[09:06] Morning all! [09:45] Happy Monday! === mandel is now known as mandel_lunch [13:54] whee === davmor2_ is now known as davmor2 === mandel_lunch is now known as makdel [15:06] dobey, hello! [15:06] hi broken keyboard mandel [15:06] or you just can't spell your own name now? :) [15:06] dobey, it seems that we are not going to be many today, karni is ill and is a holiday in ar [15:06] dobey, he, I'm stupid :P === francisco is now known as Guest73379 === makdel is now known as mandel [15:07] dobey, much better :) [15:08] hehe [15:10] hi folks [15:10] hi mmcc [15:11] mmcc, o/ [15:37] dobey, mmcc, what time do we have the standup? [15:37] if we have one, 'cause is just the 3 of us + brian (or so I think) [15:37] mandel: I'm checking on bug 1061901 with current trunk, I think brian fixed it with https://code.launchpad.net/~brian.curtin/ubuntu-sso-client/nam-setproxy-windows-workaround/+merge/130626 [15:38] Launchpad bug 1061901 in Ubuntu Single Sign On Client "webclient tests broken on windows: setProxy fails" [Undecided,New] https://launchpad.net/bugs/1061901 [15:38] mandel: in 20 minutes, according to the calendar [15:38] mandel: in ~22 minutes, but if it's just us few, i'd say skip it [15:38] dobey, I think the same [15:38] mmcc, hm... I wonder what was going on, did you see it happening on mac or linux? [15:38] * dobey needs to file his expenses today though [15:39] mandel: mvo is apparently around somewhere today too no? :) [15:40] dobey, uh, I forgot about him, sorry [15:40] dobey: hello [15:40] mandel: IIRC there was a function that only existed in some new qt version or something - I don't remember seeing it on mac, but I might not have checked [15:40] I'm here, just finished (again!) my unity sc-integration branch, lets hope its really finished this time around :) [15:40] but yeah, half the team is gone [15:40] mvo, hehe I have the problems landing things in unity [15:41] mandel: :) [15:41] * mvo crosses fingers [15:42] mvo: is software-center the only project we (client engineering) ship for the consumer apps side of things? [15:44] * mvo scratches head [15:44] dobey: I think so [15:47] mandel: confirmed that this branch https://code.launchpad.net/~brian.curtin/ubuntu-sso-client/nam-setproxy-windows-workaround/+merge/130626 fixes that bug, so I set it to fix committed. [15:48] hrmm [15:48] * mandel looks [15:50] mmcc, bummer, we have started working around bugs in diff lib versions.. I wonder if it is better to ship windows and mac with the same qt version as on linux [15:50] mmcc, will 'in theory' easy the pain with these type of things [15:51] mandel: do all the supported versions of linux have the same version of qt? [15:51] mandel: we have to support multiple qt versions on linux [15:51] mmcc, no ^ [15:51] mandel: we are going to have to do these sorts of things regardless of what versions we ship on mac/win [15:51] dobey, mmcc, which means that that should be failing on previews versions, right? [15:53] mandel: no. but that check is wrong; we should check the qt version, not the platform, to determine when to make that call [15:53] or pyqt version rather [15:53] mandel: maybe - see brian's comments on the merge, he suspected a bug in pyqt, maybe that doesn't happen on linux? [15:53] is it broken on all platforms in this version? [15:54] oh dobey's right - it's the pyqt version, not the qt version [15:54] we should double check.. I would not be surprised if it is a platform specific bug... [15:54] i suspect it's not a platform-specific bug [15:54] but a pyqt-version-specific change [15:57] dobey, we don't have tarmac running on old versions of ubuntu, do we? [15:58] would be something nice to have... [15:58] mandel: yes, but we don't land trunk on them; we do have nightlies building on them though, and the nightlies run the tests [15:58] mandel: but if the test isn't correct, it doesn't necessarily mean anything if it passes when it should fail :) [15:59] dobey, yes, that is one of the flaws of tests, you can write them to pass or they can have bugs :) [15:59] dobey, lets do tdd on our tests [15:59] hehe [16:00] mvo, dobey, mmcc, you decide if we do the standup, if we do it has to be in #u1-client due to my work :( [16:00] for the systems that run the nightlies - how do (did) they get their pyqt versions? getting old pyqt versions is a problem [16:00] mandel: I would say lets do a short one [16:01] mmcc: they use what's in the ubuntu archive [16:01] mandel: i say skip it :) [16:02] mmcc, is you call [16:02] where's brian? [16:03] I'd vote for a quick one. phone it in if you like [16:03] mmcc, well, lets do a quick one over #u1-client internal channel, is a matter of writing the notes.. [16:20] mandel, this could use a review: https://code.launchpad.net/~mikemc/ubuntuone-control-panel/launch-separate-darwin-menu/+merge/135268 [16:20] also rockstar is on deck for reviews today. do we have rockstar here? [16:20] mmcc, on it [16:21] we need a bot to remember who's on the team. [16:22] brb [16:22] mmcc: https://launchpad.net/~online-services-clienteng [16:22] mmcc: rockstar is on swap day today [16:23] dobey: ah, ok thanks [16:27] mmcc, I like the fact that we use subprocess, but is there a reason why we don't use the runner we already have under sso utils (I hate that module, but since it is there...) [16:28] mmcc, I'm talking about https://code.launchpad.net/~mikemc/ubuntuone-control-panel/launch-separate-darwin-menu/+merge/135268 [16:32] mmcc: won't that also potentially launch multiple copies of the menu? [16:32] or does the menu do ipc and keep itself unique? [16:32] good point, we might need to use the trick we have atm with control panel [16:33] and sd, right? [16:34] i'm not sure if the other runner module code helps with that case or not, either [16:34] dobey, it does not AFAIK and i'd love to kill it, is buggy [16:39] sorry, was afk there - the menu process doesn't try to keep itself unique, but the branch instantiates the menubariconlauncher after control panel has done its own unique app checking -- so if we launch CP multiple times, we don't launch the menu multiple times. [16:39] mmcc: but if you close the control panel, and then run it again, does it not open a new menu process? [16:40] mmcc: it's understandable if cp is already running, but what if it's not? [16:40] It hasn't in my testing, no. now I need to double check to see if I can explain why :) [16:45] so running the menu process twice (via popen or the shell) doesn't start two menu processes. the second one just silently does nothing. this is the expected behavior for mac apps, and it's handled by default by the cocoa framework. (which is why our qt apps don't also do it) [16:46] mmcc, go for cocoa! I wonder, is it possible to add a test for this in case this ever changes? [16:48] mmcc: ah ok. so it's just impossible to run the same cocoa process twice? [16:49] mandel: not totally impossible. you can use 'open -n' on the shell to force it to start a new process, but I'm not sure what mechanism that uses [16:49] ah ok [16:50] mandel, not sure about adding a test, what exactly do you want to test? I wouldn't expect the basic process running stuff in cocoa to ever change on us… [16:51] mmcc, that there is only one process that is ran.. maybe starting one and asserting that the second one is not.. but I don't know how [16:54] assert sys.platform == 'darwin' [17:01] alright, need to get lunch; bbiab [17:19] EOD here, mmcc I'll finish the review tom early in my morning [17:19] see you all tom! [17:20] ok, bye mandel! [18:14] who to beg for reviews now [18:15] oh brian is on holiday today [18:16] mmcc: so i guess it's just us left for today [18:16] yeah, if you have a review for me, shoot. [18:17] btw, how did you see that brian is on holiday? is there a listing somewhere I don't know about? [18:22] there's a 'holiday/leave calendar' on canonical admin [18:24] oh, yeah I guess I remember seeing that. thanks [18:31] mmcc: i have 3 branches that i proposed during the sprint… [18:31] https://code.launchpad.net/~dobey/ubuntuone-client-gnome/kill-share-contacts/+merge/134296 [18:31] https://code.launchpad.net/~dobey/ubuntuone-control-panel/no-msg-menu/+merge/134467 [18:32] https://code.launchpad.net/~dobey/ubuntuone-client/no-msg-menu/+merge/134465 [18:33] ok, I'll take a look [19:50] mmcc: also, do you know the correct process for https://one.ubuntu.com/help/faq/what-does-the-root_mismatch-error-mean/ on Mac? [19:50] hrmm, the Ubuntu version of that needs to be updated as well [19:51] dobey: yes, I recently sent chaselivingston some details about that, I'm not sure if he was planning to update that page or not. [19:51] mmcc: …oops [19:51] ah [19:51] mmcc: let me see if i can find that [19:51] hrm, actually it wasn't that specifically. I'll send an email with the full details [19:51] great [19:52] mmcc: ah ok, whew. thought i had lost it [19:52] chaselivingston: no, in retrospect I think it was just on IRC or something :) [19:53] mmcc: ah ok [20:03] ok chaselivingston , I sent you the instructions. let me know if you need extra detail or if I should send them to anyone else too [20:03] mmcc: thanks, i'll add those to the faq [20:03] dobey: ^^^ [20:03] cool, thanks [21:53] so, our proxy tunnel runner works from source but not frozen for OSX. it also doesn't log any errors and just uses the plain t.i.reactor version of connectSSL when it can't start the proxy tunnel [22:09] mmcc: did you get a chance to look at my branches btw? [22:11] dobey: still in the queue, sorry - I was thinking I'd be out of the weeds with this proxy question sooner, so I kept digging on it. I'll look now. [22:19] oh. what's wrong with the proxy? [22:28] well i think it's time for me to call it a day for now at least. later! [22:34] whoops, missed you there. bye. as for the proxy - the basic issue is that the proxy tunnel is not being started on os x, and so we don't use the system-configured proxy. also we're not logging that fact.