localename: improvements for MacOS X and Cygwin

2009-12-21 Thread Bruno Haible
This patch changes localename.[hc]. It - Introduces a function gl_locale_name_environ used by related code in libintl. - Tweaks the behaviour on MacOS X. - On Cygwin, respects the locale set by the user in the Windows "Control Panel". In parallel to this, I'm adding to libintl for MacOS X

Re: stdlib tweak

2009-12-21 Thread Eric Blake
Eric Blake byu.net> writes: > I got this failure on cygwin, when building m4 with -DGNULIB_POSIXCHECK (m4 > does not use the getsubopt module): > > gcc -std=gnu99 -I. -I../../lib -I/cygdrive/c/cygwin/usr/local/include - > DGNULIB_POSIXCHECK=1 -MT gl_avltree_oset.o -MD -MP - > MF .deps/gl_av

.gitignore and generated files (was: Re: _GL_ARG_NONNULL)

2009-12-21 Thread Bruno Haible
Eric Blake wrote: > One other thing - the generated .gitignore files are failing to pick up on > the > fact that link-warning.h and arg-nonnull.h (and soon-to-be warn-on-use.h) are > generated files: > > # Untracked files: > # (use "git add ..." to include in what will be committed) > # > #

Re: stdlib tweak

2009-12-21 Thread Eric Blake
Eric Blake byu.net> writes: > On cygwin, compiling with -DGNULIB_POSIXCHECK failed because > the getsubopt link warning definition interfered with the > inclusion of the system header. The fix, as always in these > types of problems, is to ensure that system headers are > completely included bef

stdlib tweak

2009-12-21 Thread Eric Blake
I got this failure on cygwin, when building m4 with -DGNULIB_POSIXCHECK (m4 does not use the getsubopt module): gcc -std=gnu99 -I. -I../../lib -I/cygdrive/c/cygwin/usr/local/include - DGNULIB_POSIXCHECK=1 -MT gl_avltree_oset.o -MD -MP - MF .deps/gl_avltree_oset.Tpo -c -o gl_avltree_oset.o ../

Re: _GL_ARG_NONNULL

2009-12-21 Thread Eric Blake
Bruno Haible clisp.org> writes: > Indeed, I reproduce. The reason is that the module 'arg-nonnull' ends up > among the main modules, not the tests modules. But the tests modules need > it as well. We have fixed this for 'link-warning' three months ago, but > forgot about the issue for 'arg-nonnul

Re: coreutils-8.2 misc/ls-time test failure

2009-12-21 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Eric Blake on 12/19/2009 5:34 PM: > Here's what I'm finally pushing to gnulib, and a followup to coreutils. And a followup; now that the test requires inequality in ctime, the nap() function needs to sleep longer when run on ext3 or any o