Le 2017-03-02 18:18, Leo Famulari a écrit : > On Thu, Mar 02, 2017 at 05:55:30PM +0100, Marius Bakke wrote: >> julien lepiller writes: >> >> > Le 2017-03-01 16:34, julien lepiller a écrit : >> >> I've "fixed" the glib failure by disabling a test. Attached are >> >> patches I would like to see in core-updates. If it's too late to add >> >> the patch for gd, then I would like to keep gd-for-php. This time I've >> >> tested them and php works. >> > >> > A better version of those patches. I still don't know why glib tests >> > hang. I had to add gnutls as an input because otherwise the build >> > doesn't find "-lgnutls". There's no problem with the same php version on >> > master, so I guess a dependency doesn't propagate gnutls anymore. >> >> Ok, thanks for the update! The latter two patches LGTM for >> 'core-updates'. I would hold the glib patch until we understand what's >> going on, I am not able to reproduce it. >> >> Please also update the commit message of the PHP patch to mention the >> new inputs and arguments change, as well as the removal of the >> "gd-for-php" variable. > > +1 > > Please add the link to the upstream commit in the XBM patch, as > requested in my previous message. I think they are good now. How could I investigate glib's failure?