John Hunter wrote: > On Sat, Jan 16, 2010 at 7:57 PM, <[email protected]> wrote: > >> STATUS: Failure >> revision: 8091 ( >> http://matplotlib.svn.sourceforge.net/matplotlib/?view=rev&rev=8091 ) >> >> Files: >> lib/matplotlib/backends/qt4_editor/figureoptions.py >> lib/matplotlib/backends/qt4_editor/formlayout.py >> At: Sun 17 Jan 2010 01:45:00 >> Changed By: jdh2358 >> Comments: added Goekhan's apply patch to qt editor >> > > > This is bizarre -- now the "no latex" machine is building and testing > properly, but the Ubuntu 8.04, Python 2.5, amd64 machine failed a test > on the font cache at 1:55:49, though neither the failure of this > machine or the sudden passing of the "no latex" machine would have > anything to do with this qt build. Did you do anything to the "no > latex" build machine/environment? Yes, I wiped the svn checkout, which forced the buildbot to re-download. To me, the weird thing is that this seems to have solved the problem -- the script wipes the build and install directories as a matter of course during each test cycle. In fact, looking at the difference between the contents of the wiped (I didn't delete it, I just moved it aside) and new directories, I only see a bunch of test images. How could they have caused a segfault, I wonder.
The bug related to the font cache is known and is unrelated to the CXX-induced issues. It's a race condition on multi-process simultaneous MPL use. It wouldn't come up on the buildbot slaves except I'm running two test instances on the same virtual machine and they usually get triggered simultaneously. It may be good to fix, nonetheless, though. I imagine this could be triggered in real world scenarios such as web servers spawning multiple processes in response to incoming hits. The SF tracker issue is: http://sourceforge.net/tracker/?func=detail&aid=2856125&group_id=80706&atid=560720 > And do you know why the scons > machine doesn't appear to have built or tested anything in a while? > I don't have it hooked up to an automatic scheduler... > One of my goals for the weekend is to get the OSX sage machine back > up.... so I may need to ping you (Andrew) with some questions at some > point since my buildbot knowledge has gotten a little rusty. > No problem. ------------------------------------------------------------------------------ Throughout its 18-year history, RSA Conference consistently attracts the world's best and brightest in the field, creating opportunities for Conference attendees to learn about information security's most important issues through interactions with peers, luminaries and emerging and established companies. http://p.sf.net/sfu/rsaconf-dev2dev _______________________________________________ Matplotlib-buildbot mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/matplotlib-buildbot
