This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [v3] Fix / clean-up config vs crosses (5/n): avoid a TRY_RUN
On Mon, 25 Aug 2008, Mark Mitchell wrote:
> Paolo Carlini wrote:
> > HI Mark,
> >> checking a property of the build machine, unrelated to the target
> >> machine?
> >>
> > Right. That kind of consideration led me to **not** move it outside the
> > IS_NATIVE block, where I introduced it originally. Can be considered in
> > that case approximately correct? Or we need something entirely different?
>
> We need something different. The current mechanism (test -r) means that
> a cross to (say) x86 GNU/Linux may yield different results than a native
> build, since in the cross case the answer (whatever it is) will be
> determined in some different way. That's exactly the situation we need
> to avoid. To a first approximation, as long as there are IS_NATIVE
> checks, we've not yet solved the cross-compilation problem.
>
> I would suggest either:
>
> A1. Hard-code the answer based on the target OS (e.g., GNU/Linux has
> /dev/urandom).
The cross case already hardcodes the result of this particular configure
test for GNU/Linux targets, since:
2008-06-06 Joseph Myers <joseph@codesourcery.com>
* configure.ac: Do not check for gconv.h.
* crossconfig.m4 (GLIBCXX_CROSSCONFIG): Do not test for gconv.h or
gconf.h. For glibc and uClibc systems, define
_GLIBCXX_USE_RANDOM_TR1 and HAVE_MMAP and use AC_LC_MESSAGES and
AM_ICONV.
* configure, config.h.in: Regenerate.
--
Joseph S. Myers
joseph@codesourcery.com