[PATCH] c: Fix up c_parser_has_attribute_expression [PR101176]

Marek Polacek polacek@redhat.com
Thu Jun 24 13:07:24 GMT 2021


On Thu, Jun 24, 2021 at 12:20:56PM +0200, Jakub Jelinek wrote:
> Hi!
> 
> This function keeps src_range member of the result uninitialized, which at
> least under valgrind can show up later when those uninitialized location_t's
> can make it into the IL or location_t hash tables.
> 
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
> trunk?

Ok, thanks.
 
> 2021-06-24  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR c/101176
> 	* c-parser.c (c_parser_has_attribute_expression): Set source range for
> 	the result.
> 
> --- gcc/c/c-parser.c.jj	2021-06-23 13:33:00.374434233 +0200
> +++ gcc/c/c-parser.c	2021-06-23 23:53:45.146083923 +0200
> @@ -8406,6 +8406,7 @@ c_parser_has_attribute_expression (c_par
>  {
>    gcc_assert (c_parser_next_token_is_keyword (parser,
>  					      RID_BUILTIN_HAS_ATTRIBUTE));
> +  location_t start = c_parser_peek_token (parser)->location;
>    c_parser_consume_token (parser);
>  
>    c_inhibit_evaluation_warnings++;
> @@ -8484,6 +8485,7 @@ c_parser_has_attribute_expression (c_par
>  
>    parser->translate_strings_p = save_translate_strings_p;
>  
> +  location_t finish = c_parser_peek_token (parser)->location;
>    if (c_parser_next_token_is (parser, CPP_CLOSE_PAREN))
>      c_parser_consume_token (parser);
>    else
> @@ -8512,6 +8514,7 @@ c_parser_has_attribute_expression (c_par
>    else
>      result.value =  boolean_false_node;
>  
> +  set_c_expr_source_range (&result, start, finish);
>    return result;
>  }

Marek



More information about the Gcc-patches mailing list