This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: * configure.in: Test for libintl.h.
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- To: phil at jaj dot com, libstdc++ at gcc dot gnu dot org
- Date: Thu, 1 May 2003 17:05:49 -0500 (CDT)
- Subject: Re: * configure.in: Test for libintl.h.
- References: <Pine.GSO.4.44.0305011657210.5403-100000@igloo.df.lth.se>
- Reply-to: rittle at labs dot mot dot com
>> 2003-04-30 Phil Edwards <pme@gcc.gnu.org>
>>
>> * acinclude.m4: Add bit missing from previous patch.
>> * aclocal.m4, configure: Regenerated.
> This is (still) not enough. :-) (at least not for NetBSD.)
> There are two problems:
> 1. This test is only done for USE_NLS = yes, but the gettext function is
> used in libstdc++-v3/src/functexcept.cc if _GLIBCPP_HAVE_LIBINTL_H is
> defined, which it may be regardless of the status of USE_NLS.
> 2. USE_NLS is only set to 'yes' if enable_clocale_flag = gnu.
> configure does however use enable_clocale_flag = generic for NetBSD...
Yes, Phil, I was too quick to thank you as well. ;-) The recent patch
doesn't actually fix the problem on FreeBSD either. I still
see it fail while trying to build abi_check:
[...]
/usr/users/rittle/tmp/gcc-build-latour-mainline-0425/gcc/g++ -shared-libgcc -B/usr/users/rittle/tmp/gcc-build-latour-mainline-0425/gcc/ -nostdinc++ -B/usr/local/beta-gcc/i686-unknown-freebsd4.8/bin/ -B/usr/local/beta-gcc/i686-unknown-freebsd4.8/lib/ -isystem /usr/local/beta-gcc/i686-unknown-freebsd4.8/include -isystem /usr/local/beta-gcc/i686-unknown-freebsd4.8/sys-include -g -O2 -o abi_check abi_check.o -L/usr/users/rittle/tmp/gcc-build-latour-mainline-0425/i686-unknown-freebsd4.8/libstdc++-v3/src -L/usr/users/rittle/tmp/gcc-build-latour-mainline-0425/i686-unknown-freebsd4.8/libstdc++-v3/src/.libs -lm
/usr/users/rittle/tmp/gcc-build-latour-mainline-0425/i686-unknown-freebsd4.8/libstdc++-v3/src/.libs/libstdc++.so: undefined reference to `libintl_gettext'
I would prefer if you could continue the battle to fix it as I'm
attempting to fix the i386/atomicity.h thing on 3.3 branch before
first release. We seem to have a different failure mode than NetBSD
(although Gerald and I would probably see FreeBSD fail as in case #1
above if we didn't have GNU libintl installed). - Loren