[gcc/devel/rust/master] Improve parsing of raw byte string literals

[Thomas Schwinge via Gcc-cvs]

https://gcc.gnu.org/g:67928be5b05ab927baf0a366689c246ec61ed745

commit 67928be5b05ab927baf0a366689c246ec61ed745
Author: Owen Avery <powerboat9.ga...@gmail.com>
Date:   Tue Mar 26 20:16:08 2024 -0400

    Improve parsing of raw byte string literals
    
    gcc/rust/ChangeLog:
    
            * lex/rust-lex.cc
            (Lexer::parse_raw_byte_string):
            Bring handling of edge cases to par with parse_byte_string.
    
    gcc/testsuite/ChangeLog:
    
            * rust/compile/raw-byte-string-loc.rs: New test.
    
    Signed-off-by: Owen Avery <powerboat9.ga...@gmail.com>

Diff:
---
 gcc/rust/lex/rust-lex.cc                          | 23 +++++++++++++++++++----
 gcc/testsuite/rust/compile/raw-byte-string-loc.rs |  6 ++++++
 2 files changed, 25 insertions(+), 4 deletions(-)

diff --git a/gcc/rust/lex/rust-lex.cc b/gcc/rust/lex/rust-lex.cc
index 9c2203160cd4..7c37e83d6cb7 100644
--- a/gcc/rust/lex/rust-lex.cc
+++ b/gcc/rust/lex/rust-lex.cc
@@ -1840,14 +1840,18 @@ Lexer::parse_raw_byte_string (location_t loc)
   int length = 1;
   int hash_count = 0;
 
+  const location_t string_begin_locus = get_current_location ();
+
   // get hash count at beginnning
   skip_input ();
   current_char = peek_input ();
   length++;
+  current_column++;
   while (current_char == '#')
     {
       hash_count++;
       length++;
+      current_column++;
 
       skip_input ();
       current_char = peek_input ();
@@ -1862,6 +1866,7 @@ Lexer::parse_raw_byte_string (location_t loc)
   skip_input ();
   current_char = peek_input ();
   length++;
+  current_column++;
 
   while (true)
     {
@@ -1884,27 +1889,37 @@ Lexer::parse_raw_byte_string (location_t loc)
              skip_input (hash_count);
              current_char = peek_input ();
              length += hash_count + 1;
+             current_column += hash_count + 1;
              break;
            }
        }
-
-      if (current_char.value > 127)
+      else if (current_char.value > 127)
        {
          rust_error_at (get_current_location (),
                         "character %<%s%> in raw byte string out of range",
                         current_char.as_string ().c_str ());
          current_char = 0;
        }
+      else if (current_char.is_eof ())
+       {
+         rust_error_at (string_begin_locus, "unended raw byte string literal");
+         return Token::make (END_OF_FILE, get_current_location ());
+       }
 
       length++;
+      current_column++;
+      if (current_char == '\n')
+       {
+         current_line++;
+         current_column = 1;
+         start_line (current_line, max_column_hint);
+       }
 
       str += current_char;
       skip_input ();
       current_char = peek_input ();
     }
 
-  current_column += length;
-
   loc += length - 1;
 
   str.shrink_to_fit ();
diff --git a/gcc/testsuite/rust/compile/raw-byte-string-loc.rs 
b/gcc/testsuite/rust/compile/raw-byte-string-loc.rs
new file mode 100644
index 000000000000..f37d3f9694d3
--- /dev/null
+++ b/gcc/testsuite/rust/compile/raw-byte-string-loc.rs
@@ -0,0 +1,6 @@
+const X: &'static u8 = br#"12
+12"#;
+
+BREAK
+// { dg-error "unrecognised token" "" { target *-*-* } .-1 }
+// { dg-excess-errors "error 'failed to parse item' does not have location" }