This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Bug in genextract.c
- To: richard dot earnshaw at arm dot com
- Subject: Re: Bug in genextract.c
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Fri, 22 May 1998 01:31:33 -0600
- cc: egcs at cygnus dot com, wilson at cygnus dot com
- Reply-To: law at cygnus dot com
In message <199805191134.MAA13450@sun52.NIS.cambridge>you write:
> This is a multipart MIME message.
>
> --==_Exmh_-16010448380
> Content-Type: text/plain; charset=us-ascii
>
>
> Running purify on the compiler yesterday, I noticed that insn_extract was
> occasionally putting the address of a local variable into
> recog_operand_loc. This happens, for example, when the whole pattern of
> the insn is matched with a match_parallel (as in the *ldmsi pattern on the
> ARM). A solution is to make print_path handle this as a special case and
> reference back to the pattern of the original insn.
>
> <date> Richard Earnshaw (rearnsha@arm.com)
>
> * genextract.c (print_path): Handle zero-length path as a special
> case.
It looks correct to me. I'm a little amazed we haven't tripped over
this before....
Please install this patch.
Thanks!
jeff