On Thu, Oct 22, 2015 at 05:09:14PM +0100, Peter Maydell wrote: > On 22 October 2015 at 16:43, Daniel P. Berrange <berra...@redhat.com> wrote: > > Yes, if configure finds gnutls, it tries to figure out if gnutls > > links to nettle or gcrypt, and then checks for the corresponding > > one. It fails if gnutls is found, but the corresponding nettle/gcrypt > > is not found, on the basis that that should not actually happen, as > > installing gnutls-dev should pull in nettle-dev/gcrypt-dev packages > > as appropriate. > > This doesn't appear to be true in practice: > > [pm215@gcc1-power7 all]$ yum list gnutls-devel > Installed Packages > gnutls-devel.ppc64 3.1.26-2.fc20 > @updates > Available Packages > gnutls-devel.ppc 3.1.26-2.fc20 > updates > [pm215@gcc1-power7 all]$ yum list libgcrypt-devel > Available Packages > libgcrypt-devel.ppc 1.5.3-2.fc20 > fedora > libgcrypt-devel.ppc64 1.5.3-2.fc20 > fedora > [pm215@gcc1-power7 all]$ yum list nettle-devel > Available Packages > nettle-devel.ppc 2.7.1-3.fc20 > updates > nettle-devel.ppc64 2.7.1-3.fc20 > updates > > (This is gcc110 in the FSF compile farm, if you happen to have > a compile farm account.) > > > Did you only see this with todays master ? This particular hard failure > > logic was present even before commit > > 4e2abbeacce6e12e62a0183c67936c807b19c3b9 > > so I would expect you to see it all the way back to > > ed754746fea55df726f4de3dadb5bea0b6aa7409 > > The machine in question (a ppc64be system running Fedora 20) has > had a fresh OS install and a bunch of dev packages installed > today, which is why I noticed it. I suspect it previously did > not have any of the gnutls dev packages.
Ok, I will fix this, as part of the pull request I already have pending and send a new pull request. Regards, Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
