On Fri, Sep 29, 2017 at 10:45 AM, Daniel P. Berrange <berra...@redhat.com> wrote:
> On Thu, Sep 28, 2017 at 04:41:37PM -0700, Paul Eggert wrote: > > That patch essentially negates the point of the test, which is that > getopt > > should be visible from unistd.h. I'd rather fix the problem than nuke the > > test. > > > > Could you explain what the Gnulib problem is here? I can't really see it > in > > your email. A self-contained example would help. > > > > For what it's worth, I could not reproduce the problem on Fedora 26 by > doing > > this in Gnulib (this tells 'configure' to use Gnulib-supplied getopt.h > and > > getopt.c): > > Fedora 26 only has glibc 2.25 - you need to have Fedora rawhide to get > the broken behaviour, as that has glibc 2.26.90 > > > ./gnulib-tool --create-testdir --dir foo getopt-posix > > cd foo > > ./configure gl_cv_func_getopt_posix=no ac_cv_header_getopt_h=no > > make > > make check > As Daniel said at least glibc 2.26 as in Fedora rawhide or Ubuntu Artful. But thanks for these commands that way I could reproduce without needing any of the libvirt build env. Here [1] a log of your commands on such a system showing the issue. I added a gdb to show the assert and an LD_DEBUG so you can see that getopt is not taken from gnulib as the test assumes when verifying behavior. [1]: http://paste.ubuntu.com/25638847/ > Regards, > Daniel > -- > |: https://berrange.com -o- https://www.flickr.com/photos/ > dberrange :| > |: https://libvirt.org -o- > https://fstop138.berrange.com :| > |: https://entangle-photo.org -o- https://www.instagram.com/ > dberrange :| > -- Christian Ehrhardt Software Engineer, Ubuntu Server Canonical Ltd
