This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Basic block profiler out of libgcc


On Mon, Apr 16, 2001 at 07:40:05AM -0700, Mark Mitchell wrote:
> >>>>> "Zack" == Zack Weinberg <zackw@Stanford.EDU> writes:
>     Zack> IIRC that was in a lot of config header files, that now have
>     Zack> no code matching it, so I took out the comment too.  I could
>     Zack> stick it above the block in toplev.c that now emits
>     Zack> .ident's.
> 
> In this case it was above a definition of IDENT_ASM_OP that you left
> in place; you just removed the comment.  I think it should go back.

In e.g. config/freebsd.h?  I don't see that the comment makes sense
attached to IDENT_ASM_OP.  It would have made sense attached to the
ASM_FILE_END definition in the same file, but defining IDENT_ASM_OP
really only says "the assembler for this target understands .ident"
and the trailing .ident "GCC: (GNU) ..." is not its only use.

I've now put the comment here (toplev.c):

#ifdef ASM_FILE_END
  ASM_FILE_END (asm_out_file);
#endif

  /* Attach a special .ident directive to the end of the file to identify
     the version of GCC which compiled this code.  The format of the .ident
     string is patterned after the ones produced by native SVR4 compilers.  */
#ifdef IDENT_ASM_OP
  if (!flag_no_ident)
    fprintf (asm_out_file, "%s\"GCC: (GNU) %s\"\n",
             IDENT_ASM_OP, version_string);
#endif

I feel that it doesn't need to be in tm.h too.

zw


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]