diff options
| author | Guillaume Gomez <guillaume1.gomez@gmail.com> | 2025-04-04 15:04:17 +0200 |
|---|---|---|
| committer | Guillaume Gomez <guillaume1.gomez@gmail.com> | 2025-04-04 15:04:17 +0200 |
| commit | 13400c2c10f04fe49f1765ceb2de837c86b6f915 (patch) | |
| tree | c231036187de12782b93bfa7f60e30f1e816ead6 /library/proc_macro | |
| parent | bba0d57bb2773c0298e1b37456754610f069dd4a (diff) | |
| download | rust-13400c2c10f04fe49f1765ceb2de837c86b6f915.tar.gz rust-13400c2c10f04fe49f1765ceb2de837c86b6f915.zip | |
Add `_value` methods to proc_macro lib
Diffstat (limited to 'library/proc_macro')
| -rw-r--r-- | library/proc_macro/Cargo.toml | 1 | ||||
| -rw-r--r-- | library/proc_macro/src/lib.rs | 114 |
2 files changed, 115 insertions, 0 deletions
diff --git a/library/proc_macro/Cargo.toml b/library/proc_macro/Cargo.toml index 72cb4e4166f..f65c5e1a330 100644 --- a/library/proc_macro/Cargo.toml +++ b/library/proc_macro/Cargo.toml @@ -9,3 +9,4 @@ std = { path = "../std" } # `core` when resolving doc links. Without this line a different `core` will be # loaded from sysroot causing duplicate lang items and other similar errors. core = { path = "../core" } +rustc-literal-escaper = "0.0.1" diff --git a/library/proc_macro/src/lib.rs b/library/proc_macro/src/lib.rs index d9141eab591..35a8af03ba4 100644 --- a/library/proc_macro/src/lib.rs +++ b/library/proc_macro/src/lib.rs @@ -51,11 +51,24 @@ use std::{error, fmt}; #[unstable(feature = "proc_macro_diagnostic", issue = "54140")] pub use diagnostic::{Diagnostic, Level, MultiSpan}; +#[unstable(feature = "proc_macro_value", issue = "136652")] +pub use rustc_literal_escaper::EscapeError; +use rustc_literal_escaper::{MixedUnit, Mode, byte_from_char, unescape_mixed, unescape_unicode}; #[unstable(feature = "proc_macro_totokens", issue = "130977")] pub use to_tokens::ToTokens; use crate::escape::{EscapeOptions, escape_bytes}; +/// Errors returned when trying to retrieve a literal unescaped value. +#[unstable(feature = "proc_macro_value", issue = "136652")] +#[derive(Debug, PartialEq, Eq)] +pub enum ConversionErrorKind { + /// The literal failed to be escaped, take a look at [`EscapeError`] for more information. + FailedToUnescape(EscapeError), + /// Trying to convert a literal with the wrong type. + InvalidLiteralKind, +} + /// Determines whether proc_macro has been made accessible to the currently /// running program. /// @@ -1451,6 +1464,107 @@ impl Literal { } }) } + + /// Returns the unescaped string value if the current literal is a string or a string literal. + #[unstable(feature = "proc_macro_value", issue = "136652")] + pub fn str_value(&self) -> Result<String, ConversionErrorKind> { + self.0.symbol.with(|symbol| match self.0.kind { + bridge::LitKind::Str => { + if symbol.contains('\\') { + let mut buf = String::with_capacity(symbol.len()); + let mut error = None; + // Force-inlining here is aggressive but the closure is + // called on every char in the string, so it can be hot in + // programs with many long strings containing escapes. + unescape_unicode( + symbol, + Mode::Str, + &mut #[inline(always)] + |_, c| match c { + Ok(c) => buf.push(c), + Err(err) => { + if err.is_fatal() { + error = Some(ConversionErrorKind::FailedToUnescape(err)); + } + } + }, + ); + if let Some(error) = error { Err(error) } else { Ok(buf) } + } else { + Ok(symbol.to_string()) + } + } + bridge::LitKind::StrRaw(_) => Ok(symbol.to_string()), + _ => Err(ConversionErrorKind::InvalidLiteralKind), + }) + } + + /// Returns the unescaped string value if the current literal is a c-string or a c-string + /// literal. + #[unstable(feature = "proc_macro_value", issue = "136652")] + pub fn cstr_value(&self) -> Result<Vec<u8>, ConversionErrorKind> { + self.0.symbol.with(|symbol| match self.0.kind { + bridge::LitKind::CStr => { + let mut error = None; + let mut buf = Vec::with_capacity(symbol.len()); + + unescape_mixed(symbol, Mode::CStr, &mut |_span, c| match c { + Ok(MixedUnit::Char(c)) => { + buf.extend_from_slice(c.encode_utf8(&mut [0; 4]).as_bytes()) + } + Ok(MixedUnit::HighByte(b)) => buf.push(b), + Err(err) => { + if err.is_fatal() { + error = Some(ConversionErrorKind::FailedToUnescape(err)); + } + } + }); + if let Some(error) = error { + Err(error) + } else { + buf.push(0); + Ok(buf) + } + } + bridge::LitKind::CStrRaw(_) => { + // Raw strings have no escapes so we can convert the symbol + // directly to a `Lrc<u8>` after appending the terminating NUL + // char. + let mut buf = symbol.to_owned().into_bytes(); + buf.push(0); + Ok(buf) + } + _ => Err(ConversionErrorKind::InvalidLiteralKind), + }) + } + + /// Returns the unescaped string value if the current literal is a byte string or a byte string + /// literal. + #[unstable(feature = "proc_macro_value", issue = "136652")] + pub fn byte_str_value(&self) -> Result<Vec<u8>, ConversionErrorKind> { + self.0.symbol.with(|symbol| match self.0.kind { + bridge::LitKind::ByteStr => { + let mut buf = Vec::with_capacity(symbol.len()); + let mut error = None; + + unescape_unicode(symbol, Mode::ByteStr, &mut |_, c| match c { + Ok(c) => buf.push(byte_from_char(c)), + Err(err) => { + if err.is_fatal() { + error = Some(ConversionErrorKind::FailedToUnescape(err)); + } + } + }); + if let Some(error) = error { Err(error) } else { Ok(buf) } + } + bridge::LitKind::ByteStrRaw(_) => { + // Raw strings have no escapes so we can convert the symbol + // directly to a `Lrc<u8>`. + Ok(symbol.to_owned().into_bytes()) + } + _ => Err(ConversionErrorKind::InvalidLiteralKind), + }) + } } /// Parse a single literal from its stringified representation. |
