This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: bootstrap broken on mingw
On Thu, Feb 21, 2008 at 8:40 AM, Ralf Wildenhues <Ralf.Wildenhues@gmx.de> wrote:
> > * Jakub Jelinek wrote on Wed, Feb 20, 2008 at 02:31:16PM CET:
> > > On Wed, Feb 20, 2008 at 10:02:09PM +1300, Danny Smith wrote:
> > > >
> > > > The ' CMD //c' (with 2 slashes preceding the switch rather than the
> > > > one documented for cmd by Micrsosoft) is a MSYS-ism that causes
> > > > problems with non-MSYS shells (including the one provided by MS in
> > > > its SFU package)
> >
> > > What exactly will do CMD //c echo do with non-MSYS shells?
>
> With the Cygwin shell, it outputs nothing and exits successfully.
> Thus, my first idea does not work, but the second seems to. Danny,
> could you please test this patch for me[1]? FWIW, a quick negative
> test is
> cd gcc && rm gcc-vers.texi && make info pdf
>
> If successful, OK to apply to trunk, 4.3, and 4.2 branch (in that
> order)?
>
> Thanks,
> Ralf
>
> [1] BTW, is there a w32 system in some build farm suitable for GCC
> testing?
>
> gcc/ChangeLog:
> 2008-02-20 Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
>
> PR bootstrap/35218
> * config.build (build_file_translate): Set to `CMD //c' only if
> it works.
>
This restores build with cygwin version of bash and is OK.
Perhaps the comment in Makefile.in needs to be expanded, eg,
# If necessary (eg, when using the MSYS shell on Microsoft Windows)
# translate the shell's notion of absolute pathnames to the native
# spelling.
build_file_translate = @build_file_translate@
Thanks for fixing this so promptly.
Danny