reimplement C string literals

1 files changed, 25 insertions, 1 deletions

diff --git a/compiler/rustc_parse/src/lexer/mod.rs b/compiler/rustc_parse/src/lexer/mod.rs
index c6e6b46e455..cfcc2ec42fa 100644
--- a/compiler/rustc_parse/src/lexer/mod.rs
+++ b/compiler/rustc_parse/src/lexer/mod.rs
@@ -9,8 +9,8 @@ use rustc_ast::tokenstream::TokenStream;
 use rustc_ast::util::unicode::contains_text_flow_control_chars;
 use rustc_errors::{error_code, Applicability, Diagnostic, DiagnosticBuilder, StashKey};
 use rustc_lexer::unescape::{self, EscapeError, Mode};
-use rustc_lexer::Cursor;
 use rustc_lexer::{Base, DocStyle, RawStrError};
+use rustc_lexer::{Cursor, LiteralKind};
 use rustc_session::lint::builtin::{
     RUST_2021_PREFIXES_INCOMPATIBLE_SYNTAX, TEXT_DIRECTION_CODEPOINT_IN_COMMENT,
 };
@@ -118,6 +118,7 @@ impl<'a> StringReader<'a> {
         let mut swallow_next_invalid = 0;
         // Skip trivial (whitespace & comments) tokens
         loop {
+            let str_before = self.cursor.as_str();
             let token = self.cursor.advance_token();
             let start = self.pos;
             self.pos = self.pos + BytePos(token.len);
@@ -203,6 +204,29 @@ impl<'a> StringReader<'a> {
                         .push(span);
                     token::Ident(sym, false)
                 }
+                // split up (raw) c string literals to an ident and a string literal when edition < 2021.
+                rustc_lexer::TokenKind::Literal {
+                    kind: kind @ (LiteralKind::CStr { .. } | LiteralKind::RawCStr { .. }),
+                    suffix_start: _,
+                } if !self.mk_sp(start, self.pos).edition().at_least_rust_2021() => {
+                    let prefix_len = match kind {
+                        LiteralKind::CStr { .. } => 1,
+                        LiteralKind::RawCStr { .. } => 2,
+                        _ => unreachable!(),
+                    };
+
+                    // reset the state so that only the prefix ("c" or "cr")
+                    // was consumed.
+                    let lit_start = start + BytePos(prefix_len);
+                    self.pos = lit_start;
+                    self.cursor = Cursor::new(&str_before[prefix_len as usize..]);
+
+                    self.report_unknown_prefix(start);
+                    let sym = nfc_normalize(self.str_from(start));
+                    let prefix_span = self.mk_sp(start, lit_start);
+                    self.sess.symbol_gallery.insert(sym, prefix_span);
+                    return (Token::new(token::Ident(sym, false), prefix_span), preceded_by_whitespace);
+                }
                 rustc_lexer::TokenKind::Literal { kind, suffix_start } => {
                     let suffix_start = start + BytePos(suffix_start);
                     let (kind, symbol) = self.cook_lexer_literal(start, suffix_start, kind);