Re: Bootstrap error with gcc-20020429 snapshot on mips-sgi-irix6.5

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

On Wednesday 01 May 2002 03:59, Alexandre Oliva wrote:
> On Apr 30, 2002, Franz Sirl <Franz.Sirl-kernel@lauterbach.com> wrote:
> > Probably related to my patch to merge the weak declarations better. Can
> > you run (if you have tcl/expect/dejagnu installed)
> >
> > 	make check-gcc RUNTESTFLAGS="dg.exp=weak-*.c"
> >
> > in your builddir? Or someone else on this platforms? Please post the
> > resulting gcc.log then.
>
> Here's the only failure I got from weak-*.c tests:
>
> output is:
> /home/aoliva/src/tool/egcs/gcc/testsuite/gcc.dg/weak-5.c:42: warning: weak
> declaration of `vfoo1c' after first use results in unspecified behavior
> /home/aoliva/src/tool/egcs/gcc/testsuite/gcc.dg/weak-5.c:114: warning: only
> weak aliases are supported in this configuration
>
> PASS: gcc.dg/weak-5.c weak declaration (test for warnings, line 42)
> FAIL: gcc.dg/weak-5.c (test for excess errors)
> Excess errors:
> /home/aoliva/src/tool/egcs/gcc/testsuite/gcc.dg/weak-5.c:114: warning: only
> weak aliases are supported in this configuration

Can you try if the attached patch fixes it for you?

Now that more stuff is put on the weak_decls list, the check for duplicates 
needs to be better.

Running a bootstrap now.

BTW, wondering about this assembler error, does this symbol end up as weak in 
misc-inst.o? On linux readelf tells me:

   809: 00000000     4 OBJECT  WEAK   DEFAULT 1410 
_ZNSt15basic_streambufIcSt11char_traitsIcEE13_S_pback_sizeE

Franz.

	* varasm.c (globalize_decl): Compare names to check for duplicates.

Attachment: gcc-weaksym-8.patch
Description: Text document