This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix a passel of libintl/libiconv issues
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- To: zack at codesourcery dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Tue, 8 Jul 2003 12:07:51 -0400 (EDT)
- Subject: Re: Fix a passel of libintl/libiconv issues
- References: <87brw5tvfp.fsf@egil.codesourcery.com>
> This patch corrects a bunch of issues related to the new libintl. In
> particular, it should make it possible to bootstrap again on Cygwin
> hosts, where the old gettext probe macro and the new one disagree on
> whether or not there's an acceptable gettext provided with the system.
>
> Also addressed are the failure to probe xgettext/msgfmt properly; the
> syntax errors in gcc/Makefile.in (I have no idea how that bug got
> through a bootstrap on my system, but it did); paving the way for
> proper dependencies on ../iconv/libiconv.a when we have it; proper
> setting of several #defines in libintl, the victim of an over-eager
> move to config.h; and a bunch more cruft removal.
Thanks. I can confirm (on both irix6.5 and solaris2.7) that I'm able
to link cc1 without requiring --disable-nls. So the problem where we
were missing libintl.a when linking has been fixed.
> I have tested this on i686-linux and i686-netbsd1.6, and I have been
> told that it solves the bootstrap problems on cygwin. It does not
> address the issue with system-provided iconv not knowing how to
> convert UTF-8 to UCS-[42][LB]E; that will be the subject of a patch
> which I hope to write tomorrow afternoon, when I am awake enough to do
> bit fiddling. I do not have access to any such system, so I would
> appreciate help testing that patch once it is written.
Yeah, I still get this on both boxes.
I'll help test your patch, send me a note when it's ready.
--
Kaveh R. Ghazi ghazi@caip.rutgers.edu