Re: Name mangling once again

[Martin Kahlert <martin dot kahlert at infineon dot com>]

Hallo Alexandre!

Sorry for the late reply, but i have been at home for the last few days,
so i could not check your patch until today.

Result: It works like a charm again!

Thanks,
Martin.

On Mon, Feb 05, 2001 at 04:04:05PM -0800, Alexandre Petit-Bianco wrote:
> 
> I don't feel like changing this setup now. So I'll revert part of the
> patch I wrote to fix a bogus error report because we weren't taking
> into account the `Synthetic' attribute in method names (only.) This
> patch should fix your problem, let me know.
> 
> ./A
> 
> Index: jcf-parse.c
> ===================================================================
> RCS file: /cvs/gcc/egcs/gcc/java/jcf-parse.c,v
> retrieving revision 1.74
> diff -u -p -r1.74 jcf-parse.c
> --- jcf-parse.c 2001/02/01 10:35:50     1.74
> +++ jcf-parse.c 2001/02/06 00:01:03
> @@ -224,11 +224,12 @@ set_source_filename (jcf, index)
>  
>  #define HANDLE_SYNTHETIC_ATTRIBUTE()                                   \
>  {                                                                      \
> -  /* Irrelevant decls should have been nullified by the END macros. */ \
> +  /* Irrelevant decls should have been nullified by the END macros.    \
> +     We only handle the `Synthetic' attribute on method DECLs.         \
> +     DECL_ARTIFICIAL is used for something else (See PUSH_FIELD in     \
> +     java-tree.h)  */                                                  \
>    if (current_method)                                                  \
>      DECL_ARTIFICIAL (current_method) = 1;                              \
> -  else                                                                 \
> -    DECL_ARTIFICIAL (current_field) = 1;                               \
>  }
>  
>  #include "jcf-reader.c"

-- 
The early bird gets the worm. If you want something else for       
breakfast, get up later.