On Fri, Sep 16, 2011 at 07:13:04AM -0600, Rob Savoye wrote: > On 09/16/11 03:16, Sandro Santilli wrote: > > > It turns out I was wrong. The "happy run" of buildbot was with > > a patch prepared by Gabriele truly reverting your ffmpeg commits. > > I reverted my patch fully, I saw the segfaults with the existing > patch, not the one I applied.
Where did you see it ? I see a normal build with 5c76984eea636cdf301b70012a77d8d8e7f6e611, see http://www.gnashdev.org:8010/builders/squeeze-linux-x86/builds/664 I see segfaults with f5170a8879abdeb7f99176d8444bcdd6785d120a, see http://www.gnashdev.org:8010/builders/squeeze-linux-x86/builds/665 I see segfaults with 9d4e24b875ebeba72852ede9aa90082142ffa4e3, see http://www.gnashdev.org:8010/builders/squeeze-linux-x86/builds/666 No segfaults with a patch applied over 9d4e24b, see http://www.gnashdev.org:8010/builders/squeeze-linux-x86/builds/667 I see segfaults with aaa3f2cabc126384802337c1c8f86a4c8c895b38, see http://www.gnashdev.org:8010/builders/squeeze-linux-x86/builds/668 So, first segfault ever spotted is with: commit f5170a8879abdeb7f99176d8444bcdd6785d120a Merge: b8f483b 99cedd1 Author: Rob Savoye <r...@welcomehome.org> Date: Mon Sep 12 19:53:52 2011 -0600 Merge branch 'master' of git.sv.gnu.org:/srv/git/gnash Previous successful build was with this: commit 5c76984eea636cdf301b70012a77d8d8e7f6e611 Author: Gabriele Giacone <1o5g4...@gmail.com> Date: Sun Sep 11 10:48:42 2011 +0200 Fix build on sid with libav 0.7.1. Patch from http://savannah.gnu.org/bugs/?33696 adapted to support older versions. Thanks chithead! There are no builds for in-between commits, but all in-between commits are yours (rob) and related to ffmpeg. I'd either bring the repository to 5c76984eea636cdf301b70012a77d8d8e7f6e611 (first known success) and ask you to file bugs for any issue that commit gave to you or apply the workaround patch applied with build 667 above (which likely has the same effect). In any case, Gabriele and you are the only one who pushed anything to the repository so one you (whoever broke it) of you should take care of fixing the segfaults. All clues in my hand point to you, but I might have missed something. > > What do we do now ? > > About what, a feature freeze ? About fixing the segfaults... --strk; () Free GIS & Flash consultant/developer /\ http://strk.keybit.net/services.html _______________________________________________ Gnash-dev mailing list Gnash-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnash-dev
