Extract `unescape` from `rustc_lexer` into its own crate

14 files changed, 36 insertions, 12 deletions

diff --git a/Cargo.lock b/Cargo.lock
index fc4c7c9888f..63441814d4a 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -2152,6 +2152,10 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "4ee93343901ab17bd981295f2cf0026d4ad018c7c31ba84549a4ddbb47a45104"
 
 [[package]]
+name = "literal-escaper"
+version = "0.0.1"
+
+[[package]]
 name = "lld-wrapper"
 version = "0.1.0"
 
@@ -3366,6 +3370,7 @@ name = "rustc_ast"
 version = "0.0.0"
 dependencies = [
  "bitflags",
+ "literal-escaper",
  "memchr",
  "rustc_ast_ir",
  "rustc_data_structures",
@@ -4325,6 +4330,7 @@ name = "rustc_parse"
 version = "0.0.0"
 dependencies = [
  "bitflags",
+ "literal-escaper",
  "rustc_ast",
  "rustc_ast_pretty",
  "rustc_data_structures",
@@ -4347,6 +4353,7 @@ dependencies = [
 name = "rustc_parse_format"
 version = "0.0.0"
 dependencies = [
+ "literal-escaper",
  "rustc_index",
  "rustc_lexer",
 ]
diff --git a/compiler/rustc_ast/Cargo.toml b/compiler/rustc_ast/Cargo.toml
index 34c612dac69..176c2d6f8a7 100644
--- a/compiler/rustc_ast/Cargo.toml
+++ b/compiler/rustc_ast/Cargo.toml
@@ -6,6 +6,7 @@ edition = "2021"
 [dependencies]
 # tidy-alphabetical-start
 bitflags = "2.4.1"
+literal-escaper = { path = "../../library/literal-escaper" }
 memchr = "2.7.4"
 rustc_ast_ir = { path = "../rustc_ast_ir" }
 rustc_data_structures = { path = "../rustc_data_structures" }
diff --git a/compiler/rustc_ast/src/util/literal.rs b/compiler/rustc_ast/src/util/literal.rs
index 6896ac723fa..121331ece6d 100644
--- a/compiler/rustc_ast/src/util/literal.rs
+++ b/compiler/rustc_ast/src/util/literal.rs
@@ -2,7 +2,7 @@
 
 use std::{ascii, fmt, str};
 
-use rustc_lexer::unescape::{
+use literal_escaper::{
     MixedUnit, Mode, byte_from_char, unescape_byte, unescape_char, unescape_mixed, unescape_unicode,
 };
 use rustc_span::{Span, Symbol, kw, sym};
diff --git a/compiler/rustc_lexer/src/lib.rs b/compiler/rustc_lexer/src/lib.rs
index c63ab77deca..334f451a39c 100644
--- a/compiler/rustc_lexer/src/lib.rs
+++ b/compiler/rustc_lexer/src/lib.rs
@@ -27,7 +27,6 @@
 // tidy-alphabetical-end
 
 mod cursor;
-pub mod unescape;
 
 #[cfg(test)]
 mod tests;
diff --git a/compiler/rustc_parse/Cargo.toml b/compiler/rustc_parse/Cargo.toml
index 2360914a0ab..109e1f5c093 100644
--- a/compiler/rustc_parse/Cargo.toml
+++ b/compiler/rustc_parse/Cargo.toml
@@ -6,6 +6,7 @@ edition = "2021"
 [dependencies]
 # tidy-alphabetical-start
 bitflags = "2.4.1"
+literal-escaper = { path = "../../library/literal-escaper" }
 rustc_ast = { path = "../rustc_ast" }
 rustc_ast_pretty = { path = "../rustc_ast_pretty" }
 rustc_data_structures = { path = "../rustc_data_structures" }
diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs
index 792e2cc26ef..cefaf13e31f 100644
--- a/compiler/rustc_parse/src/lexer/mod.rs
+++ b/compiler/rustc_parse/src/lexer/mod.rs
@@ -1,12 +1,12 @@
 use std::ops::Range;
 
+use literal_escaper::{self, EscapeError, Mode};
 use rustc_ast::ast::{self, AttrStyle};
 use rustc_ast::token::{self, CommentKind, Delimiter, IdentIsRaw, Token, TokenKind};
 use rustc_ast::tokenstream::TokenStream;
 use rustc_ast::util::unicode::contains_text_flow_control_chars;
 use rustc_errors::codes::*;
 use rustc_errors::{Applicability, Diag, DiagCtxtHandle, StashKey};
-use rustc_lexer::unescape::{self, EscapeError, Mode};
 use rustc_lexer::{Base, Cursor, DocStyle, LiteralKind, RawStrError};
 use rustc_session::lint::BuiltinLintDiag;
 use rustc_session::lint::builtin::{
@@ -970,7 +970,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
         postfix_len: u32,
     ) -> (token::LitKind, Symbol) {
         self.cook_common(kind, mode, start, end, prefix_len, postfix_len, |src, mode, callback| {
-            unescape::unescape_unicode(src, mode, &mut |span, result| {
+            literal_escaper::unescape_unicode(src, mode, &mut |span, result| {
                 callback(span, result.map(drop))
             })
         })
@@ -986,7 +986,7 @@ impl<'psess, 'src> Lexer<'psess, 'src> {
         postfix_len: u32,
     ) -> (token::LitKind, Symbol) {
         self.cook_common(kind, mode, start, end, prefix_len, postfix_len, |src, mode, callback| {
-            unescape::unescape_mixed(src, mode, &mut |span, result| {
+            literal_escaper::unescape_mixed(src, mode, &mut |span, result| {
                 callback(span, result.map(drop))
             })
         })
diff --git a/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs b/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs
index 2e066f0179c..e8aa400e73d 100644
--- a/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs
+++ b/compiler/rustc_parse/src/lexer/unescape_error_reporting.rs
@@ -3,8 +3,8 @@
 use std::iter::once;
 use std::ops::Range;
 
+use literal_escaper::{EscapeError, Mode};
 use rustc_errors::{Applicability, DiagCtxtHandle, ErrorGuaranteed};
-use rustc_lexer::unescape::{EscapeError, Mode};
 use rustc_span::{BytePos, Span};
 use tracing::debug;
 
diff --git a/compiler/rustc_parse/src/parser/expr.rs b/compiler/rustc_parse/src/parser/expr.rs
index e0e6c2177da..0b745217bc8 100644
--- a/compiler/rustc_parse/src/parser/expr.rs
+++ b/compiler/rustc_parse/src/parser/expr.rs
@@ -6,6 +6,7 @@ use core::ops::{Bound, ControlFlow};
 use ast::mut_visit::{self, MutVisitor};
 use ast::token::IdentIsRaw;
 use ast::{CoroutineKind, ForLoopKind, GenBlockKind, MatchKind, Pat, Path, PathSegment, Recovered};
+use literal_escaper::unescape_char;
 use rustc_ast::ptr::P;
 use rustc_ast::token::{self, Delimiter, Token, TokenKind};
 use rustc_ast::tokenstream::TokenTree;
@@ -21,7 +22,6 @@ use rustc_ast::{
 use rustc_ast_pretty::pprust;
 use rustc_data_structures::stack::ensure_sufficient_stack;
 use rustc_errors::{Applicability, Diag, PResult, StashKey, Subdiagnostic};
-use rustc_lexer::unescape::unescape_char;
 use rustc_macros::Subdiagnostic;
 use rustc_session::errors::{ExprParenthesesNeeded, report_lit_error};
 use rustc_session::lint::BuiltinLintDiag;
diff --git a/compiler/rustc_parse_format/Cargo.toml b/compiler/rustc_parse_format/Cargo.toml
index 707c4e31847..51ea46f4c9b 100644
--- a/compiler/rustc_parse_format/Cargo.toml
+++ b/compiler/rustc_parse_format/Cargo.toml
@@ -5,6 +5,7 @@ edition = "2021"
 
 [dependencies]
 # tidy-alphabetical-start
+literal-escaper = { path = "../../library/literal-escaper" }
 rustc_index = { path = "../rustc_index", default-features = false }
 rustc_lexer = { path = "../rustc_lexer" }
 # tidy-alphabetical-end
diff --git a/compiler/rustc_parse_format/src/lib.rs b/compiler/rustc_parse_format/src/lib.rs
index 3b985621b57..7e89f9b079b 100644
--- a/compiler/rustc_parse_format/src/lib.rs
+++ b/compiler/rustc_parse_format/src/lib.rs
@@ -19,7 +19,6 @@
 pub use Alignment::*;
 pub use Count::*;
 pub use Position::*;
-use rustc_lexer::unescape;
 
 // Note: copied from rustc_span
 /// Range inside of a `Span` used for diagnostics when we only have access to relative positions.
@@ -1095,12 +1094,14 @@ fn find_width_map_from_snippet(
 fn unescape_string(string: &str) -> Option<String> {
     let mut buf = String::new();
     let mut ok = true;
-    unescape::unescape_unicode(string, unescape::Mode::Str, &mut |_, unescaped_char| {
-        match unescaped_char {
+    literal_escaper::unescape_unicode(
+        string,
+        literal_escaper::Mode::Str,
+        &mut |_, unescaped_char| match unescaped_char {
             Ok(c) => buf.push(c),
             Err(_) => ok = false,
-        }
-    });
+        },
+    );
 
     ok.then_some(buf)
 }
diff --git a/library/literal-escaper/Cargo.toml b/library/literal-escaper/Cargo.toml
new file mode 100644
index 00000000000..708fcd3cacb
--- /dev/null
+++ b/library/literal-escaper/Cargo.toml
@@ -0,0 +1,10 @@
+[package]
+name = "literal-escaper"
+version = "0.0.0"
+edition = "2021"
+
+[dependencies]
+std = { version = '1.0.0', optional = true, package = 'rustc-std-workspace-std' }
+
+[features]
+rustc-dep-of-std = ["dep:std"]
diff --git a/library/literal-escaper/README.md b/library/literal-escaper/README.md
new file mode 100644
index 00000000000..5384ac4556a
--- /dev/null
+++ b/library/literal-escaper/README.md
@@ -0,0 +1,4 @@
+# literal-escaper
+
+This crate provides code to unescape string literals. It is used by `rustc_lexer`
+and `proc-macro`.
diff --git a/compiler/rustc_lexer/src/unescape.rs b/library/literal-escaper/src/lib.rs
index d6ea4249247..d6ea4249247 100644
--- a/compiler/rustc_lexer/src/unescape.rs
+++ b/library/literal-escaper/src/lib.rs
diff --git a/compiler/rustc_lexer/src/unescape/tests.rs b/library/literal-escaper/src/tests.rs
index 5b99495f475..5b99495f475 100644
--- a/compiler/rustc_lexer/src/unescape/tests.rs
+++ b/library/literal-escaper/src/tests.rs