RFA: patch to solve PR37535

[David Edelsohn]

On Fri, Sep 26, 2008 at 11:30 PM, Vladimir Makarov <vmakarov@redhat.com> wrote:

> I tried to reproduce the bootstrap failure on today trunk (@140713) on
> powerpc64-gnu-linux and failed.  I tried bootstrap with all languages and it
> failed on stage3 because there is no enough space on the machine.  So it
> went through all stage2.  I also tried the bootstrap with only C and it was
> successful.
>
> David, could you tell more info about your bootstrap parameters?  I used the
> default values (only --enable-languages=c for the second run).

Janis's powerpc-linux testresults list her configuration parameters for GCC:

--prefix=/opt/gcc-nightly/trunk-20080926 --build=powerpc64-linux
--host=powerpc64-linux --target=powerpc64-linux --with-cpu=default32
--with-as=/opt/gcc-nightly/binutils-20080904/bin/as
--with-ld=/opt/gcc-nightly/binutils-20080904/bin/ld
--enable-threads=posix --enable-shared
--enable-__cxa_atexit --enable-languages=c,c++,fortran
--enable-checking --disable-alsa
--with-gmp=/opt/gcc-nightly/gmp-mpfr-static
--with-mpfr=/opt/gcc-nightly/gmp-mpfr-static
--enable-decimal-float



Janis (on Linux) and I (on AIX) see a failure configuring stage 2 libgcc: xgcc
crashes.  Andrew sees a bootstrap comparison failure on darwin
and powerpc-linux.

Maybe --with-cpu=default32 is needed.  Janis uses that option and GCC AIX
defaults to that behavior.

Thanks, David