On Sat, Apr 26, 2008 at 10:27:40PM +0200, Mike Hommey wrote: > On Sat, Apr 26, 2008 at 09:00:03PM +0200, Frank Lichtenheld wrote: > > On Sat, Apr 26, 2008 at 05:51:40PM +0200, Mike Hommey wrote: > > > On Sat, Apr 26, 2008 at 05:38:31PM +0200, Frank Lichtenheld wrote: > > > > On Sat, Apr 26, 2008 at 04:57:36PM +0200, Mike Hommey wrote: > > > > > On Sat, Apr 26, 2008 at 03:59:54PM +0200, Lichtenheld wrote: > > > > > > Full build log(s): > > > > > > http://experimental.ftbfs.de/build.php?&ver=2.22.1.1-2&pkg=epiphany-browser&arch=sparc > > > > > > > > > > Would you be able to compile the failing program, and strace its > > > > > execution? This is likely to be a bug in the xpcom glue... > > > > > > > > $ strace -f ./conftest > > > > > > Mmmm not very interesting output :-/ > > > > > > Could you copy this conftest file in some readable place on sperger ? > > > > sperger:~djpig/gecko/ > > I'm puzzled... > Since the binary doesn't contain debugging symbols for the xpcomglue, I > built a new glue on sperger and couldn't get the same behaviour with > conftest.cpp built against this newly built glue... > > Can you test sperger:~glandium/conftest and > sperger:~glandium/conftest.debug when you have xulrunner-dev installed? > (I could only validate it seems to do the right thing) > > Then I'll have to figure out why the glue built on sperger works while > the one built on the buildd doesn't...
Good news, I could build a broken glue on sperger, too... seems there are differences between a glue built with in-tree nspr and a glue built with our system nspr. Looking into this. Mike -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
