This is the mail archive of the
gcc-help@gcc.gnu.org
mailing list for the GCC project.
Re: Cross-build for x86_64-pc-mingw32
- From: Brian Dessent <brian at dessent dot net>
- To: FX Coudert <fxcoudert at gmail dot com>
- Cc: mingw-w64-public at lists dot sourceforge dot net, gcc-help at gcc dot gnu dot org
- Date: Sat, 01 Dec 2007 09:35:54 -0800
- Subject: Re: Cross-build for x86_64-pc-mingw32
- References: <19c433eb0711300950r43f1ccffi28ef6b7a2e51548d@mail.gmail.com> <b609cb3b0711301053g178ed7c7pb7f225f3579a4591@mail.gmail.com> <15BB3B75-14C5-4E4E-8C44-EBD567E70A79@gmail.com>
- Reply-to: gcc-help at gcc dot gnu dot org
FX Coudert wrote:
> > checking dependency style of /home/FX/ibin64/./gcc/xgcc -B/home/FX/
> > ibin64/./gcc/ -L/home/FX/ibin64/x86_64-pc-mingw32/winsup/mingw -L/
> > home/FX/ibin64/x86_64-pc-mingw32/winsup/w32api/lib -isystem /home/
> > FX/trunk/winsup/mingw/include -isystem /home/FX/trunk/winsup/w32api/
> > include -B/home/FX/win64/x86_64-pc-mingw32/bin/ -B/home/FX/win64/
> > x86_64-pc-mingw32/lib/ -isystem /home/FX/win64/x86_64-pc-mingw32/
> > include -isystem /home/FX/win64/x86_64-pc-mingw32/sys-include... gcc3
> > checking whether symbol versioning is supported... configure:
> > error: Link tests are not allowed after GCC_NO_EXECUTABLES.
> > make[1]: *** [configure-target-libgfortran] Error 1
> > make[1]: Leaving directory `/home/FX/ibin64'
>
> I know I had this working at some earlier point, so it's probably
> that configure line is not the one I used last time. The most likely
> culprit is, I think, --with-sysroot, so I'm running another build
> without it, but in the meantime I'd appreciate any hint: is it the
> way it should be done? (and thus, an issue with the GCC build system)
That error means that something went wrong earlier. Look upwards at
config.log and you will see somewhere that it tried to do a link test
that fails, typically due to missing or improper crt*.o, and thus it set
the NO_EXECUTABLES flag. From that point on, the next test that tries
to link something will give that error, but the problem was whatever
test failed originally. The details should all be in config.log.
Brian