This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Testsuite regression in CVS for powerpc-linux-gnu
- To: Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>
- Subject: Re: Testsuite regression in CVS for powerpc-linux-gnu
- From: Jeffrey A Law <law at hurl dot cygnus dot com>
- Date: Wed, 03 Mar 1999 10:02:40 -0700
- cc: egcs-bugs at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
In message <4.2.0.25.19990303175434.03711c20@mail.lauterbach.com>you write:
> At 17:27 03.03.99 , Franz Sirl wrote:
> >At 19:32 26.02.99 , Franz Sirl wrote:
> >>Hi,
> >>
> >>since ~1-2 days I have a quite spectacular testsuite regression as can be
>
> >>seen at http://egcs.cygnus.com/ml/egcs-testresults/1999-02/msg00263.html
> >>
> >>This already happened before rth's flow rewrite went in.
> >
> >Jeff, it seems that your calls.c revamp is the culprit. If I revert
> >calls.c to revision 1.43, everything is fine again.
>
> To be more specific, this part causes the regressions in varargs related
> tests (eg. execute/980205.c):
>
> * calls.c (initialize_argument_information): New function extracted
> from expand_call.
> (expand_call): Use initialize_argument_information. Remove variabl
> es
> which are no longer used due to cleanups.
This doesn't actually help me since I don't know anything about the problem
you are running into. Debug the problem, or send a testcase I can run through
the compiler so that I can debug the compiler.
jeff