This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: 31 GCC regressions, 4 new, with your patch on 2002-05-22T05:55:52Z.
- From: Geoff Keating <geoffk at geoffk dot org>
- To: neil at daikokuya dot demon dot co dot uk
- Cc: gcc-regression at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org, shebs at apple dot com
- Date: Wed, 22 May 2002 17:48:01 -0700
- Subject: Re: 31 GCC regressions, 4 new, with your patch on 2002-05-22T05:55:52Z.
- References: <200205220941.g4M9fkv32691@maat.sfbay.redhat.com> <20020522225507.GA11853@daikokuya.demon.co.uk>
- Reply-to: Geoff Keating <geoffk at redhat dot com>
> Date: Wed, 22 May 2002 23:55:07 +0100
> From: Neil Booth <neil@daikokuya.demon.co.uk>
> GCC regression checker wrote:-
>
> > With your recent patch, GCC has some regression test failures, which
> > used to pass. There are 4 new failures, and 27
> > failures that existed before and after that patch; 0 failures
> > have been fixed.
> >
> > The new failures are:
> > powerpc-eabisim objc.sum objc.dg/const-str-2.m
> > mips-elf objc.sum objc.dg/const-str-2.m
> > native gdb.sum gdb.threads/linux-dp.exp:
> > native objc.sum objc.dg/const-str-2.m
>
> This fixes the objc failures (I have no idea at all what the
> debug failure is. If it's my patch I guess it's the same kind
> of thing). My patch causes the compiler to stop compiling if
> there are any errors by the time the command line options have
> been parsed. This test seems to test for a command-line error,
> but to have needed to catch some other random issue that used
> to happen too.
>
> OK Stan?
>
> Neil.
I bet the original problem is that the objc frontend was using the
parser to parse the argument, and then got confused when there was no
argument. But that's only a guess...
> testsuite:
> * objc.dg/const-str-2.m: Update now that we stop after
> a command line error.
>
> Index: objc.dg/const-str-2.m
> ===================================================================
> RCS file: /cvs/gcc/gcc/gcc/testsuite/objc.dg/const-str-2.m,v
> retrieving revision 1.2
> diff -u -p -r1.2 const-str-2.m
> --- objc.dg/const-str-2.m 7 Feb 2002 09:08:25 -0000 1.2
> +++ objc.dg/const-str-2.m 22 May 2002 22:47:08 -0000
> @@ -3,8 +3,3 @@
> /* { dg-options "-fconstant-string-class" } */
>
> { dg-error "no class name specified as argument to -fconstant-string-class" "" { target *-*-* } 0 }
> -
> -void foo () {}
> -
> -/* Seem bogus, should investigate some day. */
> -/* { dg-error "parse error" "" { target *-*-* } 5 } */
>
Um, did you really want to delete the 'void foo ()' line? It should
make no difference...
--
- Geoffrey Keating <geoffk@geoffk.org> <geoffk@redhat.com>