rhelmerchrisccoulson: btw, nigelb pointed me at https://lists.ubuntu.com/archives/ubuntu-devel/2011-August/033931.html (I work on crash-stats aka socorro)04:53
rhelmerI am interested in helping to package socorro for debian/ubuntu, we were intending on providing rpm/deb ourselves soon anyway04:54
=== asac_ is now known as asac
chrisccoulsonhi rhelmer, that's awesome10:21
chrisccoulson(sorry, i was asleep earlier)10:21
chrisccoulsonrhelmer, is socorro dependant on the client applications using Breakpad, or would it be possible to modify it to be compatible with something else? (eg, Apport, which is what we use in Ubuntu for most things)10:29
chrisccoulsonapport has some important advantages for us. the main one being that we don't need to embed it in to applications10:30
=== asac_ is now known as asac
=== m_conley` is now known as m_conley
=== m_conley is now known as m_conley_away
rhelmerchrisccoulson: socorro uses a component of breakpad called minidump_stackwalk to process incoming crashes, and the processors have access to the debug symbols used to generate the binaries that the crashes came from14:20
rhelmerchrisccoulson: I don't think there are any other explicit dependencies but there are probably some assumptions in the code, I guess things like signature generation would need to be modified a bit14:21
chrisccoulsonrhelmer, yeah, possibly14:21
chrisccoulsonit would be nice to have something like that :)14:21
chrisccoulsonwe definitely miss having a good crash DB14:21
chrisccoulsonrhelmer, bug 629306 is an example of what we currently have14:23
ubot2Launchpad bug 629306 in ubuntuone-client/stable-1-4 "gnome-settings-daemon crashed with SIGSEGV in g_return_if_fail_warning()" [High,Fix committed] https://launchpad.net/bugs/62930614:23
chrisccoulsonwhich is a problem for frequent crashes, as it generates a lot of bug mail ;)14:23
rhelmerchrisccoulson: socorro has a lot of dependencies and can be a bit tricky to get going, we've started a vagrant project which bootstraps an ubuntu VM (using puppet) https://github.com/rhelmer/socorro-vagrant14:27
rhelmerchrisccoulson: we're definitely interested in packaging it in any case, but I think we could investigate adding support for apport as well14:28
rhelmerchrisccoulson: so I accidentally woke up way too early this morning, I am going to go back to sleep, but nice to meet you :)14:29
chrisccoulsonrhelmer, excellent, thanks :)14:29
=== m_conley_away is now known as m_conley
bhearsumm_conley: nice work driving Thunderbird along!18:28
m_conleybhearsum: thanks! :)18:35
chrisccoulsonm_conley, are you coming to UDS in orlando btw?18:57
chrisccoulsonnice hotel and pool ;)18:57
m_conleychrisccoulson: Oh, good question - has registration opened?18:57
chrisccoulsonm_conley, good question :)18:57
chrisccoulsonjcastro might be able to answer that18:58
chrisccoulsonah https://launchpad.net/sprints/uds-p18:58
chrisccoulsonso the answer is probably "yes" :)18:58
m_conleyright, cool18:58
m_conleychrisccoulson: I hope I can come!  I really enjoyed the last UDS.18:59
chrisccoulsonsigh @ https://crash-stats.mozilla.com/report/index/1de1b1b4-3524-4628-8bfb-778722110810 :(19:54
chrisccoulsonall the reports from the same person19:54
chrisccoulsoni wish he would respond to my e-mail19:54
micahgm_conley: you're famous: http://feedproxy.google.com/~r/d0od/~3/OSIVlyO0s9s/20:37
m_conleymicahg: :D20:37
BUGabundoevening friends of the Dark20:45
=== m_conley is now known as m_conley_away
chrisccoulsonwoah, just had my first firefox crash in a month!22:14
chrisccoulsonnot bad for running nightly builds22:14
OmegaMine crashes about once a week22:16
chrisccoulsongah, mozilla bug 661297 totally breaks my extension :'(23:08
ubot2Mozilla bug 661297 in DOM: Events "Kill AddEventListenerByIID/RemoveEventListenerByIID" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=66129723:08
chrisccoulsonmaintaining a binary extension sucks23:09
rhelmerchrisccoulson: yeah don't do that :P23:23
chrisccoulsonit's not fun ;)23:24
rhelmerchrisccoulson: srsly though is jsctypes an option? what does it do?23:24
chrisccoulsonrhelmer, https://launchpad.net/globalmenu-extension23:24
chrisccoulsonctypes would be difficult atm, as i need access to some interfaces which aren't scriptable23:25
chrisccoulsonand there are problems with managing memory with ctypes too23:25
rhelmerchrisccoulson: hm yeah that's a tricky one23:27
rhelmerchrisccoulson: has there been any discussion of getting this upstreamed to mozilla, so our test infra etc. would be able to help?23:28
rhelmerguess since it's ubuntu-only might be a little odd23:28
chrisccoulsonrhelmer, there has been some brief talk, but nothing has materialised so far23:28
chrisccoulsoni'd certainly like that to happen23:28
rhelmerit'd be nice if we could help w/ test infra at least, I wonder if that's something we could do sooner23:30
rhelmerthere's been a lot of discussion about binary extensions in the newsgroups since we went with the rapid release cycles, that's one thing that has been discussed to help out enterprise users iirc23:30
chrisccoulsonrhelmer, i probably have more of an advantage than some other extension maintainers, as we bundle this extension with firefox23:40
chrisccoulsonwhich means it also gets built nightly23:40
chrisccoulsonso i spot breakage fairly quickly ;)23:41
rhelmeryeah that's good23:41
rhelmerchrisccoulson: do you do any kind of automated tests? we have a whole bunch of them that you could run, they have crash detection and such (I assume you mean it gets bundled and built on the ubuntu side right?)23:42
chrisccoulsonrhelmer, yeah, that's right23:43
chrisccoulsonthere aren't any tests for it atm, as it depends on other things in the environment23:43
chrisccoulson(ie, it needs to talk to the unity panel)23:43
rhelmeryeah makes sense23:43
chrisccoulsoni guess i could emulate that somehow though23:43

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!