Re: Bug#584610: [mips] gcc-4.4 build failure after upgrade to eGLIBC-2.11

[Matthias Klose <doko@debian.org>]

On 11.06.2010 15:22, Aurelien Jarno wrote:
> reassign 584610 gcc-4.4
> tag 584610 + pending
> thanks
>
> On Mon, Jun 07, 2010 at 02:08:20PM +0200, Aurelien Jarno wrote:
> > Matthias Klose a écrit :
> > > On 06.06.2010 00:51, Aurelien Jarno wrote:
> > > > On Sat, Jun 05, 2010 at 03:50:51AM +0200, Matthias Klose wrote:
> > > > > Package: eGLIBC
> > > > > Version: 2.11.1-2
> > > > > Severity: serious
> > > > >
> > > > > gcc-4.4 and gcc-4.5 fail to build after the upgrade to eGLIBC-2.11:
> > > > >
> > > > > https://buildd.debian.org/fetch.cgi?pkg=gcc-4.4;ver=4.4.4-4;arch=mips;stamp=1275677666
> > > > This FTBFS is caused by the following change:
> > > >
> > > > |2009-11-20  Jakub Jelinek<jakub@redhat.com>
> > > > |
> > > > |        PR libc/10103
> > > > |
> > > > |        * math/math.h: Provide *l long double prototypes redirecting
> > > > |        to double functions even when __NO_LONG_DOUBLE_MATH and not
> > > > |        __LDBL_COMPAT.
> > > > |        * math/complex.h: Likewise.
> > > >
> > > > These functions were present before in the library, but not exported
> > > > in the headers. This has been changed as it is required by ISO C99.
> > > >
> > > > GCC tries to find these functions in the GLIBC by compiling a program,
> > > > so it was failing before, and is successful now. When they are already
> > > > present in the GLIBC it does not re-export them.
> > > >
> > > > Strangely this should also affect ARM, but it seems to build correctly.
> > > > I haven't investigated why.
> > > >
> > > > While these functions are strictly not needed in libstdc++6 anymore, we
> > > > have two options:
> > > > - revert the GLIBC change, which means we break the C99 compatibility
> > > >     (as before)
> > > > - patch GCC to export these functions anyway.
> > > >
> > > > What's your opinion?
> > >
> > > For ARM I did choose the second option, but didn't get any feedback about it.
> > > So maybe it's time to ask the mips and arm porters?
> > >
> > > The patch applied for armel is:
> > > http://svn.debian.org/viewsvn/gcccvs/branches/sid/gcc-4.4/debian/patches/libstdc%2B%2B-arm-ldbl-compat.diff?view=log
> >
> > I think we should go for the same patch on mips, it's probably better to
> > be ISO C99 compliant on the glibc side.
>
> As we discussed, this bug has to be fixed on the gcc side. I have just
> checked-in a patch in the SVN to do that on both gcc-4.4 and gcc-4.5.
>
> I am therefore reassigning this bug on gcc-4.4 and tagging it pending. I
> don't think it is worth cloning this bug to gcc-4.5 as the bug is fixed
> in the SVN and the package is in experimental.

proposed instead
http://gcc.gnu.org/ml/gcc-patches/2010-06/msg01290.html

Joseph Meyers suggested that the symbols should only be exported for the old 
mips O32 ABI, not for the mips64 and n32 ABIs.