_point = 0x10000 + ((((lead - 0xD800) as u32) << 10) | (trail - 0xDC00) as u32); + unsafe { char::from_u32_unchecked(code_point) } +} + +/// Copied from str::is_char_boundary +#[inline] +pub fn is_code_point_boundary(slice: &Wtf8, index: usize) -> bool { + if index == 0 { + return true; + } + match slice.bytes.get(index) { + None => index == slice.len(), + Some(&b) => (b as i8) >= -0x40, + } +} + +/// Verify that `index` is at the edge of either a valid UTF-8 codepoint +/// (i.e. a codepoint that's not a surrogate) or of the whole string. +/// +/// These are the cases currently permitted by `OsStr::slice_encoded_bytes`. +/// Splitting between surrogates is valid as far as WTF-8 is concerned, but +/// we do not permit it in the public API because WTF-8 is considered an +/// implementation detail. +#[track_caller] +#[inline] +pub fn check_utf8_boundary(slice: &Wtf8, index: usize) { + if index == 0 { + return; + } + match slice.bytes.get(index) { + Some(0xED) => (), // Might be a surrogate + Some(&b) if (b as i8) >= -0x40 => return, + Some(_) => panic!("byte index {index} is not a codepoint boundary"), + None if index == slice.len() => return, + None => panic!("byte index {index} is out of bounds"), + } + if slice.bytes[index + 1] >= 0xA0 { + // There's a surrogate after index. Now check before index. + if index >= 3 && slice.bytes[index - 3] == 0xED && slice.bytes[index - 2] >= 0xA0 { + panic!("byte index {index} lies between surrogate codepoints"); + } + } +} + +/// Copied from core::str::raw::slice_unchecked +#[inline] +pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { + // SAFETY: memory layout of a &[u8] and &Wtf8 are the same + unsafe { + let len = end - begin; + let start = s.as_bytes().as_ptr().add(begin); + Wtf8::from_bytes_unchecked(slice::from_raw_parts(start, len)) + } +} + +/// Copied from core::str::raw::slice_error_fail +#[inline(never)] +pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { + assert!(begin <= end); + panic!("index {begin} and/or {end} in `{s:?}` do not lie on character boundary"); +} + +/// Iterator for the code points of a WTF-8 string. +/// +/// Created with the method `.code_points()`. +#[derive(Clone)] +pub struct Wtf8CodePoints<'a> { + bytes: slice::Iter<'a, u8>, +} + +impl Iterator for Wtf8CodePoints<'_> { + type Item = CodePoint; + + #[inline] + fn next(&mut self) -> Option { + // SAFETY: `self.bytes` has been created from a WTF-8 string + unsafe { next_code_point(&mut self.bytes).map(|c| CodePoint { value: c }) } + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + let len = self.bytes.len(); + (len.saturating_add(3) / 4, Some(len)) + } +} + +/// Generates a wide character sequence for potentially ill-formed UTF-16. +#[stable(feature = "rust1", since = "1.0.0")] +#[derive(Clone)] +pub struct EncodeWide<'a> { + code_points: Wtf8CodePoints<'a>, + extra: u16, +} + +// Copied from libunicode/u_str.rs +#[stable(feature = "rust1", since = "1.0.0")] +impl Iterator for EncodeWide<'_> { + type Item = u16; + + #[inline] + fn next(&mut self) -> Option { + if self.extra != 0 { + let tmp = self.extra; + self.extra = 0; + return Some(tmp); + } + + let mut buf = [0; MAX_LEN_UTF16]; + self.code_points.next().map(|code_point| { + let n = encode_utf16_raw(code_point.value, &mut buf).len(); + if n == 2 { + self.extra = buf[1]; + } + buf[0] + }) + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + let (low, high) = self.code_points.size_hint(); + let ext = (self.extra != 0) as usize; + // every code point gets either one u16 or two u16, + // so this iterator is between 1 or 2 times as + // long as the underlying iterator. + (low + ext, high.and_then(|n| n.checked_mul(2)).and_then(|n| n.checked_add(ext))) + } +} + +#[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")] +impl FusedIterator for EncodeWide<'_> {} + +impl Hash for CodePoint { + #[inline] + fn hash(&self, state: &mut H) { + self.value.hash(state) + } +} + +impl Hash for Wtf8Buf { + #[inline] + fn hash(&self, state: &mut H) { + state.write(&self.bytes); + 0xfeu8.hash(state) + } +} + +impl Hash for Wtf8 { + #[inline] + fn hash(&self, state: &mut H) { + state.write(&self.bytes); + 0xfeu8.hash(state) + } +} + +#[unstable(feature = "clone_to_uninit", issue = "126799")] +unsafe impl CloneToUninit for Wtf8 { + #[inline] + #[cfg_attr(debug_assertions, track_caller)] + unsafe fn clone_to_uninit(&self, dst: *mut u8) { + // SAFETY: we're just a transparent wrapper around [u8] + unsafe { self.bytes.clone_to_uninit(dst) } + } +} diff --git a/library/std/src/sys_common/wtf8.rs b/library/std/src/sys_common/wtf8.rs deleted file mode 100644 index 50bde88b5a4..00000000000 --- a/library/std/src/sys_common/wtf8.rs +++ /dev/null @@ -1,1083 +0,0 @@ -//! Implementation of [the WTF-8 encoding](https://simonsapin.github.io/wtf-8/). -//! -//! This library uses Rust’s type system to maintain -//! [well-formedness](https://simonsapin.github.io/wtf-8/#well-formed), -//! like the `String` and `&str` types do for UTF-8. -//! -//! Since [WTF-8 must not be used -//! for interchange](https://simonsapin.github.io/wtf-8/#intended-audience), -//! this library deliberately does not provide access to the underlying bytes -//! of WTF-8 strings, -//! nor can it decode WTF-8 from arbitrary bytes. -//! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points. - -// this module is imported from @SimonSapin's repo and has tons of dead code on -// unix (it's mostly used on windows), so don't worry about dead code here. -#![allow(dead_code)] - -#[cfg(test)] -mod tests; - -use core::char::{MAX_LEN_UTF8, MAX_LEN_UTF16, encode_utf8_raw, encode_utf16_raw}; -use core::clone::CloneToUninit; -use core::str::next_code_point; - -use crate::borrow::Cow; -use crate::collections::TryReserveError; -use crate::hash::{Hash, Hasher}; -use crate::iter::FusedIterator; -use crate::rc::Rc; -use crate::sync::Arc; -use crate::sys_common::AsInner; -use crate::{fmt, mem, ops, slice, str}; - -const UTF8_REPLACEMENT_CHARACTER: &str = "\u{FFFD}"; - -/// A Unicode code point: from U+0000 to U+10FFFF. -/// -/// Compares with the `char` type, -/// which represents a Unicode scalar value: -/// a code point that is not a surrogate (U+D800 to U+DFFF). -#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)] -pub struct CodePoint { - value: u32, -} - -/// Format the code point as `U+` followed by four to six hexadecimal digits. -/// Example: `U+1F4A9` -impl fmt::Debug for CodePoint { - #[inline] - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - write!(formatter, "U+{:04X}", self.value) - } -} - -impl CodePoint { - /// Unsafely creates a new `CodePoint` without checking the value. - /// - /// Only use when `value` is known to be less than or equal to 0x10FFFF. - #[inline] - pub unsafe fn from_u32_unchecked(value: u32) -> CodePoint { - CodePoint { value } - } - - /// Creates a new `CodePoint` if the value is a valid code point. - /// - /// Returns `None` if `value` is above 0x10FFFF. - #[inline] - pub fn from_u32(value: u32) -> Option { - match value { - 0..=0x10FFFF => Some(CodePoint { value }), - _ => None, - } - } - - /// Creates a new `CodePoint` from a `char`. - /// - /// Since all Unicode scalar values are code points, this always succeeds. - #[inline] - pub fn from_char(value: char) -> CodePoint { - CodePoint { value: value as u32 } - } - - /// Returns the numeric value of the code point. - #[inline] - pub fn to_u32(&self) -> u32 { - self.value - } - - /// Returns the numeric value of the code point if it is a leading surrogate. - #[inline] - pub fn to_lead_surrogate(&self) -> Option { - match self.value { - lead @ 0xD800..=0xDBFF => Some(lead as u16), - _ => None, - } - } - - /// Returns the numeric value of the code point if it is a trailing surrogate. - #[inline] - pub fn to_trail_surrogate(&self) -> Option { - match self.value { - trail @ 0xDC00..=0xDFFF => Some(trail as u16), - _ => None, - } - } - - /// Optionally returns a Unicode scalar value for the code point. - /// - /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF). - #[inline] - pub fn to_char(&self) -> Option { - match self.value { - 0xD800..=0xDFFF => None, - _ => Some(unsafe { char::from_u32_unchecked(self.value) }), - } - } - - /// Returns a Unicode scalar value for the code point. - /// - /// Returns `'\u{FFFD}'` (the replacement character “�”) - /// if the code point is a surrogate (from U+D800 to U+DFFF). - #[inline] - pub fn to_char_lossy(&self) -> char { - self.to_char().unwrap_or('\u{FFFD}') - } -} - -/// An owned, growable string of well-formed WTF-8 data. -/// -/// Similar to `String`, but can additionally contain surrogate code points -/// if they’re not in a surrogate pair. -#[derive(Eq, PartialEq, Ord, PartialOrd, Clone)] -pub struct Wtf8Buf { - bytes: Vec, - - /// Do we know that `bytes` holds a valid UTF-8 encoding? We can easily - /// know this if we're constructed from a `String` or `&str`. - /// - /// It is possible for `bytes` to have valid UTF-8 without this being - /// set, such as when we're concatenating `&Wtf8`'s and surrogates become - /// paired, as we don't bother to rescan the entire string. - is_known_utf8: bool, -} - -impl ops::Deref for Wtf8Buf { - type Target = Wtf8; - - fn deref(&self) -> &Wtf8 { - self.as_slice() - } -} - -impl ops::DerefMut for Wtf8Buf { - fn deref_mut(&mut self) -> &mut Wtf8 { - self.as_mut_slice() - } -} - -/// Formats the string in double quotes, with characters escaped according to -/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, -/// where each `x` is a hexadecimal digit. -/// -/// For example, the code units [U+0061, U+D800, U+000A] are formatted as -/// `"a\u{D800}\n"`. -impl fmt::Debug for Wtf8Buf { - #[inline] - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - fmt::Debug::fmt(&**self, formatter) - } -} - -/// Formats the string with unpaired surrogates substituted with the replacement -/// character, U+FFFD. -impl fmt::Display for Wtf8Buf { - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - if let Some(s) = self.as_known_utf8() { - fmt::Display::fmt(s, formatter) - } else { - fmt::Display::fmt(&**self, formatter) - } - } -} - -impl Wtf8Buf { - /// Creates a new, empty WTF-8 string. - #[inline] - pub fn new() -> Wtf8Buf { - Wtf8Buf { bytes: Vec::new(), is_known_utf8: true } - } - - /// Creates a new, empty WTF-8 string with pre-allocated capacity for `capacity` bytes. - #[inline] - pub fn with_capacity(capacity: usize) -> Wtf8Buf { - Wtf8Buf { bytes: Vec::with_capacity(capacity), is_known_utf8: true } - } - - /// Creates a WTF-8 string from a WTF-8 byte vec. - /// - /// Since the byte vec is not checked for valid WTF-8, this function is - /// marked unsafe. - #[inline] - pub unsafe fn from_bytes_unchecked(value: Vec) -> Wtf8Buf { - Wtf8Buf { bytes: value, is_known_utf8: false } - } - - /// Creates a WTF-8 string from a UTF-8 `String`. - /// - /// This takes ownership of the `String` and does not copy. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. - #[inline] - pub const fn from_string(string: String) -> Wtf8Buf { - Wtf8Buf { bytes: string.into_bytes(), is_known_utf8: true } - } - - /// Creates a WTF-8 string from a UTF-8 `&str` slice. - /// - /// This copies the content of the slice. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. - #[inline] - pub fn from_str(s: &str) -> Wtf8Buf { - Wtf8Buf { bytes: s.as_bytes().to_vec(), is_known_utf8: true } - } - - pub fn clear(&mut self) { - self.bytes.clear(); - self.is_known_utf8 = true; - } - - /// Creates a WTF-8 string from a potentially ill-formed UTF-16 slice of 16-bit code units. - /// - /// This is lossless: calling `.encode_wide()` on the resulting string - /// will always return the original code units. - pub fn from_wide(v: &[u16]) -> Wtf8Buf { - let mut string = Wtf8Buf::with_capacity(v.len()); - for item in char::decode_utf16(v.iter().cloned()) { - match item { - Ok(ch) => string.push_char(ch), - Err(surrogate) => { - let surrogate = surrogate.unpaired_surrogate(); - // Surrogates are known to be in the code point range. - let code_point = unsafe { CodePoint::from_u32_unchecked(surrogate as u32) }; - // The string will now contain an unpaired surrogate. - string.is_known_utf8 = false; - // Skip the WTF-8 concatenation check, - // surrogate pairs are already decoded by decode_utf16 - string.push_code_point_unchecked(code_point); - } - } - } - string - } - - /// Appends the given `char` to the end of this string. - /// This does **not** include the WTF-8 concatenation check or `is_known_utf8` check. - /// Copied from String::push. - fn push_code_point_unchecked(&mut self, code_point: CodePoint) { - let mut bytes = [0; MAX_LEN_UTF8]; - let bytes = encode_utf8_raw(code_point.value, &mut bytes); - self.bytes.extend_from_slice(bytes) - } - - #[inline] - pub fn as_slice(&self) -> &Wtf8 { - unsafe { Wtf8::from_bytes_unchecked(&self.bytes) } - } - - #[inline] - pub fn as_mut_slice(&mut self) -> &mut Wtf8 { - // Safety: `Wtf8` doesn't expose any way to mutate the bytes that would - // cause them to change from well-formed UTF-8 to ill-formed UTF-8, - // which would break the assumptions of the `is_known_utf8` field. - unsafe { Wtf8::from_mut_bytes_unchecked(&mut self.bytes) } - } - - /// Converts the string to UTF-8 without validation, if it was created from - /// valid UTF-8. - #[inline] - fn as_known_utf8(&self) -> Option<&str> { - if self.is_known_utf8 { - // SAFETY: The buffer is known to be valid UTF-8. - Some(unsafe { str::from_utf8_unchecked(self.as_bytes()) }) - } else { - None - } - } - - /// Reserves capacity for at least `additional` more bytes to be inserted - /// in the given `Wtf8Buf`. - /// The collection may reserve more space to avoid frequent reallocations. - /// - /// # Panics - /// - /// Panics if the new capacity exceeds `isize::MAX` bytes. - #[inline] - pub fn reserve(&mut self, additional: usize) { - self.bytes.reserve(additional) - } - - /// Tries to reserve capacity for at least `additional` more bytes to be - /// inserted in the given `Wtf8Buf`. The `Wtf8Buf` may reserve more space to - /// avoid frequent reallocations. After calling `try_reserve`, capacity will - /// be greater than or equal to `self.len() + additional`. Does nothing if - /// capacity is already sufficient. This method preserves the contents even - /// if an error occurs. - /// - /// # Errors - /// - /// If the capacity overflows, or the allocator reports a failure, then an error - /// is returned. - #[inline] - pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { - self.bytes.try_reserve(additional) - } - - #[inline] - pub fn reserve_exact(&mut self, additional: usize) { - self.bytes.reserve_exact(additional) - } - - /// Tries to reserve the minimum capacity for exactly `additional` more - /// bytes to be inserted in the given `Wtf8Buf`. After calling - /// `try_reserve_exact`, capacity will be greater than or equal to - /// `self.len() + additional` if it returns `Ok(())`. - /// Does nothing if the capacity is already sufficient. - /// - /// Note that the allocator may give the `Wtf8Buf` more space than it - /// requests. Therefore, capacity can not be relied upon to be precisely - /// minimal. Prefer [`try_reserve`] if future insertions are expected. - /// - /// [`try_reserve`]: Wtf8Buf::try_reserve - /// - /// # Errors - /// - /// If the capacity overflows, or the allocator reports a failure, then an error - /// is returned. - #[inline] - pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { - self.bytes.try_reserve_exact(additional) - } - - #[inline] - pub fn shrink_to_fit(&mut self) { - self.bytes.shrink_to_fit() - } - - #[inline] - pub fn shrink_to(&mut self, min_capacity: usize) { - self.bytes.shrink_to(min_capacity) - } - - #[inline] - pub fn leak<'a>(self) -> &'a mut Wtf8 { - unsafe { Wtf8::from_mut_bytes_unchecked(self.bytes.leak()) } - } - - /// Returns the number of bytes that this string buffer can hold without reallocating. - #[inline] - pub fn capacity(&self) -> usize { - self.bytes.capacity() - } - - /// Append a UTF-8 slice at the end of the string. - #[inline] - pub fn push_str(&mut self, other: &str) { - self.bytes.extend_from_slice(other.as_bytes()) - } - - /// Append a WTF-8 slice at the end of the string. - /// - /// This replaces newly paired surrogates at the boundary - /// with a supplementary code point, - /// like concatenating ill-formed UTF-16 strings effectively would. - #[inline] - pub fn push_wtf8(&mut self, other: &Wtf8) { - match ((&*self).final_lead_surrogate(), other.initial_trail_surrogate()) { - // Replace newly paired surrogates by a supplementary code point. - (Some(lead), Some(trail)) => { - let len_without_lead_surrogate = self.len() - 3; - self.bytes.truncate(len_without_lead_surrogate); - let other_without_trail_surrogate = &other.bytes[3..]; - // 4 bytes for the supplementary code point - self.bytes.reserve(4 + other_without_trail_surrogate.len()); - self.push_char(decode_surrogate_pair(lead, trail)); - self.bytes.extend_from_slice(other_without_trail_surrogate); - } - _ => { - // If we'll be pushing a string containing a surrogate, we may - // no longer have UTF-8. - if self.is_known_utf8 && other.next_surrogate(0).is_some() { - self.is_known_utf8 = false; - } - - self.bytes.extend_from_slice(&other.bytes); - } - } - } - - /// Append a Unicode scalar value at the end of the string. - #[inline] - pub fn push_char(&mut self, c: char) { - self.push_code_point_unchecked(CodePoint::from_char(c)) - } - - /// Append a code point at the end of the string. - /// - /// This replaces newly paired surrogates at the boundary - /// with a supplementary code point, - /// like concatenating ill-formed UTF-16 strings effectively would. - #[inline] - pub fn push(&mut self, code_point: CodePoint) { - if let Some(trail) = code_point.to_trail_surrogate() { - if let Some(lead) = (&*self).final_lead_surrogate() { - let len_without_lead_surrogate = self.len() - 3; - self.bytes.truncate(len_without_lead_surrogate); - self.push_char(decode_surrogate_pair(lead, trail)); - return; - } - - // We're pushing a trailing surrogate. - self.is_known_utf8 = false; - } else if code_point.to_lead_surrogate().is_some() { - // We're pushing a leading surrogate. - self.is_known_utf8 = false; - } - - // No newly paired surrogates at the boundary. - self.push_code_point_unchecked(code_point) - } - - /// Shortens a string to the specified length. - /// - /// # Panics - /// - /// Panics if `new_len` > current length, - /// or if `new_len` is not a code point boundary. - #[inline] - pub fn truncate(&mut self, new_len: usize) { - assert!(is_code_point_boundary(self, new_len)); - self.bytes.truncate(new_len) - } - - /// Consumes the WTF-8 string and tries to convert it to a vec of bytes. - #[inline] - pub fn into_bytes(self) -> Vec { - self.bytes - } - - /// Consumes the WTF-8 string and tries to convert it to UTF-8. - /// - /// This does not copy the data. - /// - /// If the contents are not well-formed UTF-8 - /// (that is, if the string contains surrogates), - /// the original WTF-8 string is returned instead. - pub fn into_string(self) -> Result { - if self.is_known_utf8 || self.next_surrogate(0).is_none() { - Ok(unsafe { String::from_utf8_unchecked(self.bytes) }) - } else { - Err(self) - } - } - - /// Consumes the WTF-8 string and converts it lossily to UTF-8. - /// - /// This does not copy the data (but may overwrite parts of it in place). - /// - /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”) - pub fn into_string_lossy(mut self) -> String { - if !self.is_known_utf8 { - let mut pos = 0; - while let Some((surrogate_pos, _)) = self.next_surrogate(pos) { - pos = surrogate_pos + 3; - // Surrogates and the replacement character are all 3 bytes, so - // they can substituted in-place. - self.bytes[surrogate_pos..pos] - .copy_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); - } - } - unsafe { String::from_utf8_unchecked(self.bytes) } - } - - /// Converts this `Wtf8Buf` into a boxed `Wtf8`. - #[inline] - pub fn into_box(self) -> Box { - // SAFETY: relies on `Wtf8` being `repr(transparent)`. - unsafe { mem::transmute(self.bytes.into_boxed_slice()) } - } - - /// Converts a `Box` into a `Wtf8Buf`. - pub fn from_box(boxed: Box) -> Wtf8Buf { - let bytes: Box<[u8]> = unsafe { mem::transmute(boxed) }; - Wtf8Buf { bytes: bytes.into_vec(), is_known_utf8: false } - } - - /// Provides plumbing to core `Vec::extend_from_slice`. - /// More well behaving alternative to allowing outer types - /// full mutable access to the core `Vec`. - #[inline] - pub(crate) fn extend_from_slice(&mut self, other: &[u8]) { - self.bytes.extend_from_slice(other); - self.is_known_utf8 = false; - } -} - -/// Creates a new WTF-8 string from an iterator of code points. -/// -/// This replaces surrogate code point pairs with supplementary code points, -/// like concatenating ill-formed UTF-16 strings effectively would. -impl FromIterator for Wtf8Buf { - fn from_iter>(iter: T) -> Wtf8Buf { - let mut string = Wtf8Buf::new(); - string.extend(iter); - string - } -} - -/// Append code points from an iterator to the string. -/// -/// This replaces surrogate code point pairs with supplementary code points, -/// like concatenating ill-formed UTF-16 strings effectively would. -impl Extend for Wtf8Buf { - fn extend>(&mut self, iter: T) { - let iterator = iter.into_iter(); - let (low, _high) = iterator.size_hint(); - // Lower bound of one byte per code point (ASCII only) - self.bytes.reserve(low); - iterator.for_each(move |code_point| self.push(code_point)); - } - - #[inline] - fn extend_one(&mut self, code_point: CodePoint) { - self.push(code_point); - } - - #[inline] - fn extend_reserve(&mut self, additional: usize) { - // Lower bound of one byte per code point (ASCII only) - self.bytes.reserve(additional); - } -} - -/// A borrowed slice of well-formed WTF-8 data. -/// -/// Similar to `&str`, but can additionally contain surrogate code points -/// if they’re not in a surrogate pair. -#[derive(Eq, Ord, PartialEq, PartialOrd)] -#[repr(transparent)] -pub struct Wtf8 { - bytes: [u8], -} - -impl AsInner<[u8]> for Wtf8 { - #[inline] - fn as_inner(&self) -> &[u8] { - &self.bytes - } -} - -/// Formats the string in double quotes, with characters escaped according to -/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, -/// where each `x` is a hexadecimal digit. -impl fmt::Debug for Wtf8 { - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - fn write_str_escaped(f: &mut fmt::Formatter<'_>, s: &str) -> fmt::Result { - use crate::fmt::Write; - for c in s.chars().flat_map(|c| c.escape_debug()) { - f.write_char(c)? - } - Ok(()) - } - - formatter.write_str("\"")?; - let mut pos = 0; - while let Some((surrogate_pos, surrogate)) = self.next_surrogate(pos) { - write_str_escaped(formatter, unsafe { - str::from_utf8_unchecked(&self.bytes[pos..surrogate_pos]) - })?; - write!(formatter, "\\u{{{:x}}}", surrogate)?; - pos = surrogate_pos + 3; - } - write_str_escaped(formatter, unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) })?; - formatter.write_str("\"") - } -} - -/// Formats the string with unpaired surrogates substituted with the replacement -/// character, U+FFFD. -impl fmt::Display for Wtf8 { - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - let wtf8_bytes = &self.bytes; - let mut pos = 0; - loop { - match self.next_surrogate(pos) { - Some((surrogate_pos, _)) => { - formatter.write_str(unsafe { - str::from_utf8_unchecked(&wtf8_bytes[pos..surrogate_pos]) - })?; - formatter.write_str(UTF8_REPLACEMENT_CHARACTER)?; - pos = surrogate_pos + 3; - } - None => { - let s = unsafe { str::from_utf8_unchecked(&wtf8_bytes[pos..]) }; - if pos == 0 { return s.fmt(formatter) } else { return formatter.write_str(s) } - } - } - } - } -} - -impl Wtf8 { - /// Creates a WTF-8 slice from a UTF-8 `&str` slice. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. - #[inline] - pub fn from_str(value: &str) -> &Wtf8 { - unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) } - } - - /// Creates a WTF-8 slice from a WTF-8 byte slice. - /// - /// Since the byte slice is not checked for valid WTF-8, this functions is - /// marked unsafe. - #[inline] - pub unsafe fn from_bytes_unchecked(value: &[u8]) -> &Wtf8 { - // SAFETY: start with &[u8], end with fancy &[u8] - unsafe { &*(value as *const [u8] as *const Wtf8) } - } - - /// Creates a mutable WTF-8 slice from a mutable WTF-8 byte slice. - /// - /// Since the byte slice is not checked for valid WTF-8, this functions is - /// marked unsafe. - #[inline] - unsafe fn from_mut_bytes_unchecked(value: &mut [u8]) -> &mut Wtf8 { - // SAFETY: start with &mut [u8], end with fancy &mut [u8] - unsafe { &mut *(value as *mut [u8] as *mut Wtf8) } - } - - /// Returns the length, in WTF-8 bytes. - #[inline] - pub fn len(&self) -> usize { - self.bytes.len() - } - - #[inline] - pub fn is_empty(&self) -> bool { - self.bytes.is_empty() - } - - /// Returns the code point at `position` if it is in the ASCII range, - /// or `b'\xFF'` otherwise. - /// - /// # Panics - /// - /// Panics if `position` is beyond the end of the string. - #[inline] - pub fn ascii_byte_at(&self, position: usize) -> u8 { - match self.bytes[position] { - ascii_byte @ 0x00..=0x7F => ascii_byte, - _ => 0xFF, - } - } - - /// Returns an iterator for the string’s code points. - #[inline] - pub fn code_points(&self) -> Wtf8CodePoints<'_> { - Wtf8CodePoints { bytes: self.bytes.iter() } - } - - /// Access raw bytes of WTF-8 data - #[inline] - pub fn as_bytes(&self) -> &[u8] { - &self.bytes - } - - /// Tries to convert the string to UTF-8 and return a `&str` slice. - /// - /// Returns `None` if the string contains surrogates. - /// - /// This does not copy the data. - #[inline] - pub fn as_str(&self) -> Result<&str, str::Utf8Error> { - str::from_utf8(&self.bytes) - } - - /// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. - pub fn to_owned(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_vec(), is_known_utf8: false } - } - - /// Lossily converts the string to UTF-8. - /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. - /// - /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). - /// - /// This only copies the data if necessary (if it contains any surrogate). - pub fn to_string_lossy(&self) -> Cow<'_, str> { - let Some((surrogate_pos, _)) = self.next_surrogate(0) else { - return Cow::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }); - }; - let wtf8_bytes = &self.bytes; - let mut utf8_bytes = Vec::with_capacity(self.len()); - utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); - utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); - let mut pos = surrogate_pos + 3; - loop { - match self.next_surrogate(pos) { - Some((surrogate_pos, _)) => { - utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); - utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); - pos = surrogate_pos + 3; - } - None => { - utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); - return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); - } - } - } - } - - /// Converts the WTF-8 string to potentially ill-formed UTF-16 - /// and return an iterator of 16-bit code units. - /// - /// This is lossless: - /// calling `Wtf8Buf::from_ill_formed_utf16` on the resulting code units - /// would always return the original WTF-8 string. - #[inline] - pub fn encode_wide(&self) -> EncodeWide<'_> { - EncodeWide { code_points: self.code_points(), extra: 0 } - } - - #[inline] - fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { - let mut iter = self.bytes[pos..].iter(); - loop { - let b = *iter.next()?; - if b < 0x80 { - pos += 1; - } else if b < 0xE0 { - iter.next(); - pos += 2; - } else if b == 0xED { - match (iter.next(), iter.next()) { - (Some(&b2), Some(&b3)) if b2 >= 0xA0 => { - return Some((pos, decode_surrogate(b2, b3))); - } - _ => pos += 3, - } - } else if b < 0xF0 { - iter.next(); - iter.next(); - pos += 3; - } else { - iter.next(); - iter.next(); - iter.next(); - pos += 4; - } - } - } - - #[inline] - fn final_lead_surrogate(&self) -> Option { - match self.bytes { - [.., 0xED, b2 @ 0xA0..=0xAF, b3] => Some(decode_surrogate(b2, b3)), - _ => None, - } - } - - #[inline] - fn initial_trail_surrogate(&self) -> Option { - match self.bytes { - [0xED, b2 @ 0xB0..=0xBF, b3, ..] => Some(decode_surrogate(b2, b3)), - _ => None, - } - } - - pub fn clone_into(&self, buf: &mut Wtf8Buf) { - buf.is_known_utf8 = false; - self.bytes.clone_into(&mut buf.bytes); - } - - /// Boxes this `Wtf8`. - #[inline] - pub fn into_box(&self) -> Box { - let boxed: Box<[u8]> = self.bytes.into(); - unsafe { mem::transmute(boxed) } - } - - /// Creates a boxed, empty `Wtf8`. - pub fn empty_box() -> Box { - let boxed: Box<[u8]> = Default::default(); - unsafe { mem::transmute(boxed) } - } - - #[inline] - pub fn into_arc(&self) -> Arc { - let arc: Arc<[u8]> = Arc::from(&self.bytes); - unsafe { Arc::from_raw(Arc::into_raw(arc) as *const Wtf8) } - } - - #[inline] - pub fn into_rc(&self) -> Rc { - let rc: Rc<[u8]> = Rc::from(&self.bytes); - unsafe { Rc::from_raw(Rc::into_raw(rc) as *const Wtf8) } - } - - #[inline] - pub fn make_ascii_lowercase(&mut self) { - self.bytes.make_ascii_lowercase() - } - - #[inline] - pub fn make_ascii_uppercase(&mut self) { - self.bytes.make_ascii_uppercase() - } - - #[inline] - pub fn to_ascii_lowercase(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_ascii_lowercase(), is_known_utf8: false } - } - - #[inline] - pub fn to_ascii_uppercase(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_ascii_uppercase(), is_known_utf8: false } - } - - #[inline] - pub fn is_ascii(&self) -> bool { - self.bytes.is_ascii() - } - - #[inline] - pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { - self.bytes.eq_ignore_ascii_case(&other.bytes) - } -} - -/// Returns a slice of the given string for the byte range \[`begin`..`end`). -/// -/// # Panics -/// -/// Panics when `begin` and `end` do not point to code point boundaries, -/// or point beyond the end of the string. -impl ops::Index> for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, range: ops::Range) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if range.start <= range.end - && is_code_point_boundary(self, range.start) - && is_code_point_boundary(self, range.end) - { - unsafe { slice_unchecked(self, range.start, range.end) } - } else { - slice_error_fail(self, range.start, range.end) - } - } -} - -/// Returns a slice of the given string from byte `begin` to its end. -/// -/// # Panics -/// -/// Panics when `begin` is not at a code point boundary, -/// or is beyond the end of the string. -impl ops::Index> for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, range: ops::RangeFrom) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if is_code_point_boundary(self, range.start) { - unsafe { slice_unchecked(self, range.start, self.len()) } - } else { - slice_error_fail(self, range.start, self.len()) - } - } -} - -/// Returns a slice of the given string from its beginning to byte `end`. -/// -/// # Panics -/// -/// Panics when `end` is not at a code point boundary, -/// or is beyond the end of the string. -impl ops::Index> for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, range: ops::RangeTo) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if is_code_point_boundary(self, range.end) { - unsafe { slice_unchecked(self, 0, range.end) } - } else { - slice_error_fail(self, 0, range.end) - } - } -} - -impl ops::Index for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, _range: ops::RangeFull) -> &Wtf8 { - self - } -} - -#[inline] -fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { - // The first byte is assumed to be 0xED - 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F -} - -#[inline] -fn decode_surrogate_pair(lead: u16, trail: u16) -> char { - let code_point = 0x10000 + ((((lead - 0xD800) as u32) << 10) | (trail - 0xDC00) as u32); - unsafe { char::from_u32_unchecked(code_point) } -} - -/// Copied from str::is_char_boundary -#[inline] -pub fn is_code_point_boundary(slice: &Wtf8, index: usize) -> bool { - if index == 0 { - return true; - } - match slice.bytes.get(index) { - None => index == slice.len(), - Some(&b) => (b as i8) >= -0x40, - } -} - -/// Verify that `index` is at the edge of either a valid UTF-8 codepoint -/// (i.e. a codepoint that's not a surrogate) or of the whole string. -/// -/// These are the cases currently permitted by `OsStr::slice_encoded_bytes`. -/// Splitting between surrogates is valid as far as WTF-8 is concerned, but -/// we do not permit it in the public API because WTF-8 is considered an -/// implementation detail. -#[track_caller] -#[inline] -pub fn check_utf8_boundary(slice: &Wtf8, index: usize) { - if index == 0 { - return; - } - match slice.bytes.get(index) { - Some(0xED) => (), // Might be a surrogate - Some(&b) if (b as i8) >= -0x40 => return, - Some(_) => panic!("byte index {index} is not a codepoint boundary"), - None if index == slice.len() => return, - None => panic!("byte index {index} is out of bounds"), - } - if slice.bytes[index + 1] >= 0xA0 { - // There's a surrogate after index. Now check before index. - if index >= 3 && slice.bytes[index - 3] == 0xED && slice.bytes[index - 2] >= 0xA0 { - panic!("byte index {index} lies between surrogate codepoints"); - } - } -} - -/// Copied from core::str::raw::slice_unchecked -#[inline] -pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { - // SAFETY: memory layout of a &[u8] and &Wtf8 are the same - unsafe { - let len = end - begin; - let start = s.as_bytes().as_ptr().add(begin); - Wtf8::from_bytes_unchecked(slice::from_raw_parts(start, len)) - } -} - -/// Copied from core::str::raw::slice_error_fail -#[inline(never)] -pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { - assert!(begin <= end); - panic!("index {begin} and/or {end} in `{s:?}` do not lie on character boundary"); -} - -/// Iterator for the code points of a WTF-8 string. -/// -/// Created with the method `.code_points()`. -#[derive(Clone)] -pub struct Wtf8CodePoints<'a> { - bytes: slice::Iter<'a, u8>, -} - -impl Iterator for Wtf8CodePoints<'_> { - type Item = CodePoint; - - #[inline] - fn next(&mut self) -> Option { - // SAFETY: `self.bytes` has been created from a WTF-8 string - unsafe { next_code_point(&mut self.bytes).map(|c| CodePoint { value: c }) } - } - - #[inline] - fn size_hint(&self) -> (usize, Option) { - let len = self.bytes.len(); - (len.saturating_add(3) / 4, Some(len)) - } -} - -/// Generates a wide character sequence for potentially ill-formed UTF-16. -#[stable(feature = "rust1", since = "1.0.0")] -#[derive(Clone)] -pub struct EncodeWide<'a> { - code_points: Wtf8CodePoints<'a>, - extra: u16, -} - -// Copied from libunicode/u_str.rs -#[stable(feature = "rust1", since = "1.0.0")] -impl Iterator for EncodeWide<'_> { - type Item = u16; - - #[inline] - fn next(&mut self) -> Option { - if self.extra != 0 { - let tmp = self.extra; - self.extra = 0; - return Some(tmp); - } - - let mut buf = [0; MAX_LEN_UTF16]; - self.code_points.next().map(|code_point| { - let n = encode_utf16_raw(code_point.value, &mut buf).len(); - if n == 2 { - self.extra = buf[1]; - } - buf[0] - }) - } - - #[inline] - fn size_hint(&self) -> (usize, Option) { - let (low, high) = self.code_points.size_hint(); - let ext = (self.extra != 0) as usize; - // every code point gets either one u16 or two u16, - // so this iterator is between 1 or 2 times as - // long as the underlying iterator. - (low + ext, high.and_then(|n| n.checked_mul(2)).and_then(|n| n.checked_add(ext))) - } -} - -#[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")] -impl FusedIterator for EncodeWide<'_> {} - -impl Hash for CodePoint { - #[inline] - fn hash(&self, state: &mut H) { - self.value.hash(state) - } -} - -impl Hash for Wtf8Buf { - #[inline] - fn hash(&self, state: &mut H) { - state.write(&self.bytes); - 0xfeu8.hash(state) - } -} - -impl Hash for Wtf8 { - #[inline] - fn hash(&self, state: &mut H) { - state.write(&self.bytes); - 0xfeu8.hash(state) - } -} - -#[unstable(feature = "clone_to_uninit", issue = "126799")] -unsafe impl CloneToUninit for Wtf8 { - #[inline] - #[cfg_attr(debug_assertions, track_caller)] - unsafe fn clone_to_uninit(&self, dst: *mut u8) { - // SAFETY: we're just a transparent wrapper around [u8] - unsafe { self.bytes.clone_to_uninit(dst) } - } -} diff --git a/library/std/src/sys_common/wtf8/tests.rs b/library/std/src/sys_common/wtf8/tests.rs deleted file mode 100644 index b57c99a8452..00000000000 --- a/library/std/src/sys_common/wtf8/tests.rs +++ /dev/null @@ -1,751 +0,0 @@ -use super::*; - -#[test] -fn code_point_from_u32() { - assert!(CodePoint::from_u32(0).is_some()); - assert!(CodePoint::from_u32(0xD800).is_some()); - assert!(CodePoint::from_u32(0x10FFFF).is_some()); - assert!(CodePoint::from_u32(0x110000).is_none()); -} - -#[test] -fn code_point_to_u32() { - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - assert_eq!(c(0).to_u32(), 0); - assert_eq!(c(0xD800).to_u32(), 0xD800); - assert_eq!(c(0x10FFFF).to_u32(), 0x10FFFF); -} - -#[test] -fn code_point_to_lead_surrogate() { - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - assert_eq!(c(0).to_lead_surrogate(), None); - assert_eq!(c(0xE9).to_lead_surrogate(), None); - assert_eq!(c(0xD800).to_lead_surrogate(), Some(0xD800)); - assert_eq!(c(0xDBFF).to_lead_surrogate(), Some(0xDBFF)); - assert_eq!(c(0xDC00).to_lead_surrogate(), None); - assert_eq!(c(0xDFFF).to_lead_surrogate(), None); - assert_eq!(c(0x1F4A9).to_lead_surrogate(), None); - assert_eq!(c(0x10FFFF).to_lead_surrogate(), None); -} - -#[test] -fn code_point_to_trail_surrogate() { - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - assert_eq!(c(0).to_trail_surrogate(), None); - assert_eq!(c(0xE9).to_trail_surrogate(), None); - assert_eq!(c(0xD800).to_trail_surrogate(), None); - assert_eq!(c(0xDBFF).to_trail_surrogate(), None); - assert_eq!(c(0xDC00).to_trail_surrogate(), Some(0xDC00)); - assert_eq!(c(0xDFFF).to_trail_surrogate(), Some(0xDFFF)); - assert_eq!(c(0x1F4A9).to_trail_surrogate(), None); - assert_eq!(c(0x10FFFF).to_trail_surrogate(), None); -} - -#[test] -fn code_point_from_char() { - assert_eq!(CodePoint::from_char('a').to_u32(), 0x61); - assert_eq!(CodePoint::from_char('💩').to_u32(), 0x1F4A9); -} - -#[test] -fn code_point_to_string() { - assert_eq!(format!("{:?}", CodePoint::from_char('a')), "U+0061"); - assert_eq!(format!("{:?}", CodePoint::from_char('💩')), "U+1F4A9"); -} - -#[test] -fn code_point_to_char() { - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - assert_eq!(c(0x61).to_char(), Some('a')); - assert_eq!(c(0x1F4A9).to_char(), Some('💩')); - assert_eq!(c(0xD800).to_char(), None); -} - -#[test] -fn code_point_to_char_lossy() { - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - assert_eq!(c(0x61).to_char_lossy(), 'a'); - assert_eq!(c(0x1F4A9).to_char_lossy(), '💩'); - assert_eq!(c(0xD800).to_char_lossy(), '\u{FFFD}'); -} - -#[test] -fn wtf8buf_new() { - assert_eq!(Wtf8Buf::new().bytes, b""); -} - -#[test] -fn wtf8buf_from_str() { - assert_eq!(Wtf8Buf::from_str("").bytes, b""); - assert_eq!(Wtf8Buf::from_str("aé 💩").bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); -} - -#[test] -fn wtf8buf_from_string() { - assert_eq!(Wtf8Buf::from_string(String::from("")).bytes, b""); - assert_eq!(Wtf8Buf::from_string(String::from("aé 💩")).bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); -} - -#[test] -fn wtf8buf_from_wide() { - let buf = Wtf8Buf::from_wide(&[]); - assert_eq!(buf.bytes, b""); - assert!(buf.is_known_utf8); - - let buf = Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xDCA9]); - assert_eq!(buf.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); - assert!(buf.is_known_utf8); - - let buf = Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]); - assert_eq!(buf.bytes, b"a\xC3\xA9 \xED\xA0\xBD\xF0\x9F\x92\xA9"); - assert!(!buf.is_known_utf8); - - let buf = Wtf8Buf::from_wide(&[0xD800]); - assert_eq!(buf.bytes, b"\xED\xA0\x80"); - assert!(!buf.is_known_utf8); - - let buf = Wtf8Buf::from_wide(&[0xDBFF]); - assert_eq!(buf.bytes, b"\xED\xAF\xBF"); - assert!(!buf.is_known_utf8); - - let buf = Wtf8Buf::from_wide(&[0xDC00]); - assert_eq!(buf.bytes, b"\xED\xB0\x80"); - assert!(!buf.is_known_utf8); - - let buf = Wtf8Buf::from_wide(&[0xDFFF]); - assert_eq!(buf.bytes, b"\xED\xBF\xBF"); - assert!(!buf.is_known_utf8); -} - -#[test] -fn wtf8buf_push_str() { - let mut string = Wtf8Buf::new(); - assert_eq!(string.bytes, b""); - assert!(string.is_known_utf8); - - string.push_str("aé 💩"); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); - assert!(string.is_known_utf8); -} - -#[test] -fn wtf8buf_push_char() { - let mut string = Wtf8Buf::from_str("aé "); - assert_eq!(string.bytes, b"a\xC3\xA9 "); - assert!(string.is_known_utf8); - - string.push_char('💩'); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); - assert!(string.is_known_utf8); -} - -#[test] -fn wtf8buf_push() { - let mut string = Wtf8Buf::from_str("aé "); - assert_eq!(string.bytes, b"a\xC3\xA9 "); - assert!(string.is_known_utf8); - - string.push(CodePoint::from_char('💩')); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); - assert!(string.is_known_utf8); - - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - - let mut string = Wtf8Buf::new(); - string.push(c(0xD83D)); // lead - assert!(!string.is_known_utf8); - string.push(c(0xDCA9)); // trail - assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! - - let mut string = Wtf8Buf::new(); - string.push(c(0xD83D)); // lead - assert!(!string.is_known_utf8); - string.push(c(0x20)); // not surrogate - string.push(c(0xDCA9)); // trail - assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); - - let mut string = Wtf8Buf::new(); - string.push(c(0xD800)); // lead - assert!(!string.is_known_utf8); - string.push(c(0xDBFF)); // lead - assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); - - let mut string = Wtf8Buf::new(); - string.push(c(0xD800)); // lead - assert!(!string.is_known_utf8); - string.push(c(0xE000)); // not surrogate - assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); - - let mut string = Wtf8Buf::new(); - string.push(c(0xD7FF)); // not surrogate - assert!(string.is_known_utf8); - string.push(c(0xDC00)); // trail - assert!(!string.is_known_utf8); - assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); - - let mut string = Wtf8Buf::new(); - string.push(c(0x61)); // not surrogate, < 3 bytes - assert!(string.is_known_utf8); - string.push(c(0xDC00)); // trail - assert!(!string.is_known_utf8); - assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); - - let mut string = Wtf8Buf::new(); - string.push(c(0xDC00)); // trail - assert!(!string.is_known_utf8); - assert_eq!(string.bytes, b"\xED\xB0\x80"); -} - -#[test] -fn wtf8buf_push_wtf8() { - let mut string = Wtf8Buf::from_str("aé"); - assert_eq!(string.bytes, b"a\xC3\xA9"); - string.push_wtf8(Wtf8::from_str(" 💩")); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); - assert!(string.is_known_utf8); - - fn w(v: &[u8]) -> &Wtf8 { - unsafe { Wtf8::from_bytes_unchecked(v) } - } - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"\xED\xA0\xBD")); // lead - string.push_wtf8(w(b"\xED\xB2\xA9")); // trail - assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"\xED\xA0\xBD")); // lead - string.push_wtf8(w(b" ")); // not surrogate - string.push_wtf8(w(b"\xED\xB2\xA9")); // trail - assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); - assert!(!string.is_known_utf8); - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"\xED\xA0\x80")); // lead - string.push_wtf8(w(b"\xED\xAF\xBF")); // lead - assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); - assert!(!string.is_known_utf8); - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"\xED\xA0\x80")); // lead - string.push_wtf8(w(b"\xEE\x80\x80")); // not surrogate - assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); - assert!(!string.is_known_utf8); - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"\xED\x9F\xBF")); // not surrogate - string.push_wtf8(w(b"\xED\xB0\x80")); // trail - assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); - assert!(!string.is_known_utf8); - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"a")); // not surrogate, < 3 bytes - string.push_wtf8(w(b"\xED\xB0\x80")); // trail - assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); - assert!(!string.is_known_utf8); - - let mut string = Wtf8Buf::new(); - string.push_wtf8(w(b"\xED\xB0\x80")); // trail - assert_eq!(string.bytes, b"\xED\xB0\x80"); - assert!(!string.is_known_utf8); -} - -#[test] -fn wtf8buf_truncate() { - let mut string = Wtf8Buf::from_str("aé"); - assert!(string.is_known_utf8); - - string.truncate(3); - assert_eq!(string.bytes, b"a\xC3\xA9"); - assert!(string.is_known_utf8); - - string.truncate(1); - assert_eq!(string.bytes, b"a"); - assert!(string.is_known_utf8); - - string.truncate(0); - assert_eq!(string.bytes, b""); - assert!(string.is_known_utf8); -} - -#[test] -fn wtf8buf_truncate_around_non_bmp() { - let mut string = Wtf8Buf::from_str("💩"); - assert!(string.is_known_utf8); - - string.truncate(4); - assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); - assert!(string.is_known_utf8); - - string.truncate(0); - assert_eq!(string.bytes, b""); - assert!(string.is_known_utf8); -} - -#[test] -#[should_panic] -fn wtf8buf_truncate_fail_code_point_boundary() { - let mut string = Wtf8Buf::from_str("aé"); - string.truncate(2); -} - -#[test] -#[should_panic] -fn wtf8buf_truncate_fail_longer() { - let mut string = Wtf8Buf::from_str("aé"); - string.truncate(4); -} - -#[test] -#[should_panic] -fn wtf8buf_truncate_splitting_non_bmp3() { - let mut string = Wtf8Buf::from_str("💩"); - assert!(string.is_known_utf8); - string.truncate(3); -} - -#[test] -#[should_panic] -fn wtf8buf_truncate_splitting_non_bmp2() { - let mut string = Wtf8Buf::from_str("💩"); - assert!(string.is_known_utf8); - string.truncate(2); -} - -#[test] -#[should_panic] -fn wtf8buf_truncate_splitting_non_bmp1() { - let mut string = Wtf8Buf::from_str("💩"); - assert!(string.is_known_utf8); - string.truncate(1); -} - -#[test] -fn wtf8buf_into_string() { - let mut string = Wtf8Buf::from_str("aé 💩"); - assert!(string.is_known_utf8); - assert_eq!(string.clone().into_string(), Ok(String::from("aé 💩"))); - string.push(CodePoint::from_u32(0xD800).unwrap()); - assert!(!string.is_known_utf8); - assert_eq!(string.clone().into_string(), Err(string)); -} - -#[test] -fn wtf8buf_into_string_lossy() { - let mut string = Wtf8Buf::from_str("aé 💩"); - assert_eq!(string.clone().into_string_lossy(), String::from("aé 💩")); - string.push(CodePoint::from_u32(0xD800).unwrap()); - assert_eq!(string.clone().into_string_lossy(), String::from("aé 💩�")); -} - -#[test] -fn wtf8buf_from_iterator() { - fn f(values: &[u32]) -> Wtf8Buf { - values.iter().map(|&c| CodePoint::from_u32(c).unwrap()).collect::() - } - assert_eq!( - f(&[0x61, 0xE9, 0x20, 0x1F4A9]), - Wtf8Buf { bytes: b"a\xC3\xA9 \xF0\x9F\x92\xA9".to_vec(), is_known_utf8: true } - ); - - assert_eq!(f(&[0xD83D, 0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! - assert_eq!( - f(&[0xD83D, 0x20, 0xDCA9]), - Wtf8Buf { bytes: b"\xED\xA0\xBD \xED\xB2\xA9".to_vec(), is_known_utf8: false } - ); - assert_eq!( - f(&[0xD800, 0xDBFF]), - Wtf8Buf { bytes: b"\xED\xA0\x80\xED\xAF\xBF".to_vec(), is_known_utf8: false } - ); - assert_eq!( - f(&[0xD800, 0xE000]), - Wtf8Buf { bytes: b"\xED\xA0\x80\xEE\x80\x80".to_vec(), is_known_utf8: false } - ); - assert_eq!( - f(&[0xD7FF, 0xDC00]), - Wtf8Buf { bytes: b"\xED\x9F\xBF\xED\xB0\x80".to_vec(), is_known_utf8: false } - ); - assert_eq!( - f(&[0x61, 0xDC00]), - Wtf8Buf { bytes: b"\x61\xED\xB0\x80".to_vec(), is_known_utf8: false } - ); - assert_eq!(f(&[0xDC00]), Wtf8Buf { bytes: b"\xED\xB0\x80".to_vec(), is_known_utf8: false }); -} - -#[test] -fn wtf8buf_extend() { - fn e(initial: &[u32], extended: &[u32]) -> Wtf8Buf { - fn c(value: &u32) -> CodePoint { - CodePoint::from_u32(*value).unwrap() - } - let mut string = initial.iter().map(c).collect::(); - string.extend(extended.iter().map(c)); - string - } - - assert_eq!( - e(&[0x61, 0xE9], &[0x20, 0x1F4A9]), - Wtf8Buf { bytes: b"a\xC3\xA9 \xF0\x9F\x92\xA9".to_vec(), is_known_utf8: true } - ); - - assert_eq!(e(&[0xD83D], &[0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! - assert_eq!( - e(&[0xD83D, 0x20], &[0xDCA9]), - Wtf8Buf { bytes: b"\xED\xA0\xBD \xED\xB2\xA9".to_vec(), is_known_utf8: false } - ); - assert_eq!( - e(&[0xD800], &[0xDBFF]), - Wtf8Buf { bytes: b"\xED\xA0\x80\xED\xAF\xBF".to_vec(), is_known_utf8: false } - ); - assert_eq!( - e(&[0xD800], &[0xE000]), - Wtf8Buf { bytes: b"\xED\xA0\x80\xEE\x80\x80".to_vec(), is_known_utf8: false } - ); - assert_eq!( - e(&[0xD7FF], &[0xDC00]), - Wtf8Buf { bytes: b"\xED\x9F\xBF\xED\xB0\x80".to_vec(), is_known_utf8: false } - ); - assert_eq!( - e(&[0x61], &[0xDC00]), - Wtf8Buf { bytes: b"\x61\xED\xB0\x80".to_vec(), is_known_utf8: false } - ); - assert_eq!( - e(&[], &[0xDC00]), - Wtf8Buf { bytes: b"\xED\xB0\x80".to_vec(), is_known_utf8: false } - ); -} - -#[test] -fn wtf8buf_show() { - let mut string = Wtf8Buf::from_str("a\té \u{7f}💩\r"); - string.push(CodePoint::from_u32(0xD800).unwrap()); - assert_eq!(format!("{string:?}"), "\"a\\té \\u{7f}\u{1f4a9}\\r\\u{d800}\""); -} - -#[test] -fn wtf8buf_as_slice() { - assert_eq!(Wtf8Buf::from_str("aé").as_slice(), Wtf8::from_str("aé")); -} - -#[test] -fn wtf8buf_show_str() { - let text = "a\té 💩\r"; - let string = Wtf8Buf::from_str(text); - assert_eq!(format!("{text:?}"), format!("{string:?}")); -} - -#[test] -fn wtf8_from_str() { - assert_eq!(&Wtf8::from_str("").bytes, b""); - assert_eq!(&Wtf8::from_str("aé 💩").bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); -} - -#[test] -fn wtf8_len() { - assert_eq!(Wtf8::from_str("").len(), 0); - assert_eq!(Wtf8::from_str("aé 💩").len(), 8); -} - -#[test] -fn wtf8_slice() { - assert_eq!(&Wtf8::from_str("aé 💩")[1..4].bytes, b"\xC3\xA9 "); -} - -#[test] -#[should_panic] -fn wtf8_slice_not_code_point_boundary() { - let _ = &Wtf8::from_str("aé 💩")[2..4]; -} - -#[test] -fn wtf8_slice_from() { - assert_eq!(&Wtf8::from_str("aé 💩")[1..].bytes, b"\xC3\xA9 \xF0\x9F\x92\xA9"); -} - -#[test] -#[should_panic] -fn wtf8_slice_from_not_code_point_boundary() { - let _ = &Wtf8::from_str("aé 💩")[2..]; -} - -#[test] -fn wtf8_slice_to() { - assert_eq!(&Wtf8::from_str("aé 💩")[..4].bytes, b"a\xC3\xA9 "); -} - -#[test] -#[should_panic] -fn wtf8_slice_to_not_code_point_boundary() { - let _ = &Wtf8::from_str("aé 💩")[5..]; -} - -#[test] -fn wtf8_ascii_byte_at() { - let slice = Wtf8::from_str("aé 💩"); - assert_eq!(slice.ascii_byte_at(0), b'a'); - assert_eq!(slice.ascii_byte_at(1), b'\xFF'); - assert_eq!(slice.ascii_byte_at(2), b'\xFF'); - assert_eq!(slice.ascii_byte_at(3), b' '); - assert_eq!(slice.ascii_byte_at(4), b'\xFF'); -} - -#[test] -fn wtf8_code_points() { - fn c(value: u32) -> CodePoint { - CodePoint::from_u32(value).unwrap() - } - fn cp(string: &Wtf8Buf) -> Vec> { - string.code_points().map(|c| c.to_char()).collect::>() - } - let mut string = Wtf8Buf::from_str("é "); - assert_eq!(cp(&string), [Some('é'), Some(' ')]); - string.push(c(0xD83D)); - assert_eq!(cp(&string), [Some('é'), Some(' '), None]); - string.push(c(0xDCA9)); - assert_eq!(cp(&string), [Some('é'), Some(' '), Some('💩')]); -} - -#[test] -fn wtf8_as_str() { - assert_eq!(Wtf8::from_str("").as_str(), Ok("")); - assert_eq!(Wtf8::from_str("aé 💩").as_str(), Ok("aé 💩")); - let mut string = Wtf8Buf::new(); - string.push(CodePoint::from_u32(0xD800).unwrap()); - assert!(string.as_str().is_err()); -} - -#[test] -fn wtf8_to_string_lossy() { - assert_eq!(Wtf8::from_str("").to_string_lossy(), Cow::Borrowed("")); - assert_eq!(Wtf8::from_str("aé 💩").to_string_lossy(), Cow::Borrowed("aé 💩")); - let mut string = Wtf8Buf::from_str("aé 💩"); - string.push(CodePoint::from_u32(0xD800).unwrap()); - let expected: Cow<'_, str> = Cow::Owned(String::from("aé 💩�")); - assert_eq!(string.to_string_lossy(), expected); -} - -#[test] -fn wtf8_display() { - fn d(b: &[u8]) -> String { - (&unsafe { Wtf8::from_bytes_unchecked(b) }).to_string() - } - - assert_eq!("", d("".as_bytes())); - assert_eq!("aé 💩", d("aé 💩".as_bytes())); - - let mut string = Wtf8Buf::from_str("aé 💩"); - string.push(CodePoint::from_u32(0xD800).unwrap()); - assert_eq!("aé 💩�", d(string.as_inner())); -} - -#[test] -fn wtf8_encode_wide() { - let mut string = Wtf8Buf::from_str("aé "); - string.push(CodePoint::from_u32(0xD83D).unwrap()); - string.push_char('💩'); - assert_eq!( - string.encode_wide().collect::>(), - vec![0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9] - ); -} - -#[test] -fn wtf8_encode_wide_size_hint() { - let string = Wtf8Buf::from_str("\u{12345}"); - let mut iter = string.encode_wide(); - assert_eq!((1, Some(8)), iter.size_hint()); - iter.next().unwrap(); - assert_eq!((1, Some(1)), iter.size_hint()); - iter.next().unwrap(); - assert_eq!((0, Some(0)), iter.size_hint()); - assert!(iter.next().is_none()); -} - -#[test] -fn wtf8_clone_into() { - let mut string = Wtf8Buf::new(); - Wtf8::from_str("green").clone_into(&mut string); - assert_eq!(string.bytes, b"green"); - - let mut string = Wtf8Buf::from_str("green"); - Wtf8::from_str("").clone_into(&mut string); - assert_eq!(string.bytes, b""); - - let mut string = Wtf8Buf::from_str("red"); - Wtf8::from_str("green").clone_into(&mut string); - assert_eq!(string.bytes, b"green"); - - let mut string = Wtf8Buf::from_str("green"); - Wtf8::from_str("red").clone_into(&mut string); - assert_eq!(string.bytes, b"red"); - - let mut string = Wtf8Buf::from_str("green"); - assert!(string.is_known_utf8); - unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").clone_into(&mut string) }; - assert_eq!(string.bytes, b"\xED\xA0\x80"); - assert!(!string.is_known_utf8); -} - -#[test] -fn wtf8_to_ascii_lowercase() { - let lowercase = Wtf8::from_str("").to_ascii_lowercase(); - assert_eq!(lowercase.bytes, b""); - - let lowercase = Wtf8::from_str("GrEeN gRaPeS! 🍇").to_ascii_lowercase(); - assert_eq!(lowercase.bytes, b"green grapes! \xf0\x9f\x8d\x87"); - - let lowercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_ascii_lowercase() }; - assert_eq!(lowercase.bytes, b"\xED\xA0\x80"); - assert!(!lowercase.is_known_utf8); -} - -#[test] -fn wtf8_to_ascii_uppercase() { - let uppercase = Wtf8::from_str("").to_ascii_uppercase(); - assert_eq!(uppercase.bytes, b""); - - let uppercase = Wtf8::from_str("GrEeN gRaPeS! 🍇").to_ascii_uppercase(); - assert_eq!(uppercase.bytes, b"GREEN GRAPES! \xf0\x9f\x8d\x87"); - - let uppercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_ascii_uppercase() }; - assert_eq!(uppercase.bytes, b"\xED\xA0\x80"); - assert!(!uppercase.is_known_utf8); -} - -#[test] -fn wtf8_make_ascii_lowercase() { - let mut lowercase = Wtf8Buf::from_str(""); - lowercase.make_ascii_lowercase(); - assert_eq!(lowercase.bytes, b""); - - let mut lowercase = Wtf8Buf::from_str("GrEeN gRaPeS! 🍇"); - lowercase.make_ascii_lowercase(); - assert_eq!(lowercase.bytes, b"green grapes! \xf0\x9f\x8d\x87"); - - let mut lowercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; - lowercase.make_ascii_lowercase(); - assert_eq!(lowercase.bytes, b"\xED\xA0\x80"); - assert!(!lowercase.is_known_utf8); -} - -#[test] -fn wtf8_make_ascii_uppercase() { - let mut uppercase = Wtf8Buf::from_str(""); - uppercase.make_ascii_uppercase(); - assert_eq!(uppercase.bytes, b""); - - let mut uppercase = Wtf8Buf::from_str("GrEeN gRaPeS! 🍇"); - uppercase.make_ascii_uppercase(); - assert_eq!(uppercase.bytes, b"GREEN GRAPES! \xf0\x9f\x8d\x87"); - - let mut uppercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; - uppercase.make_ascii_uppercase(); - assert_eq!(uppercase.bytes, b"\xED\xA0\x80"); - assert!(!uppercase.is_known_utf8); -} - -#[test] -fn wtf8_to_owned() { - let string = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; - assert_eq!(string.bytes, b"\xED\xA0\x80"); - assert!(!string.is_known_utf8); -} - -#[test] -fn wtf8_valid_utf8_boundaries() { - let mut string = Wtf8Buf::from_str("aé 💩"); - string.push(CodePoint::from_u32(0xD800).unwrap()); - string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 0); - check_utf8_boundary(&string, 1); - check_utf8_boundary(&string, 3); - check_utf8_boundary(&string, 4); - check_utf8_boundary(&string, 8); - check_utf8_boundary(&string, 14); - assert_eq!(string.len(), 14); - - string.push_char('a'); - check_utf8_boundary(&string, 14); - check_utf8_boundary(&string, 15); - - let mut string = Wtf8Buf::from_str("a"); - string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 1); - - let mut string = Wtf8Buf::from_str("\u{D7FF}"); - string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 3); - - let mut string = Wtf8Buf::new(); - string.push(CodePoint::from_u32(0xD800).unwrap()); - string.push_char('\u{D7FF}'); - check_utf8_boundary(&string, 3); -} - -#[test] -#[should_panic(expected = "byte index 4 is out of bounds")] -fn wtf8_utf8_boundary_out_of_bounds() { - let string = Wtf8::from_str("aé"); - check_utf8_boundary(&string, 4); -} - -#[test] -#[should_panic(expected = "byte index 1 is not a codepoint boundary")] -fn wtf8_utf8_boundary_inside_codepoint() { - let string = Wtf8::from_str("é"); - check_utf8_boundary(&string, 1); -} - -#[test] -#[should_panic(expected = "byte index 1 is not a codepoint boundary")] -fn wtf8_utf8_boundary_inside_surrogate() { - let mut string = Wtf8Buf::new(); - string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 1); -} - -#[test] -#[should_panic(expected = "byte index 3 lies between surrogate codepoints")] -fn wtf8_utf8_boundary_between_surrogates() { - let mut string = Wtf8Buf::new(); - string.push(CodePoint::from_u32(0xD800).unwrap()); - string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 3); -} - -#[test] -fn wobbled_wtf8_plus_bytes_isnt_utf8() { - let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; - assert!(!string.is_known_utf8); - string.extend_from_slice(b"some utf-8"); - assert!(!string.is_known_utf8); -} - -#[test] -fn wobbled_wtf8_plus_str_isnt_utf8() { - let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; - assert!(!string.is_known_utf8); - string.push_str("some utf-8"); - assert!(!string.is_known_utf8); -} - -#[test] -fn unwobbly_wtf8_plus_utf8_is_utf8() { - let mut string: Wtf8Buf = Wtf8Buf::from_str("hello world"); - assert!(string.is_known_utf8); - string.push_str("some utf-8"); - assert!(string.is_known_utf8); -} -- cgit 1.4.1-3-g733a5 From 2914291e09cb13aab64207f9e11f2aaf74de3904 Mon Sep 17 00:00:00 2001 From: ltdk Date: Wed, 13 Aug 2025 01:16:42 -0400 Subject: Move WTF-8 code from std to core/alloc --- library/alloc/src/lib.rs | 3 + library/alloc/src/wtf8/mod.rs | 674 +++++------------------------------- library/alloc/src/wtf8/tests.rs | 209 +++++------ library/alloctests/lib.rs | 12 +- library/alloctests/tests/lib.rs | 1 + library/core/src/lib.rs | 2 + library/core/src/num/niche_types.rs | 15 + library/core/src/wtf8.rs | 642 +++++----------------------------- library/coretests/tests/lib.rs | 1 + library/coretests/tests/wtf8.rs | 1 + library/std/src/lib.rs | 1 + library/std/src/os/windows/ffi.rs | 38 +- library/std/src/sys/os_str/wtf8.rs | 8 +- library/std/src/sys_common/mod.rs | 1 - src/tools/tidy/src/unit_tests.rs | 1 + 15 files changed, 331 insertions(+), 1278 deletions(-) create mode 100644 library/coretests/tests/wtf8.rs diff --git a/library/alloc/src/lib.rs b/library/alloc/src/lib.rs index 639c5d4c930..711092ae8eb 100644 --- a/library/alloc/src/lib.rs +++ b/library/alloc/src/lib.rs @@ -159,6 +159,7 @@ #![feature(unicode_internals)] #![feature(unsize)] #![feature(unwrap_infallible)] +#![feature(wtf8_internals)] // tidy-alphabetical-end // // Language features: @@ -232,6 +233,8 @@ pub mod sync; #[cfg(all(not(no_global_oom_handling), not(no_rc), not(no_sync)))] pub mod task; pub mod vec; +#[cfg(all(not(no_rc), not(no_sync), not(no_global_oom_handling)))] +pub mod wtf8; #[doc(hidden)] #[unstable(feature = "liballoc_internals", issue = "none", reason = "implementation detail")] diff --git a/library/alloc/src/wtf8/mod.rs b/library/alloc/src/wtf8/mod.rs index 50bde88b5a4..95d317a5efb 100644 --- a/library/alloc/src/wtf8/mod.rs +++ b/library/alloc/src/wtf8/mod.rs @@ -1,135 +1,42 @@ -//! Implementation of [the WTF-8 encoding](https://simonsapin.github.io/wtf-8/). -//! -//! This library uses Rust’s type system to maintain -//! [well-formedness](https://simonsapin.github.io/wtf-8/#well-formed), -//! like the `String` and `&str` types do for UTF-8. -//! -//! Since [WTF-8 must not be used -//! for interchange](https://simonsapin.github.io/wtf-8/#intended-audience), -//! this library deliberately does not provide access to the underlying bytes -//! of WTF-8 strings, -//! nor can it decode WTF-8 from arbitrary bytes. -//! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points. - -// this module is imported from @SimonSapin's repo and has tons of dead code on -// unix (it's mostly used on windows), so don't worry about dead code here. -#![allow(dead_code)] +//! Heap-allocated counterpart to core `wtf8` module. +#![unstable( + feature = "wtf8_internals", + issue = "none", + reason = "this is internal code for representing OsStr on some platforms and not a public API" +)] +// rustdoc bug: doc(hidden) on the module won't stop types in the module from showing up in trait +// implementations, so, we'll have to add more doc(hidden)s anyway +#![doc(hidden)] + +// Note: This module is also included in the alloctests crate using #[path] to +// run the tests. See the comment there for an explanation why this is the case. #[cfg(test)] mod tests; -use core::char::{MAX_LEN_UTF8, MAX_LEN_UTF16, encode_utf8_raw, encode_utf16_raw}; -use core::clone::CloneToUninit; -use core::str::next_code_point; +use core::char::{MAX_LEN_UTF8, encode_utf8_raw}; +use core::hash::{Hash, Hasher}; +pub use core::wtf8::{CodePoint, Wtf8}; +#[cfg(not(test))] +pub use core::wtf8::{EncodeWide, Wtf8CodePoints}; +use core::{fmt, mem, ops, str}; -use crate::borrow::Cow; +use crate::borrow::{Cow, ToOwned}; +use crate::boxed::Box; use crate::collections::TryReserveError; -use crate::hash::{Hash, Hasher}; -use crate::iter::FusedIterator; +#[cfg(not(test))] use crate::rc::Rc; +use crate::string::String; +#[cfg(all(not(test), target_has_atomic = "ptr"))] use crate::sync::Arc; -use crate::sys_common::AsInner; -use crate::{fmt, mem, ops, slice, str}; - -const UTF8_REPLACEMENT_CHARACTER: &str = "\u{FFFD}"; - -/// A Unicode code point: from U+0000 to U+10FFFF. -/// -/// Compares with the `char` type, -/// which represents a Unicode scalar value: -/// a code point that is not a surrogate (U+D800 to U+DFFF). -#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)] -pub struct CodePoint { - value: u32, -} - -/// Format the code point as `U+` followed by four to six hexadecimal digits. -/// Example: `U+1F4A9` -impl fmt::Debug for CodePoint { - #[inline] - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - write!(formatter, "U+{:04X}", self.value) - } -} - -impl CodePoint { - /// Unsafely creates a new `CodePoint` without checking the value. - /// - /// Only use when `value` is known to be less than or equal to 0x10FFFF. - #[inline] - pub unsafe fn from_u32_unchecked(value: u32) -> CodePoint { - CodePoint { value } - } - - /// Creates a new `CodePoint` if the value is a valid code point. - /// - /// Returns `None` if `value` is above 0x10FFFF. - #[inline] - pub fn from_u32(value: u32) -> Option { - match value { - 0..=0x10FFFF => Some(CodePoint { value }), - _ => None, - } - } - - /// Creates a new `CodePoint` from a `char`. - /// - /// Since all Unicode scalar values are code points, this always succeeds. - #[inline] - pub fn from_char(value: char) -> CodePoint { - CodePoint { value: value as u32 } - } - - /// Returns the numeric value of the code point. - #[inline] - pub fn to_u32(&self) -> u32 { - self.value - } - - /// Returns the numeric value of the code point if it is a leading surrogate. - #[inline] - pub fn to_lead_surrogate(&self) -> Option { - match self.value { - lead @ 0xD800..=0xDBFF => Some(lead as u16), - _ => None, - } - } - - /// Returns the numeric value of the code point if it is a trailing surrogate. - #[inline] - pub fn to_trail_surrogate(&self) -> Option { - match self.value { - trail @ 0xDC00..=0xDFFF => Some(trail as u16), - _ => None, - } - } - - /// Optionally returns a Unicode scalar value for the code point. - /// - /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF). - #[inline] - pub fn to_char(&self) -> Option { - match self.value { - 0xD800..=0xDFFF => None, - _ => Some(unsafe { char::from_u32_unchecked(self.value) }), - } - } - - /// Returns a Unicode scalar value for the code point. - /// - /// Returns `'\u{FFFD}'` (the replacement character “�”) - /// if the code point is a surrogate (from U+D800 to U+DFFF). - #[inline] - pub fn to_char_lossy(&self) -> char { - self.to_char().unwrap_or('\u{FFFD}') - } -} +use crate::vec::Vec; /// An owned, growable string of well-formed WTF-8 data. /// /// Similar to `String`, but can additionally contain surrogate code points /// if they’re not in a surrogate pair. #[derive(Eq, PartialEq, Ord, PartialOrd, Clone)] +#[doc(hidden)] pub struct Wtf8Buf { bytes: Vec, @@ -181,6 +88,7 @@ impl fmt::Display for Wtf8Buf { } } +#[cfg_attr(test, allow(dead_code))] impl Wtf8Buf { /// Creates a new, empty WTF-8 string. #[inline] @@ -245,7 +153,9 @@ impl Wtf8Buf { string.is_known_utf8 = false; // Skip the WTF-8 concatenation check, // surrogate pairs are already decoded by decode_utf16 - string.push_code_point_unchecked(code_point); + unsafe { + string.push_code_point_unchecked(code_point); + } } } } @@ -255,9 +165,9 @@ impl Wtf8Buf { /// Appends the given `char` to the end of this string. /// This does **not** include the WTF-8 concatenation check or `is_known_utf8` check. /// Copied from String::push. - fn push_code_point_unchecked(&mut self, code_point: CodePoint) { + unsafe fn push_code_point_unchecked(&mut self, code_point: CodePoint) { let mut bytes = [0; MAX_LEN_UTF8]; - let bytes = encode_utf8_raw(code_point.value, &mut bytes); + let bytes = encode_utf8_raw(code_point.to_u32(), &mut bytes); self.bytes.extend_from_slice(bytes) } @@ -379,7 +289,7 @@ impl Wtf8Buf { (Some(lead), Some(trail)) => { let len_without_lead_surrogate = self.len() - 3; self.bytes.truncate(len_without_lead_surrogate); - let other_without_trail_surrogate = &other.bytes[3..]; + let other_without_trail_surrogate = &other.as_bytes()[3..]; // 4 bytes for the supplementary code point self.bytes.reserve(4 + other_without_trail_surrogate.len()); self.push_char(decode_surrogate_pair(lead, trail)); @@ -392,7 +302,7 @@ impl Wtf8Buf { self.is_known_utf8 = false; } - self.bytes.extend_from_slice(&other.bytes); + self.bytes.extend_from_slice(other.as_bytes()); } } } @@ -400,7 +310,8 @@ impl Wtf8Buf { /// Append a Unicode scalar value at the end of the string. #[inline] pub fn push_char(&mut self, c: char) { - self.push_code_point_unchecked(CodePoint::from_char(c)) + // SAFETY: It's always safe to push a char. + unsafe { self.push_code_point_unchecked(CodePoint::from_char(c)) } } /// Append a code point at the end of the string. @@ -426,7 +337,7 @@ impl Wtf8Buf { } // No newly paired surrogates at the boundary. - self.push_code_point_unchecked(code_point) + unsafe { self.push_code_point_unchecked(code_point) } } /// Shortens a string to the specified length. @@ -437,7 +348,7 @@ impl Wtf8Buf { /// or if `new_len` is not a code point boundary. #[inline] pub fn truncate(&mut self, new_len: usize) { - assert!(is_code_point_boundary(self, new_len)); + assert!(self.is_code_point_boundary(new_len)); self.bytes.truncate(new_len) } @@ -474,8 +385,7 @@ impl Wtf8Buf { pos = surrogate_pos + 3; // Surrogates and the replacement character are all 3 bytes, so // they can substituted in-place. - self.bytes[surrogate_pos..pos] - .copy_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + self.bytes[surrogate_pos..pos].copy_from_slice("\u{FFFD}".as_bytes()); } } unsafe { String::from_utf8_unchecked(self.bytes) } @@ -498,7 +408,7 @@ impl Wtf8Buf { /// More well behaving alternative to allowing outer types /// full mutable access to the core `Vec`. #[inline] - pub(crate) fn extend_from_slice(&mut self, other: &[u8]) { + pub unsafe fn extend_from_slice_unchecked(&mut self, other: &[u8]) { self.bytes.extend_from_slice(other); self.is_known_utf8 = false; } @@ -541,153 +451,13 @@ impl Extend for Wtf8Buf { } } -/// A borrowed slice of well-formed WTF-8 data. -/// -/// Similar to `&str`, but can additionally contain surrogate code points -/// if they’re not in a surrogate pair. -#[derive(Eq, Ord, PartialEq, PartialOrd)] -#[repr(transparent)] -pub struct Wtf8 { - bytes: [u8], -} - -impl AsInner<[u8]> for Wtf8 { - #[inline] - fn as_inner(&self) -> &[u8] { - &self.bytes - } -} - -/// Formats the string in double quotes, with characters escaped according to -/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, -/// where each `x` is a hexadecimal digit. -impl fmt::Debug for Wtf8 { - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - fn write_str_escaped(f: &mut fmt::Formatter<'_>, s: &str) -> fmt::Result { - use crate::fmt::Write; - for c in s.chars().flat_map(|c| c.escape_debug()) { - f.write_char(c)? - } - Ok(()) - } - - formatter.write_str("\"")?; - let mut pos = 0; - while let Some((surrogate_pos, surrogate)) = self.next_surrogate(pos) { - write_str_escaped(formatter, unsafe { - str::from_utf8_unchecked(&self.bytes[pos..surrogate_pos]) - })?; - write!(formatter, "\\u{{{:x}}}", surrogate)?; - pos = surrogate_pos + 3; - } - write_str_escaped(formatter, unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) })?; - formatter.write_str("\"") - } -} - -/// Formats the string with unpaired surrogates substituted with the replacement -/// character, U+FFFD. -impl fmt::Display for Wtf8 { - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - let wtf8_bytes = &self.bytes; - let mut pos = 0; - loop { - match self.next_surrogate(pos) { - Some((surrogate_pos, _)) => { - formatter.write_str(unsafe { - str::from_utf8_unchecked(&wtf8_bytes[pos..surrogate_pos]) - })?; - formatter.write_str(UTF8_REPLACEMENT_CHARACTER)?; - pos = surrogate_pos + 3; - } - None => { - let s = unsafe { str::from_utf8_unchecked(&wtf8_bytes[pos..]) }; - if pos == 0 { return s.fmt(formatter) } else { return formatter.write_str(s) } - } - } - } - } -} - -impl Wtf8 { - /// Creates a WTF-8 slice from a UTF-8 `&str` slice. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. - #[inline] - pub fn from_str(value: &str) -> &Wtf8 { - unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) } - } - - /// Creates a WTF-8 slice from a WTF-8 byte slice. - /// - /// Since the byte slice is not checked for valid WTF-8, this functions is - /// marked unsafe. - #[inline] - pub unsafe fn from_bytes_unchecked(value: &[u8]) -> &Wtf8 { - // SAFETY: start with &[u8], end with fancy &[u8] - unsafe { &*(value as *const [u8] as *const Wtf8) } - } - - /// Creates a mutable WTF-8 slice from a mutable WTF-8 byte slice. - /// - /// Since the byte slice is not checked for valid WTF-8, this functions is - /// marked unsafe. - #[inline] - unsafe fn from_mut_bytes_unchecked(value: &mut [u8]) -> &mut Wtf8 { - // SAFETY: start with &mut [u8], end with fancy &mut [u8] - unsafe { &mut *(value as *mut [u8] as *mut Wtf8) } - } - - /// Returns the length, in WTF-8 bytes. - #[inline] - pub fn len(&self) -> usize { - self.bytes.len() - } - - #[inline] - pub fn is_empty(&self) -> bool { - self.bytes.is_empty() - } - - /// Returns the code point at `position` if it is in the ASCII range, - /// or `b'\xFF'` otherwise. - /// - /// # Panics - /// - /// Panics if `position` is beyond the end of the string. - #[inline] - pub fn ascii_byte_at(&self, position: usize) -> u8 { - match self.bytes[position] { - ascii_byte @ 0x00..=0x7F => ascii_byte, - _ => 0xFF, - } - } - - /// Returns an iterator for the string’s code points. - #[inline] - pub fn code_points(&self) -> Wtf8CodePoints<'_> { - Wtf8CodePoints { bytes: self.bytes.iter() } - } - - /// Access raw bytes of WTF-8 data - #[inline] - pub fn as_bytes(&self) -> &[u8] { - &self.bytes - } - - /// Tries to convert the string to UTF-8 and return a `&str` slice. - /// - /// Returns `None` if the string contains surrogates. - /// - /// This does not copy the data. - #[inline] - pub fn as_str(&self) -> Result<&str, str::Utf8Error> { - str::from_utf8(&self.bytes) - } +// helps diff +mod wtf8 { + use super::*; /// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. - pub fn to_owned(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_vec(), is_known_utf8: false } + pub(super) fn to_owned(slice: &Wtf8) -> Wtf8Buf { + Wtf8Buf { bytes: slice.as_bytes().to_vec(), is_known_utf8: false } } /// Lossily converts the string to UTF-8. @@ -696,20 +466,20 @@ impl Wtf8 { /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). /// /// This only copies the data if necessary (if it contains any surrogate). - pub fn to_string_lossy(&self) -> Cow<'_, str> { - let Some((surrogate_pos, _)) = self.next_surrogate(0) else { - return Cow::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }); + pub(super) fn to_string_lossy(slice: &Wtf8) -> Cow<'_, str> { + let Some((surrogate_pos, _)) = slice.next_surrogate(0) else { + return Cow::Borrowed(unsafe { str::from_utf8_unchecked(slice.as_bytes()) }); }; - let wtf8_bytes = &self.bytes; - let mut utf8_bytes = Vec::with_capacity(self.len()); + let wtf8_bytes = slice.as_bytes(); + let mut utf8_bytes = Vec::with_capacity(slice.len()); utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); - utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + utf8_bytes.extend_from_slice("\u{FFFD}".as_bytes()); let mut pos = surrogate_pos + 3; loop { - match self.next_surrogate(pos) { + match slice.next_surrogate(pos) { Some((surrogate_pos, _)) => { utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); - utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); + utf8_bytes.extend_from_slice("\u{FFFD}".as_bytes()); pos = surrogate_pos + 3; } None => { @@ -720,342 +490,76 @@ impl Wtf8 { } } - /// Converts the WTF-8 string to potentially ill-formed UTF-16 - /// and return an iterator of 16-bit code units. - /// - /// This is lossless: - /// calling `Wtf8Buf::from_ill_formed_utf16` on the resulting code units - /// would always return the original WTF-8 string. #[inline] - pub fn encode_wide(&self) -> EncodeWide<'_> { - EncodeWide { code_points: self.code_points(), extra: 0 } + pub(super) fn clone_into(slice: &Wtf8, buf: &mut Wtf8Buf) { + buf.is_known_utf8 = false; + slice.as_bytes().clone_into(&mut buf.bytes); } +} - #[inline] - fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { - let mut iter = self.bytes[pos..].iter(); - loop { - let b = *iter.next()?; - if b < 0x80 { - pos += 1; - } else if b < 0xE0 { - iter.next(); - pos += 2; - } else if b == 0xED { - match (iter.next(), iter.next()) { - (Some(&b2), Some(&b3)) if b2 >= 0xA0 => { - return Some((pos, decode_surrogate(b2, b3))); - } - _ => pos += 3, - } - } else if b < 0xF0 { - iter.next(); - iter.next(); - pos += 3; - } else { - iter.next(); - iter.next(); - iter.next(); - pos += 4; - } - } - } +use self::wtf8::{to_owned, to_string_lossy, clone_into}; - #[inline] - fn final_lead_surrogate(&self) -> Option { - match self.bytes { - [.., 0xED, b2 @ 0xA0..=0xAF, b3] => Some(decode_surrogate(b2, b3)), - _ => None, - } +#[cfg(not(test))] +impl Wtf8 { + #[rustc_allow_incoherent_impl] + pub fn to_owned(&self) -> Wtf8Buf { + to_owned(self) } - #[inline] - fn initial_trail_surrogate(&self) -> Option { - match self.bytes { - [0xED, b2 @ 0xB0..=0xBF, b3, ..] => Some(decode_surrogate(b2, b3)), - _ => None, - } + #[rustc_allow_incoherent_impl] + pub fn clone_into(&self, buf: &mut Wtf8Buf) { + clone_into(self, buf) } - pub fn clone_into(&self, buf: &mut Wtf8Buf) { - buf.is_known_utf8 = false; - self.bytes.clone_into(&mut buf.bytes); + #[rustc_allow_incoherent_impl] + pub fn to_string_lossy(&self) -> Cow<'_, str> { + to_string_lossy(self) } - /// Boxes this `Wtf8`. - #[inline] + #[rustc_allow_incoherent_impl] pub fn into_box(&self) -> Box { - let boxed: Box<[u8]> = self.bytes.into(); + let boxed: Box<[u8]> = self.as_bytes().into(); unsafe { mem::transmute(boxed) } } - /// Creates a boxed, empty `Wtf8`. + #[rustc_allow_incoherent_impl] pub fn empty_box() -> Box { let boxed: Box<[u8]> = Default::default(); unsafe { mem::transmute(boxed) } } - #[inline] + #[cfg(target_has_atomic = "ptr")] + #[rustc_allow_incoherent_impl] pub fn into_arc(&self) -> Arc { - let arc: Arc<[u8]> = Arc::from(&self.bytes); + let arc: Arc<[u8]> = Arc::from(self.as_bytes()); unsafe { Arc::from_raw(Arc::into_raw(arc) as *const Wtf8) } } - #[inline] + #[rustc_allow_incoherent_impl] pub fn into_rc(&self) -> Rc { - let rc: Rc<[u8]> = Rc::from(&self.bytes); + let rc: Rc<[u8]> = Rc::from(self.as_bytes()); unsafe { Rc::from_raw(Rc::into_raw(rc) as *const Wtf8) } } #[inline] - pub fn make_ascii_lowercase(&mut self) { - self.bytes.make_ascii_lowercase() - } - - #[inline] - pub fn make_ascii_uppercase(&mut self) { - self.bytes.make_ascii_uppercase() - } - - #[inline] + #[rustc_allow_incoherent_impl] pub fn to_ascii_lowercase(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_ascii_lowercase(), is_known_utf8: false } + Wtf8Buf { bytes: self.as_bytes().to_ascii_lowercase(), is_known_utf8: false } } #[inline] + #[rustc_allow_incoherent_impl] pub fn to_ascii_uppercase(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_ascii_uppercase(), is_known_utf8: false } - } - - #[inline] - pub fn is_ascii(&self) -> bool { - self.bytes.is_ascii() - } - - #[inline] - pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { - self.bytes.eq_ignore_ascii_case(&other.bytes) - } -} - -/// Returns a slice of the given string for the byte range \[`begin`..`end`). -/// -/// # Panics -/// -/// Panics when `begin` and `end` do not point to code point boundaries, -/// or point beyond the end of the string. -impl ops::Index> for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, range: ops::Range) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if range.start <= range.end - && is_code_point_boundary(self, range.start) - && is_code_point_boundary(self, range.end) - { - unsafe { slice_unchecked(self, range.start, range.end) } - } else { - slice_error_fail(self, range.start, range.end) - } - } -} - -/// Returns a slice of the given string from byte `begin` to its end. -/// -/// # Panics -/// -/// Panics when `begin` is not at a code point boundary, -/// or is beyond the end of the string. -impl ops::Index> for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, range: ops::RangeFrom) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if is_code_point_boundary(self, range.start) { - unsafe { slice_unchecked(self, range.start, self.len()) } - } else { - slice_error_fail(self, range.start, self.len()) - } + Wtf8Buf { bytes: self.as_bytes().to_ascii_uppercase(), is_known_utf8: false } } } -/// Returns a slice of the given string from its beginning to byte `end`. -/// -/// # Panics -/// -/// Panics when `end` is not at a code point boundary, -/// or is beyond the end of the string. -impl ops::Index> for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, range: ops::RangeTo) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if is_code_point_boundary(self, range.end) { - unsafe { slice_unchecked(self, 0, range.end) } - } else { - slice_error_fail(self, 0, range.end) - } - } -} - -impl ops::Index for Wtf8 { - type Output = Wtf8; - - #[inline] - fn index(&self, _range: ops::RangeFull) -> &Wtf8 { - self - } -} - -#[inline] -fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { - // The first byte is assumed to be 0xED - 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F -} - #[inline] fn decode_surrogate_pair(lead: u16, trail: u16) -> char { let code_point = 0x10000 + ((((lead - 0xD800) as u32) << 10) | (trail - 0xDC00) as u32); unsafe { char::from_u32_unchecked(code_point) } } -/// Copied from str::is_char_boundary -#[inline] -pub fn is_code_point_boundary(slice: &Wtf8, index: usize) -> bool { - if index == 0 { - return true; - } - match slice.bytes.get(index) { - None => index == slice.len(), - Some(&b) => (b as i8) >= -0x40, - } -} - -/// Verify that `index` is at the edge of either a valid UTF-8 codepoint -/// (i.e. a codepoint that's not a surrogate) or of the whole string. -/// -/// These are the cases currently permitted by `OsStr::slice_encoded_bytes`. -/// Splitting between surrogates is valid as far as WTF-8 is concerned, but -/// we do not permit it in the public API because WTF-8 is considered an -/// implementation detail. -#[track_caller] -#[inline] -pub fn check_utf8_boundary(slice: &Wtf8, index: usize) { - if index == 0 { - return; - } - match slice.bytes.get(index) { - Some(0xED) => (), // Might be a surrogate - Some(&b) if (b as i8) >= -0x40 => return, - Some(_) => panic!("byte index {index} is not a codepoint boundary"), - None if index == slice.len() => return, - None => panic!("byte index {index} is out of bounds"), - } - if slice.bytes[index + 1] >= 0xA0 { - // There's a surrogate after index. Now check before index. - if index >= 3 && slice.bytes[index - 3] == 0xED && slice.bytes[index - 2] >= 0xA0 { - panic!("byte index {index} lies between surrogate codepoints"); - } - } -} - -/// Copied from core::str::raw::slice_unchecked -#[inline] -pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { - // SAFETY: memory layout of a &[u8] and &Wtf8 are the same - unsafe { - let len = end - begin; - let start = s.as_bytes().as_ptr().add(begin); - Wtf8::from_bytes_unchecked(slice::from_raw_parts(start, len)) - } -} - -/// Copied from core::str::raw::slice_error_fail -#[inline(never)] -pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { - assert!(begin <= end); - panic!("index {begin} and/or {end} in `{s:?}` do not lie on character boundary"); -} - -/// Iterator for the code points of a WTF-8 string. -/// -/// Created with the method `.code_points()`. -#[derive(Clone)] -pub struct Wtf8CodePoints<'a> { - bytes: slice::Iter<'a, u8>, -} - -impl Iterator for Wtf8CodePoints<'_> { - type Item = CodePoint; - - #[inline] - fn next(&mut self) -> Option { - // SAFETY: `self.bytes` has been created from a WTF-8 string - unsafe { next_code_point(&mut self.bytes).map(|c| CodePoint { value: c }) } - } - - #[inline] - fn size_hint(&self) -> (usize, Option) { - let len = self.bytes.len(); - (len.saturating_add(3) / 4, Some(len)) - } -} - -/// Generates a wide character sequence for potentially ill-formed UTF-16. -#[stable(feature = "rust1", since = "1.0.0")] -#[derive(Clone)] -pub struct EncodeWide<'a> { - code_points: Wtf8CodePoints<'a>, - extra: u16, -} - -// Copied from libunicode/u_str.rs -#[stable(feature = "rust1", since = "1.0.0")] -impl Iterator for EncodeWide<'_> { - type Item = u16; - - #[inline] - fn next(&mut self) -> Option { - if self.extra != 0 { - let tmp = self.extra; - self.extra = 0; - return Some(tmp); - } - - let mut buf = [0; MAX_LEN_UTF16]; - self.code_points.next().map(|code_point| { - let n = encode_utf16_raw(code_point.value, &mut buf).len(); - if n == 2 { - self.extra = buf[1]; - } - buf[0] - }) - } - - #[inline] - fn size_hint(&self) -> (usize, Option) { - let (low, high) = self.code_points.size_hint(); - let ext = (self.extra != 0) as usize; - // every code point gets either one u16 or two u16, - // so this iterator is between 1 or 2 times as - // long as the underlying iterator. - (low + ext, high.and_then(|n| n.checked_mul(2)).and_then(|n| n.checked_add(ext))) - } -} - -#[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")] -impl FusedIterator for EncodeWide<'_> {} - -impl Hash for CodePoint { - #[inline] - fn hash(&self, state: &mut H) { - self.value.hash(state) - } -} - impl Hash for Wtf8Buf { #[inline] fn hash(&self, state: &mut H) { @@ -1063,21 +567,3 @@ impl Hash for Wtf8Buf { 0xfeu8.hash(state) } } - -impl Hash for Wtf8 { - #[inline] - fn hash(&self, state: &mut H) { - state.write(&self.bytes); - 0xfeu8.hash(state) - } -} - -#[unstable(feature = "clone_to_uninit", issue = "126799")] -unsafe impl CloneToUninit for Wtf8 { - #[inline] - #[cfg_attr(debug_assertions, track_caller)] - unsafe fn clone_to_uninit(&self, dst: *mut u8) { - // SAFETY: we're just a transparent wrapper around [u8] - unsafe { self.bytes.clone_to_uninit(dst) } - } -} diff --git a/library/alloc/src/wtf8/tests.rs b/library/alloc/src/wtf8/tests.rs index b57c99a8452..291f63f9f9e 100644 --- a/library/alloc/src/wtf8/tests.rs +++ b/library/alloc/src/wtf8/tests.rs @@ -1,3 +1,5 @@ +use realalloc::string::ToString; + use super::*; #[test] @@ -82,82 +84,85 @@ fn code_point_to_char_lossy() { #[test] fn wtf8buf_new() { - assert_eq!(Wtf8Buf::new().bytes, b""); + assert_eq!(Wtf8Buf::new().as_bytes(), b""); } #[test] fn wtf8buf_from_str() { - assert_eq!(Wtf8Buf::from_str("").bytes, b""); - assert_eq!(Wtf8Buf::from_str("aé 💩").bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(Wtf8Buf::from_str("").as_bytes(), b""); + assert_eq!(Wtf8Buf::from_str("aé 💩").as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); } #[test] fn wtf8buf_from_string() { - assert_eq!(Wtf8Buf::from_string(String::from("")).bytes, b""); - assert_eq!(Wtf8Buf::from_string(String::from("aé 💩")).bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(Wtf8Buf::from_string(String::from("")).as_bytes(), b""); + assert_eq!( + Wtf8Buf::from_string(String::from("aé 💩")).as_bytes(), + b"a\xC3\xA9 \xF0\x9F\x92\xA9" + ); } #[test] fn wtf8buf_from_wide() { let buf = Wtf8Buf::from_wide(&[]); - assert_eq!(buf.bytes, b""); + assert_eq!(buf.as_bytes(), b""); assert!(buf.is_known_utf8); let buf = Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xDCA9]); - assert_eq!(buf.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(buf.as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); assert!(buf.is_known_utf8); let buf = Wtf8Buf::from_wide(&[0x61, 0xE9, 0x20, 0xD83D, 0xD83D, 0xDCA9]); - assert_eq!(buf.bytes, b"a\xC3\xA9 \xED\xA0\xBD\xF0\x9F\x92\xA9"); + assert_eq!(buf.as_bytes(), b"a\xC3\xA9 \xED\xA0\xBD\xF0\x9F\x92\xA9"); assert!(!buf.is_known_utf8); let buf = Wtf8Buf::from_wide(&[0xD800]); - assert_eq!(buf.bytes, b"\xED\xA0\x80"); + assert_eq!(buf.as_bytes(), b"\xED\xA0\x80"); assert!(!buf.is_known_utf8); let buf = Wtf8Buf::from_wide(&[0xDBFF]); - assert_eq!(buf.bytes, b"\xED\xAF\xBF"); + assert_eq!(buf.as_bytes(), b"\xED\xAF\xBF"); assert!(!buf.is_known_utf8); let buf = Wtf8Buf::from_wide(&[0xDC00]); - assert_eq!(buf.bytes, b"\xED\xB0\x80"); + assert_eq!(buf.as_bytes(), b"\xED\xB0\x80"); assert!(!buf.is_known_utf8); let buf = Wtf8Buf::from_wide(&[0xDFFF]); - assert_eq!(buf.bytes, b"\xED\xBF\xBF"); + assert_eq!(buf.as_bytes(), b"\xED\xBF\xBF"); assert!(!buf.is_known_utf8); } #[test] fn wtf8buf_push_str() { let mut string = Wtf8Buf::new(); - assert_eq!(string.bytes, b""); + assert_eq!(string.as_bytes(), b""); assert!(string.is_known_utf8); string.push_str("aé 💩"); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(string.as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); assert!(string.is_known_utf8); } #[test] fn wtf8buf_push_char() { let mut string = Wtf8Buf::from_str("aé "); - assert_eq!(string.bytes, b"a\xC3\xA9 "); + assert_eq!(string.as_bytes(), b"a\xC3\xA9 "); assert!(string.is_known_utf8); string.push_char('💩'); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(string.as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); assert!(string.is_known_utf8); } #[test] fn wtf8buf_push() { let mut string = Wtf8Buf::from_str("aé "); - assert_eq!(string.bytes, b"a\xC3\xA9 "); + assert_eq!(string.as_bytes(), b"a\xC3\xA9 "); assert!(string.is_known_utf8); string.push(CodePoint::from_char('💩')); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(string.as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); assert!(string.is_known_utf8); fn c(value: u32) -> CodePoint { @@ -168,53 +173,53 @@ fn wtf8buf_push() { string.push(c(0xD83D)); // lead assert!(!string.is_known_utf8); string.push(c(0xDCA9)); // trail - assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!(string.as_bytes(), b"\xF0\x9F\x92\xA9"); // Magic! let mut string = Wtf8Buf::new(); string.push(c(0xD83D)); // lead assert!(!string.is_known_utf8); string.push(c(0x20)); // not surrogate string.push(c(0xDCA9)); // trail - assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); + assert_eq!(string.as_bytes(), b"\xED\xA0\xBD \xED\xB2\xA9"); let mut string = Wtf8Buf::new(); string.push(c(0xD800)); // lead assert!(!string.is_known_utf8); string.push(c(0xDBFF)); // lead - assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + assert_eq!(string.as_bytes(), b"\xED\xA0\x80\xED\xAF\xBF"); let mut string = Wtf8Buf::new(); string.push(c(0xD800)); // lead assert!(!string.is_known_utf8); string.push(c(0xE000)); // not surrogate - assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); + assert_eq!(string.as_bytes(), b"\xED\xA0\x80\xEE\x80\x80"); let mut string = Wtf8Buf::new(); string.push(c(0xD7FF)); // not surrogate assert!(string.is_known_utf8); string.push(c(0xDC00)); // trail assert!(!string.is_known_utf8); - assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + assert_eq!(string.as_bytes(), b"\xED\x9F\xBF\xED\xB0\x80"); let mut string = Wtf8Buf::new(); string.push(c(0x61)); // not surrogate, < 3 bytes assert!(string.is_known_utf8); string.push(c(0xDC00)); // trail assert!(!string.is_known_utf8); - assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); + assert_eq!(string.as_bytes(), b"\x61\xED\xB0\x80"); let mut string = Wtf8Buf::new(); string.push(c(0xDC00)); // trail assert!(!string.is_known_utf8); - assert_eq!(string.bytes, b"\xED\xB0\x80"); + assert_eq!(string.as_bytes(), b"\xED\xB0\x80"); } #[test] fn wtf8buf_push_wtf8() { let mut string = Wtf8Buf::from_str("aé"); - assert_eq!(string.bytes, b"a\xC3\xA9"); + assert_eq!(string.as_bytes(), b"a\xC3\xA9"); string.push_wtf8(Wtf8::from_str(" 💩")); - assert_eq!(string.bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(string.as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); assert!(string.is_known_utf8); fn w(v: &[u8]) -> &Wtf8 { @@ -224,42 +229,42 @@ fn wtf8buf_push_wtf8() { let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"\xED\xA0\xBD")); // lead string.push_wtf8(w(b"\xED\xB2\xA9")); // trail - assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!(string.as_bytes(), b"\xF0\x9F\x92\xA9"); // Magic! let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"\xED\xA0\xBD")); // lead string.push_wtf8(w(b" ")); // not surrogate string.push_wtf8(w(b"\xED\xB2\xA9")); // trail - assert_eq!(string.bytes, b"\xED\xA0\xBD \xED\xB2\xA9"); + assert_eq!(string.as_bytes(), b"\xED\xA0\xBD \xED\xB2\xA9"); assert!(!string.is_known_utf8); let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"\xED\xA0\x80")); // lead string.push_wtf8(w(b"\xED\xAF\xBF")); // lead - assert_eq!(string.bytes, b"\xED\xA0\x80\xED\xAF\xBF"); + assert_eq!(string.as_bytes(), b"\xED\xA0\x80\xED\xAF\xBF"); assert!(!string.is_known_utf8); let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"\xED\xA0\x80")); // lead string.push_wtf8(w(b"\xEE\x80\x80")); // not surrogate - assert_eq!(string.bytes, b"\xED\xA0\x80\xEE\x80\x80"); + assert_eq!(string.as_bytes(), b"\xED\xA0\x80\xEE\x80\x80"); assert!(!string.is_known_utf8); let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"\xED\x9F\xBF")); // not surrogate string.push_wtf8(w(b"\xED\xB0\x80")); // trail - assert_eq!(string.bytes, b"\xED\x9F\xBF\xED\xB0\x80"); + assert_eq!(string.as_bytes(), b"\xED\x9F\xBF\xED\xB0\x80"); assert!(!string.is_known_utf8); let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"a")); // not surrogate, < 3 bytes string.push_wtf8(w(b"\xED\xB0\x80")); // trail - assert_eq!(string.bytes, b"\x61\xED\xB0\x80"); + assert_eq!(string.as_bytes(), b"\x61\xED\xB0\x80"); assert!(!string.is_known_utf8); let mut string = Wtf8Buf::new(); string.push_wtf8(w(b"\xED\xB0\x80")); // trail - assert_eq!(string.bytes, b"\xED\xB0\x80"); + assert_eq!(string.as_bytes(), b"\xED\xB0\x80"); assert!(!string.is_known_utf8); } @@ -269,15 +274,15 @@ fn wtf8buf_truncate() { assert!(string.is_known_utf8); string.truncate(3); - assert_eq!(string.bytes, b"a\xC3\xA9"); + assert_eq!(string.as_bytes(), b"a\xC3\xA9"); assert!(string.is_known_utf8); string.truncate(1); - assert_eq!(string.bytes, b"a"); + assert_eq!(string.as_bytes(), b"a"); assert!(string.is_known_utf8); string.truncate(0); - assert_eq!(string.bytes, b""); + assert_eq!(string.as_bytes(), b""); assert!(string.is_known_utf8); } @@ -287,11 +292,11 @@ fn wtf8buf_truncate_around_non_bmp() { assert!(string.is_known_utf8); string.truncate(4); - assert_eq!(string.bytes, b"\xF0\x9F\x92\xA9"); + assert_eq!(string.as_bytes(), b"\xF0\x9F\x92\xA9"); assert!(string.is_known_utf8); string.truncate(0); - assert_eq!(string.bytes, b""); + assert_eq!(string.as_bytes(), b""); assert!(string.is_known_utf8); } @@ -361,7 +366,7 @@ fn wtf8buf_from_iterator() { Wtf8Buf { bytes: b"a\xC3\xA9 \xF0\x9F\x92\xA9".to_vec(), is_known_utf8: true } ); - assert_eq!(f(&[0xD83D, 0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!(f(&[0xD83D, 0xDCA9]).as_bytes(), b"\xF0\x9F\x92\xA9"); // Magic! assert_eq!( f(&[0xD83D, 0x20, 0xDCA9]), Wtf8Buf { bytes: b"\xED\xA0\xBD \xED\xB2\xA9".to_vec(), is_known_utf8: false } @@ -401,7 +406,7 @@ fn wtf8buf_extend() { Wtf8Buf { bytes: b"a\xC3\xA9 \xF0\x9F\x92\xA9".to_vec(), is_known_utf8: true } ); - assert_eq!(e(&[0xD83D], &[0xDCA9]).bytes, b"\xF0\x9F\x92\xA9"); // Magic! + assert_eq!(e(&[0xD83D], &[0xDCA9]).as_bytes(), b"\xF0\x9F\x92\xA9"); // Magic! assert_eq!( e(&[0xD83D, 0x20], &[0xDCA9]), Wtf8Buf { bytes: b"\xED\xA0\xBD \xED\xB2\xA9".to_vec(), is_known_utf8: false } @@ -449,8 +454,8 @@ fn wtf8buf_show_str() { #[test] fn wtf8_from_str() { - assert_eq!(&Wtf8::from_str("").bytes, b""); - assert_eq!(&Wtf8::from_str("aé 💩").bytes, b"a\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(&Wtf8::from_str("").as_bytes(), b""); + assert_eq!(&Wtf8::from_str("aé 💩").as_bytes(), b"a\xC3\xA9 \xF0\x9F\x92\xA9"); } #[test] @@ -461,7 +466,7 @@ fn wtf8_len() { #[test] fn wtf8_slice() { - assert_eq!(&Wtf8::from_str("aé 💩")[1..4].bytes, b"\xC3\xA9 "); + assert_eq!(&Wtf8::from_str("aé 💩")[1..4].as_bytes(), b"\xC3\xA9 "); } #[test] @@ -472,7 +477,7 @@ fn wtf8_slice_not_code_point_boundary() { #[test] fn wtf8_slice_from() { - assert_eq!(&Wtf8::from_str("aé 💩")[1..].bytes, b"\xC3\xA9 \xF0\x9F\x92\xA9"); + assert_eq!(&Wtf8::from_str("aé 💩")[1..].as_bytes(), b"\xC3\xA9 \xF0\x9F\x92\xA9"); } #[test] @@ -483,7 +488,7 @@ fn wtf8_slice_from_not_code_point_boundary() { #[test] fn wtf8_slice_to() { - assert_eq!(&Wtf8::from_str("aé 💩")[..4].bytes, b"a\xC3\xA9 "); + assert_eq!(&Wtf8::from_str("aé 💩")[..4].as_bytes(), b"a\xC3\xA9 "); } #[test] @@ -529,12 +534,12 @@ fn wtf8_as_str() { #[test] fn wtf8_to_string_lossy() { - assert_eq!(Wtf8::from_str("").to_string_lossy(), Cow::Borrowed("")); - assert_eq!(Wtf8::from_str("aé 💩").to_string_lossy(), Cow::Borrowed("aé 💩")); + assert_eq!(to_string_lossy(Wtf8::from_str("")), Cow::Borrowed("")); + assert_eq!(to_string_lossy(Wtf8::from_str("aé 💩")), Cow::Borrowed("aé 💩")); let mut string = Wtf8Buf::from_str("aé 💩"); string.push(CodePoint::from_u32(0xD800).unwrap()); let expected: Cow<'_, str> = Cow::Owned(String::from("aé 💩�")); - assert_eq!(string.to_string_lossy(), expected); + assert_eq!(to_string_lossy(&string), expected); } #[test] @@ -548,7 +553,7 @@ fn wtf8_display() { let mut string = Wtf8Buf::from_str("aé 💩"); string.push(CodePoint::from_u32(0xD800).unwrap()); - assert_eq!("aé 💩�", d(string.as_inner())); + assert_eq!("aé 💩�", d(string.as_ref())); } #[test] @@ -577,67 +582,41 @@ fn wtf8_encode_wide_size_hint() { #[test] fn wtf8_clone_into() { let mut string = Wtf8Buf::new(); - Wtf8::from_str("green").clone_into(&mut string); - assert_eq!(string.bytes, b"green"); + clone_into(Wtf8::from_str("green"), &mut string); + assert_eq!(string.as_bytes(), b"green"); let mut string = Wtf8Buf::from_str("green"); - Wtf8::from_str("").clone_into(&mut string); - assert_eq!(string.bytes, b""); + clone_into(Wtf8::from_str(""), &mut string); + assert_eq!(string.as_bytes(), b""); let mut string = Wtf8Buf::from_str("red"); - Wtf8::from_str("green").clone_into(&mut string); - assert_eq!(string.bytes, b"green"); + clone_into(Wtf8::from_str("green"), &mut string); + assert_eq!(string.as_bytes(), b"green"); let mut string = Wtf8Buf::from_str("green"); - Wtf8::from_str("red").clone_into(&mut string); - assert_eq!(string.bytes, b"red"); + clone_into(Wtf8::from_str("red"), &mut string); + assert_eq!(string.as_bytes(), b"red"); let mut string = Wtf8Buf::from_str("green"); assert!(string.is_known_utf8); - unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").clone_into(&mut string) }; - assert_eq!(string.bytes, b"\xED\xA0\x80"); + clone_into(unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80") }, &mut string); + assert_eq!(string.as_bytes(), b"\xED\xA0\x80"); assert!(!string.is_known_utf8); } -#[test] -fn wtf8_to_ascii_lowercase() { - let lowercase = Wtf8::from_str("").to_ascii_lowercase(); - assert_eq!(lowercase.bytes, b""); - - let lowercase = Wtf8::from_str("GrEeN gRaPeS! 🍇").to_ascii_lowercase(); - assert_eq!(lowercase.bytes, b"green grapes! \xf0\x9f\x8d\x87"); - - let lowercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_ascii_lowercase() }; - assert_eq!(lowercase.bytes, b"\xED\xA0\x80"); - assert!(!lowercase.is_known_utf8); -} - -#[test] -fn wtf8_to_ascii_uppercase() { - let uppercase = Wtf8::from_str("").to_ascii_uppercase(); - assert_eq!(uppercase.bytes, b""); - - let uppercase = Wtf8::from_str("GrEeN gRaPeS! 🍇").to_ascii_uppercase(); - assert_eq!(uppercase.bytes, b"GREEN GRAPES! \xf0\x9f\x8d\x87"); - - let uppercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_ascii_uppercase() }; - assert_eq!(uppercase.bytes, b"\xED\xA0\x80"); - assert!(!uppercase.is_known_utf8); -} - #[test] fn wtf8_make_ascii_lowercase() { let mut lowercase = Wtf8Buf::from_str(""); lowercase.make_ascii_lowercase(); - assert_eq!(lowercase.bytes, b""); + assert_eq!(lowercase.as_bytes(), b""); let mut lowercase = Wtf8Buf::from_str("GrEeN gRaPeS! 🍇"); lowercase.make_ascii_lowercase(); - assert_eq!(lowercase.bytes, b"green grapes! \xf0\x9f\x8d\x87"); + assert_eq!(lowercase.as_bytes(), b"green grapes! \xf0\x9f\x8d\x87"); - let mut lowercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + let mut lowercase = to_owned(unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80") }); lowercase.make_ascii_lowercase(); - assert_eq!(lowercase.bytes, b"\xED\xA0\x80"); + assert_eq!(lowercase.as_bytes(), b"\xED\xA0\x80"); assert!(!lowercase.is_known_utf8); } @@ -645,22 +624,22 @@ fn wtf8_make_ascii_lowercase() { fn wtf8_make_ascii_uppercase() { let mut uppercase = Wtf8Buf::from_str(""); uppercase.make_ascii_uppercase(); - assert_eq!(uppercase.bytes, b""); + assert_eq!(uppercase.as_bytes(), b""); let mut uppercase = Wtf8Buf::from_str("GrEeN gRaPeS! 🍇"); uppercase.make_ascii_uppercase(); - assert_eq!(uppercase.bytes, b"GREEN GRAPES! \xf0\x9f\x8d\x87"); + assert_eq!(uppercase.as_bytes(), b"GREEN GRAPES! \xf0\x9f\x8d\x87"); - let mut uppercase = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + let mut uppercase = to_owned(unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80") }); uppercase.make_ascii_uppercase(); - assert_eq!(uppercase.bytes, b"\xED\xA0\x80"); + assert_eq!(uppercase.as_bytes(), b"\xED\xA0\x80"); assert!(!uppercase.is_known_utf8); } #[test] fn wtf8_to_owned() { - let string = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; - assert_eq!(string.bytes, b"\xED\xA0\x80"); + let string = to_owned(unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80") }); + assert_eq!(string.as_bytes(), b"\xED\xA0\x80"); assert!(!string.is_known_utf8); } @@ -669,44 +648,44 @@ fn wtf8_valid_utf8_boundaries() { let mut string = Wtf8Buf::from_str("aé 💩"); string.push(CodePoint::from_u32(0xD800).unwrap()); string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 0); - check_utf8_boundary(&string, 1); - check_utf8_boundary(&string, 3); - check_utf8_boundary(&string, 4); - check_utf8_boundary(&string, 8); - check_utf8_boundary(&string, 14); + string.check_utf8_boundary(0); + string.check_utf8_boundary(1); + string.check_utf8_boundary(3); + string.check_utf8_boundary(4); + string.check_utf8_boundary(8); + string.check_utf8_boundary(14); assert_eq!(string.len(), 14); string.push_char('a'); - check_utf8_boundary(&string, 14); - check_utf8_boundary(&string, 15); + string.check_utf8_boundary(14); + string.check_utf8_boundary(15); let mut string = Wtf8Buf::from_str("a"); string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 1); + string.check_utf8_boundary(1); let mut string = Wtf8Buf::from_str("\u{D7FF}"); string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 3); + string.check_utf8_boundary(3); let mut string = Wtf8Buf::new(); string.push(CodePoint::from_u32(0xD800).unwrap()); string.push_char('\u{D7FF}'); - check_utf8_boundary(&string, 3); + string.check_utf8_boundary(3); } #[test] #[should_panic(expected = "byte index 4 is out of bounds")] fn wtf8_utf8_boundary_out_of_bounds() { let string = Wtf8::from_str("aé"); - check_utf8_boundary(&string, 4); + string.check_utf8_boundary(4); } #[test] #[should_panic(expected = "byte index 1 is not a codepoint boundary")] fn wtf8_utf8_boundary_inside_codepoint() { let string = Wtf8::from_str("é"); - check_utf8_boundary(&string, 1); + string.check_utf8_boundary(1); } #[test] @@ -714,7 +693,7 @@ fn wtf8_utf8_boundary_inside_codepoint() { fn wtf8_utf8_boundary_inside_surrogate() { let mut string = Wtf8Buf::new(); string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 1); + string.check_utf8_boundary(1); } #[test] @@ -723,20 +702,22 @@ fn wtf8_utf8_boundary_between_surrogates() { let mut string = Wtf8Buf::new(); string.push(CodePoint::from_u32(0xD800).unwrap()); string.push(CodePoint::from_u32(0xD800).unwrap()); - check_utf8_boundary(&string, 3); + string.check_utf8_boundary(3); } #[test] fn wobbled_wtf8_plus_bytes_isnt_utf8() { - let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + let mut string: Wtf8Buf = to_owned(unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80") }); assert!(!string.is_known_utf8); - string.extend_from_slice(b"some utf-8"); + unsafe { + string.extend_from_slice_unchecked(b"some utf-8"); + } assert!(!string.is_known_utf8); } #[test] fn wobbled_wtf8_plus_str_isnt_utf8() { - let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() }; + let mut string: Wtf8Buf = to_owned(unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80") }); assert!(!string.is_known_utf8); string.push_str("some utf-8"); assert!(!string.is_known_utf8); diff --git a/library/alloctests/lib.rs b/library/alloctests/lib.rs index 3241b4b0045..97de7d6a701 100644 --- a/library/alloctests/lib.rs +++ b/library/alloctests/lib.rs @@ -18,6 +18,8 @@ #![feature(allocator_api)] #![feature(array_into_iter_constructors)] #![feature(assert_matches)] +#![feature(char_internals)] +#![feature(char_max_len)] #![feature(core_intrinsics)] #![feature(exact_size_is_empty)] #![feature(extend_one)] @@ -41,6 +43,7 @@ #![feature(trusted_random_access)] #![feature(try_reserve_kind)] #![feature(try_trait_v2)] +#![feature(wtf8_internals)] // tidy-alphabetical-end // // Language features: @@ -68,15 +71,18 @@ extern crate test; mod testing; use realalloc::*; -// We are directly including collections and raw_vec here as both use non-public -// methods and fields in tests and as such need to have the types to test in the -// same crate as the tests themself. +// We are directly including collections, raw_vec, and wtf8 here as they use non-public +// methods and fields in tests and as such need to have the types to test in the same +// crate as the tests themself. #[path = "../alloc/src/collections/mod.rs"] mod collections; #[path = "../alloc/src/raw_vec/mod.rs"] mod raw_vec; +#[path = "../alloc/src/wtf8/mod.rs"] +mod wtf8; + #[allow(dead_code)] // Not used in all configurations pub(crate) mod test_helpers { /// Copied from `std::test_helpers::test_rng`, since these tests rely on the diff --git a/library/alloctests/tests/lib.rs b/library/alloctests/tests/lib.rs index fcfc7f8dd29..447af240a4b 100644 --- a/library/alloctests/tests/lib.rs +++ b/library/alloctests/tests/lib.rs @@ -3,6 +3,7 @@ #![feature(iter_array_chunks)] #![feature(assert_matches)] #![feature(btree_extract_if)] +#![feature(wtf8_internals)] #![feature(char_max_len)] #![feature(cow_is_borrowed)] #![feature(core_intrinsics)] diff --git a/library/core/src/lib.rs b/library/core/src/lib.rs index 71abd707374..2b6ea7d9afc 100644 --- a/library/core/src/lib.rs +++ b/library/core/src/lib.rs @@ -362,6 +362,8 @@ pub mod slice; pub mod str; pub mod time; +pub mod wtf8; + pub mod unicode; /* Async */ diff --git a/library/core/src/num/niche_types.rs b/library/core/src/num/niche_types.rs index d57b1d433e5..610d9d8cf92 100644 --- a/library/core/src/num/niche_types.rs +++ b/library/core/src/num/niche_types.rs @@ -178,3 +178,18 @@ impl NotAllOnesHelper for u64 { impl NotAllOnesHelper for i64 { type Type = I64NotAllOnes; } + +define_valid_range_type! { + pub struct CodePointInner(u32 as u32 in 0..=0x10ffff); +} + +impl CodePointInner { + pub const ZERO: Self = CodePointInner::new(0).unwrap(); +} + +impl Default for CodePointInner { + #[inline] + fn default() -> Self { + Self::ZERO + } +} diff --git a/library/core/src/wtf8.rs b/library/core/src/wtf8.rs index 50bde88b5a4..5631993dea2 100644 --- a/library/core/src/wtf8.rs +++ b/library/core/src/wtf8.rs @@ -10,28 +10,23 @@ //! of WTF-8 strings, //! nor can it decode WTF-8 from arbitrary bytes. //! WTF-8 strings can be obtained from UTF-8, UTF-16, or code points. - -// this module is imported from @SimonSapin's repo and has tons of dead code on -// unix (it's mostly used on windows), so don't worry about dead code here. -#![allow(dead_code)] - -#[cfg(test)] -mod tests; - -use core::char::{MAX_LEN_UTF8, MAX_LEN_UTF16, encode_utf8_raw, encode_utf16_raw}; -use core::clone::CloneToUninit; -use core::str::next_code_point; - -use crate::borrow::Cow; -use crate::collections::TryReserveError; +#![unstable( + feature = "wtf8_internals", + issue = "none", + reason = "this is internal code for representing OsStr on some platforms and not a public API" +)] +// rustdoc bug: doc(hidden) on the module won't stop types in the module from showing up in trait +// implementations, so, we'll have to add more doc(hidden)s anyway +#![doc(hidden)] + +use crate::char::{MAX_LEN_UTF16, encode_utf16_raw}; +use crate::clone::CloneToUninit; +use crate::fmt::{self, Write}; use crate::hash::{Hash, Hasher}; use crate::iter::FusedIterator; -use crate::rc::Rc; -use crate::sync::Arc; -use crate::sys_common::AsInner; -use crate::{fmt, mem, ops, slice, str}; - -const UTF8_REPLACEMENT_CHARACTER: &str = "\u{FFFD}"; +use crate::num::niche_types::CodePointInner; +use crate::str::next_code_point; +use crate::{ops, slice, str}; /// A Unicode code point: from U+0000 to U+10FFFF. /// @@ -39,16 +34,15 @@ const UTF8_REPLACEMENT_CHARACTER: &str = "\u{FFFD}"; /// which represents a Unicode scalar value: /// a code point that is not a surrogate (U+D800 to U+DFFF). #[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Copy)] -pub struct CodePoint { - value: u32, -} +#[doc(hidden)] +pub struct CodePoint(CodePointInner); /// Format the code point as `U+` followed by four to six hexadecimal digits. /// Example: `U+1F4A9` impl fmt::Debug for CodePoint { #[inline] fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - write!(formatter, "U+{:04X}", self.value) + write!(formatter, "U+{:04X}", self.0.as_inner()) } } @@ -58,7 +52,8 @@ impl CodePoint { /// Only use when `value` is known to be less than or equal to 0x10FFFF. #[inline] pub unsafe fn from_u32_unchecked(value: u32) -> CodePoint { - CodePoint { value } + // SAFETY: Guaranteed by caller. + CodePoint(unsafe { CodePointInner::new_unchecked(value) }) } /// Creates a new `CodePoint` if the value is a valid code point. @@ -66,10 +61,7 @@ impl CodePoint { /// Returns `None` if `value` is above 0x10FFFF. #[inline] pub fn from_u32(value: u32) -> Option { - match value { - 0..=0x10FFFF => Some(CodePoint { value }), - _ => None, - } + Some(CodePoint(CodePointInner::new(value)?)) } /// Creates a new `CodePoint` from a `char`. @@ -77,19 +69,20 @@ impl CodePoint { /// Since all Unicode scalar values are code points, this always succeeds. #[inline] pub fn from_char(value: char) -> CodePoint { - CodePoint { value: value as u32 } + // SAFETY: All char are valid for this type. + unsafe { CodePoint::from_u32_unchecked(value as u32) } } /// Returns the numeric value of the code point. #[inline] pub fn to_u32(&self) -> u32 { - self.value + self.0.as_inner() } /// Returns the numeric value of the code point if it is a leading surrogate. #[inline] pub fn to_lead_surrogate(&self) -> Option { - match self.value { + match self.to_u32() { lead @ 0xD800..=0xDBFF => Some(lead as u16), _ => None, } @@ -98,7 +91,7 @@ impl CodePoint { /// Returns the numeric value of the code point if it is a trailing surrogate. #[inline] pub fn to_trail_surrogate(&self) -> Option { - match self.value { + match self.to_u32() { trail @ 0xDC00..=0xDFFF => Some(trail as u16), _ => None, } @@ -109,9 +102,10 @@ impl CodePoint { /// Returns `None` if the code point is a surrogate (from U+D800 to U+DFFF). #[inline] pub fn to_char(&self) -> Option { - match self.value { + match self.to_u32() { 0xD800..=0xDFFF => None, - _ => Some(unsafe { char::from_u32_unchecked(self.value) }), + // SAFETY: We explicitly check that the char is valid. + valid => Some(unsafe { char::from_u32_unchecked(valid) }), } } @@ -121,423 +115,7 @@ impl CodePoint { /// if the code point is a surrogate (from U+D800 to U+DFFF). #[inline] pub fn to_char_lossy(&self) -> char { - self.to_char().unwrap_or('\u{FFFD}') - } -} - -/// An owned, growable string of well-formed WTF-8 data. -/// -/// Similar to `String`, but can additionally contain surrogate code points -/// if they’re not in a surrogate pair. -#[derive(Eq, PartialEq, Ord, PartialOrd, Clone)] -pub struct Wtf8Buf { - bytes: Vec, - - /// Do we know that `bytes` holds a valid UTF-8 encoding? We can easily - /// know this if we're constructed from a `String` or `&str`. - /// - /// It is possible for `bytes` to have valid UTF-8 without this being - /// set, such as when we're concatenating `&Wtf8`'s and surrogates become - /// paired, as we don't bother to rescan the entire string. - is_known_utf8: bool, -} - -impl ops::Deref for Wtf8Buf { - type Target = Wtf8; - - fn deref(&self) -> &Wtf8 { - self.as_slice() - } -} - -impl ops::DerefMut for Wtf8Buf { - fn deref_mut(&mut self) -> &mut Wtf8 { - self.as_mut_slice() - } -} - -/// Formats the string in double quotes, with characters escaped according to -/// [`char::escape_debug`] and unpaired surrogates represented as `\u{xxxx}`, -/// where each `x` is a hexadecimal digit. -/// -/// For example, the code units [U+0061, U+D800, U+000A] are formatted as -/// `"a\u{D800}\n"`. -impl fmt::Debug for Wtf8Buf { - #[inline] - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - fmt::Debug::fmt(&**self, formatter) - } -} - -/// Formats the string with unpaired surrogates substituted with the replacement -/// character, U+FFFD. -impl fmt::Display for Wtf8Buf { - fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result { - if let Some(s) = self.as_known_utf8() { - fmt::Display::fmt(s, formatter) - } else { - fmt::Display::fmt(&**self, formatter) - } - } -} - -impl Wtf8Buf { - /// Creates a new, empty WTF-8 string. - #[inline] - pub fn new() -> Wtf8Buf { - Wtf8Buf { bytes: Vec::new(), is_known_utf8: true } - } - - /// Creates a new, empty WTF-8 string with pre-allocated capacity for `capacity` bytes. - #[inline] - pub fn with_capacity(capacity: usize) -> Wtf8Buf { - Wtf8Buf { bytes: Vec::with_capacity(capacity), is_known_utf8: true } - } - - /// Creates a WTF-8 string from a WTF-8 byte vec. - /// - /// Since the byte vec is not checked for valid WTF-8, this function is - /// marked unsafe. - #[inline] - pub unsafe fn from_bytes_unchecked(value: Vec) -> Wtf8Buf { - Wtf8Buf { bytes: value, is_known_utf8: false } - } - - /// Creates a WTF-8 string from a UTF-8 `String`. - /// - /// This takes ownership of the `String` and does not copy. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. - #[inline] - pub const fn from_string(string: String) -> Wtf8Buf { - Wtf8Buf { bytes: string.into_bytes(), is_known_utf8: true } - } - - /// Creates a WTF-8 string from a UTF-8 `&str` slice. - /// - /// This copies the content of the slice. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. - #[inline] - pub fn from_str(s: &str) -> Wtf8Buf { - Wtf8Buf { bytes: s.as_bytes().to_vec(), is_known_utf8: true } - } - - pub fn clear(&mut self) { - self.bytes.clear(); - self.is_known_utf8 = true; - } - - /// Creates a WTF-8 string from a potentially ill-formed UTF-16 slice of 16-bit code units. - /// - /// This is lossless: calling `.encode_wide()` on the resulting string - /// will always return the original code units. - pub fn from_wide(v: &[u16]) -> Wtf8Buf { - let mut string = Wtf8Buf::with_capacity(v.len()); - for item in char::decode_utf16(v.iter().cloned()) { - match item { - Ok(ch) => string.push_char(ch), - Err(surrogate) => { - let surrogate = surrogate.unpaired_surrogate(); - // Surrogates are known to be in the code point range. - let code_point = unsafe { CodePoint::from_u32_unchecked(surrogate as u32) }; - // The string will now contain an unpaired surrogate. - string.is_known_utf8 = false; - // Skip the WTF-8 concatenation check, - // surrogate pairs are already decoded by decode_utf16 - string.push_code_point_unchecked(code_point); - } - } - } - string - } - - /// Appends the given `char` to the end of this string. - /// This does **not** include the WTF-8 concatenation check or `is_known_utf8` check. - /// Copied from String::push. - fn push_code_point_unchecked(&mut self, code_point: CodePoint) { - let mut bytes = [0; MAX_LEN_UTF8]; - let bytes = encode_utf8_raw(code_point.value, &mut bytes); - self.bytes.extend_from_slice(bytes) - } - - #[inline] - pub fn as_slice(&self) -> &Wtf8 { - unsafe { Wtf8::from_bytes_unchecked(&self.bytes) } - } - - #[inline] - pub fn as_mut_slice(&mut self) -> &mut Wtf8 { - // Safety: `Wtf8` doesn't expose any way to mutate the bytes that would - // cause them to change from well-formed UTF-8 to ill-formed UTF-8, - // which would break the assumptions of the `is_known_utf8` field. - unsafe { Wtf8::from_mut_bytes_unchecked(&mut self.bytes) } - } - - /// Converts the string to UTF-8 without validation, if it was created from - /// valid UTF-8. - #[inline] - fn as_known_utf8(&self) -> Option<&str> { - if self.is_known_utf8 { - // SAFETY: The buffer is known to be valid UTF-8. - Some(unsafe { str::from_utf8_unchecked(self.as_bytes()) }) - } else { - None - } - } - - /// Reserves capacity for at least `additional` more bytes to be inserted - /// in the given `Wtf8Buf`. - /// The collection may reserve more space to avoid frequent reallocations. - /// - /// # Panics - /// - /// Panics if the new capacity exceeds `isize::MAX` bytes. - #[inline] - pub fn reserve(&mut self, additional: usize) { - self.bytes.reserve(additional) - } - - /// Tries to reserve capacity for at least `additional` more bytes to be - /// inserted in the given `Wtf8Buf`. The `Wtf8Buf` may reserve more space to - /// avoid frequent reallocations. After calling `try_reserve`, capacity will - /// be greater than or equal to `self.len() + additional`. Does nothing if - /// capacity is already sufficient. This method preserves the contents even - /// if an error occurs. - /// - /// # Errors - /// - /// If the capacity overflows, or the allocator reports a failure, then an error - /// is returned. - #[inline] - pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError> { - self.bytes.try_reserve(additional) - } - - #[inline] - pub fn reserve_exact(&mut self, additional: usize) { - self.bytes.reserve_exact(additional) - } - - /// Tries to reserve the minimum capacity for exactly `additional` more - /// bytes to be inserted in the given `Wtf8Buf`. After calling - /// `try_reserve_exact`, capacity will be greater than or equal to - /// `self.len() + additional` if it returns `Ok(())`. - /// Does nothing if the capacity is already sufficient. - /// - /// Note that the allocator may give the `Wtf8Buf` more space than it - /// requests. Therefore, capacity can not be relied upon to be precisely - /// minimal. Prefer [`try_reserve`] if future insertions are expected. - /// - /// [`try_reserve`]: Wtf8Buf::try_reserve - /// - /// # Errors - /// - /// If the capacity overflows, or the allocator reports a failure, then an error - /// is returned. - #[inline] - pub fn try_reserve_exact(&mut self, additional: usize) -> Result<(), TryReserveError> { - self.bytes.try_reserve_exact(additional) - } - - #[inline] - pub fn shrink_to_fit(&mut self) { - self.bytes.shrink_to_fit() - } - - #[inline] - pub fn shrink_to(&mut self, min_capacity: usize) { - self.bytes.shrink_to(min_capacity) - } - - #[inline] - pub fn leak<'a>(self) -> &'a mut Wtf8 { - unsafe { Wtf8::from_mut_bytes_unchecked(self.bytes.leak()) } - } - - /// Returns the number of bytes that this string buffer can hold without reallocating. - #[inline] - pub fn capacity(&self) -> usize { - self.bytes.capacity() - } - - /// Append a UTF-8 slice at the end of the string. - #[inline] - pub fn push_str(&mut self, other: &str) { - self.bytes.extend_from_slice(other.as_bytes()) - } - - /// Append a WTF-8 slice at the end of the string. - /// - /// This replaces newly paired surrogates at the boundary - /// with a supplementary code point, - /// like concatenating ill-formed UTF-16 strings effectively would. - #[inline] - pub fn push_wtf8(&mut self, other: &Wtf8) { - match ((&*self).final_lead_surrogate(), other.initial_trail_surrogate()) { - // Replace newly paired surrogates by a supplementary code point. - (Some(lead), Some(trail)) => { - let len_without_lead_surrogate = self.len() - 3; - self.bytes.truncate(len_without_lead_surrogate); - let other_without_trail_surrogate = &other.bytes[3..]; - // 4 bytes for the supplementary code point - self.bytes.reserve(4 + other_without_trail_surrogate.len()); - self.push_char(decode_surrogate_pair(lead, trail)); - self.bytes.extend_from_slice(other_without_trail_surrogate); - } - _ => { - // If we'll be pushing a string containing a surrogate, we may - // no longer have UTF-8. - if self.is_known_utf8 && other.next_surrogate(0).is_some() { - self.is_known_utf8 = false; - } - - self.bytes.extend_from_slice(&other.bytes); - } - } - } - - /// Append a Unicode scalar value at the end of the string. - #[inline] - pub fn push_char(&mut self, c: char) { - self.push_code_point_unchecked(CodePoint::from_char(c)) - } - - /// Append a code point at the end of the string. - /// - /// This replaces newly paired surrogates at the boundary - /// with a supplementary code point, - /// like concatenating ill-formed UTF-16 strings effectively would. - #[inline] - pub fn push(&mut self, code_point: CodePoint) { - if let Some(trail) = code_point.to_trail_surrogate() { - if let Some(lead) = (&*self).final_lead_surrogate() { - let len_without_lead_surrogate = self.len() - 3; - self.bytes.truncate(len_without_lead_surrogate); - self.push_char(decode_surrogate_pair(lead, trail)); - return; - } - - // We're pushing a trailing surrogate. - self.is_known_utf8 = false; - } else if code_point.to_lead_surrogate().is_some() { - // We're pushing a leading surrogate. - self.is_known_utf8 = false; - } - - // No newly paired surrogates at the boundary. - self.push_code_point_unchecked(code_point) - } - - /// Shortens a string to the specified length. - /// - /// # Panics - /// - /// Panics if `new_len` > current length, - /// or if `new_len` is not a code point boundary. - #[inline] - pub fn truncate(&mut self, new_len: usize) { - assert!(is_code_point_boundary(self, new_len)); - self.bytes.truncate(new_len) - } - - /// Consumes the WTF-8 string and tries to convert it to a vec of bytes. - #[inline] - pub fn into_bytes(self) -> Vec { - self.bytes - } - - /// Consumes the WTF-8 string and tries to convert it to UTF-8. - /// - /// This does not copy the data. - /// - /// If the contents are not well-formed UTF-8 - /// (that is, if the string contains surrogates), - /// the original WTF-8 string is returned instead. - pub fn into_string(self) -> Result { - if self.is_known_utf8 || self.next_surrogate(0).is_none() { - Ok(unsafe { String::from_utf8_unchecked(self.bytes) }) - } else { - Err(self) - } - } - - /// Consumes the WTF-8 string and converts it lossily to UTF-8. - /// - /// This does not copy the data (but may overwrite parts of it in place). - /// - /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”) - pub fn into_string_lossy(mut self) -> String { - if !self.is_known_utf8 { - let mut pos = 0; - while let Some((surrogate_pos, _)) = self.next_surrogate(pos) { - pos = surrogate_pos + 3; - // Surrogates and the replacement character are all 3 bytes, so - // they can substituted in-place. - self.bytes[surrogate_pos..pos] - .copy_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); - } - } - unsafe { String::from_utf8_unchecked(self.bytes) } - } - - /// Converts this `Wtf8Buf` into a boxed `Wtf8`. - #[inline] - pub fn into_box(self) -> Box { - // SAFETY: relies on `Wtf8` being `repr(transparent)`. - unsafe { mem::transmute(self.bytes.into_boxed_slice()) } - } - - /// Converts a `Box` into a `Wtf8Buf`. - pub fn from_box(boxed: Box) -> Wtf8Buf { - let bytes: Box<[u8]> = unsafe { mem::transmute(boxed) }; - Wtf8Buf { bytes: bytes.into_vec(), is_known_utf8: false } - } - - /// Provides plumbing to core `Vec::extend_from_slice`. - /// More well behaving alternative to allowing outer types - /// full mutable access to the core `Vec`. - #[inline] - pub(crate) fn extend_from_slice(&mut self, other: &[u8]) { - self.bytes.extend_from_slice(other); - self.is_known_utf8 = false; - } -} - -/// Creates a new WTF-8 string from an iterator of code points. -/// -/// This replaces surrogate code point pairs with supplementary code points, -/// like concatenating ill-formed UTF-16 strings effectively would. -impl FromIterator for Wtf8Buf { - fn from_iter>(iter: T) -> Wtf8Buf { - let mut string = Wtf8Buf::new(); - string.extend(iter); - string - } -} - -/// Append code points from an iterator to the string. -/// -/// This replaces surrogate code point pairs with supplementary code points, -/// like concatenating ill-formed UTF-16 strings effectively would. -impl Extend for Wtf8Buf { - fn extend>(&mut self, iter: T) { - let iterator = iter.into_iter(); - let (low, _high) = iterator.size_hint(); - // Lower bound of one byte per code point (ASCII only) - self.bytes.reserve(low); - iterator.for_each(move |code_point| self.push(code_point)); - } - - #[inline] - fn extend_one(&mut self, code_point: CodePoint) { - self.push(code_point); - } - - #[inline] - fn extend_reserve(&mut self, additional: usize) { - // Lower bound of one byte per code point (ASCII only) - self.bytes.reserve(additional); + self.to_char().unwrap_or(char::REPLACEMENT_CHARACTER) } } @@ -547,13 +125,15 @@ impl Extend for Wtf8Buf { /// if they’re not in a surrogate pair. #[derive(Eq, Ord, PartialEq, PartialOrd)] #[repr(transparent)] +#[rustc_has_incoherent_inherent_impls] +#[doc(hidden)] pub struct Wtf8 { bytes: [u8], } -impl AsInner<[u8]> for Wtf8 { +impl AsRef<[u8]> for Wtf8 { #[inline] - fn as_inner(&self) -> &[u8] { + fn as_ref(&self) -> &[u8] { &self.bytes } } @@ -574,12 +154,15 @@ impl fmt::Debug for Wtf8 { formatter.write_str("\"")?; let mut pos = 0; while let Some((surrogate_pos, surrogate)) = self.next_surrogate(pos) { + // SAFETY: next_surrogate provides an index for a range of valid UTF-8 bytes. write_str_escaped(formatter, unsafe { str::from_utf8_unchecked(&self.bytes[pos..surrogate_pos]) })?; write!(formatter, "\\u{{{:x}}}", surrogate)?; pos = surrogate_pos + 3; } + + // SAFETY: after next_surrogate returns None, the remainder is valid UTF-8. write_str_escaped(formatter, unsafe { str::from_utf8_unchecked(&self.bytes[pos..]) })?; formatter.write_str("\"") } @@ -594,13 +177,15 @@ impl fmt::Display for Wtf8 { loop { match self.next_surrogate(pos) { Some((surrogate_pos, _)) => { + // SAFETY: next_surrogate provides an index for a range of valid UTF-8 bytes. formatter.write_str(unsafe { str::from_utf8_unchecked(&wtf8_bytes[pos..surrogate_pos]) })?; - formatter.write_str(UTF8_REPLACEMENT_CHARACTER)?; + formatter.write_char(char::REPLACEMENT_CHARACTER)?; pos = surrogate_pos + 3; } None => { + // SAFETY: after next_surrogate returns None, the remainder is valid UTF-8. let s = unsafe { str::from_utf8_unchecked(&wtf8_bytes[pos..]) }; if pos == 0 { return s.fmt(formatter) } else { return formatter.write_str(s) } } @@ -611,10 +196,9 @@ impl fmt::Display for Wtf8 { impl Wtf8 { /// Creates a WTF-8 slice from a UTF-8 `&str` slice. - /// - /// Since WTF-8 is a superset of UTF-8, this always succeeds. #[inline] pub fn from_str(value: &str) -> &Wtf8 { + // SAFETY: Since WTF-8 is a superset of UTF-8, this always is valid. unsafe { Wtf8::from_bytes_unchecked(value.as_bytes()) } } @@ -633,7 +217,7 @@ impl Wtf8 { /// Since the byte slice is not checked for valid WTF-8, this functions is /// marked unsafe. #[inline] - unsafe fn from_mut_bytes_unchecked(value: &mut [u8]) -> &mut Wtf8 { + pub unsafe fn from_mut_bytes_unchecked(value: &mut [u8]) -> &mut Wtf8 { // SAFETY: start with &mut [u8], end with fancy &mut [u8] unsafe { &mut *(value as *mut [u8] as *mut Wtf8) } } @@ -685,41 +269,6 @@ impl Wtf8 { str::from_utf8(&self.bytes) } - /// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. - pub fn to_owned(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_vec(), is_known_utf8: false } - } - - /// Lossily converts the string to UTF-8. - /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. - /// - /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). - /// - /// This only copies the data if necessary (if it contains any surrogate). - pub fn to_string_lossy(&self) -> Cow<'_, str> { - let Some((surrogate_pos, _)) = self.next_surrogate(0) else { - return Cow::Borrowed(unsafe { str::from_utf8_unchecked(&self.bytes) }); - }; - let wtf8_bytes = &self.bytes; - let mut utf8_bytes = Vec::with_capacity(self.len()); - utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); - utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); - let mut pos = surrogate_pos + 3; - loop { - match self.next_surrogate(pos) { - Some((surrogate_pos, _)) => { - utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); - utf8_bytes.extend_from_slice(UTF8_REPLACEMENT_CHARACTER.as_bytes()); - pos = surrogate_pos + 3; - } - None => { - utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); - return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); - } - } - } - } - /// Converts the WTF-8 string to potentially ill-formed UTF-16 /// and return an iterator of 16-bit code units. /// @@ -732,7 +281,7 @@ impl Wtf8 { } #[inline] - fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { + pub fn next_surrogate(&self, mut pos: usize) -> Option<(usize, u16)> { let mut iter = self.bytes[pos..].iter(); loop { let b = *iter.next()?; @@ -762,7 +311,7 @@ impl Wtf8 { } #[inline] - fn final_lead_surrogate(&self) -> Option { + pub fn final_lead_surrogate(&self) -> Option { match self.bytes { [.., 0xED, b2 @ 0xA0..=0xAF, b3] => Some(decode_surrogate(b2, b3)), _ => None, @@ -770,43 +319,13 @@ impl Wtf8 { } #[inline] - fn initial_trail_surrogate(&self) -> Option { + pub fn initial_trail_surrogate(&self) -> Option { match self.bytes { [0xED, b2 @ 0xB0..=0xBF, b3, ..] => Some(decode_surrogate(b2, b3)), _ => None, } } - pub fn clone_into(&self, buf: &mut Wtf8Buf) { - buf.is_known_utf8 = false; - self.bytes.clone_into(&mut buf.bytes); - } - - /// Boxes this `Wtf8`. - #[inline] - pub fn into_box(&self) -> Box { - let boxed: Box<[u8]> = self.bytes.into(); - unsafe { mem::transmute(boxed) } - } - - /// Creates a boxed, empty `Wtf8`. - pub fn empty_box() -> Box { - let boxed: Box<[u8]> = Default::default(); - unsafe { mem::transmute(boxed) } - } - - #[inline] - pub fn into_arc(&self) -> Arc { - let arc: Arc<[u8]> = Arc::from(&self.bytes); - unsafe { Arc::from_raw(Arc::into_raw(arc) as *const Wtf8) } - } - - #[inline] - pub fn into_rc(&self) -> Rc { - let rc: Rc<[u8]> = Rc::from(&self.bytes); - unsafe { Rc::from_raw(Rc::into_raw(rc) as *const Wtf8) } - } - #[inline] pub fn make_ascii_lowercase(&mut self) { self.bytes.make_ascii_lowercase() @@ -818,23 +337,23 @@ impl Wtf8 { } #[inline] - pub fn to_ascii_lowercase(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_ascii_lowercase(), is_known_utf8: false } + pub fn is_ascii(&self) -> bool { + self.bytes.is_ascii() } #[inline] - pub fn to_ascii_uppercase(&self) -> Wtf8Buf { - Wtf8Buf { bytes: self.bytes.to_ascii_uppercase(), is_known_utf8: false } + pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { + self.bytes.eq_ignore_ascii_case(&other.bytes) } #[inline] - pub fn is_ascii(&self) -> bool { - self.bytes.is_ascii() + pub fn is_code_point_boundary(&self, index: usize) -> bool { + is_code_point_boundary(self, index) } #[inline] - pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { - self.bytes.eq_ignore_ascii_case(&other.bytes) + pub fn check_utf8_boundary(&self, index: usize) { + check_utf8_boundary(self, index) } } @@ -849,11 +368,11 @@ impl ops::Index> for Wtf8 { #[inline] fn index(&self, range: ops::Range) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] if range.start <= range.end - && is_code_point_boundary(self, range.start) - && is_code_point_boundary(self, range.end) + && self.is_code_point_boundary(range.start) + && self.is_code_point_boundary(range.end) { + // SAFETY: is_code_point_boundary checks that the index is valid unsafe { slice_unchecked(self, range.start, range.end) } } else { slice_error_fail(self, range.start, range.end) @@ -872,8 +391,8 @@ impl ops::Index> for Wtf8 { #[inline] fn index(&self, range: ops::RangeFrom) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if is_code_point_boundary(self, range.start) { + if self.is_code_point_boundary(range.start) { + // SAFETY: is_code_point_boundary checks that the index is valid unsafe { slice_unchecked(self, range.start, self.len()) } } else { slice_error_fail(self, range.start, self.len()) @@ -892,8 +411,8 @@ impl ops::Index> for Wtf8 { #[inline] fn index(&self, range: ops::RangeTo) -> &Wtf8 { - // is_code_point_boundary checks that the index is in [0, .len()] - if is_code_point_boundary(self, range.end) { + if self.is_code_point_boundary(range.end) { + // SAFETY: is_code_point_boundary checks that the index is valid unsafe { slice_unchecked(self, 0, range.end) } } else { slice_error_fail(self, 0, range.end) @@ -916,11 +435,7 @@ fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F } -#[inline] -fn decode_surrogate_pair(lead: u16, trail: u16) -> char { - let code_point = 0x10000 + ((((lead - 0xD800) as u32) << 10) | (trail - 0xDC00) as u32); - unsafe { char::from_u32_unchecked(code_point) } -} +// helps diff to be unindented /// Copied from str::is_char_boundary #[inline] @@ -964,7 +479,7 @@ pub fn check_utf8_boundary(slice: &Wtf8, index: usize) { /// Copied from core::str::raw::slice_unchecked #[inline] -pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { +unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { // SAFETY: memory layout of a &[u8] and &Wtf8 are the same unsafe { let len = end - begin; @@ -975,7 +490,7 @@ pub unsafe fn slice_unchecked(s: &Wtf8, begin: usize, end: usize) -> &Wtf8 { /// Copied from core::str::raw::slice_error_fail #[inline(never)] -pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { +fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { assert!(begin <= end); panic!("index {begin} and/or {end} in `{s:?}` do not lie on character boundary"); } @@ -984,6 +499,7 @@ pub fn slice_error_fail(s: &Wtf8, begin: usize, end: usize) -> ! { /// /// Created with the method `.code_points()`. #[derive(Clone)] +#[doc(hidden)] pub struct Wtf8CodePoints<'a> { bytes: slice::Iter<'a, u8>, } @@ -994,7 +510,7 @@ impl Iterator for Wtf8CodePoints<'_> { #[inline] fn next(&mut self) -> Option { // SAFETY: `self.bytes` has been created from a WTF-8 string - unsafe { next_code_point(&mut self.bytes).map(|c| CodePoint { value: c }) } + unsafe { next_code_point(&mut self.bytes).map(|c| CodePoint::from_u32_unchecked(c)) } } #[inline] @@ -1004,9 +520,19 @@ impl Iterator for Wtf8CodePoints<'_> { } } +impl fmt::Debug for Wtf8CodePoints<'_> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_tuple("Wtf8CodePoints") + // SAFETY: We always leave the string in a valid state after each iteration. + .field(&unsafe { Wtf8::from_bytes_unchecked(self.bytes.as_slice()) }) + .finish() + } +} + /// Generates a wide character sequence for potentially ill-formed UTF-16. #[stable(feature = "rust1", since = "1.0.0")] #[derive(Clone)] +#[doc(hidden)] pub struct EncodeWide<'a> { code_points: Wtf8CodePoints<'a>, extra: u16, @@ -1027,7 +553,7 @@ impl Iterator for EncodeWide<'_> { let mut buf = [0; MAX_LEN_UTF16]; self.code_points.next().map(|code_point| { - let n = encode_utf16_raw(code_point.value, &mut buf).len(); + let n = encode_utf16_raw(code_point.to_u32(), &mut buf).len(); if n == 2 { self.extra = buf[1]; } @@ -1046,21 +572,19 @@ impl Iterator for EncodeWide<'_> { } } +impl fmt::Debug for EncodeWide<'_> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + f.debug_struct("EncodeWide").finish_non_exhaustive() + } +} + #[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")] impl FusedIterator for EncodeWide<'_> {} impl Hash for CodePoint { #[inline] fn hash(&self, state: &mut H) { - self.value.hash(state) - } -} - -impl Hash for Wtf8Buf { - #[inline] - fn hash(&self, state: &mut H) { - state.write(&self.bytes); - 0xfeu8.hash(state) + self.0.hash(state) } } diff --git a/library/coretests/tests/lib.rs b/library/coretests/tests/lib.rs index d2281b1df2f..4415b614817 100644 --- a/library/coretests/tests/lib.rs +++ b/library/coretests/tests/lib.rs @@ -194,6 +194,7 @@ mod time; mod tuple; mod unicode; mod waker; +mod wtf8; /// Copied from `std::test_helpers::test_rng`, see that function for rationale. #[track_caller] diff --git a/library/coretests/tests/wtf8.rs b/library/coretests/tests/wtf8.rs new file mode 100644 index 00000000000..9f187e70630 --- /dev/null +++ b/library/coretests/tests/wtf8.rs @@ -0,0 +1 @@ +// All `wtf8` tests live in library/alloctests/tests/wtf8.rs diff --git a/library/std/src/lib.rs b/library/std/src/lib.rs index ab417b6c72f..ca76dcc5147 100644 --- a/library/std/src/lib.rs +++ b/library/std/src/lib.rs @@ -391,6 +391,7 @@ #![feature(try_with_capacity)] #![feature(unique_rc_arc)] #![feature(vec_into_raw_parts)] +#![feature(wtf8_internals)] // tidy-alphabetical-end // // Library features (unwind): diff --git a/library/std/src/os/windows/ffi.rs b/library/std/src/os/windows/ffi.rs index 496443dbbc3..345d5b74285 100644 --- a/library/std/src/os/windows/ffi.rs +++ b/library/std/src/os/windows/ffi.rs @@ -53,12 +53,13 @@ #![stable(feature = "rust1", since = "1.0.0")] +use alloc::wtf8::Wtf8Buf; + use crate::ffi::{OsStr, OsString}; +use crate::fmt; +use crate::iter::FusedIterator; use crate::sealed::Sealed; use crate::sys::os_str::Buf; -#[stable(feature = "rust1", since = "1.0.0")] -pub use crate::sys_common::wtf8::EncodeWide; -use crate::sys_common::wtf8::Wtf8Buf; use crate::sys_common::{AsInner, FromInner}; /// Windows-specific extensions to [`OsString`]. @@ -130,6 +131,35 @@ pub trait OsStrExt: Sealed { impl OsStrExt for OsStr { #[inline] fn encode_wide(&self) -> EncodeWide<'_> { - self.as_inner().inner.encode_wide() + EncodeWide { inner: self.as_inner().inner.encode_wide() } + } +} + +/// Iterator returned by [`OsStrExt::encode_wide`]. +#[stable(feature = "rust1", since = "1.0.0")] +#[derive(Clone)] +pub struct EncodeWide<'a> { + inner: alloc::wtf8::EncodeWide<'a>, +} +#[stable(feature = "encode_wide_debug", since = "CURRENT_RUSTC_VERSION")] +impl fmt::Debug for EncodeWide<'_> { + fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { + fmt::Debug::fmt(&self.inner, f) + } +} +#[stable(feature = "rust1", since = "1.0.0")] +impl Iterator for EncodeWide<'_> { + type Item = u16; + + #[inline] + fn next(&mut self) -> Option { + self.inner.next() + } + + #[inline] + fn size_hint(&self) -> (usize, Option) { + self.inner.size_hint() } } +#[stable(feature = "encode_wide_fused_iterator", since = "1.62.0")] +impl FusedIterator for EncodeWide<'_> {} diff --git a/library/std/src/sys/os_str/wtf8.rs b/library/std/src/sys/os_str/wtf8.rs index bbc704ebf86..96da891874e 100644 --- a/library/std/src/sys/os_str/wtf8.rs +++ b/library/std/src/sys/os_str/wtf8.rs @@ -1,12 +1,12 @@ //! The underlying OsString/OsStr implementation on Windows is a //! wrapper around the "WTF-8" encoding; see the `wtf8` module for more. +use alloc::wtf8::{Wtf8, Wtf8Buf}; use core::clone::CloneToUninit; use crate::borrow::Cow; use crate::collections::TryReserveError; use crate::rc::Rc; use crate::sync::Arc; -use crate::sys_common::wtf8::{Wtf8, Wtf8Buf, check_utf8_boundary}; use crate::sys_common::{AsInner, FromInner, IntoInner}; use crate::{fmt, mem}; @@ -220,7 +220,9 @@ impl Buf { /// trailing surrogate half. #[inline] pub unsafe fn extend_from_slice_unchecked(&mut self, other: &[u8]) { - self.inner.extend_from_slice(other); + unsafe { + self.inner.extend_from_slice_unchecked(other); + } } } @@ -238,7 +240,7 @@ impl Slice { #[track_caller] #[inline] pub fn check_public_boundary(&self, index: usize) { - check_utf8_boundary(&self.inner, index); + self.inner.check_utf8_boundary(index); } #[inline] diff --git a/library/std/src/sys_common/mod.rs b/library/std/src/sys_common/mod.rs index 24b6cff1309..ec45c723e0d 100644 --- a/library/std/src/sys_common/mod.rs +++ b/library/std/src/sys_common/mod.rs @@ -21,7 +21,6 @@ mod tests; pub mod wstr; -pub mod wtf8; // common error constructors diff --git a/src/tools/tidy/src/unit_tests.rs b/src/tools/tidy/src/unit_tests.rs index 3d14a467319..7396310ed37 100644 --- a/src/tools/tidy/src/unit_tests.rs +++ b/src/tools/tidy/src/unit_tests.rs @@ -61,6 +61,7 @@ pub fn check(root_path: &Path, stdlib: bool, bad: &mut bool) { || path.ends_with("library/alloc/src/collections/linked_list/tests.rs") || path.ends_with("library/alloc/src/collections/vec_deque/tests.rs") || path.ends_with("library/alloc/src/raw_vec/tests.rs") + || path.ends_with("library/alloc/src/wtf8/tests.rs") } }; -- cgit 1.4.1-3-g733a5 From 7c81a067ea4cfd289d30c3903ac60b113f481c87 Mon Sep 17 00:00:00 2001 From: ltdk Date: Wed, 13 Aug 2025 18:54:48 -0400 Subject: Diff-massaging commit --- library/alloc/src/wtf8/mod.rs | 75 ++++++++++++++++++--------------------- library/core/src/wtf8.rs | 82 +++++++++++++++++++------------------------ 2 files changed, 70 insertions(+), 87 deletions(-) diff --git a/library/alloc/src/wtf8/mod.rs b/library/alloc/src/wtf8/mod.rs index 95d317a5efb..047994adc44 100644 --- a/library/alloc/src/wtf8/mod.rs +++ b/library/alloc/src/wtf8/mod.rs @@ -451,53 +451,46 @@ impl Extend for Wtf8Buf { } } -// helps diff -mod wtf8 { - use super::*; - - /// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. - pub(super) fn to_owned(slice: &Wtf8) -> Wtf8Buf { - Wtf8Buf { bytes: slice.as_bytes().to_vec(), is_known_utf8: false } - } +/// Creates an owned `Wtf8Buf` from a borrowed `Wtf8`. +pub(super) fn to_owned(slice: &Wtf8) -> Wtf8Buf { + Wtf8Buf { bytes: slice.as_bytes().to_vec(), is_known_utf8: false } +} - /// Lossily converts the string to UTF-8. - /// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. - /// - /// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). - /// - /// This only copies the data if necessary (if it contains any surrogate). - pub(super) fn to_string_lossy(slice: &Wtf8) -> Cow<'_, str> { - let Some((surrogate_pos, _)) = slice.next_surrogate(0) else { - return Cow::Borrowed(unsafe { str::from_utf8_unchecked(slice.as_bytes()) }); - }; - let wtf8_bytes = slice.as_bytes(); - let mut utf8_bytes = Vec::with_capacity(slice.len()); - utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); - utf8_bytes.extend_from_slice("\u{FFFD}".as_bytes()); - let mut pos = surrogate_pos + 3; - loop { - match slice.next_surrogate(pos) { - Some((surrogate_pos, _)) => { - utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); - utf8_bytes.extend_from_slice("\u{FFFD}".as_bytes()); - pos = surrogate_pos + 3; - } - None => { - utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); - return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); - } +/// Lossily converts the string to UTF-8. +/// Returns a UTF-8 `&str` slice if the contents are well-formed in UTF-8. +/// +/// Surrogates are replaced with `"\u{FFFD}"` (the replacement character “�”). +/// +/// This only copies the data if necessary (if it contains any surrogate). +pub(super) fn to_string_lossy(slice: &Wtf8) -> Cow<'_, str> { + let Some((surrogate_pos, _)) = slice.next_surrogate(0) else { + return Cow::Borrowed(unsafe { str::from_utf8_unchecked(slice.as_bytes()) }); + }; + let wtf8_bytes = slice.as_bytes(); + let mut utf8_bytes = Vec::with_capacity(slice.len()); + utf8_bytes.extend_from_slice(&wtf8_bytes[..surrogate_pos]); + utf8_bytes.extend_from_slice("\u{FFFD}".as_bytes()); + let mut pos = surrogate_pos + 3; + loop { + match slice.next_surrogate(pos) { + Some((surrogate_pos, _)) => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..surrogate_pos]); + utf8_bytes.extend_from_slice("\u{FFFD}".as_bytes()); + pos = surrogate_pos + 3; + } + None => { + utf8_bytes.extend_from_slice(&wtf8_bytes[pos..]); + return Cow::Owned(unsafe { String::from_utf8_unchecked(utf8_bytes) }); } } } - - #[inline] - pub(super) fn clone_into(slice: &Wtf8, buf: &mut Wtf8Buf) { - buf.is_known_utf8 = false; - slice.as_bytes().clone_into(&mut buf.bytes); - } } -use self::wtf8::{to_owned, to_string_lossy, clone_into}; +#[inline] +pub(super) fn clone_into(slice: &Wtf8, buf: &mut Wtf8Buf) { + buf.is_known_utf8 = false; + slice.as_bytes().clone_into(&mut buf.bytes); +} #[cfg(not(test))] impl Wtf8 { diff --git a/library/core/src/wtf8.rs b/library/core/src/wtf8.rs index 5631993dea2..de0dfa560a3 100644 --- a/library/core/src/wtf8.rs +++ b/library/core/src/wtf8.rs @@ -345,16 +345,6 @@ impl Wtf8 { pub fn eq_ignore_ascii_case(&self, other: &Self) -> bool { self.bytes.eq_ignore_ascii_case(&other.bytes) } - - #[inline] - pub fn is_code_point_boundary(&self, index: usize) -> bool { - is_code_point_boundary(self, index) - } - - #[inline] - pub fn check_utf8_boundary(&self, index: usize) { - check_utf8_boundary(self, index) - } } /// Returns a slice of the given string for the byte range \[`begin`..`end`). @@ -435,44 +425,44 @@ fn decode_surrogate(second_byte: u8, third_byte: u8) -> u16 { 0xD800 | (second_byte as u16 & 0x3F) << 6 | third_byte as u16 & 0x3F } -// helps diff to be unindented - -/// Copied from str::is_char_boundary -#[inline] -pub fn is_code_point_boundary(slice: &Wtf8, index: usize) -> bool { - if index == 0 { - return true; - } - match slice.bytes.get(index) { - None => index == slice.len(), - Some(&b) => (b as i8) >= -0x40, +impl Wtf8 { + /// Copied from str::is_char_boundary + #[inline] + pub fn is_code_point_boundary(&self, index: usize) -> bool { + if index == 0 { + return true; + } + match self.bytes.get(index) { + None => index == self.len(), + Some(&b) => (b as i8) >= -0x40, + } } -} -/// Verify that `index` is at the edge of either a valid UTF-8 codepoint -/// (i.e. a codepoint that's not a surrogate) or of the whole string. -/// -/// These are the cases currently permitted by `OsStr::slice_encoded_bytes`. -/// Splitting between surrogates is valid as far as WTF-8 is concerned, but -/// we do not permit it in the public API because WTF-8 is considered an -/// implementation detail. -#[track_caller] -#[inline] -pub fn check_utf8_boundary(slice: &Wtf8, index: usize) { - if index == 0 { - return; - } - match slice.bytes.get(index) { - Some(0xED) => (), // Might be a surrogate - Some(&b) if (b as i8) >= -0x40 => return, - Some(_) => panic!("byte index {index} is not a codepoint boundary"), - None if index == slice.len() => return, - None => panic!("byte index {index} is out of bounds"), - } - if slice.bytes[index + 1] >= 0xA0 { - // There's a surrogate after index. Now check before index. - if index >= 3 && slice.bytes[index - 3] == 0xED && slice.bytes[index - 2] >= 0xA0 { - panic!("byte index {index} lies between surrogate codepoints"); + /// Verify that `index` is at the edge of either a valid UTF-8 codepoint + /// (i.e. a codepoint that's not a surrogate) or of the whole string. + /// + /// These are the cases currently permitted by `OsStr::self_encoded_bytes`. + /// Splitting between surrogates is valid as far as WTF-8 is concerned, but + /// we do not permit it in the public API because WTF-8 is considered an + /// implementation detail. + #[track_caller] + #[inline] + pub fn check_utf8_boundary(&self, index: usize) { + if index == 0 { + return; + } + match self.bytes.get(index) { + Some(0xED) => (), // Might be a surrogate + Some(&b) if (b as i8) >= -0x40 => return, + Some(_) => panic!("byte index {index} is not a codepoint boundary"), + None if index == self.len() => return, + None => panic!("byte index {index} is out of bounds"), + } + if self.bytes[index + 1] >= 0xA0 { + // There's a surrogate after index. Now check before index. + if index >= 3 && self.bytes[index - 3] == 0xED && self.bytes[index - 2] >= 0xA0 { + panic!("byte index {index} lies between surrogate codepoints"); + } } } } -- cgit 1.4.1-3-g733a5 From 220249eebb4cbd96fd77abb5878727f4ae4f6a6f Mon Sep 17 00:00:00 2001 From: yanglsh Date: Wed, 18 Jun 2025 21:45:47 +0800 Subject: fix: `redundant_closure` suggests wrongly with deref overload --- clippy_lints/src/eta_reduction.rs | 21 ++++++++++++++---- tests/ui/eta.fixed | 45 +++++++++++++++++++++++++++++++++++++++ tests/ui/eta.rs | 45 +++++++++++++++++++++++++++++++++++++++ tests/ui/eta.stderr | 28 +++++++++++++++++++----- 4 files changed, 130 insertions(+), 9 deletions(-) diff --git a/clippy_lints/src/eta_reduction.rs b/clippy_lints/src/eta_reduction.rs index 0eefc2f6109..2da1c2bad11 100644 --- a/clippy_lints/src/eta_reduction.rs +++ b/clippy_lints/src/eta_reduction.rs @@ -12,6 +12,7 @@ use rustc_hir::attrs::AttributeKind; use rustc_hir::{BindingMode, Expr, ExprKind, FnRetTy, GenericArgs, Param, PatKind, QPath, Safety, TyKind, find_attr}; use rustc_infer::infer::TyCtxtInferExt; use rustc_lint::{LateContext, LateLintPass}; +use rustc_middle::ty::adjustment::Adjust; use rustc_middle::ty::{ self, Binder, ClosureKind, FnSig, GenericArg, GenericArgKind, List, Region, Ty, TypeVisitableExt, TypeckResults, }; @@ -148,10 +149,9 @@ fn check_closure<'tcx>(cx: &LateContext<'tcx>, outer_receiver: Option<&Expr<'tcx { return; } - let callee_ty_adjusted = typeck - .expr_adjustments(callee) - .last() - .map_or(callee_ty, |a| a.target.peel_refs()); + + let callee_ty_adjustments = typeck.expr_adjustments(callee); + let callee_ty_adjusted = callee_ty_adjustments.last().map_or(callee_ty, |a| a.target); let sig = match callee_ty_adjusted.kind() { ty::FnDef(def, _) => { @@ -230,7 +230,20 @@ fn check_closure<'tcx>(cx: &LateContext<'tcx>, outer_receiver: Option<&Expr<'tcx }, _ => (), } + } else if let n_refs = + callee_ty_adjustments + .iter() + .rev() + .fold(0, |acc, adjustment| match adjustment.kind { + Adjust::Deref(Some(_)) => acc + 1, + Adjust::Deref(_) if acc > 0 => acc + 1, + _ => acc, + }) + && n_refs > 0 + { + snippet = format!("{}{snippet}", "*".repeat(n_refs)); } + let replace_with = match callee_ty_adjusted.kind() { ty::FnDef(def, _) => cx.tcx.def_descr(*def), _ => "function", diff --git a/tests/ui/eta.fixed b/tests/ui/eta.fixed index 3d2b41b8fb8..6944a979c05 100644 --- a/tests/ui/eta.fixed +++ b/tests/ui/eta.fixed @@ -565,6 +565,51 @@ fn issue_14789() { ); } +fn issue_15072() { + use std::ops::Deref; + + struct Foo; + impl Deref for Foo { + type Target = fn() -> &'static str; + + fn deref(&self) -> &Self::Target { + fn hello() -> &'static str { + "Hello, world!" + } + &(hello as fn() -> &'static str) + } + } + + fn accepts_fn(f: impl Fn() -> &'static str) { + println!("{}", f()); + } + + fn some_fn() -> &'static str { + todo!() + } + + let f = &Foo; + accepts_fn(**f); + //~^ redundant_closure + + let g = &some_fn; + accepts_fn(g); + //~^ redundant_closure + + struct Bar(Foo); + impl Deref for Bar { + type Target = Foo; + + fn deref(&self) -> &Self::Target { + &self.0 + } + } + + let b = &Bar(Foo); + accepts_fn(***b); + //~^ redundant_closure +} + fn issue8817() { fn f(_: u32) -> u32 { todo!() diff --git a/tests/ui/eta.rs b/tests/ui/eta.rs index 79d1103410d..5bcc1cb26fd 100644 --- a/tests/ui/eta.rs +++ b/tests/ui/eta.rs @@ -565,6 +565,51 @@ fn issue_14789() { ); } +fn issue_15072() { + use std::ops::Deref; + + struct Foo; + impl Deref for Foo { + type Target = fn() -> &'static str; + + fn deref(&self) -> &Self::Target { + fn hello() -> &'static str { + "Hello, world!" + } + &(hello as fn() -> &'static str) + } + } + + fn accepts_fn(f: impl Fn() -> &'static str) { + println!("{}", f()); + } + + fn some_fn() -> &'static str { + todo!() + } + + let f = &Foo; + accepts_fn(|| f()); + //~^ redundant_closure + + let g = &some_fn; + accepts_fn(|| g()); + //~^ redundant_closure + + struct Bar(Foo); + impl Deref for Bar { + type Target = Foo; + + fn deref(&self) -> &Self::Target { + &self.0 + } + } + + let b = &Bar(Foo); + accepts_fn(|| b()); + //~^ redundant_closure +} + fn issue8817() { fn f(_: u32) -> u32 { todo!() diff --git a/tests/ui/eta.stderr b/tests/ui/eta.stderr index aa32ed1a38e..0b401cdea98 100644 --- a/tests/ui/eta.stderr +++ b/tests/ui/eta.stderr @@ -215,28 +215,46 @@ LL | let _field = bind.or_else(|| get_default()).unwrap(); | ^^^^^^^^^^^^^^^^ help: replace the closure with the function itself: `get_default` error: redundant closure - --> tests/ui/eta.rs:588:14 + --> tests/ui/eta.rs:592:16 + | +LL | accepts_fn(|| f()); + | ^^^^^^ help: replace the closure with the function itself: `**f` + +error: redundant closure + --> tests/ui/eta.rs:596:16 + | +LL | accepts_fn(|| g()); + | ^^^^^^ help: replace the closure with the function itself: `g` + +error: redundant closure + --> tests/ui/eta.rs:609:16 + | +LL | accepts_fn(|| b()); + | ^^^^^^ help: replace the closure with the function itself: `***b` + +error: redundant closure + --> tests/ui/eta.rs:633:14 | LL | .map(|n| MyError::A(n)) | ^^^^^^^^^^^^^^^^^ help: replace the closure with the tuple variant itself: `MyError::A` error: redundant closure - --> tests/ui/eta.rs:585:14 + --> tests/ui/eta.rs:630:14 | LL | .map(|n| S(n)) | ^^^^^^^^ help: replace the closure with the tuple struct itself: `S` error: redundant closure - --> tests/ui/eta.rs:582:14 + --> tests/ui/eta.rs:627:14 | LL | .map(|n| g(n)) | ^^^^^^^^ help: replace the closure with the function itself: `g` error: redundant closure - --> tests/ui/eta.rs:579:14 + --> tests/ui/eta.rs:624:14 | LL | .map(|n| f(n)) | ^^^^^^^^ help: replace the closure with the function itself: `f` -error: aborting due to 39 previous errors +error: aborting due to 42 previous errors -- cgit 1.4.1-3-g733a5 From f11d51273ccb540154748d9668c069bb8e50de3a Mon Sep 17 00:00:00 2001 From: yanglsh Date: Wed, 13 Aug 2025 22:30:35 +0800 Subject: fix: `semicolon_inside_block` FP when attribute over expr is not enabled --- clippy_lints/src/semicolon_block.rs | 5 +++++ tests/ui/semicolon_inside_block.fixed | 5 +++++ tests/ui/semicolon_inside_block.rs | 5 +++++ tests/ui/semicolon_inside_block_stmt_expr_attrs.fixed | 11 +++++++++++ tests/ui/semicolon_inside_block_stmt_expr_attrs.rs | 11 +++++++++++ tests/ui/semicolon_inside_block_stmt_expr_attrs.stderr | 16 ++++++++++++++++ 6 files changed, 53 insertions(+) create mode 100644 tests/ui/semicolon_inside_block_stmt_expr_attrs.fixed create mode 100644 tests/ui/semicolon_inside_block_stmt_expr_attrs.rs create mode 100644 tests/ui/semicolon_inside_block_stmt_expr_attrs.stderr diff --git a/clippy_lints/src/semicolon_block.rs b/clippy_lints/src/semicolon_block.rs index db91c57b181..1dea8f17c34 100644 --- a/clippy_lints/src/semicolon_block.rs +++ b/clippy_lints/src/semicolon_block.rs @@ -155,6 +155,11 @@ impl LateLintPass<'_> for SemicolonBlock { kind: ExprKind::Block(block, _), .. }) if !block.span.from_expansion() => { + let attrs = cx.tcx.hir_attrs(stmt.hir_id); + if !attrs.is_empty() && !cx.tcx.features().stmt_expr_attributes() { + return; + } + if let Some(tail) = block.expr { self.semicolon_inside_block(cx, block, tail, stmt.span); } diff --git a/tests/ui/semicolon_inside_block.fixed b/tests/ui/semicolon_inside_block.fixed index 7eb53e733ad..7308e78aae2 100644 --- a/tests/ui/semicolon_inside_block.fixed +++ b/tests/ui/semicolon_inside_block.fixed @@ -86,3 +86,8 @@ fn main() { unit_fn_block() } + +pub fn issue15388() { + #[rustfmt::skip] + {0; 0}; +} diff --git a/tests/ui/semicolon_inside_block.rs b/tests/ui/semicolon_inside_block.rs index 9fa5b117194..467bf4d779f 100644 --- a/tests/ui/semicolon_inside_block.rs +++ b/tests/ui/semicolon_inside_block.rs @@ -86,3 +86,8 @@ fn main() { unit_fn_block() } + +pub fn issue15388() { + #[rustfmt::skip] + {0; 0}; +} diff --git a/tests/ui/semicolon_inside_block_stmt_expr_attrs.fixed b/tests/ui/semicolon_inside_block_stmt_expr_attrs.fixed new file mode 100644 index 00000000000..5b93a91da00 --- /dev/null +++ b/tests/ui/semicolon_inside_block_stmt_expr_attrs.fixed @@ -0,0 +1,11 @@ +// Test when the feature `stmt_expr_attributes` is enabled + +#![feature(stmt_expr_attributes)] +#![allow(clippy::no_effect)] +#![warn(clippy::semicolon_inside_block)] + +pub fn issue15388() { + #[rustfmt::skip] + {0; 0;} + //~^ semicolon_inside_block +} diff --git a/tests/ui/semicolon_inside_block_stmt_expr_attrs.rs b/tests/ui/semicolon_inside_block_stmt_expr_attrs.rs new file mode 100644 index 00000000000..aa2c0cb5029 --- /dev/null +++ b/tests/ui/semicolon_inside_block_stmt_expr_attrs.rs @@ -0,0 +1,11 @@ +// Test when the feature `stmt_expr_attributes` is enabled + +#![feature(stmt_expr_attributes)] +#![allow(clippy::no_effect)] +#![warn(clippy::semicolon_inside_block)] + +pub fn issue15388() { + #[rustfmt::skip] + {0; 0}; + //~^ semicolon_inside_block +} diff --git a/tests/ui/semicolon_inside_block_stmt_expr_attrs.stderr b/tests/ui/semicolon_inside_block_stmt_expr_attrs.stderr new file mode 100644 index 00000000000..5bb91915a5f --- /dev/null +++ b/tests/ui/semicolon_inside_block_stmt_expr_attrs.stderr @@ -0,0 +1,16 @@ +error: consider moving the `;` inside the block for consistent formatting + --> tests/ui/semicolon_inside_block_stmt_expr_attrs.rs:9:5 + | +LL | {0; 0}; + | ^^^^^^^ + | + = note: `-D clippy::semicolon-inside-block` implied by `-D warnings` + = help: to override `-D warnings` add `#[allow(clippy::semicolon_inside_block)]` +help: put the `;` here + | +LL - {0; 0}; +LL + {0; 0;} + | + +error: aborting due to 1 previous error + -- cgit 1.4.1-3-g733a5 From 07aff76dda8fe022052de2ad7624df7de3306529 Mon Sep 17 00:00:00 2001 From: The Miri Cronjob Bot Date: Thu, 21 Aug 2025 04:53:54 +0000 Subject: Prepare for merging from rust-lang/rust This updates the rust-version file to 125ff8a788c5d6a66917f499abdc00051afe6886. --- src/tools/miri/rust-version | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/tools/miri/rust-version b/src/tools/miri/rust-version index 59adc572eaa..85ce9ed79f4 100644 --- a/src/tools/miri/rust-version +++ b/src/tools/miri/rust-version @@ -1 +1 @@ -f605b57042ffeb320d7ae44490113a827139b766 +125ff8a788c5d6a66917f499abdc00051afe6886 -- cgit 1.4.1-3-g733a5 From 6d637dfeccd7d22523e584203e666535db878165 Mon Sep 17 00:00:00 2001 From: Andrew Zhogin Date: Thu, 20 Mar 2025 12:43:44 +0700 Subject: -Zsanitize and -Zsanitizer-cfi-normalize-integers flags are now target modifiers with custom consistency check function --- compiler/rustc_metadata/src/creader.rs | 18 ++++-- compiler/rustc_session/src/options.rs | 71 +++++++++++++++++++++- .../address-sanitizer-globals-tracking.rs | 2 +- .../cfi/add-canonical-jump-tables-flag.rs | 2 +- .../cfi/add-cfi-normalize-integers-flag.rs | 2 +- .../cfi/add-enable-split-lto-unit-flag.rs | 2 +- .../sanitizer/cfi/dbg-location-on-cfi-blocks.rs | 2 +- .../cfi/emit-type-checks-attr-sanitize-off.rs | 2 +- .../codegen-llvm/sanitizer/cfi/emit-type-checks.rs | 2 +- .../cfi/emit-type-metadata-attr-cfi-encoding.rs | 2 +- ...e-metadata-id-itanium-cxx-abi-const-generics.rs | 2 +- ...pe-metadata-id-itanium-cxx-abi-drop-in-place.rs | 2 +- ...e-metadata-id-itanium-cxx-abi-function-types.rs | 2 +- ...t-type-metadata-id-itanium-cxx-abi-lifetimes.rs | 2 +- ...a-id-itanium-cxx-abi-method-secondary-typeid.rs | 2 +- .../emit-type-metadata-id-itanium-cxx-abi-paths.rs | 2 +- ...pe-metadata-id-itanium-cxx-abi-pointer-types.rs | 2 +- ...-metadata-id-itanium-cxx-abi-primitive-types.rs | 2 +- ...ta-id-itanium-cxx-abi-repr-transparent-types.rs | 2 +- ...e-metadata-id-itanium-cxx-abi-sequence-types.rs | 2 +- ...type-metadata-id-itanium-cxx-abi-trait-types.rs | 2 +- ...tadata-id-itanium-cxx-abi-user-defined-types.rs | 2 +- ...it-type-metadata-itanium-cxx-abi-generalized.rs | 2 +- ...adata-itanium-cxx-abi-normalized-generalized.rs | 2 +- ...mit-type-metadata-itanium-cxx-abi-normalized.rs | 2 +- .../cfi/emit-type-metadata-itanium-cxx-abi.rs | 2 +- .../cfi/emit-type-metadata-trait-objects.rs | 2 +- .../sanitizer/cfi/external_weak_symbols.rs | 2 +- .../sanitizer/cfi/generalize-pointers.rs | 2 +- .../sanitizer/cfi/normalize-integers.rs | 2 +- .../sanitizer/dataflow-instrument-functions.rs | 2 +- .../codegen-llvm/sanitizer/memory-track-origins.rs | 2 +- tests/codegen-llvm/sanitizer/memtag-attr-check.rs | 2 +- .../codegen-llvm/sanitizer/safestack-attr-check.rs | 2 +- .../sanitizer/sanitize-off-inlining.rs | 2 +- .../sanitizer/sanitize-off-kasan-asan.rs | 2 +- tests/codegen-llvm/sanitizer/sanitizer-recover.rs | 2 +- tests/run-make/sanitizer-cdylib-link/rmake.rs | 2 + tests/run-make/sanitizer-dylib-link/rmake.rs | 2 + tests/run-make/sanitizer-staticlib-link/rmake.rs | 2 + tests/rustdoc/sanitizer-option.rs | 2 +- tests/ui/asm/global-asm-isnt-really-a-mir-body.rs | 2 + tests/ui/lto/issue-100772.rs | 2 +- tests/ui/sanitizer/address.rs | 2 +- tests/ui/sanitizer/asan_odr_windows.rs | 2 +- tests/ui/sanitizer/badfree.rs | 2 +- .../cfi/assoc-ty-lifetime-issue-123053.rs | 2 +- tests/ui/sanitizer/cfi/async-closures.rs | 2 +- tests/ui/sanitizer/cfi/can-reveal-opaques.rs | 2 +- tests/ui/sanitizer/cfi/closures.rs | 2 +- tests/ui/sanitizer/cfi/complex-receiver.rs | 2 +- tests/ui/sanitizer/cfi/coroutine.rs | 1 + tests/ui/sanitizer/cfi/drop-in-place.rs | 1 + tests/ui/sanitizer/cfi/drop-no-principal.rs | 2 +- tests/ui/sanitizer/cfi/fn-ptr.rs | 1 + .../sanitizer/cfi/generalize-pointers-attr-cfg.rs | 1 + .../sanitizer/cfi/normalize-integers-attr-cfg.rs | 2 +- tests/ui/sanitizer/cfi/self-ref.rs | 2 +- tests/ui/sanitizer/cfi/sized-associated-ty.rs | 2 +- tests/ui/sanitizer/cfi/supertraits.rs | 2 +- tests/ui/sanitizer/cfi/transparent-has-regions.rs | 2 +- tests/ui/sanitizer/cfi/virtual-auto.rs | 2 +- tests/ui/sanitizer/dataflow.rs | 2 +- tests/ui/sanitizer/hwaddress.rs | 2 +- .../issue-111184-cfi-coroutine-witness.rs | 2 +- .../issue-114275-cfi-const-expr-in-arry-len.rs | 2 +- .../issue-72154-address-lifetime-markers.rs | 2 +- tests/ui/sanitizer/kcfi-mangling.rs | 2 +- tests/ui/sanitizer/leak.rs | 2 +- tests/ui/sanitizer/memory-eager.rs | 2 + tests/ui/sanitizer/memory-passing.rs | 2 + tests/ui/sanitizer/memory.rs | 2 + .../ui/sanitizer/new-llvm-pass-manager-thin-lto.rs | 2 + tests/ui/sanitizer/thread.rs | 2 +- tests/ui/sanitizer/use-after-scope.rs | 2 +- .../auxiliary/kcfi-normalize-ints.rs | 7 +++ .../ui/target_modifiers/auxiliary/no-sanitizers.rs | 6 ++ .../auxiliary/safestack-and-kcfi.rs | 10 +++ .../sanitizer-kcfi-normalize-ints.rs | 18 ++++++ ...sanitizer-kcfi-normalize-ints.wrong_flag.stderr | 13 ++++ ...izer-kcfi-normalize-ints.wrong_sanitizer.stderr | 13 ++++ .../sanitizers-good-for-inconsistency.rs | 19 ++++++ ...anitizers-safestack-and-kcfi.missed_both.stderr | 13 ++++ ...anitizers-safestack-and-kcfi.missed_kcfi.stderr | 13 ++++ ...zers-safestack-and-kcfi.missed_safestack.stderr | 13 ++++ .../sanitizers-safestack-and-kcfi.rs | 23 +++++++ 86 files changed, 311 insertions(+), 68 deletions(-) create mode 100644 tests/ui/target_modifiers/auxiliary/kcfi-normalize-ints.rs create mode 100644 tests/ui/target_modifiers/auxiliary/no-sanitizers.rs create mode 100644 tests/ui/target_modifiers/auxiliary/safestack-and-kcfi.rs create mode 100644 tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.rs create mode 100644 tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_flag.stderr create mode 100644 tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_sanitizer.stderr create mode 100644 tests/ui/target_modifiers/sanitizers-good-for-inconsistency.rs create mode 100644 tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_both.stderr create mode 100644 tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_kcfi.stderr create mode 100644 tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_safestack.stderr create mode 100644 tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.rs diff --git a/compiler/rustc_metadata/src/creader.rs b/compiler/rustc_metadata/src/creader.rs index 5d776ea581d..9e23da88f5e 100644 --- a/compiler/rustc_metadata/src/creader.rs +++ b/compiler/rustc_metadata/src/creader.rs @@ -412,7 +412,7 @@ impl CStore { match (&left_name_val, &right_name_val) { (Some(l), Some(r)) => match l.1.opt.cmp(&r.1.opt) { cmp::Ordering::Equal => { - if l.0.tech_value != r.0.tech_value { + if !l.1.consistent(&tcx.sess.opts, Some(&r.1)) { report_diff( &l.0.prefix, &l.0.name, @@ -424,20 +424,28 @@ impl CStore { right_name_val = None; } cmp::Ordering::Greater => { - report_diff(&r.0.prefix, &r.0.name, None, Some(&r.1.value_name)); + if !r.1.consistent(&tcx.sess.opts, None) { + report_diff(&r.0.prefix, &r.0.name, None, Some(&r.1.value_name)); + } right_name_val = None; } cmp::Ordering::Less => { - report_diff(&l.0.prefix, &l.0.name, Some(&l.1.value_name), None); + if !l.1.consistent(&tcx.sess.opts, None) { + report_diff(&l.0.prefix, &l.0.name, Some(&l.1.value_name), None); + } left_name_val = None; } }, (Some(l), None) => { - report_diff(&l.0.prefix, &l.0.name, Some(&l.1.value_name), None); + if !l.1.consistent(&tcx.sess.opts, None) { + report_diff(&l.0.prefix, &l.0.name, Some(&l.1.value_name), None); + } left_name_val = None; } (None, Some(r)) => { - report_diff(&r.0.prefix, &r.0.name, None, Some(&r.1.value_name)); + if !r.1.consistent(&tcx.sess.opts, None) { + report_diff(&r.0.prefix, &r.0.name, None, Some(&r.1.value_name)); + } right_name_val = None; } (None, None) => break, diff --git a/compiler/rustc_session/src/options.rs b/compiler/rustc_session/src/options.rs index 6d5be2d92cd..69facde6936 100644 --- a/compiler/rustc_session/src/options.rs +++ b/compiler/rustc_session/src/options.rs @@ -83,10 +83,77 @@ pub struct TargetModifier { pub value_name: String, } +mod target_modifier_consistency_check { + use super::*; + pub(super) fn sanitizer(l: &TargetModifier, r: Option<&TargetModifier>) -> bool { + let mut lparsed: SanitizerSet = Default::default(); + let lval = if l.value_name.is_empty() { None } else { Some(l.value_name.as_str()) }; + parse::parse_sanitizers(&mut lparsed, lval); + + let mut rparsed: SanitizerSet = Default::default(); + let rval = r.filter(|v| !v.value_name.is_empty()).map(|v| v.value_name.as_str()); + parse::parse_sanitizers(&mut rparsed, rval); + + // Some sanitizers need to be target modifiers, and some do not. + // For now, we should mark all sanitizers as target modifiers except for these: + // AddressSanitizer, LeakSanitizer + let tmod_sanitizers = SanitizerSet::MEMORY + | SanitizerSet::THREAD + | SanitizerSet::HWADDRESS + | SanitizerSet::CFI + | SanitizerSet::MEMTAG + | SanitizerSet::SHADOWCALLSTACK + | SanitizerSet::KCFI + | SanitizerSet::KERNELADDRESS + | SanitizerSet::SAFESTACK + | SanitizerSet::DATAFLOW; + + lparsed & tmod_sanitizers == rparsed & tmod_sanitizers + } + pub(super) fn sanitizer_cfi_normalize_integers( + opts: &Options, + l: &TargetModifier, + r: Option<&TargetModifier>, + ) -> bool { + // For kCFI, the helper flag -Zsanitizer-cfi-normalize-integers should also be a target modifier + if opts.unstable_opts.sanitizer.contains(SanitizerSet::KCFI) { + if let Some(r) = r { + return l.extend().tech_value == r.extend().tech_value; + } else { + return false; + } + } + true + } +} + impl TargetModifier { pub fn extend(&self) -> ExtendedTargetModifierInfo { self.opt.reparse(&self.value_name) } + // Custom consistency check for target modifiers (or default `l.tech_value == r.tech_value`) + // When other is None, consistency with default value is checked + pub fn consistent(&self, opts: &Options, other: Option<&TargetModifier>) -> bool { + assert!(other.is_none() || self.opt == other.unwrap().opt); + match self.opt { + OptionsTargetModifiers::UnstableOptions(unstable) => match unstable { + UnstableOptionsTargetModifiers::sanitizer => { + return target_modifier_consistency_check::sanitizer(self, other); + } + UnstableOptionsTargetModifiers::sanitizer_cfi_normalize_integers => { + return target_modifier_consistency_check::sanitizer_cfi_normalize_integers( + opts, self, other, + ); + } + _ => {} + }, + _ => {} + }; + match other { + Some(other) => self.extend().tech_value == other.extend().tech_value, + None => false, + } + } } fn tmod_push_impl( @@ -2504,13 +2571,13 @@ written to standard error output)"), retpoline_external_thunk: bool = (false, parse_bool, [TRACKED TARGET_MODIFIER], "enables retpoline-external-thunk, retpoline-indirect-branches and retpoline-indirect-calls \ target features (default: no)"), - sanitizer: SanitizerSet = (SanitizerSet::empty(), parse_sanitizers, [TRACKED], + sanitizer: SanitizerSet = (SanitizerSet::empty(), parse_sanitizers, [TRACKED TARGET_MODIFIER], "use a sanitizer"), sanitizer_cfi_canonical_jump_tables: Option = (Some(true), parse_opt_bool, [TRACKED], "enable canonical jump tables (default: yes)"), sanitizer_cfi_generalize_pointers: Option = (None, parse_opt_bool, [TRACKED], "enable generalizing pointer types (default: no)"), - sanitizer_cfi_normalize_integers: Option = (None, parse_opt_bool, [TRACKED], + sanitizer_cfi_normalize_integers: Option = (None, parse_opt_bool, [TRACKED TARGET_MODIFIER], "enable normalizing integer types (default: no)"), sanitizer_dataflow_abilist: Vec = (Vec::new(), parse_comma_list, [TRACKED], "additional ABI list files that control how shadow parameters are passed (comma separated)"), diff --git a/tests/codegen-llvm/sanitizer/address-sanitizer-globals-tracking.rs b/tests/codegen-llvm/sanitizer/address-sanitizer-globals-tracking.rs index 642bf5e7576..ada525b6c80 100644 --- a/tests/codegen-llvm/sanitizer/address-sanitizer-globals-tracking.rs +++ b/tests/codegen-llvm/sanitizer/address-sanitizer-globals-tracking.rs @@ -19,7 +19,7 @@ //@ only-linux // //@ revisions:ASAN ASAN-FAT-LTO -//@ compile-flags: -Zsanitizer=address -Ctarget-feature=-crt-static +//@ compile-flags: -Zsanitizer=address -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer // [ASAN] no extra compile-flags //@[ASAN-FAT-LTO] compile-flags: -Cprefer-dynamic=false -Clto=fat diff --git a/tests/codegen-llvm/sanitizer/cfi/add-canonical-jump-tables-flag.rs b/tests/codegen-llvm/sanitizer/cfi/add-canonical-jump-tables-flag.rs index 22577e2a3c4..77857ca4ccb 100644 --- a/tests/codegen-llvm/sanitizer/cfi/add-canonical-jump-tables-flag.rs +++ b/tests/codegen-llvm/sanitizer/cfi/add-canonical-jump-tables-flag.rs @@ -1,7 +1,7 @@ // Verifies that "CFI Canonical Jump Tables" module flag is added. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/add-cfi-normalize-integers-flag.rs b/tests/codegen-llvm/sanitizer/cfi/add-cfi-normalize-integers-flag.rs index a54a6d84a80..6cf9a72b748 100644 --- a/tests/codegen-llvm/sanitizer/cfi/add-cfi-normalize-integers-flag.rs +++ b/tests/codegen-llvm/sanitizer/cfi/add-cfi-normalize-integers-flag.rs @@ -1,7 +1,7 @@ // Verifies that "cfi-normalize-integers" module flag is added. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers +//@ compile-flags: -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -C unsafe-allow-abi-mismatch=sanitizer,sanitizer-cfi-normalize-integers #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/add-enable-split-lto-unit-flag.rs b/tests/codegen-llvm/sanitizer/cfi/add-enable-split-lto-unit-flag.rs index 283b8f26102..0bfdbfba5d2 100644 --- a/tests/codegen-llvm/sanitizer/cfi/add-enable-split-lto-unit-flag.rs +++ b/tests/codegen-llvm/sanitizer/cfi/add-enable-split-lto-unit-flag.rs @@ -1,7 +1,7 @@ // Verifies that "EnableSplitLTOUnit" module flag is added. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/dbg-location-on-cfi-blocks.rs b/tests/codegen-llvm/sanitizer/cfi/dbg-location-on-cfi-blocks.rs index df65960dfe0..2a18e30e2b0 100644 --- a/tests/codegen-llvm/sanitizer/cfi/dbg-location-on-cfi-blocks.rs +++ b/tests/codegen-llvm/sanitizer/cfi/dbg-location-on-cfi-blocks.rs @@ -1,7 +1,7 @@ // Verifies that the parent block's debug information are assigned to the inserted cfi block. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -Cdebuginfo=1 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -Cdebuginfo=1 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-checks-attr-sanitize-off.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-checks-attr-sanitize-off.rs index 651afb33228..c49438f4318 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-checks-attr-sanitize-off.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-checks-attr-sanitize-off.rs @@ -1,7 +1,7 @@ // Verifies that pointer type membership tests for indirect calls are omitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(sanitize)] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-checks.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-checks.rs index ebc66a015df..9cad88f6518 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-checks.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-checks.rs @@ -1,7 +1,7 @@ // Verifies that pointer type membership tests for indirect calls are emitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-attr-cfi-encoding.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-attr-cfi-encoding.rs index 9bc2e42db0f..cd9088f58af 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-attr-cfi-encoding.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-attr-cfi-encoding.rs @@ -1,7 +1,7 @@ // Verifies that user-defined CFI encoding for types are emitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(cfi_encoding, extern_types)] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-const-generics.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-const-generics.rs index 9048c6a1f18..cf26c17af1e 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-const-generics.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-const-generics.rs @@ -2,7 +2,7 @@ // for const generics. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(type_alias_impl_trait)] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-drop-in-place.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-drop-in-place.rs index 8fec275fd06..279350d20c5 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-drop-in-place.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-drop-in-place.rs @@ -5,7 +5,7 @@ // future. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-function-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-function-types.rs index 7e60aafff68..047b532e994 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-function-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-function-types.rs @@ -2,7 +2,7 @@ // for function types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-lifetimes.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-lifetimes.rs index 36d2e8c9f25..92b2ab32ea0 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-lifetimes.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-lifetimes.rs @@ -2,7 +2,7 @@ // for lifetimes/regions. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(type_alias_impl_trait)] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-method-secondary-typeid.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-method-secondary-typeid.rs index 9d611777ff0..5de39dc85c1 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-method-secondary-typeid.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-method-secondary-typeid.rs @@ -2,7 +2,7 @@ // self so they can be used as function pointers. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-paths.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-paths.rs index a8ba8db1be3..f5846713bd5 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-paths.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-paths.rs @@ -2,7 +2,7 @@ // for paths. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(type_alias_impl_trait)] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-pointer-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-pointer-types.rs index d37bb740f55..ad4fe11d087 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-pointer-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-pointer-types.rs @@ -2,7 +2,7 @@ // for pointer types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-primitive-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-primitive-types.rs index 7d9e4d05872..93845d05195 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-primitive-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-primitive-types.rs @@ -2,7 +2,7 @@ // for primitive types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-repr-transparent-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-repr-transparent-types.rs index 0f97c70f3f9..025aa902658 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-repr-transparent-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-repr-transparent-types.rs @@ -2,7 +2,7 @@ // for repr transparent types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-sequence-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-sequence-types.rs index bdee3f47a83..76c8150b778 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-sequence-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-sequence-types.rs @@ -2,7 +2,7 @@ // for sequence types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-trait-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-trait-types.rs index 55e816178f8..4fafdd2f040 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-trait-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-trait-types.rs @@ -2,7 +2,7 @@ // for trait types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-user-defined-types.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-user-defined-types.rs index c1f3ca61afe..91351096ca2 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-user-defined-types.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-id-itanium-cxx-abi-user-defined-types.rs @@ -2,7 +2,7 @@ // for user-defined types. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(extern_types)] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-generalized.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-generalized.rs index 32637b64b3e..22d518cca74 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-generalized.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-generalized.rs @@ -1,7 +1,7 @@ // Verifies that generalized type metadata for functions are emitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-generalize-pointers +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-generalize-pointers -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized-generalized.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized-generalized.rs index 51121b0aef1..7639ce7b104 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized-generalized.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized-generalized.rs @@ -1,7 +1,7 @@ // Verifies that normalized and generalized type metadata for functions are emitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -Zsanitizer-cfi-generalize-pointers +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -Zsanitizer-cfi-generalize-pointers -C unsafe-allow-abi-mismatch=sanitizer,sanitizer-cfi-normalize-integers #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized.rs index 1cfdd23006e..acd72b0ca3c 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi-normalized.rs @@ -1,7 +1,7 @@ // Verifies that normalized type metadata for functions are emitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -C unsafe-allow-abi-mismatch=sanitizer,sanitizer-cfi-normalize-integers #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi.rs index 56ab1ce4b35..fa5cd471466 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-itanium-cxx-abi.rs @@ -1,7 +1,7 @@ // Verifies that type metadata for functions are emitted. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-trait-objects.rs b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-trait-objects.rs index 0e57ce322d1..82873e935b2 100644 --- a/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-trait-objects.rs +++ b/tests/codegen-llvm/sanitizer/cfi/emit-type-metadata-trait-objects.rs @@ -1,7 +1,7 @@ // Verifies that type metadata identifiers for trait objects are emitted correctly. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Clto -Cno-prepopulate-passes -Copt-level=0 -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/external_weak_symbols.rs b/tests/codegen-llvm/sanitizer/cfi/external_weak_symbols.rs index b3cb6dfdd37..893b0167693 100644 --- a/tests/codegen-llvm/sanitizer/cfi/external_weak_symbols.rs +++ b/tests/codegen-llvm/sanitizer/cfi/external_weak_symbols.rs @@ -2,7 +2,7 @@ // emitted correctly. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clinker-plugin-lto -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Clinker-plugin-lto -Copt-level=0 -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "bin"] #![feature(linkage)] diff --git a/tests/codegen-llvm/sanitizer/cfi/generalize-pointers.rs b/tests/codegen-llvm/sanitizer/cfi/generalize-pointers.rs index 57004da6f8e..caa2f258f8f 100644 --- a/tests/codegen-llvm/sanitizer/cfi/generalize-pointers.rs +++ b/tests/codegen-llvm/sanitizer/cfi/generalize-pointers.rs @@ -1,7 +1,7 @@ // Verifies that pointer types are generalized. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-generalize-pointers -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-generalize-pointers -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/cfi/normalize-integers.rs b/tests/codegen-llvm/sanitizer/cfi/normalize-integers.rs index 770ee4e64e0..16f76adafb8 100644 --- a/tests/codegen-llvm/sanitizer/cfi/normalize-integers.rs +++ b/tests/codegen-llvm/sanitizer/cfi/normalize-integers.rs @@ -1,7 +1,7 @@ // Verifies that integer types are normalized. // //@ needs-sanitizer-cfi -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -Copt-level=0 +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer,sanitizer-cfi-normalize-integers #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/dataflow-instrument-functions.rs b/tests/codegen-llvm/sanitizer/dataflow-instrument-functions.rs index a2d0d63cc17..cd8c2c48ea4 100644 --- a/tests/codegen-llvm/sanitizer/dataflow-instrument-functions.rs +++ b/tests/codegen-llvm/sanitizer/dataflow-instrument-functions.rs @@ -1,7 +1,7 @@ // Verifies that functions are instrumented. // //@ needs-sanitizer-dataflow -//@ compile-flags: -Copt-level=0 -Zsanitizer=dataflow +//@ compile-flags: -Copt-level=0 -Zsanitizer=dataflow -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/memory-track-origins.rs b/tests/codegen-llvm/sanitizer/memory-track-origins.rs index 5eb5b356b05..a72e523c4e1 100644 --- a/tests/codegen-llvm/sanitizer/memory-track-origins.rs +++ b/tests/codegen-llvm/sanitizer/memory-track-origins.rs @@ -4,7 +4,7 @@ //@ needs-sanitizer-memory //@ revisions:MSAN-0 MSAN-1 MSAN-2 MSAN-1-LTO MSAN-2-LTO // -//@ compile-flags: -Zsanitizer=memory -Ctarget-feature=-crt-static +//@ compile-flags: -Zsanitizer=memory -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer // [MSAN-0] no extra compile-flags //@[MSAN-1] compile-flags: -Zsanitizer-memory-track-origins=1 //@[MSAN-2] compile-flags: -Zsanitizer-memory-track-origins diff --git a/tests/codegen-llvm/sanitizer/memtag-attr-check.rs b/tests/codegen-llvm/sanitizer/memtag-attr-check.rs index ffe3a2322a2..fc430f3a570 100644 --- a/tests/codegen-llvm/sanitizer/memtag-attr-check.rs +++ b/tests/codegen-llvm/sanitizer/memtag-attr-check.rs @@ -2,7 +2,7 @@ // applied when enabling the memtag sanitizer. // //@ needs-sanitizer-memtag -//@ compile-flags: -Zsanitizer=memtag -Ctarget-feature=+mte -Copt-level=0 +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer -Zsanitizer=memtag -Ctarget-feature=+mte -Copt-level=0 #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/safestack-attr-check.rs b/tests/codegen-llvm/sanitizer/safestack-attr-check.rs index 050a60333af..414dd89a580 100644 --- a/tests/codegen-llvm/sanitizer/safestack-attr-check.rs +++ b/tests/codegen-llvm/sanitizer/safestack-attr-check.rs @@ -1,7 +1,7 @@ // This tests that the safestack attribute is applied when enabling the safe-stack sanitizer. // //@ needs-sanitizer-safestack -//@ compile-flags: -Zsanitizer=safestack -Copt-level=0 +//@ compile-flags: -Zsanitizer=safestack -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] diff --git a/tests/codegen-llvm/sanitizer/sanitize-off-inlining.rs b/tests/codegen-llvm/sanitizer/sanitize-off-inlining.rs index 69771827c3a..0f43e6b8393 100644 --- a/tests/codegen-llvm/sanitizer/sanitize-off-inlining.rs +++ b/tests/codegen-llvm/sanitizer/sanitize-off-inlining.rs @@ -1,9 +1,9 @@ // Verifies that sanitize(xyz = "off") attribute prevents inlining when // given sanitizer is enabled, but has no effect on inlining otherwise. -// //@ needs-sanitizer-address //@ needs-sanitizer-leak //@ revisions: ASAN LSAN +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer //@ compile-flags: -Copt-level=3 -Zmir-opt-level=4 -Ctarget-feature=-crt-static //@[ASAN] compile-flags: -Zsanitizer=address //@[LSAN] compile-flags: -Zsanitizer=leak diff --git a/tests/codegen-llvm/sanitizer/sanitize-off-kasan-asan.rs b/tests/codegen-llvm/sanitizer/sanitize-off-kasan-asan.rs index 94945f2b2e6..61ad0ba7d90 100644 --- a/tests/codegen-llvm/sanitizer/sanitize-off-kasan-asan.rs +++ b/tests/codegen-llvm/sanitizer/sanitize-off-kasan-asan.rs @@ -2,7 +2,7 @@ // the address sanitizer. // //@ needs-sanitizer-address -//@ compile-flags: -Zsanitizer=address -Ctarget-feature=-crt-static -Copt-level=0 +//@ compile-flags: -Zsanitizer=address -Ctarget-feature=-crt-static -Copt-level=0 -C unsafe-allow-abi-mismatch=sanitizer #![crate_type = "lib"] #![feature(sanitize)] diff --git a/tests/codegen-llvm/sanitizer/sanitizer-recover.rs b/tests/codegen-llvm/sanitizer/sanitizer-recover.rs index 6b659320481..b8a24e31c30 100644 --- a/tests/codegen-llvm/sanitizer/sanitizer-recover.rs +++ b/tests/codegen-llvm/sanitizer/sanitizer-recover.rs @@ -5,7 +5,7 @@ //@ needs-sanitizer-memory //@ revisions:ASAN ASAN-RECOVER MSAN MSAN-RECOVER MSAN-RECOVER-LTO //@ no-prefer-dynamic -// +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer //@ compile-flags: -Ctarget-feature=-crt-static //@[ASAN] compile-flags: -Zsanitizer=address -Copt-level=0 //@[ASAN-RECOVER] compile-flags: -Zsanitizer=address -Zsanitizer-recover=address -Copt-level=0 diff --git a/tests/run-make/sanitizer-cdylib-link/rmake.rs b/tests/run-make/sanitizer-cdylib-link/rmake.rs index f9d7fd98789..8ff58594d10 100644 --- a/tests/run-make/sanitizer-cdylib-link/rmake.rs +++ b/tests/run-make/sanitizer-cdylib-link/rmake.rs @@ -8,6 +8,8 @@ //@ needs-sanitizer-support //@ needs-sanitizer-address +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer + use run_make_support::{run_fail, rustc}; fn main() { diff --git a/tests/run-make/sanitizer-dylib-link/rmake.rs b/tests/run-make/sanitizer-dylib-link/rmake.rs index b43420adc72..bae3240a236 100644 --- a/tests/run-make/sanitizer-dylib-link/rmake.rs +++ b/tests/run-make/sanitizer-dylib-link/rmake.rs @@ -7,6 +7,8 @@ //@ needs-sanitizer-support //@ needs-sanitizer-address +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer + use run_make_support::{run_fail, rustc}; fn main() { diff --git a/tests/run-make/sanitizer-staticlib-link/rmake.rs b/tests/run-make/sanitizer-staticlib-link/rmake.rs index e38d15a8a3c..dda92626189 100644 --- a/tests/run-make/sanitizer-staticlib-link/rmake.rs +++ b/tests/run-make/sanitizer-staticlib-link/rmake.rs @@ -11,6 +11,8 @@ //@ needs-sanitizer-support //@ needs-sanitizer-address +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer + use run_make_support::{cc, extra_c_flags, extra_cxx_flags, run_fail, rustc, static_lib_name}; fn main() { diff --git a/tests/rustdoc/sanitizer-option.rs b/tests/rustdoc/sanitizer-option.rs index 2adf1be51fd..7b0038138f0 100644 --- a/tests/rustdoc/sanitizer-option.rs +++ b/tests/rustdoc/sanitizer-option.rs @@ -1,6 +1,6 @@ //@ needs-sanitizer-support //@ needs-sanitizer-address -//@ compile-flags: --test -Z sanitizer=address +//@ compile-flags: --test -Z sanitizer=address -C unsafe-allow-abi-mismatch=sanitizer // // #43031: Verify that rustdoc passes `-Z` options to rustc. Use an extern // function that is provided by the sanitizer runtime, if flag is not passed diff --git a/tests/ui/asm/global-asm-isnt-really-a-mir-body.rs b/tests/ui/asm/global-asm-isnt-really-a-mir-body.rs index b7636d116ec..aef25d057d4 100644 --- a/tests/ui/asm/global-asm-isnt-really-a-mir-body.rs +++ b/tests/ui/asm/global-asm-isnt-really-a-mir-body.rs @@ -1,5 +1,7 @@ //@ revisions: emit_mir instrument cfi +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer + // Make sure we don't try to emit MIR for it. //@[emit_mir] compile-flags: --emit=mir diff --git a/tests/ui/lto/issue-100772.rs b/tests/ui/lto/issue-100772.rs index 29ec5b9bf96..9468e20894a 100644 --- a/tests/ui/lto/issue-100772.rs +++ b/tests/ui/lto/issue-100772.rs @@ -1,6 +1,6 @@ //@ build-pass //@ needs-sanitizer-cfi -//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer //@ no-prefer-dynamic //@ only-x86_64-unknown-linux-gnu diff --git a/tests/ui/sanitizer/address.rs b/tests/ui/sanitizer/address.rs index 704d84764c1..1688f46c2d5 100644 --- a/tests/ui/sanitizer/address.rs +++ b/tests/ui/sanitizer/address.rs @@ -2,7 +2,7 @@ //@ needs-sanitizer-address //@ ignore-cross-compile // -//@ compile-flags: -Z sanitizer=address -O -g +//@ compile-flags: -Z sanitizer=address -O -g -C unsafe-allow-abi-mismatch=sanitizer // //@ run-fail-or-crash //@ error-pattern: AddressSanitizer: stack-buffer-overflow diff --git a/tests/ui/sanitizer/asan_odr_windows.rs b/tests/ui/sanitizer/asan_odr_windows.rs index 28c24716761..b638d6eb969 100644 --- a/tests/ui/sanitizer/asan_odr_windows.rs +++ b/tests/ui/sanitizer/asan_odr_windows.rs @@ -2,7 +2,7 @@ //! See . //@ run-pass -//@ compile-flags:-Zsanitizer=address +//@ compile-flags:-Zsanitizer=address -C unsafe-allow-abi-mismatch=sanitizer //@ aux-build: asan_odr_win-2.rs //@ only-windows-msvc //@ needs-sanitizer-support diff --git a/tests/ui/sanitizer/badfree.rs b/tests/ui/sanitizer/badfree.rs index 6b3aea7239c..b1b02649dcc 100644 --- a/tests/ui/sanitizer/badfree.rs +++ b/tests/ui/sanitizer/badfree.rs @@ -2,7 +2,7 @@ //@ needs-sanitizer-address //@ ignore-cross-compile // -//@ compile-flags: -Z sanitizer=address -O +//@ compile-flags: -Z sanitizer=address -O -C unsafe-allow-abi-mismatch=sanitizer // //@ run-fail-or-crash //@ regex-error-pattern: AddressSanitizer: (SEGV|attempting free on address which was not malloc) diff --git a/tests/ui/sanitizer/cfi/assoc-ty-lifetime-issue-123053.rs b/tests/ui/sanitizer/cfi/assoc-ty-lifetime-issue-123053.rs index fad57198dfb..5de5e99f6d0 100644 --- a/tests/ui/sanitizer/cfi/assoc-ty-lifetime-issue-123053.rs +++ b/tests/ui/sanitizer/cfi/assoc-ty-lifetime-issue-123053.rs @@ -2,7 +2,7 @@ // trait object type to fail, causing an ICE. // //@ needs-sanitizer-cfi -//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer //@ edition: 2021 //@ no-prefer-dynamic //@ only-x86_64-unknown-linux-gnu diff --git a/tests/ui/sanitizer/cfi/async-closures.rs b/tests/ui/sanitizer/cfi/async-closures.rs index 9b099263000..621a0882c91 100644 --- a/tests/ui/sanitizer/cfi/async-closures.rs +++ b/tests/ui/sanitizer/cfi/async-closures.rs @@ -7,7 +7,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/can-reveal-opaques.rs b/tests/ui/sanitizer/cfi/can-reveal-opaques.rs index a881c6b92b2..310ce04c552 100644 --- a/tests/ui/sanitizer/cfi/can-reveal-opaques.rs +++ b/tests/ui/sanitizer/cfi/can-reveal-opaques.rs @@ -1,5 +1,5 @@ //@ needs-sanitizer-cfi -//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer //@ no-prefer-dynamic //@ only-x86_64-unknown-linux-gnu //@ ignore-backends: gcc diff --git a/tests/ui/sanitizer/cfi/closures.rs b/tests/ui/sanitizer/cfi/closures.rs index fc9718faa28..7493dba4928 100644 --- a/tests/ui/sanitizer/cfi/closures.rs +++ b/tests/ui/sanitizer/cfi/closures.rs @@ -6,7 +6,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/complex-receiver.rs b/tests/ui/sanitizer/cfi/complex-receiver.rs index aac44c33227..adacc0d6c5d 100644 --- a/tests/ui/sanitizer/cfi/complex-receiver.rs +++ b/tests/ui/sanitizer/cfi/complex-receiver.rs @@ -8,7 +8,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/coroutine.rs b/tests/ui/sanitizer/cfi/coroutine.rs index 084a521f597..d85615b597d 100644 --- a/tests/ui/sanitizer/cfi/coroutine.rs +++ b/tests/ui/sanitizer/cfi/coroutine.rs @@ -8,6 +8,7 @@ //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi //@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/drop-in-place.rs b/tests/ui/sanitizer/cfi/drop-in-place.rs index 160338a5b54..fe59d546312 100644 --- a/tests/ui/sanitizer/cfi/drop-in-place.rs +++ b/tests/ui/sanitizer/cfi/drop-in-place.rs @@ -5,6 +5,7 @@ //@ ignore-backends: gcc //@ needs-sanitizer-cfi //@ compile-flags: -Clto -Copt-level=0 -Cprefer-dynamic=off -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer //@ run-pass struct EmptyDrop; diff --git a/tests/ui/sanitizer/cfi/drop-no-principal.rs b/tests/ui/sanitizer/cfi/drop-no-principal.rs index e3a46fe9d75..4fb905eb51d 100644 --- a/tests/ui/sanitizer/cfi/drop-no-principal.rs +++ b/tests/ui/sanitizer/cfi/drop-no-principal.rs @@ -4,7 +4,7 @@ // FIXME(#122848) Remove only-linux once OSX CFI binaries works //@ only-linux //@ ignore-backends: gcc -//@ compile-flags: --crate-type=bin -Cprefer-dynamic=off -Clto -Zsanitizer=cfi +//@ compile-flags: --crate-type=bin -Cprefer-dynamic=off -Clto -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer //@ compile-flags: -C target-feature=-crt-static -C codegen-units=1 -C opt-level=0 // FIXME(#118761) Should be run-pass once the labels on drop are compatible. // This test is being landed ahead of that to test that the compiler doesn't ICE while labeling the diff --git a/tests/ui/sanitizer/cfi/fn-ptr.rs b/tests/ui/sanitizer/cfi/fn-ptr.rs index d3209c62ddf..bdb8c7ceb32 100644 --- a/tests/ui/sanitizer/cfi/fn-ptr.rs +++ b/tests/ui/sanitizer/cfi/fn-ptr.rs @@ -7,6 +7,7 @@ //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi //@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C opt-level=0 -C codegen-units=1 -C lto //@ [cfi] compile-flags: -C prefer-dynamic=off //@ [cfi] compile-flags: -Z sanitizer=cfi diff --git a/tests/ui/sanitizer/cfi/generalize-pointers-attr-cfg.rs b/tests/ui/sanitizer/cfi/generalize-pointers-attr-cfg.rs index d46002c69fd..44cdcb250e7 100644 --- a/tests/ui/sanitizer/cfi/generalize-pointers-attr-cfg.rs +++ b/tests/ui/sanitizer/cfi/generalize-pointers-attr-cfg.rs @@ -4,6 +4,7 @@ //@ needs-sanitizer-cfi //@ check-pass //@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-generalize-pointers +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer #![feature(cfg_sanitizer_cfi)] diff --git a/tests/ui/sanitizer/cfi/normalize-integers-attr-cfg.rs b/tests/ui/sanitizer/cfi/normalize-integers-attr-cfg.rs index 24c2c2c13da..ce4e31eb69b 100644 --- a/tests/ui/sanitizer/cfi/normalize-integers-attr-cfg.rs +++ b/tests/ui/sanitizer/cfi/normalize-integers-attr-cfg.rs @@ -3,7 +3,7 @@ // //@ needs-sanitizer-cfi //@ check-pass -//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers +//@ compile-flags: -Clto -Cno-prepopulate-passes -Ctarget-feature=-crt-static -Zsanitizer=cfi -Zsanitizer-cfi-normalize-integers -C unsafe-allow-abi-mismatch=sanitizer,sanitizer-cfi-normalize-integers #![feature(cfg_sanitizer_cfi)] diff --git a/tests/ui/sanitizer/cfi/self-ref.rs b/tests/ui/sanitizer/cfi/self-ref.rs index b93d49296b6..827610a2610 100644 --- a/tests/ui/sanitizer/cfi/self-ref.rs +++ b/tests/ui/sanitizer/cfi/self-ref.rs @@ -6,7 +6,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/sized-associated-ty.rs b/tests/ui/sanitizer/cfi/sized-associated-ty.rs index eefc3e2e9db..da8c385c6fc 100644 --- a/tests/ui/sanitizer/cfi/sized-associated-ty.rs +++ b/tests/ui/sanitizer/cfi/sized-associated-ty.rs @@ -7,7 +7,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/supertraits.rs b/tests/ui/sanitizer/cfi/supertraits.rs index f5a984b9583..b2782dff5d5 100644 --- a/tests/ui/sanitizer/cfi/supertraits.rs +++ b/tests/ui/sanitizer/cfi/supertraits.rs @@ -6,7 +6,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/cfi/transparent-has-regions.rs b/tests/ui/sanitizer/cfi/transparent-has-regions.rs index b70e1ea1791..b82850133c1 100644 --- a/tests/ui/sanitizer/cfi/transparent-has-regions.rs +++ b/tests/ui/sanitizer/cfi/transparent-has-regions.rs @@ -1,5 +1,5 @@ //@ needs-sanitizer-cfi -//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer //@ no-prefer-dynamic //@ only-x86_64-unknown-linux-gnu //@ build-pass diff --git a/tests/ui/sanitizer/cfi/virtual-auto.rs b/tests/ui/sanitizer/cfi/virtual-auto.rs index e6f2ebd4b2c..d3a715c079a 100644 --- a/tests/ui/sanitizer/cfi/virtual-auto.rs +++ b/tests/ui/sanitizer/cfi/virtual-auto.rs @@ -6,7 +6,7 @@ //@ ignore-backends: gcc //@ [cfi] needs-sanitizer-cfi //@ [kcfi] needs-sanitizer-kcfi -//@ compile-flags: -C target-feature=-crt-static +//@ compile-flags: -C target-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ [cfi] compile-flags: -C codegen-units=1 -C lto -C prefer-dynamic=off -C opt-level=0 //@ [cfi] compile-flags: -Z sanitizer=cfi //@ [kcfi] compile-flags: -Z sanitizer=kcfi diff --git a/tests/ui/sanitizer/dataflow.rs b/tests/ui/sanitizer/dataflow.rs index 658a9e48086..d99a1a6ac24 100644 --- a/tests/ui/sanitizer/dataflow.rs +++ b/tests/ui/sanitizer/dataflow.rs @@ -4,7 +4,7 @@ //@ needs-sanitizer-support //@ needs-sanitizer-dataflow //@ run-pass -//@ compile-flags: -Zsanitizer=dataflow -Zsanitizer-dataflow-abilist={{src-base}}/sanitizer/dataflow-abilist.txt +//@ compile-flags: -Zsanitizer=dataflow -Zsanitizer-dataflow-abilist={{src-base}}/sanitizer/dataflow-abilist.txt -C unsafe-allow-abi-mismatch=sanitizer use std::mem::size_of; use std::os::raw::{c_int, c_long, c_void}; diff --git a/tests/ui/sanitizer/hwaddress.rs b/tests/ui/sanitizer/hwaddress.rs index e5939eb734b..05fcab17506 100644 --- a/tests/ui/sanitizer/hwaddress.rs +++ b/tests/ui/sanitizer/hwaddress.rs @@ -5,7 +5,7 @@ //@ ignore-aarch64-unknown-linux-gnu // // FIXME(#83989): codegen-units=1 triggers linker errors on aarch64-gnu -//@ compile-flags: -Z sanitizer=hwaddress -O -g -C codegen-units=16 +//@ compile-flags: -Z sanitizer=hwaddress -O -g -C codegen-units=16 -C unsafe-allow-abi-mismatch=sanitizer // //@ run-fail //@ error-pattern: HWAddressSanitizer: tag-mismatch diff --git a/tests/ui/sanitizer/issue-111184-cfi-coroutine-witness.rs b/tests/ui/sanitizer/issue-111184-cfi-coroutine-witness.rs index 7d0c73c2841..be81c7bd0ca 100644 --- a/tests/ui/sanitizer/issue-111184-cfi-coroutine-witness.rs +++ b/tests/ui/sanitizer/issue-111184-cfi-coroutine-witness.rs @@ -2,7 +2,7 @@ // encode_ty and caused the compiler to ICE. // //@ needs-sanitizer-cfi -//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi +//@ compile-flags: -Ccodegen-units=1 -Clto -Ctarget-feature=-crt-static -Zsanitizer=cfi -C unsafe-allow-abi-mismatch=sanitizer //@ edition: 2021 //@ no-prefer-dynamic //@ only-x86_64-unknown-linux-gnu diff --git a/tests/ui/sanitizer/issue-114275-cfi-const-expr-in-arry-len.rs b/tests/ui/sanitizer/issue-114275-cfi-const-expr-in-arry-len.rs index b1b7487fa2a..f7af2842ad6 100644 --- a/tests/ui/sanitizer/issue-114275-cfi-const-expr-in-arry-len.rs +++ b/tests/ui/sanitizer/issue-114275-cfi-const-expr-in-arry-len.rs @@ -2,7 +2,7 @@ // was expecting array type lengths to be evaluated, this was causing an ICE. // //@ build-pass -//@ compile-flags: -Ccodegen-units=1 -Clto -Zsanitizer=cfi -Ctarget-feature=-crt-static +//@ compile-flags: -Ccodegen-units=1 -Clto -Zsanitizer=cfi -Ctarget-feature=-crt-static -C unsafe-allow-abi-mismatch=sanitizer //@ needs-sanitizer-cfi #![crate_type = "lib"] diff --git a/tests/ui/sanitizer/issue-72154-address-lifetime-markers.rs b/tests/ui/sanitizer/issue-72154-address-lifetime-markers.rs index aa0c19db9a1..edeb1b0bf94 100644 --- a/tests/ui/sanitizer/issue-72154-address-lifetime-markers.rs +++ b/tests/ui/sanitizer/issue-72154-address-lifetime-markers.rs @@ -7,7 +7,7 @@ //@ needs-sanitizer-address //@ ignore-cross-compile // -//@ compile-flags: -Copt-level=0 -Zsanitizer=address +//@ compile-flags: -Copt-level=0 -Zsanitizer=address -C unsafe-allow-abi-mismatch=sanitizer //@ run-pass pub struct Wrap { diff --git a/tests/ui/sanitizer/kcfi-mangling.rs b/tests/ui/sanitizer/kcfi-mangling.rs index ba36dc184ec..371f34ba72a 100644 --- a/tests/ui/sanitizer/kcfi-mangling.rs +++ b/tests/ui/sanitizer/kcfi-mangling.rs @@ -2,7 +2,7 @@ //@ needs-sanitizer-kcfi //@ no-prefer-dynamic -//@ compile-flags: -C panic=abort -Zsanitizer=kcfi -C symbol-mangling-version=v0 +//@ compile-flags: -C panic=abort -Zsanitizer=kcfi -C symbol-mangling-version=v0 -C unsafe-allow-abi-mismatch=sanitizer //@ build-pass //@ ignore-backends: gcc diff --git a/tests/ui/sanitizer/leak.rs b/tests/ui/sanitizer/leak.rs index 65915ec24b7..4ce3f7ab1b9 100644 --- a/tests/ui/sanitizer/leak.rs +++ b/tests/ui/sanitizer/leak.rs @@ -1,7 +1,7 @@ //@ needs-sanitizer-support //@ needs-sanitizer-leak // -//@ compile-flags: -Z sanitizer=leak -O +//@ compile-flags: -Z sanitizer=leak -O -C unsafe-allow-abi-mismatch=sanitizer // //@ run-fail //@ error-pattern: LeakSanitizer: detected memory leaks diff --git a/tests/ui/sanitizer/memory-eager.rs b/tests/ui/sanitizer/memory-eager.rs index 709299f87d4..9498336abb7 100644 --- a/tests/ui/sanitizer/memory-eager.rs +++ b/tests/ui/sanitizer/memory-eager.rs @@ -1,6 +1,8 @@ //@ needs-sanitizer-support //@ needs-sanitizer-memory // +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer +// //@ revisions: unoptimized optimized // //@ [optimized]compile-flags: -Z sanitizer=memory -Zsanitizer-memory-track-origins -O diff --git a/tests/ui/sanitizer/memory-passing.rs b/tests/ui/sanitizer/memory-passing.rs index 96a4cd909c7..6ac41b178fe 100644 --- a/tests/ui/sanitizer/memory-passing.rs +++ b/tests/ui/sanitizer/memory-passing.rs @@ -1,6 +1,8 @@ //@ needs-sanitizer-support //@ needs-sanitizer-memory // +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer +// //@ revisions: unoptimized optimized // //@ [optimized]compile-flags: -Z sanitizer=memory -Zsanitizer-memory-track-origins -O diff --git a/tests/ui/sanitizer/memory.rs b/tests/ui/sanitizer/memory.rs index a91fefe4d16..1566637acd2 100644 --- a/tests/ui/sanitizer/memory.rs +++ b/tests/ui/sanitizer/memory.rs @@ -1,6 +1,8 @@ //@ needs-sanitizer-support //@ needs-sanitizer-memory // +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer +// //@ revisions: unoptimized optimized // //@ [optimized]compile-flags: -Z sanitizer=memory -Zsanitizer-memory-track-origins -O diff --git a/tests/ui/sanitizer/new-llvm-pass-manager-thin-lto.rs b/tests/ui/sanitizer/new-llvm-pass-manager-thin-lto.rs index c1a2c2f26ac..b6656847427 100644 --- a/tests/ui/sanitizer/new-llvm-pass-manager-thin-lto.rs +++ b/tests/ui/sanitizer/new-llvm-pass-manager-thin-lto.rs @@ -6,6 +6,8 @@ //@ needs-sanitizer-address //@ ignore-cross-compile // +//@ compile-flags: -C unsafe-allow-abi-mismatch=sanitizer +// //@ no-prefer-dynamic //@ revisions: opt0 opt1 //@ compile-flags: -Zsanitizer=address -Clto=thin diff --git a/tests/ui/sanitizer/thread.rs b/tests/ui/sanitizer/thread.rs index 9073124d1bd..8f72b1b41a8 100644 --- a/tests/ui/sanitizer/thread.rs +++ b/tests/ui/sanitizer/thread.rs @@ -13,7 +13,7 @@ //@ needs-sanitizer-support //@ needs-sanitizer-thread // -//@ compile-flags: -Z sanitizer=thread -O +//@ compile-flags: -Z sanitizer=thread -O -C unsafe-allow-abi-mismatch=sanitizer // //@ run-fail-or-crash //@ error-pattern: WARNING: ThreadSanitizer: data race diff --git a/tests/ui/sanitizer/use-after-scope.rs b/tests/ui/sanitizer/use-after-scope.rs index 106dc6466d6..1c477d0be14 100644 --- a/tests/ui/sanitizer/use-after-scope.rs +++ b/tests/ui/sanitizer/use-after-scope.rs @@ -2,7 +2,7 @@ //@ needs-sanitizer-address //@ ignore-cross-compile // -//@ compile-flags: -Zsanitizer=address +//@ compile-flags: -Zsanitizer=address -C unsafe-allow-abi-mismatch=sanitizer //@ run-fail-or-crash //@ error-pattern: ERROR: AddressSanitizer: stack-use-after-scope diff --git a/tests/ui/target_modifiers/auxiliary/kcfi-normalize-ints.rs b/tests/ui/target_modifiers/auxiliary/kcfi-normalize-ints.rs new file mode 100644 index 00000000000..f97005a1450 --- /dev/null +++ b/tests/ui/target_modifiers/auxiliary/kcfi-normalize-ints.rs @@ -0,0 +1,7 @@ +//@ no-prefer-dynamic +//@ needs-sanitizer-kcfi +//@ compile-flags: -C panic=abort -Zsanitizer=kcfi -Zsanitizer-cfi-normalize-integers + +#![feature(no_core)] +#![crate_type = "rlib"] +#![no_core] diff --git a/tests/ui/target_modifiers/auxiliary/no-sanitizers.rs b/tests/ui/target_modifiers/auxiliary/no-sanitizers.rs new file mode 100644 index 00000000000..1d47ff6826d --- /dev/null +++ b/tests/ui/target_modifiers/auxiliary/no-sanitizers.rs @@ -0,0 +1,6 @@ +//@ no-prefer-dynamic +//@ compile-flags: -C panic=abort + +#![feature(no_core)] +#![crate_type = "rlib"] +#![no_core] diff --git a/tests/ui/target_modifiers/auxiliary/safestack-and-kcfi.rs b/tests/ui/target_modifiers/auxiliary/safestack-and-kcfi.rs new file mode 100644 index 00000000000..029744d2a1f --- /dev/null +++ b/tests/ui/target_modifiers/auxiliary/safestack-and-kcfi.rs @@ -0,0 +1,10 @@ +//@ no-prefer-dynamic + +//@ needs-sanitizer-kcfi +//@ needs-sanitizer-safestack + +//@ compile-flags: -C panic=abort -Zsanitizer=safestack,kcfi + +#![feature(no_core)] +#![crate_type = "rlib"] +#![no_core] diff --git a/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.rs b/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.rs new file mode 100644 index 00000000000..cb9f701349a --- /dev/null +++ b/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.rs @@ -0,0 +1,18 @@ +// For kCFI, the helper flag -Zsanitizer-cfi-normalize-integers should also be a target modifier. + +//@ needs-sanitizer-kcfi +//@ aux-build:kcfi-normalize-ints.rs +//@ compile-flags: -Cpanic=abort + +//@ revisions: ok wrong_flag wrong_sanitizer +//@[ok] compile-flags: -Zsanitizer=kcfi -Zsanitizer-cfi-normalize-integers +//@[wrong_flag] compile-flags: -Zsanitizer=kcfi +//@[ok] check-pass + +#![feature(no_core)] +//[wrong_flag]~^ ERROR mixing `-Zsanitizer-cfi-normalize-integers` will cause an ABI mismatch in crate `sanitizer_kcfi_normalize_ints` +//[wrong_sanitizer]~^^ ERROR mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizer_kcfi_normalize_ints` +#![crate_type = "rlib"] +#![no_core] + +extern crate kcfi_normalize_ints; diff --git a/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_flag.stderr b/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_flag.stderr new file mode 100644 index 00000000000..1db79b025e9 --- /dev/null +++ b/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_flag.stderr @@ -0,0 +1,13 @@ +error: mixing `-Zsanitizer-cfi-normalize-integers` will cause an ABI mismatch in crate `sanitizer_kcfi_normalize_ints` + --> $DIR/sanitizer-kcfi-normalize-ints.rs:12:1 + | +LL | #![feature(no_core)] + | ^ + | + = help: the `-Zsanitizer-cfi-normalize-integers` flag modifies the ABI so Rust crates compiled with different values of this flag cannot be used together safely + = note: unset `-Zsanitizer-cfi-normalize-integers` in this crate is incompatible with `-Zsanitizer-cfi-normalize-integers=` in dependency `kcfi_normalize_ints` + = help: set `-Zsanitizer-cfi-normalize-integers=` in this crate or unset `-Zsanitizer-cfi-normalize-integers` in `kcfi_normalize_ints` + = help: if you are sure this will not cause problems, you may use `-Cunsafe-allow-abi-mismatch=sanitizer-cfi-normalize-integers` to silence this error + +error: aborting due to 1 previous error + diff --git a/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_sanitizer.stderr b/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_sanitizer.stderr new file mode 100644 index 00000000000..5b949c87350 --- /dev/null +++ b/tests/ui/target_modifiers/sanitizer-kcfi-normalize-ints.wrong_sanitizer.stderr @@ -0,0 +1,13 @@ +error: mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizer_kcfi_normalize_ints` + --> $DIR/sanitizer-kcfi-normalize-ints.rs:12:1 + | +LL | #![feature(no_core)] + | ^ + | + = help: the `-Zsanitizer` flag modifies the ABI so Rust crates compiled with different values of this flag cannot be used together safely + = note: unset `-Zsanitizer` in this crate is incompatible with `-Zsanitizer=kcfi` in dependency `kcfi_normalize_ints` + = help: set `-Zsanitizer=kcfi` in this crate or unset `-Zsanitizer` in `kcfi_normalize_ints` + = help: if you are sure this will not cause problems, you may use `-Cunsafe-allow-abi-mismatch=sanitizer` to silence this error + +error: aborting due to 1 previous error + diff --git a/tests/ui/target_modifiers/sanitizers-good-for-inconsistency.rs b/tests/ui/target_modifiers/sanitizers-good-for-inconsistency.rs new file mode 100644 index 00000000000..abda7be9e40 --- /dev/null +++ b/tests/ui/target_modifiers/sanitizers-good-for-inconsistency.rs @@ -0,0 +1,19 @@ +// AddressSanitizer, LeakSanitizer are good to be inconsistent (they are not a target modifiers) + +//@ revisions: wrong_address_san wrong_leak_san + +//@[wrong_address_san] needs-sanitizer-address +//@[wrong_leak_san] needs-sanitizer-leak + +//@ aux-build:no-sanitizers.rs +//@ compile-flags: -Cpanic=abort -C target-feature=-crt-static + +//@[wrong_address_san] compile-flags: -Zsanitizer=address +//@[wrong_leak_san] compile-flags: -Zsanitizer=leak +//@ check-pass + +#![feature(no_core)] +#![crate_type = "rlib"] +#![no_core] + +extern crate no_sanitizers; diff --git a/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_both.stderr b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_both.stderr new file mode 100644 index 00000000000..440a91c7707 --- /dev/null +++ b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_both.stderr @@ -0,0 +1,13 @@ +error: mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizers_safestack_and_kcfi` + --> $DIR/sanitizers-safestack-and-kcfi.rs:16:1 + | +LL | #![feature(no_core)] + | ^ + | + = help: the `-Zsanitizer` flag modifies the ABI so Rust crates compiled with different values of this flag cannot be used together safely + = note: unset `-Zsanitizer` in this crate is incompatible with `-Zsanitizer=safestack,kcfi` in dependency `safestack_and_kcfi` + = help: set `-Zsanitizer=safestack,kcfi` in this crate or unset `-Zsanitizer` in `safestack_and_kcfi` + = help: if you are sure this will not cause problems, you may use `-Cunsafe-allow-abi-mismatch=sanitizer` to silence this error + +error: aborting due to 1 previous error + diff --git a/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_kcfi.stderr b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_kcfi.stderr new file mode 100644 index 00000000000..6cdd7facc37 --- /dev/null +++ b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_kcfi.stderr @@ -0,0 +1,13 @@ +error: mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizers_safestack_and_kcfi` + --> $DIR/sanitizers-safestack-and-kcfi.rs:16:1 + | +LL | #![feature(no_core)] + | ^ + | + = help: the `-Zsanitizer` flag modifies the ABI so Rust crates compiled with different values of this flag cannot be used together safely + = note: `-Zsanitizer=safestack` in this crate is incompatible with `-Zsanitizer=safestack,kcfi` in dependency `safestack_and_kcfi` + = help: set `-Zsanitizer=safestack,kcfi` in this crate or `-Zsanitizer=safestack` in `safestack_and_kcfi` + = help: if you are sure this will not cause problems, you may use `-Cunsafe-allow-abi-mismatch=sanitizer` to silence this error + +error: aborting due to 1 previous error + diff --git a/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_safestack.stderr b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_safestack.stderr new file mode 100644 index 00000000000..ecfbcace39f --- /dev/null +++ b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.missed_safestack.stderr @@ -0,0 +1,13 @@ +error: mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizers_safestack_and_kcfi` + --> $DIR/sanitizers-safestack-and-kcfi.rs:16:1 + | +LL | #![feature(no_core)] + | ^ + | + = help: the `-Zsanitizer` flag modifies the ABI so Rust crates compiled with different values of this flag cannot be used together safely + = note: `-Zsanitizer=kcfi` in this crate is incompatible with `-Zsanitizer=safestack,kcfi` in dependency `safestack_and_kcfi` + = help: set `-Zsanitizer=safestack,kcfi` in this crate or `-Zsanitizer=kcfi` in `safestack_and_kcfi` + = help: if you are sure this will not cause problems, you may use `-Cunsafe-allow-abi-mismatch=sanitizer` to silence this error + +error: aborting due to 1 previous error + diff --git a/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.rs b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.rs new file mode 100644 index 00000000000..6c3ceb7e910 --- /dev/null +++ b/tests/ui/target_modifiers/sanitizers-safestack-and-kcfi.rs @@ -0,0 +1,23 @@ +//@ needs-sanitizer-kcfi +//@ needs-sanitizer-safestack + +//@ aux-build:safestack-and-kcfi.rs +//@ compile-flags: -Cpanic=abort + +//@ revisions: good good_reverted missed_safestack missed_kcfi missed_both +//@[good] compile-flags: -Zsanitizer=safestack,kcfi +//@[good_reverted] compile-flags: -Zsanitizer=kcfi,safestack +//@[missed_safestack] compile-flags: -Zsanitizer=kcfi +//@[missed_kcfi] compile-flags: -Zsanitizer=safestack +// [missed_both] no additional compile-flags: +//@[good] check-pass +//@[good_reverted] check-pass + +#![feature(no_core)] +//[missed_safestack]~^ ERROR mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizers_safestack_and_kcfi` +//[missed_kcfi]~^^ ERROR mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizers_safestack_and_kcfi` +//[missed_both]~^^^ ERROR mixing `-Zsanitizer` will cause an ABI mismatch in crate `sanitizers_safestack_and_kcfi` +#![crate_type = "rlib"] +#![no_core] + +extern crate safestack_and_kcfi; -- cgit 1.4.1-3-g733a5 From b65a177b63237398dc3f49cb5598b16bb9318136 Mon Sep 17 00:00:00 2001 From: WANG Rui Date: Thu, 21 Aug 2025 14:15:08 +0800 Subject: Fix LoongArch C function ABI when passing/returning structs containing floats --- compiler/rustc_target/src/callconv/loongarch.rs | 137 ++++++++++++++++----- tests/assembly-llvm/loongarch-float-struct-abi.rs | 134 ++++++++++++++++++++ tests/codegen-llvm/cast-target-abi.rs | 52 ++++++-- .../loongarch-abi/cast-local-large-enough.rs | 44 +++++++ 4 files changed, 327 insertions(+), 40 deletions(-) create mode 100644 tests/assembly-llvm/loongarch-float-struct-abi.rs create mode 100644 tests/codegen-llvm/loongarch-abi/cast-local-large-enough.rs diff --git a/compiler/rustc_target/src/callconv/loongarch.rs b/compiler/rustc_target/src/callconv/loongarch.rs index d567ad401bb..9213d73e24e 100644 --- a/compiler/rustc_target/src/callconv/loongarch.rs +++ b/compiler/rustc_target/src/callconv/loongarch.rs @@ -8,16 +8,16 @@ use crate::spec::HasTargetSpec; #[derive(Copy, Clone)] enum RegPassKind { - Float(Reg), - Integer(Reg), + Float { offset_from_start: Size, ty: Reg }, + Integer { offset_from_start: Size, ty: Reg }, Unknown, } #[derive(Copy, Clone)] enum FloatConv { - FloatPair(Reg, Reg), + FloatPair { first_ty: Reg, second_ty_offset_from_start: Size, second_ty: Reg }, Float(Reg), - MixedPair(Reg, Reg), + MixedPair { first_ty: Reg, second_ty_offset_from_start: Size, second_ty: Reg }, } #[derive(Copy, Clone)] @@ -37,6 +37,7 @@ fn should_use_fp_conv_helper<'a, Ty, C>( flen: u64, field1_kind: &mut RegPassKind, field2_kind: &mut RegPassKind, + offset_from_start: Size, ) -> Result<(), CannotUseFpConv> where Ty: TyAbiInterface<'a, C> + Copy, @@ -49,16 +50,16 @@ where } match (*field1_kind, *field2_kind) { (RegPassKind::Unknown, _) => { - *field1_kind = RegPassKind::Integer(Reg { - kind: RegKind::Integer, - size: arg_layout.size, - }); + *field1_kind = RegPassKind::Integer { + offset_from_start, + ty: Reg { kind: RegKind::Integer, size: arg_layout.size }, + }; } - (RegPassKind::Float(_), RegPassKind::Unknown) => { - *field2_kind = RegPassKind::Integer(Reg { - kind: RegKind::Integer, - size: arg_layout.size, - }); + (RegPassKind::Float { .. }, RegPassKind::Unknown) => { + *field2_kind = RegPassKind::Integer { + offset_from_start, + ty: Reg { kind: RegKind::Integer, size: arg_layout.size }, + }; } _ => return Err(CannotUseFpConv), } @@ -69,12 +70,16 @@ where } match (*field1_kind, *field2_kind) { (RegPassKind::Unknown, _) => { - *field1_kind = - RegPassKind::Float(Reg { kind: RegKind::Float, size: arg_layout.size }); + *field1_kind = RegPassKind::Float { + offset_from_start, + ty: Reg { kind: RegKind::Float, size: arg_layout.size }, + }; } (_, RegPassKind::Unknown) => { - *field2_kind = - RegPassKind::Float(Reg { kind: RegKind::Float, size: arg_layout.size }); + *field2_kind = RegPassKind::Float { + offset_from_start, + ty: Reg { kind: RegKind::Float, size: arg_layout.size }, + }; } _ => return Err(CannotUseFpConv), } @@ -96,13 +101,14 @@ where flen, field1_kind, field2_kind, + offset_from_start, ); } return Err(CannotUseFpConv); } } FieldsShape::Array { count, .. } => { - for _ in 0..count { + for i in 0..count { let elem_layout = arg_layout.field(cx, 0); should_use_fp_conv_helper( cx, @@ -111,6 +117,7 @@ where flen, field1_kind, field2_kind, + offset_from_start + elem_layout.size * i, )?; } } @@ -121,7 +128,15 @@ where } for i in arg_layout.fields.index_by_increasing_offset() { let field = arg_layout.field(cx, i); - should_use_fp_conv_helper(cx, &field, xlen, flen, field1_kind, field2_kind)?; + should_use_fp_conv_helper( + cx, + &field, + xlen, + flen, + field1_kind, + field2_kind, + offset_from_start + arg_layout.fields.offset(i), + )?; } } }, @@ -140,14 +155,52 @@ where { let mut field1_kind = RegPassKind::Unknown; let mut field2_kind = RegPassKind::Unknown; - if should_use_fp_conv_helper(cx, arg, xlen, flen, &mut field1_kind, &mut field2_kind).is_err() { + if should_use_fp_conv_helper( + cx, + arg, + xlen, + flen, + &mut field1_kind, + &mut field2_kind, + Size::ZERO, + ) + .is_err() + { return None; } match (field1_kind, field2_kind) { - (RegPassKind::Integer(l), RegPassKind::Float(r)) => Some(FloatConv::MixedPair(l, r)), - (RegPassKind::Float(l), RegPassKind::Integer(r)) => Some(FloatConv::MixedPair(l, r)), - (RegPassKind::Float(l), RegPassKind::Float(r)) => Some(FloatConv::FloatPair(l, r)), - (RegPassKind::Float(f), RegPassKind::Unknown) => Some(FloatConv::Float(f)), + ( + RegPassKind::Integer { offset_from_start, .. } + | RegPassKind::Float { offset_from_start, .. }, + _, + ) if offset_from_start != Size::ZERO => { + panic!("type {:?} has a first field with non-zero offset {offset_from_start:?}", arg.ty) + } + ( + RegPassKind::Integer { ty: first_ty, .. }, + RegPassKind::Float { offset_from_start, ty: second_ty }, + ) => Some(FloatConv::MixedPair { + first_ty, + second_ty_offset_from_start: offset_from_start, + second_ty, + }), + ( + RegPassKind::Float { ty: first_ty, .. }, + RegPassKind::Integer { offset_from_start, ty: second_ty }, + ) => Some(FloatConv::MixedPair { + first_ty, + second_ty_offset_from_start: offset_from_start, + second_ty, + }), + ( + RegPassKind::Float { ty: first_ty, .. }, + RegPassKind::Float { offset_from_start, ty: second_ty }, + ) => Some(FloatConv::FloatPair { + first_ty, + second_ty_offset_from_start: offset_from_start, + second_ty, + }), + (RegPassKind::Float { ty, .. }, RegPassKind::Unknown) => Some(FloatConv::Float(ty)), _ => None, } } @@ -165,11 +218,19 @@ where FloatConv::Float(f) => { arg.cast_to(f); } - FloatConv::FloatPair(l, r) => { - arg.cast_to(CastTarget::pair(l, r)); + FloatConv::FloatPair { first_ty, second_ty_offset_from_start, second_ty } => { + arg.cast_to(CastTarget::offset_pair( + first_ty, + second_ty_offset_from_start, + second_ty, + )); } - FloatConv::MixedPair(l, r) => { - arg.cast_to(CastTarget::pair(l, r)); + FloatConv::MixedPair { first_ty, second_ty_offset_from_start, second_ty } => { + arg.cast_to(CastTarget::offset_pair( + first_ty, + second_ty_offset_from_start, + second_ty, + )); } } return false; @@ -233,15 +294,27 @@ fn classify_arg<'a, Ty, C>( arg.cast_to(f); return; } - Some(FloatConv::FloatPair(l, r)) if *avail_fprs >= 2 => { + Some(FloatConv::FloatPair { first_ty, second_ty_offset_from_start, second_ty }) + if *avail_fprs >= 2 => + { *avail_fprs -= 2; - arg.cast_to(CastTarget::pair(l, r)); + arg.cast_to(CastTarget::offset_pair( + first_ty, + second_ty_offset_from_start, + second_ty, + )); return; } - Some(FloatConv::MixedPair(l, r)) if *avail_fprs >= 1 && *avail_gprs >= 1 => { + Some(FloatConv::MixedPair { first_ty, second_ty_offset_from_start, second_ty }) + if *avail_fprs >= 1 && *avail_gprs >= 1 => + { *avail_gprs -= 1; *avail_fprs -= 1; - arg.cast_to(CastTarget::pair(l, r)); + arg.cast_to(CastTarget::offset_pair( + first_ty, + second_ty_offset_from_start, + second_ty, + )); return; } _ => (), diff --git a/tests/assembly-llvm/loongarch-float-struct-abi.rs b/tests/assembly-llvm/loongarch-float-struct-abi.rs new file mode 100644 index 00000000000..4991004fc05 --- /dev/null +++ b/tests/assembly-llvm/loongarch-float-struct-abi.rs @@ -0,0 +1,134 @@ +//@ add-core-stubs +//@ assembly-output: emit-asm +//@ compile-flags: -Copt-level=3 --target loongarch64-unknown-linux-gnu +//@ needs-llvm-components: loongarch + +#![feature(no_core, lang_items)] +#![no_std] +#![no_core] +#![crate_type = "lib"] + +extern crate minicore; +use minicore::*; + +#[repr(C, align(64))] +struct Aligned(f64); + +#[repr(C)] +struct Padded(u8, Aligned); + +#[repr(C, packed)] +struct Packed(u8, f32); + +impl Copy for Aligned {} +impl Copy for Padded {} +impl Copy for Packed {} + +extern "C" { + fn take_padded(x: Padded); + fn get_padded() -> Padded; + fn take_packed(x: Packed); + fn get_packed() -> Packed; +} + +// CHECK-LABEL: pass_padded +#[unsafe(no_mangle)] +extern "C" fn pass_padded(out: &mut Padded, x: Padded) { + // CHECK: st.b $a1, $a0, 0 + // CHECK-NEXT: fst.d $fa0, $a0, 64 + // CHECK-NEXT: ret + *out = x; +} + +// CHECK-LABEL: ret_padded +#[unsafe(no_mangle)] +extern "C" fn ret_padded(x: &Padded) -> Padded { + // CHECK: fld.d $fa0, $a0, 64 + // CHECK-NEXT: ld.b $a0, $a0, 0 + // CHECK-NEXT: ret + *x +} + +#[unsafe(no_mangle)] +extern "C" fn call_padded(x: &Padded) { + // CHECK: fld.d $fa0, $a0, 64 + // CHECK-NEXT: ld.b $a0, $a0, 0 + // CHECK-NEXT: pcaddu18i $t8, %call36(take_padded) + // CHECK-NEXT: jr $t8 + unsafe { + take_padded(*x); + } +} + +#[unsafe(no_mangle)] +extern "C" fn receive_padded(out: &mut Padded) { + // CHECK: addi.d $sp, $sp, -16 + // CHECK-NEXT: .cfi_def_cfa_offset 16 + // CHECK-NEXT: st.d $ra, $sp, [[#%d,RA_SPILL:]] + // CHECK-NEXT: st.d [[TEMP:.*]], $sp, [[#%d,TEMP_SPILL:]] + // CHECK-NEXT: .cfi_offset 1, [[#%d,RA_SPILL - 16]] + // CHECK-NEXT: .cfi_offset [[#%d,TEMP_NUM:]], [[#%d,TEMP_SPILL - 16]] + // CHECK-NEXT: move [[TEMP]], $a0 + // CHECK-NEXT: pcaddu18i $ra, %call36(get_padded) + // CHECK-NEXT: jirl $ra, $ra, 0 + // CHECK-NEXT: st.b $a0, [[TEMP]], 0 + // CHECK-NEXT: fst.d $fa0, [[TEMP]], 64 + // CHECK-NEXT: ld.d [[TEMP]], $sp, [[#%d,TEMP_SPILL]] + // CHECK-NEXT: ld.d $ra, $sp, [[#%d,RA_SPILL]] + // CHECK: addi.d $sp, $sp, 16 + // CHECK: ret + unsafe { + *out = get_padded(); + } +} + +// CHECK-LABEL: pass_packed +#[unsafe(no_mangle)] +extern "C" fn pass_packed(out: &mut Packed, x: Packed) { + // CHECK: st.b $a1, $a0, 0 + // CHECK-NEXT: fst.s $fa0, $a0, 1 + // CHECK-NEXT: ret + *out = x; +} + +// CHECK-LABEL: ret_packed +#[unsafe(no_mangle)] +extern "C" fn ret_packed(x: &Packed) -> Packed { + // CHECK: fld.s $fa0, $a0, 1 + // CHECK-NEXT: ld.b $a0, $a0, 0 + // CHECK-NEXT: ret + *x +} + +#[unsafe(no_mangle)] +extern "C" fn call_packed(x: &Packed) { + // CHECK: fld.s $fa0, $a0, 1 + // CHECK-NEXT: ld.b $a0, $a0, 0 + // CHECK-NEXT: pcaddu18i $t8, %call36(take_packed) + // CHECK-NEXT: jr $t8 + unsafe { + take_packed(*x); + } +} + +#[unsafe(no_mangle)] +extern "C" fn receive_packed(out: &mut Packed) { + // CHECK: addi.d $sp, $sp, -16 + // CHECK-NEXT: .cfi_def_cfa_offset 16 + // CHECK-NEXT: st.d $ra, $sp, [[#%d,RA_SPILL:]] + // CHECK-NEXT: st.d [[TEMP:.*]], $sp, [[#%d,TEMP_SPILL:]] + // CHECK-NEXT: .cfi_offset 1, [[#%d,RA_SPILL - 16]] + // CHECK-NEXT: .cfi_offset [[#%d,TEMP_NUM:]], [[#%d,TEMP_SPILL - 16]] + // CHECK-NEXT: move [[TEMP]], $a0 + // CHECK-NEXT: pcaddu18i $ra, %call36(get_packed) + // CHECK-NEXT: jirl $ra, $ra, 0 + // CHECK-NEXT: st.b $a0, [[TEMP]], 0 + // CHECK-NEXT: fst.s $fa0, [[TEMP]], 1 + // CHECK-NEXT: ld.d [[TEMP]], $sp, [[#%d,TEMP_SPILL]] + // CHECK-NEXT: ld.d $ra, $sp, [[#%d,RA_SPILL]] + // CHECK: addi.d $sp, $sp, 16 + // CHECK: ret + unsafe { + *out = get_packed(); + } +} diff --git a/tests/codegen-llvm/cast-target-abi.rs b/tests/codegen-llvm/cast-target-abi.rs index 28d3ad5f61c..090de00df93 100644 --- a/tests/codegen-llvm/cast-target-abi.rs +++ b/tests/codegen-llvm/cast-target-abi.rs @@ -171,7 +171,15 @@ pub extern "C" fn receives_doubledouble(x: DoubleDouble) { // CHECK: [[RUST_ALLOCA:%.+]] = alloca [16 x i8], align [[RUST_ALIGN:8]] - // CHECK: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // aarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = extractvalue [[ABI_TYPE]] [[ABI_VALUE:%.+]], 0 + // loongarch64: [[ABI_VALUE_1:%.+]] = extractvalue [[ABI_TYPE]] [[ABI_VALUE:%.+]], 1 + // loongarch64: store double [[ABI_VALUE_0]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: store double [[ABI_VALUE_1]], ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] + // powerpc64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // sparc64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // x86_64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // CHECK: call void @llvm.memcpy.{{.+}}(ptr align [[RUST_ALIGN]] [[RUST_ALLOCA]], ptr align [[ABI_ALIGN]] [[ABI_ALLOCA]], i64 16, i1 false) } @@ -190,7 +198,11 @@ pub extern "C" fn returns_doubledouble() -> DoubleDouble { // x86_64: [[ABI_ALLOCA:%.+]] = alloca [16 x i8], align [[ABI_ALIGN:8]] // aarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:\[2 x double\]]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = load double, ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: [[ABI_VALUE_1:%.+]] = load double, ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_2:%.+]] = insertvalue [[ABI_TYPE:{ double, double }]] poison, double [[ABI_VALUE_0]], 0 + // loongarch64: [[ABI_VALUE:%.+]] = insertvalue { double, double } [[ABI_VALUE_2]], double [[ABI_VALUE_1]], 1 // sparc64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // x86_64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] @@ -269,7 +281,11 @@ pub extern "C" fn receives_doublefloat(x: DoubleFloat) { // x86_64: [[RUST_ALLOCA:%.+]] = alloca [16 x i8], align [[RUST_ALIGN:8]] // aarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = extractvalue { double, float } [[ABI_VALUE]], 0 + // loongarch64: [[ABI_VALUE_1:%.+]] = extractvalue { double, float } [[ABI_VALUE]], 1 + // loongarch64: store double [[ABI_VALUE_0]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: store float [[ABI_VALUE_1]], ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] // powerpc64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // x86_64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] @@ -297,7 +313,11 @@ pub extern "C" fn returns_doublefloat() -> DoubleFloat { // x86_64: [[ABI_ALLOCA:%.+]] = alloca [16 x i8], align [[ABI_ALIGN:8]] // aarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:\[2 x i64\]]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, float }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = load double, ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: [[ABI_VALUE_1:%.+]] = load float, ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_2:%.+]] = insertvalue [[ABI_TYPE:{ double, float }]] poison, double [[ABI_VALUE_0]], 0 + // loongarch64: [[ABI_VALUE:%.+]] = insertvalue { double, float } [[ABI_VALUE_2]], float [[ABI_VALUE_1]], 1 // x86_64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // aarch64: ret [[ABI_TYPE]] [[ABI_VALUE]] @@ -429,7 +449,11 @@ pub fn call_doubledouble() { // CHECK: call void @llvm.memcpy.{{.+}}(ptr align [[ABI_ALIGN]] [[ABI_ALLOCA]], ptr align [[RUST_ALIGN]] [[RUST_ALLOCA]], i64 16, i1 false) // aarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:\[2 x double\]]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = load double, ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: [[ABI_VALUE_1:%.+]] = load double, ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_2:%.+]] = insertvalue [[ABI_TYPE:{ double, double }]] poison, double [[ABI_VALUE_0]], 0 + // loongarch64: [[ABI_VALUE:%.+]] = insertvalue { double, double } [[ABI_VALUE_2]], double [[ABI_VALUE_1]], 1 // powerpc64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:\[2 x i64\]]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // sparc64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // x86_64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] @@ -465,7 +489,11 @@ pub fn return_doubledouble() -> DoubleDouble { // x86_64: [[ABI_VALUE:%.+]] = call [[ABI_TYPE:{ double, double }]] @returns_doubledouble() // aarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = extractvalue { double, double } [[ABI_VALUE]], 0 + // loongarch64: [[ABI_VALUE_1:%.+]] = extractvalue { double, double } [[ABI_VALUE]], 1 + // loongarch64: store double [[ABI_VALUE_0]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: store double [[ABI_VALUE_1]], ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] // sparc64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // x86_64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] @@ -500,7 +528,11 @@ pub fn call_doublefloat() { // x86_64: call void @llvm.memcpy.{{.+}}(ptr align [[ABI_ALIGN]] [[ABI_ALLOCA]], ptr align [[RUST_ALIGN]] [[RUST_ALLOCA]], i64 16, i1 false) // aarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:\[2 x i64\]]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, float }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = load double, ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: [[ABI_VALUE_1:%.+]] = load float, ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_2:%.+]] = insertvalue [[ABI_TYPE:{ double, float }]] poison, double [[ABI_VALUE_0]], 0 + // loongarch64: [[ABI_VALUE:%.+]] = insertvalue { double, float } [[ABI_VALUE_2]], float [[ABI_VALUE_1]], 1 // powerpc64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:\[2 x i64\]]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // x86_64: [[ABI_VALUE:%.+]] = load [[ABI_TYPE:{ double, double }]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] @@ -540,7 +572,11 @@ pub fn return_doublefloat() -> DoubleFloat { // x86_64: [[ABI_VALUE:%.+]] = call [[ABI_TYPE:{ double, double }]] @returns_doublefloat() // aarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] - // loongarch64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_VALUE_0:%.+]] = extractvalue { double, float } [[ABI_VALUE]], 0 + // loongarch64: [[ABI_VALUE_1:%.+]] = extractvalue { double, float } [[ABI_VALUE]], 1 + // loongarch64: store double [[ABI_VALUE_0]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] + // loongarch64: [[ABI_ALLOCA_1:%.+]] = getelementptr inbounds i8, ptr [[ABI_ALLOCA]], i64 8 + // loongarch64: store float [[ABI_VALUE_1]], ptr [[ABI_ALLOCA_1]], align [[ABI_ALIGN]] // x86_64: store [[ABI_TYPE]] [[ABI_VALUE]], ptr [[ABI_ALLOCA]], align [[ABI_ALIGN]] // aarch64: call void @llvm.memcpy.{{.+}}(ptr align [[RUST_ALIGN]] [[RUST_ALLOCA]], ptr align [[ABI_ALIGN]] [[ABI_ALLOCA]], i64 16, i1 false) diff --git a/tests/codegen-llvm/loongarch-abi/cast-local-large-enough.rs b/tests/codegen-llvm/loongarch-abi/cast-local-large-enough.rs new file mode 100644 index 00000000000..e5a0e4cd3a2 --- /dev/null +++ b/tests/codegen-llvm/loongarch-abi/cast-local-large-enough.rs @@ -0,0 +1,44 @@ +//@ add-core-stubs +//@ compile-flags: -Copt-level=0 -Cdebuginfo=0 --target loongarch64-unknown-linux-gnu +//@ needs-llvm-components: loongarch + +#![feature(no_core, lang_items)] +#![no_std] +#![no_core] +#![crate_type = "lib"] + +extern crate minicore; +use minicore::*; + +#[repr(C, align(64))] +struct Aligned(f64); + +#[repr(C, align(64))] +struct AlignedPair(f32, f64); + +impl Copy for Aligned {} +impl Copy for AlignedPair {} + +// CHECK-LABEL: define double @read_aligned +#[unsafe(no_mangle)] +pub extern "C" fn read_aligned(x: &Aligned) -> Aligned { + // CHECK: %[[TEMP:.*]] = alloca [64 x i8], align 64 + // CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 64 %[[TEMP]], ptr align 64 %[[PTR:.*]], i64 64, i1 false) + // CHECK-NEXT: %[[RES:.*]] = load double, ptr %[[TEMP]], align 64 + // CHECK-NEXT: ret double %[[RES]] + *x +} + +// CHECK-LABEL: define { float, double } @read_aligned_pair +#[unsafe(no_mangle)] +pub extern "C" fn read_aligned_pair(x: &AlignedPair) -> AlignedPair { + // CHECK: %[[TEMP:.*]] = alloca [64 x i8], align 64 + // CHECK-NEXT: call void @llvm.memcpy.p0.p0.i64(ptr align 64 %[[TEMP]], ptr align 64 %[[PTR:.*]], i64 64, i1 false) + // CHECK-NEXT: %[[FIRST:.*]] = load float, ptr %[[TEMP]], align 64 + // CHECK-NEXT: %[[SECOND_PTR:.*]] = getelementptr inbounds i8, ptr %[[TEMP]], i64 8 + // CHECK-NEXT: %[[SECOND:.*]] = load double, ptr %[[SECOND_PTR]], align 8 + // CHECK-NEXT: %[[RES1:.*]] = insertvalue { float, double } poison, float %[[FIRST]], 0 + // CHECK-NEXT: %[[RES2:.*]] = insertvalue { float, double } %[[RES1]], double %[[SECOND]], 1 + // CHECK-NEXT: ret { float, double } %[[RES2]] + *x +} -- cgit 1.4.1-3-g733a5 From 70e7c058a0c8848a4ae3f4736b2a979e821a8e78 Mon Sep 17 00:00:00 2001 From: lumiscosity Date: Tue, 19 Aug 2025 11:15:20 +0200 Subject: Losslessly optimize PNG files Losslessly optimizes all of the PNG files in the repo. Done with: ``` oxipng -o max -a -s oxipng -o max --zopfli -a -s ``` --- src/doc/rustc/src/images/image1.png | Bin 164896 -> 112780 bytes src/doc/rustc/src/images/image2.png | Bin 155307 -> 107858 bytes src/doc/rustc/src/images/image3.png | Bin 19936 -> 15559 bytes src/doc/rustc/src/images/llvm-cov-show-01.png | Bin 416748 -> 206904 bytes src/doc/rustdoc/src/images/collapsed-long-item.png | Bin 17017 -> 11156 bytes .../rustdoc/src/images/collapsed-trait-impls.png | Bin 44225 -> 31081 bytes src/etc/installer/gfx/rust-logo.png | Bin 3909 -> 3261 bytes .../html/static/images/favicon-32x32.png | Bin 1125 -> 690 bytes 8 files changed, 0 insertions(+), 0 deletions(-) diff --git a/src/doc/rustc/src/images/image1.png b/src/doc/rustc/src/images/image1.png index 0da45e56620..3aad6359389 100644 Binary files a/src/doc/rustc/src/images/image1.png and b/src/doc/rustc/src/images/image1.png differ diff --git a/src/doc/rustc/src/images/image2.png b/src/doc/rustc/src/images/image2.png index a9cf23f8737..085b1c490b8 100644 Binary files a/src/doc/rustc/src/images/image2.png and b/src/doc/rustc/src/images/image2.png differ diff --git a/src/doc/rustc/src/images/image3.png b/src/doc/rustc/src/images/image3.png index 844a2fe6747..ee332f51055 100644 Binary files a/src/doc/rustc/src/images/image3.png and b/src/doc/rustc/src/images/image3.png differ diff --git a/src/doc/rustc/src/images/llvm-cov-show-01.png b/src/doc/rustc/src/images/llvm-cov-show-01.png index 35f04594347..ce4dec128b6 100644 Binary files a/src/doc/rustc/src/images/llvm-cov-show-01.png and b/src/doc/rustc/src/images/llvm-cov-show-01.png differ diff --git a/src/doc/rustdoc/src/images/collapsed-long-item.png b/src/doc/rustdoc/src/images/collapsed-long-item.png index c382870c64a..6de759fbeb9 100644 Binary files a/src/doc/rustdoc/src/images/collapsed-long-item.png and b/src/doc/rustdoc/src/images/collapsed-long-item.png differ diff --git a/src/doc/rustdoc/src/images/collapsed-trait-impls.png b/src/doc/rustdoc/src/images/collapsed-trait-impls.png index f685656e09a..96cc7db6798 100644 Binary files a/src/doc/rustdoc/src/images/collapsed-trait-impls.png and b/src/doc/rustdoc/src/images/collapsed-trait-impls.png differ diff --git a/src/etc/installer/gfx/rust-logo.png b/src/etc/installer/gfx/rust-logo.png index 99ee7507fa2..49d8d0d9485 100644 Binary files a/src/etc/installer/gfx/rust-logo.png and b/src/etc/installer/gfx/rust-logo.png differ diff --git a/src/librustdoc/html/static/images/favicon-32x32.png b/src/librustdoc/html/static/images/favicon-32x32.png index 69b8613ce15..0670c4dabb0 100644 Binary files a/src/librustdoc/html/static/images/favicon-32x32.png and b/src/librustdoc/html/static/images/favicon-32x32.png differ -- cgit 1.4.1-3-g733a5 From f702219ba61d94a50ecb581a4e0ab51dc459e99e Mon Sep 17 00:00:00 2001 From: Ralf Jung Date: Thu, 21 Aug 2025 16:12:43 +0200 Subject: update rustc-build-sysroot --- src/tools/miri/cargo-miri/Cargo.lock | 4 ++-- src/tools/miri/cargo-miri/Cargo.toml | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/tools/miri/cargo-miri/Cargo.lock b/src/tools/miri/cargo-miri/Cargo.lock index b3f5dafab64..ea9c04a3cb5 100644 --- a/src/tools/miri/cargo-miri/Cargo.lock +++ b/src/tools/miri/cargo-miri/Cargo.lock @@ -429,9 +429,9 @@ dependencies = [ [[package]] name = "rustc-build-sysroot" -version = "0.5.9" +version = "0.5.10" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "fdb13874a0e55baf4ac3d49d38206aecb31a55b75d6c4d04fd850b53942c8cc8" +checksum = "dd41ead66a69880951b2f7df3139db401d44451b4da123344d27eaa791b89c95" dependencies = [ "anyhow", "rustc_version", diff --git a/src/tools/miri/cargo-miri/Cargo.toml b/src/tools/miri/cargo-miri/Cargo.toml index 77cb1df8e74..64b56ea114e 100644 --- a/src/tools/miri/cargo-miri/Cargo.toml +++ b/src/tools/miri/cargo-miri/Cargo.toml @@ -18,7 +18,7 @@ directories = "6" rustc_version = "0.4" serde_json = "1.0.40" cargo_metadata = "0.21" -rustc-build-sysroot = "0.5.8" +rustc-build-sysroot = "0.5.10" # Enable some feature flags that dev-dependencies need but dependencies # do not. This makes `./miri install` after `./miri build` faster. -- cgit 1.4.1-3-g733a5 From 0a5383c35a56886a6638832862132eafb288f94b Mon Sep 17 00:00:00 2001 From: Daniel Paoliello Date: Wed, 20 Aug 2025 14:39:50 -0700 Subject: Promote aarch64-pc-windows-msvc to Tier 1 --- compiler/rustc_target/src/spec/targets/aarch64_pc_windows_msvc.rs | 2 +- src/bootstrap/src/core/build_steps/llvm.rs | 2 +- src/doc/rustc/src/platform-support.md | 2 +- src/doc/rustc/src/platform-support/windows-msvc.md | 5 +---- 4 files changed, 4 insertions(+), 7 deletions(-) diff --git a/compiler/rustc_target/src/spec/targets/aarch64_pc_windows_msvc.rs b/compiler/rustc_target/src/spec/targets/aarch64_pc_windows_msvc.rs index f1b6fa123de..cd55576ef81 100644 --- a/compiler/rustc_target/src/spec/targets/aarch64_pc_windows_msvc.rs +++ b/compiler/rustc_target/src/spec/targets/aarch64_pc_windows_msvc.rs @@ -15,7 +15,7 @@ pub(crate) fn target() -> Target { llvm_target: "aarch64-pc-windows-msvc".into(), metadata: TargetMetadata { description: Some("ARM64 Windows MSVC".into()), - tier: Some(2), + tier: Some(1), host_tools: Some(true), std: Some(true), }, diff --git a/src/bootstrap/src/core/build_steps/llvm.rs b/src/bootstrap/src/core/build_steps/llvm.rs index 260108292e0..303b185919e 100644 --- a/src/bootstrap/src/core/build_steps/llvm.rs +++ b/src/bootstrap/src/core/build_steps/llvm.rs @@ -205,6 +205,7 @@ pub(crate) fn is_ci_llvm_available_for_target( // tier 1 ("aarch64-unknown-linux-gnu", false), ("aarch64-apple-darwin", false), + ("aarch64-pc-windows-msvc", false), ("i686-pc-windows-gnu", false), ("i686-pc-windows-msvc", false), ("i686-unknown-linux-gnu", false), @@ -213,7 +214,6 @@ pub(crate) fn is_ci_llvm_available_for_target( ("x86_64-pc-windows-gnu", true), ("x86_64-pc-windows-msvc", true), // tier 2 with host tools - ("aarch64-pc-windows-msvc", false), ("aarch64-unknown-linux-musl", false), ("arm-unknown-linux-gnueabi", false), ("arm-unknown-linux-gnueabihf", false), diff --git a/src/doc/rustc/src/platform-support.md b/src/doc/rustc/src/platform-support.md index c039517a970..13f03ac7c42 100644 --- a/src/doc/rustc/src/platform-support.md +++ b/src/doc/rustc/src/platform-support.md @@ -33,6 +33,7 @@ All tier 1 targets with host tools support the full standard library. target | notes -------|------- [`aarch64-apple-darwin`](platform-support/apple-darwin.md) | ARM64 macOS (11.0+, Big Sur+) +[`aarch64-pc-windows-msvc`](platform-support/windows-msvc.md) | ARM64 Windows MSVC `aarch64-unknown-linux-gnu` | ARM64 Linux (kernel 4.1+, glibc 2.17+) [`i686-pc-windows-msvc`](platform-support/windows-msvc.md) | 32-bit MSVC (Windows 10+, Windows Server 2016+, Pentium 4) [^x86_32-floats-return-ABI] [^win32-msvc-alignment] `i686-unknown-linux-gnu` | 32-bit Linux (kernel 3.2+, glibc 2.17+, Pentium 4) [^x86_32-floats-return-ABI] @@ -88,7 +89,6 @@ so Rustup may install the documentation for a similar tier 1 target instead. target | notes -------|------- [`aarch64-pc-windows-gnullvm`](platform-support/windows-gnullvm.md) | ARM64 MinGW (Windows 10+), LLVM ABI -[`aarch64-pc-windows-msvc`](platform-support/windows-msvc.md) | ARM64 Windows MSVC [`aarch64-unknown-linux-musl`](platform-support/aarch64-unknown-linux-musl.md) | ARM64 Linux with musl 1.2.3 [`aarch64-unknown-linux-ohos`](platform-support/openharmony.md) | ARM64 OpenHarmony `arm-unknown-linux-gnueabi` | Armv6 Linux (kernel 3.2+, glibc 2.17) diff --git a/src/doc/rustc/src/platform-support/windows-msvc.md b/src/doc/rustc/src/platform-support/windows-msvc.md index 71dc4ddc2e6..826c75b79c5 100644 --- a/src/doc/rustc/src/platform-support/windows-msvc.md +++ b/src/doc/rustc/src/platform-support/windows-msvc.md @@ -4,13 +4,10 @@ Windows MSVC targets. **Tier 1 with host tools:** +- `aarch64-pc-windows-msvc`: Windows on ARM64. - `i686-pc-windows-msvc`: Windows on 32-bit x86. - `x86_64-pc-windows-msvc`: Windows on 64-bit x86. -**Tier 2 with host tools:** - -- `aarch64-pc-windows-msvc`: Windows on ARM64. - ## Target maintainers [@ChrisDenton](https://github.com/ChrisDenton) -- cgit 1.4.1-3-g733a5 From 5c3f317187c86fe06cd40a81b7072323a2f8ec55 Mon Sep 17 00:00:00 2001 From: Ralf Jung Date: Thu, 21 Aug 2025 21:20:28 +0200 Subject: CI: also test on powerpc --- src/tools/miri/.github/workflows/ci.yml | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/tools/miri/.github/workflows/ci.yml b/src/tools/miri/.github/workflows/ci.yml index 7d79c384f85..c0fed96d4e6 100644 --- a/src/tools/miri/.github/workflows/ci.yml +++ b/src/tools/miri/.github/workflows/ci.yml @@ -41,6 +41,11 @@ jobs: multiarch: s390x gcc_cross: s390x-linux-gnu qemu: true + - host_target: powerpc64le-unknown-linux-gnu + os: ubuntu-latest + multiarch: ppc64el + gcc_cross: powerpc64le-linux-gnu + qemu: true - host_target: aarch64-apple-darwin os: macos-latest - host_target: i686-pc-windows-msvc -- cgit 1.4.1-3-g733a5 From e50fed79a857ccf3ea516fb4d7fdab665b7a573c Mon Sep 17 00:00:00 2001 From: binarycat Date: Sun, 17 Aug 2025 12:35:54 -0500 Subject: add regression test for #145529 --- tests/rustdoc-ui/lints/invalid-html-tags.rs | 31 +++++++++++++++++++++++++ tests/rustdoc-ui/lints/invalid-html-tags.stderr | 14 ++++++++++- 2 files changed, 44 insertions(+), 1 deletion(-) diff --git a/tests/rustdoc-ui/lints/invalid-html-tags.rs b/tests/rustdoc-ui/lints/invalid-html-tags.rs index 317f1fd1d46..b2f516725fa 100644 --- a/tests/rustdoc-ui/lints/invalid-html-tags.rs +++ b/tests/rustdoc-ui/lints/invalid-html-tags.rs @@ -121,3 +121,34 @@ pub fn no_error_1() {} /// backslashed \< //~^ ERROR unclosed HTML tag `a` pub fn p() {} + +/// +/// +/// +pub fn no_error_2() {} + +///

+pub fn no_error_3() {} + +/// unfinished ALLOWED_UNCLOSED +/// +///
+/// +//~^ ERROR unclosed HTML tag `img` +pub fn r() {} diff --git a/tests/rustdoc-ui/lints/invalid-html-tags.stderr b/tests/rustdoc-ui/lints/invalid-html-tags.stderr index 9c2bfcf2c3d..fc9849ff23c 100644 --- a/tests/rustdoc-ui/lints/invalid-html-tags.stderr +++ b/tests/rustdoc-ui/lints/invalid-html-tags.stderr @@ -100,5 +100,17 @@ error: unclosed HTML tag `a` LL | /// backslashed \<
| ^^ -error: aborting due to 16 previous errors +error: unclosed HTML tag `img` + --> $DIR/invalid-html-tags.rs:147:5 + | +LL | /// $DIR/invalid-html-tags.rs:152:8 + | +LL | ///

+ | ^^^^ + +error: aborting due to 18 previous errors -- cgit 1.4.1-3-g733a5 From 15a8999aedc0e3a0eb7b4e956dd79b3a2f28f2a3 Mon Sep 17 00:00:00 2001 From: binarycat Date: Sun, 17 Aug 2025 11:23:19 -0500 Subject: refactor rustdoc::invalid_html_tags tag parser previously, this lint did not distinguish between ``, and since the latter should be accepted under html5, the former was also accepted. the parser now also handles multi-line tags and multi-line attributes. --- src/librustdoc/lib.rs | 1 + src/librustdoc/passes/lint/html_tags.rs | 476 +++++++++++++++--------- src/librustdoc/passes/lint/html_tags/tests.rs | 73 ++++ tests/rustdoc-ui/lints/invalid-html-tags.rs | 67 +++- tests/rustdoc-ui/lints/invalid-html-tags.stderr | 69 +++- 5 files changed, 498 insertions(+), 188 deletions(-) create mode 100644 src/librustdoc/passes/lint/html_tags/tests.rs diff --git a/src/librustdoc/lib.rs b/src/librustdoc/lib.rs index 28dbd8ba7d3..62e1ad2444d 100644 --- a/src/librustdoc/lib.rs +++ b/src/librustdoc/lib.rs @@ -11,6 +11,7 @@ #![feature(file_buffered)] #![feature(format_args_nl)] #![feature(if_let_guard)] +#![feature(iter_advance_by)] #![feature(iter_intersperse)] #![feature(round_char_boundary)] #![feature(rustc_private)] diff --git a/src/librustdoc/passes/lint/html_tags.rs b/src/librustdoc/passes/lint/html_tags.rs index 19cf15d40a3..b9c9279daec 100644 --- a/src/librustdoc/passes/lint/html_tags.rs +++ b/src/librustdoc/passes/lint/html_tags.rs @@ -1,9 +1,11 @@ //! Detects invalid HTML (like an unclosed ``) in doc comments. +use std::borrow::Cow; use std::iter::Peekable; use std::ops::Range; use std::str::CharIndices; +use itertools::Itertools as _; use pulldown_cmark::{BrokenLink, Event, LinkType, Parser, Tag, TagEnd}; use rustc_hir::HirId; use rustc_resolve::rustdoc::source_span_for_markdown_range; @@ -101,7 +103,7 @@ pub(crate) fn visit_item(cx: &DocContext<'_>, item: &Item, hir_id: HirId, dox: & }); }; - let mut tags = Vec::new(); + let mut tagp = TagParser::new(); let mut is_in_comment = None; let mut in_code_block = false; @@ -126,70 +128,65 @@ pub(crate) fn visit_item(cx: &DocContext<'_>, item: &Item, hir_id: HirId, dox: & }; let p = Parser::new_with_broken_link_callback(dox, main_body_opts(), Some(&mut replacer)) - .into_offset_iter(); + .into_offset_iter() + .coalesce(|a, b| { + // for some reason, pulldown-cmark splits html blocks into separate events for each line. + // we undo this, in order to handle multi-line tags. + match (a, b) { + ((Event::Html(_), ra), (Event::Html(_), rb)) if ra.end == rb.start => { + let merged = ra.start..rb.end; + Ok((Event::Html(Cow::Borrowed(&dox[merged.clone()]).into()), merged)) + } + x => Err(x), + } + }); for (event, range) in p { match event { Event::Start(Tag::CodeBlock(_)) => in_code_block = true, Event::Html(text) | Event::InlineHtml(text) if !in_code_block => { - extract_tags(&mut tags, &text, range, &mut is_in_comment, &report_diag) + tagp.extract_tags(&text, range, &mut is_in_comment, &report_diag) } Event::End(TagEnd::CodeBlock) => in_code_block = false, _ => {} } } - for (tag, range) in tags.iter().filter(|(t, _)| { - let t = t.to_lowercase(); - !ALLOWED_UNCLOSED.contains(&t.as_str()) - }) { - report_diag(format!("unclosed HTML tag `{tag}`"), range, true); - } - if let Some(range) = is_in_comment { report_diag("Unclosed HTML comment".to_string(), &range, false); + } else if let &Some(quote_pos) = &tagp.quote_pos { + let qr = Range { start: quote_pos, end: quote_pos }; + report_diag( + format!("unclosed quoted HTML attribute on tag `{}`", &tagp.tag_name), + &qr, + false, + ); + } else { + if !tagp.tag_name.is_empty() { + report_diag( + format!("incomplete HTML tag `{}`", &tagp.tag_name), + &(tagp.tag_start_pos..dox.len()), + false, + ); + } + for (tag, range) in tagp.tags.iter().filter(|(t, _)| { + let t = t.to_lowercase(); + !is_implicitly_self_closing(&t) + }) { + report_diag(format!("unclosed HTML tag `{tag}`"), range, true); + } } } +/// These tags are interpreted as self-closing if they lack an explicit closing tag. const ALLOWED_UNCLOSED: &[&str] = &[ "area", "base", "br", "col", "embed", "hr", "img", "input", "keygen", "link", "meta", "param", "source", "track", "wbr", ]; -fn drop_tag( - tags: &mut Vec<(String, Range)>, - tag_name: String, - range: Range, - f: &impl Fn(String, &Range, bool), -) { - let tag_name_low = tag_name.to_lowercase(); - if let Some(pos) = tags.iter().rposition(|(t, _)| t.to_lowercase() == tag_name_low) { - // If the tag is nested inside a "` (the `h2` tag isn't required - // but it helps for the visualization). - f(format!("unopened HTML tag `{tag_name}`"), &range, false); - } +/// Allows constructs like ``, but not ` bool { + ALLOWED_UNCLOSED.contains(&tag_name) } fn extract_path_backwards(text: &str, end_pos: usize) -> Option { @@ -252,151 +249,292 @@ fn is_valid_for_html_tag_name(c: char, is_empty: bool) -> bool { c.is_ascii_alphabetic() || !is_empty && (c == '-' || c.is_ascii_digit()) } -fn extract_html_tag( - tags: &mut Vec<(String, Range)>, - text: &str, - range: &Range, - start_pos: usize, - iter: &mut Peekable>, - f: &impl Fn(String, &Range, bool), -) { - let mut tag_name = String::new(); - let mut is_closing = false; - let mut prev_pos = start_pos; +/// Parse html tags to ensure they are well-formed +#[derive(Debug, Clone)] +struct TagParser { + tags: Vec<(String, Range)>, + /// Name of the tag that is being parsed, if we are within a tag. + /// + /// Since the `<` and name of a tag must appear on the same line with no whitespace, + /// if this is the empty string, we are not in a tag. + tag_name: String, + tag_start_pos: usize, + is_closing: bool, + /// `true` if we are within a tag, but not within its name. + in_attrs: bool, + /// If we are in a quoted attribute, what quote char does it use? + /// + /// This needs to be stored in the struct since HTML5 allows newlines in quoted attrs. + quote: Option, + quote_pos: Option, + after_eq: bool, +} - loop { - let (pos, c) = match iter.peek() { - Some((pos, c)) => (*pos, *c), - // In case we reached the of the doc comment, we want to check that it's an - // unclosed HTML tag. For example "/// (prev_pos, '\0'), - }; - prev_pos = pos; - // Checking if this is a closing tag (like `` for ``). - if c == '/' && tag_name.is_empty() { - is_closing = true; - } else if is_valid_for_html_tag_name(c, tag_name.is_empty()) { - tag_name.push(c); - } else { - if !tag_name.is_empty() { - let mut r = Range { start: range.start + start_pos, end: range.start + pos }; - if c == '>' { - // In case we have a tag without attribute, we can consider the span to - // refer to it fully. - r.end += 1; +impl TagParser { + fn new() -> Self { + Self { + tags: Vec::new(), + tag_name: String::with_capacity(8), + tag_start_pos: 0, + is_closing: false, + in_attrs: false, + quote: None, + quote_pos: None, + after_eq: false, + } + } + + fn drop_tag(&mut self, range: Range, f: &impl Fn(String, &Range, bool)) { + let tag_name_low = self.tag_name.to_lowercase(); + if let Some(pos) = self.tags.iter().rposition(|(t, _)| t.to_lowercase() == tag_name_low) { + // If the tag is nested inside a "` (the `h2` tag isn't required + // but it helps for the visualization). + f(format!("unopened HTML tag `{}`", &self.tag_name), &range, false); + } + } + + /// Handle a `<` that appeared while parsing a tag. + fn handle_lt_in_tag( + &mut self, + range: Range, + lt_pos: usize, + f: &impl Fn(String, &Range, bool), + ) { + let global_pos = range.start + lt_pos; + // is this check needed? + if global_pos == self.tag_start_pos { + // `<` is in the tag because it is the start. + return; + } + // tried to start a new tag while in a tag + f( + format!("incomplete HTML tag `{}`", &self.tag_name), + &(self.tag_start_pos..global_pos), + false, + ); + self.tag_parsed(); + } + + fn extract_html_tag( + &mut self, + text: &str, + range: &Range, + start_pos: usize, + iter: &mut Peekable>, + f: &impl Fn(String, &Range, bool), + ) { + let mut prev_pos = start_pos; + + 'outer_loop: loop { + let (pos, c) = match iter.peek() { + Some((pos, c)) => (*pos, *c), + // In case we reached the of the doc comment, we want to check that it's an + // unclosed HTML tag. For example "/// (prev_pos, '\0'), + None => break, + }; + prev_pos = pos; + if c == '/' && self.tag_name.is_empty() { + // Checking if this is a closing tag (like `` for ``). + self.is_closing = true; + } else if !self.in_attrs && is_valid_for_html_tag_name(c, self.tag_name.is_empty()) { + self.tag_name.push(c); + } else { + if !self.tag_name.is_empty() { + self.in_attrs = true; + let mut r = Range { start: range.start + start_pos, end: range.start + pos }; + if c == '>' { + // In case we have a tag without attribute, we can consider the span to + // refer to it fully. + r.end += 1; + } + if self.is_closing { + // In case we have "" or even "". + if c != '>' { if !c.is_whitespace() { - if c == '>' { - r.end = range.start + new_pos + 1; - found = true; - } + // It seems like it's not a valid HTML tag. break; } - } - if !found { - break; - } - } - drop_tag(tags, tag_name, r, f); - } else { - let mut is_self_closing = false; - let mut quote_pos = None; - if c != '>' { - let mut quote = None; - let mut after_eq = false; - for (i, c) in text[pos..].char_indices() { - if !c.is_whitespace() { - if let Some(q) = quote { - if c == q { - quote = None; - quote_pos = None; - after_eq = false; + let mut found = false; + for (new_pos, c) in text[pos..].char_indices() { + if !c.is_whitespace() { + if c == '>' { + r.end = range.start + new_pos + 1; + found = true; + } else if c == '<' { + self.handle_lt_in_tag(range.clone(), pos + new_pos, f); } - } else if c == '>' { break; - } else if c == '/' && !after_eq { - is_self_closing = true; - } else { - if is_self_closing { - is_self_closing = false; - } - if (c == '"' || c == '\'') && after_eq { - quote = Some(c); - quote_pos = Some(pos + i); - } else if c == '=' { - after_eq = true; - } } - } else if quote.is_none() { - after_eq = false; + } + if !found { + break 'outer_loop; } } - } - if let Some(quote_pos) = quote_pos { - let qr = Range { start: quote_pos, end: quote_pos }; - f( - format!("unclosed quoted HTML attribute on tag `{tag_name}`"), - &qr, - false, - ); - } - if is_self_closing { - // https://html.spec.whatwg.org/#parse-error-non-void-html-element-start-tag-with-trailing-solidus - let valid = ALLOWED_UNCLOSED.contains(&&tag_name[..]) - || tags.iter().take(pos + 1).any(|(at, _)| { - let at = at.to_lowercase(); - at == "svg" || at == "math" - }); - if !valid { - f(format!("invalid self-closing HTML tag `{tag_name}`"), &r, false); - } + self.drop_tag(r, f); + self.tag_parsed(); } else { - tags.push((tag_name, r)); + self.extract_opening_tag(text, range, r, pos, c, iter, f) } } + break; } - break; + iter.next(); } - iter.next(); } -} - -fn extract_tags( - tags: &mut Vec<(String, Range)>, - text: &str, - range: Range, - is_in_comment: &mut Option>, - f: &impl Fn(String, &Range, bool), -) { - let mut iter = text.char_indices().peekable(); - while let Some((start_pos, c)) = iter.next() { - if is_in_comment.is_some() { - if text[start_pos..].starts_with("-->") { - *is_in_comment = None; + fn extract_opening_tag( + &mut self, + text: &str, + range: &Range, + r: Range, + pos: usize, + c: char, + iter: &mut Peekable>, + f: &impl Fn(String, &Range, bool), + ) { + // we can store this as a local, since html5 does require the `/` and `>` + // to not be separated by whitespace. + let mut is_self_closing = false; + if c != '>' { + 'parse_til_gt: { + for (i, c) in text[pos..].char_indices() { + if !c.is_whitespace() { + debug_assert_eq!(self.quote_pos.is_some(), self.quote.is_some()); + if let Some(q) = self.quote { + if c == q { + self.quote = None; + self.quote_pos = None; + self.after_eq = false; + } + } else if c == '>' { + break 'parse_til_gt; + } else if c == '<' { + self.handle_lt_in_tag(range.clone(), pos + i, f); + } else if c == '/' && !self.after_eq { + is_self_closing = true; + } else { + if is_self_closing { + is_self_closing = false; + } + if (c == '"' || c == '\'') && self.after_eq { + self.quote = Some(c); + self.quote_pos = Some(pos + i); + } else if c == '=' { + self.after_eq = true; + } + } + } else if self.quote.is_none() { + self.after_eq = false; + } + if !is_self_closing && !self.tag_name.is_empty() { + iter.next(); + } + } + // if we've run out of text but still haven't found a `>`, + // return early without calling `tag_parsed` or emitting lints. + // this allows us to either find the `>` in a later event + // or emit a lint about it being missing. + return; } - } else if c == '<' { - if text[start_pos..].starts_with("") { + *is_in_comment = None; + } + } else if c == '<' { + // " @@ -105,7 +106,7 @@ pub fn j() {} /// uiapp.run(&env::args().collect::>()); /// ``` /// -/// shouldn't warn! +// shouldn't warn! /// `````` pub fn k() {} @@ -141,14 +142,72 @@ pub fn no_error_2() {} /// pub fn no_error_3() {} +/// >

+pub fn no_error_4() {} + /// unfinished ALLOWED_UNCLOSED /// +/// note: CommonMark doesn't allow an html block to start with a multiline tag, +/// so we use `
` a bunch to force these to be parsed as html blocks. +/// ///
/// -//~^ ERROR unclosed HTML tag `img` +//~^ ERROR incomplete HTML tag `img` pub fn r() {} + +/// >
+/// > href="#broken" +pub fn s() {} + +///
+/// +//~^ ERROR incomplete HTML tag `br` +pub fn t() {} + +///
+///
html5 allows this
+pub fn no_error_5() {} + +///
+/// +pub fn no_error_6() {} + +///
+/// what +pub fn no_error_7() {} + +/// Technically this is allowed per the html5 spec, +/// but there's basically no legitemate reason to do it, +/// so we don't allow it. +/// +///

foobar

+//~^ ERROR Unclosed HTML comment +//~| ERROR incomplete HTML tag `p` +pub fn v() {} diff --git a/tests/rustdoc-ui/lints/invalid-html-tags.stderr b/tests/rustdoc-ui/lints/invalid-html-tags.stderr index fc9849ff23c..b6ec22c2479 100644 --- a/tests/rustdoc-ui/lints/invalid-html-tags.stderr +++ b/tests/rustdoc-ui/lints/invalid-html-tags.stderr @@ -52,6 +52,12 @@ error: unclosed HTML tag `p` LL | ///

| ^^^ +error: incomplete HTML tag `script` + --> $DIR/invalid-html-tags.rs:45:5 + | +LL | ///