=== m_conley_away is now known as m_conley [01:09] mmcc_: calling IPC after the reactor has finished? it can certainly work on linux, since we are using dbus, but it's no wonder that it does nothing on darwin and windows, since the ipc needs the twisted reactor to be running. Let's discuss about this tomorrow in your AM. === eu is now known as Guest78687 === m_conley is now known as m_conley_away [07:10] Tried to send this earlier but had connection issues: I tried again to create a menu item that overrides the Qt default item that calls Qapplication::quit(). This time it worked, so here's a diff that catches cmd-q and kills syncdaemon then cleans up: http://paste.ubuntu.com/1254978/ - I think this is the right way to go, but I'm curious if anyone thinks main is the wrong place to put it. [07:50] morning all! [08:13] Good morning all! :) [11:13] good morning! [11:38] gatox, buenas :) [11:44] mandel, elo! [12:09] * gatox is not starting very well the morning [12:10] if i execute the tests in mac, everything is ok..... i with the same pythonpath i try to execute control panel.... it says it can't find PyQt [12:13] gatox, are you using the python installed by brew? [12:13] gatox, if not, it will be using other path or the other way around, you installed pyqt with brew and you are using the python found inthe system [12:14] mandel, you are awesome! jejeej i was executing u1-cp as ./bin/ubu... if i do: python bin/ubu... it works! [12:14] gatox, glad to help :) [12:36] meh [12:50] lunch time for me === mandel is now known as mandel|lunch [13:31] Hi all, [13:31] alecu, hi [13:32] Gatox, hi! I'm a bit trashed today [13:32] alecu, what happend? [13:32] :S [13:32] gatox, just feeling sick [13:33] alecu, sick day was invented for that :D [13:33] Gatox, I have already written ralsina about it [13:33] Gatox, and yes, I'm taking a sick day because of that [13:34] alecu, good..... have some rest sir! :P [13:34] Gatox, I came here to see if everything was okay [13:34] fml [13:34] Gatox, Ok... Thanks! [13:34] alecu, at least for me..... i was kind of blocked with mac..... but i figure it out some stuff that was failing and moving on now [13:35] Dobey, what happened? [13:36] alecu: i think a qt update in quantal broke ubuntu-sso-client :( [13:36] Ouch [13:37] In how many pieces? [13:37] testing with the previous version (4.8.2) now, to see [13:38] I mean, dobey: what part of it is broken? [13:38] indeed, downgrading qt, and the failing tests pass now [13:38] alecu: at least qtnetwork [13:39] Ugh [13:39] alecu: https://code.launchpad.net/~dobey/ubuntu-sso-client/update-4-0/+merge/127358/comments/274187/+download === m_conley_away is now known as m_conley [13:41] * dobey wonders why these tests are failing exactly [13:43] dobey: It looks like every proxy test that is related to authentication is broken. [13:44] well, and ubuntu_sso.utils.webclient.tests.test_webclient.BasicProxyTestCase.test_anonymous_proxy_is_used [13:44] dobey: that's related to authentication too [13:46] dobey: I'm sure Mandel may be able to give you a hand with those tests, since he worked on all of them afaicr [13:48] I'm sorry I can't be of much further help: I've been feeling dizzy the whole morning, so I'm taking a sick day [13:49] * alecu will be back later today [13:50] alecu: go rest [14:09] wtf [14:09] the tests pass when i run them on my laptop with the newer qt [14:09] but are failing in tarmac === mandel|lunch is now known as mandel [14:15] alecu, ouch, get better! [14:16] mandel: ^^ any idea about those failing tests? [14:16] dobey, let me take a look [14:20] dobey, what version of squid are you running? it looks like the tests are not going throw squid [14:20] dobey, to be prices the via/squid header is missing [14:21] 3.1.20-1ubuntu1 [14:22] mandel: right, the header is missing; but it's a change in qt that affected this [14:22] alecu: go away, you are sick ;-) [14:22] what's weird though, is that the tests pass on my laptop, but not in tarmac :( [14:22] dobey, hm.. let me try to run the tests in Q [14:22] dobey, is your desktop Q too? [14:22] mandel: no, my workstation is still precise [14:23] ack [14:27] dobey, I can reproduce the failure in ! [14:27] dobey, seems like a Q problem, give me some mins and I'll try to debug what is going on [14:28] yes, like i said, it's due to a change in qt that happened in the upload for 4.8.3 [14:28] oh, on my laptop the tests were just skipped it seems :-/ [14:37] dobey, it looks like the rest of the headers are correctly added but the one from squid [14:37] gatox, do you remember the doc we had to set up the mac env? [14:38] mandel, yes... pm with that [14:38] thx [14:40] mandel: probably because qt isn't connecting to it [14:40] dobey, as in going straight to the localhost.. probably, let me check that, is quite simple [14:54] dobey, could it be that the via header is disabled by default in squid? [14:54] dobey, which might have been updated [14:55] mandel: since you are the one that has something the closest to working on the in-dash stuff: can you contact John lea and show it to him? [14:55] mandel: to make sure we are not missing something from the design and such. [14:55] ralsina, sure, I can send him a video etc.. [14:55] * ralsina carefully avoids explaining how to "show" that [14:55] mandel: no, squid hasn't changed at all [14:55] dobey, not event the default conf? [14:56] mandel: an interactive demo would be ideal, but I suppose building the whole stack is near-impossible, right? [14:56] mandel: no, the package hasn't changed in months [14:56] dobey, ok.. must be something else [14:56] ralsina, yes, that is not feasible [14:56] mandel: like i said, it's qt. if i install the old version (4.8.2) the tests pass fine, and they fail with 4.8.3 [14:56] dobey, ok.. I wonder why.. [14:57] mandel: so video it is. [14:57] ralsina, let me help dobey first [14:57] mandel: no rush, tomorrow's fine [15:00] me [15:00] me [15:00] me [15:01] me === mmcc_ is now known as mmcc [15:01] mandel, ralsina: standup? [15:01] me [15:01] sorry, otp [15:01] me [15:01] DONE: [15:02] Working on shares tab issue on mac, tracking why the ipc communication or something is failing and the response is not being received. [15:02] TODO: [15:02] Finish with that issue and move on to fix another mac stuff. [15:02] BLOCKED: [15:02] No [15:02] dobey, go [15:02] DONE: releases, uploads [15:02] TODO: finish 4.0.0 releases/uploads, poke py3 twisted guys about packaging [15:02] BLCK: None [15:02] briancurtin: go [15:02] DONE: debugging, ended up taking a half day for the john hunter memorial service [15:02] TODO: windows u1cp env debugging [15:02] NEXT: mmcc_ [15:02] DONE: figured out how to catch cmd-q [15:02] TODO: polish up cmd-q, work on daemon mem usage [15:02] BLOCK: no [15:02] next: mandel [15:02] DONE: More work on unity not compiling. Splitted most of the code out of the preview. [15:02] TODO: more splitting. try to land it. help dobey with proxy. [15:02] BLOCKED:no [15:02] ralsina, please [15:03] mmcc: please share how you caught cmd-q :-D [15:03] DONE: calls, reviews, 1-1s, planning TODO: calls, reviews, try to actually code! BLOCKED: too many of the former to do much of the latter NEXT: EOM [15:03] ralsina: from last night: "I tried again to create a menu item that overrides the Qt default item that calls Qapplication::quit(). This time it worked, so here's a diff that catches cmd-q and kills syncdaemon then cleans up: http://paste.ubuntu.com/1254978/ - I think this is the right way to go, but I'm curious if anyone thinks main is the wrong place to put it." [15:04] mmcc: so the problem was that we did *not* have our own menu, but something Qt adds to keep the mac global menu bar happy? [15:05] ralsina: yes, right. [15:05] mmcc: awesome catch [15:06] :\ - I tried exactly this earlier but just messed up creating the menu item so it didn't get picked up [15:07] of course, no useful error messages that time, so I just moved on [15:08] mandel: is there any good way to figure out what's really happeneing here? [15:09] mmcc: if it makes you feel any better, we have been asked to do an actual menu a few dozen times [15:10] mmcc: so, I am not terribly happy about it being in main, but that file already has per-platform code anyway [15:11] dobey, is not easy what so ever, we need to look at what is going on in the wire [15:12] ralsina: yeah, and the shutdown code is duplicated between there and systray, which I don't like, but I didn't think of a good place to consolidate it [15:13] mmcc: we could start a "shutdown.py" but not now [15:14] * gatox lunch === gatox is now known as gatox_lunch [15:14] ralsina: yeah, maybe just one of those great comments in the code that "if you change this, also change this file over there" :) [15:18] mmcc: I think main import systray, so we could put it in systray even if that makes very little sense [15:18] mmcc: oh, no it doesn't [15:20] dobey, I'm trying to see if I can connect with wiresshark and take a look at what is going on in the wire [15:20] dobey, should be better than adding prints everywhere [15:23] mandel: yeah, wireshark should work, since it's plain http [15:35] dobey, weird, I'm connecting wireshark to lo and get no traffic.. wtf [15:36] mandel: strange, that shuld work on linux [15:36] ralsina, indeed [15:37] mandel: try tcpdump and see if it catches packets [15:37] mandel: it's probably not connecting to localhost; try the "grab from all interfaces" and see if you can find the requests [15:46] dobey, ralsina, very strange, applying the following filter in wireshark: http.server contains "TwistedWeb" and get no requests when listening to all interfaces [15:46] but we do get our server to be twistedweb.. [15:48] ralsina, fyi I talked with johnlea, a video is enough [15:49] mandel: awesome! [15:50] mandel: don't filter, get all the traffic, then look for the connction and rebuild the TCP stream [15:51] ralsina, I captured everything, then filtered the traffic after I stoped recording [15:57] brb, coffee time === gatox_lunch is now known as gatox [16:04] agh.. I wonder what changed in qt.. [16:08] mandel: http://launchpadlibrarian.net/116637150/qt4-x11_4%3A4.8.2%2Bdfsg-2ubuntu1_4%3A4.8.3%2Bdfsg-0ubuntu1.diff.gz [16:09] dobey, what was the command to upgrade to 12.10...... i remember once you told that to ralsina..... but i can't remember the exact command [16:09] dobey, maybe => Make sure QGlobalNetworkProxy is created for QNetworkProxy [16:09] gatox: update-manager -d ? [16:09] gatox: upgrading from 12.04 to 12.10 is not likely to work right now, I heard [16:09] dobey, that one! thx! [16:10] mandel: could be [16:10] gatox: let me know if it explodes ;-) [16:10] ralsina, ah ok....... i was going to do it, because i'm working on mac now [16:10] dobey, which could mean that is not going through the proxy [16:10] mandel: right. which seems to be the case, given the missing Via header. [16:11] dobey, I'll take a look at what force_proxy is doing to solve that [16:12] mandel: great, thanks [16:12] i need to get some lunch [16:12] bbiab [16:29] ok, EOD I'll fix that bug tom morning [16:29] see you all tom! [16:32] i'm waiting for the time a person named tom joins the team and mandel has to type out "tomorrow" :) [16:32] briancurtin, I'll claim that nickname so he has to be called tom_ [16:32] haha === yofel_ is now known as yofel [18:04] mmcc: care to review https://code.launchpad.net/~dobey/ubuntuone-control-panel/update-4-0/+merge/127559 ? [18:05] dobey: sure. [18:10] +1 dobey [18:12] thanks === gatox_ is now known as gatox [18:29] brb, gotta run a couple errands [18:40] ok… after upgrade i can't connect to internet with that machine :S [18:52] any advice on testing code that adds an action and menu to Qmenubar? I wanted to just call it and test that the action is in the menu bar, but it doesn't look like I can get a list of the current menus from the menu bar… [18:53] mmcc: let me check the docs [18:53] mmcc: you can get a list of actions http://qt-project.org/doc/qt-4.8/qwidget.html#actions [18:54] aha, thanks ralsina [19:08] gatox_mac: any idea why these failures would happen as they do? https://code.launchpad.net/~dobey/ubuntuone-control-panel/update-4-0/+merge/127559/comments/274671/+download [19:09] * dobey wonders if they're also due to qt changes [19:09] nope [19:19] Any recent changes to U1CP in how its started on the command line? i'm trying to login and get CredentialsError when choosing "Sign me in with my existing account" on trunk [19:20] briancurtin: sounds like it's not able to start sso? [19:21] dobey: SSO is already started and U1CP appears to interact with it when i start it up. i'll dig in if this doesn't ring any bells [19:22] briancurtin: are you getting the dialog to enter your email/password from sso? or just the error? [19:22] briancurtin: no bells ringing [19:22] dobey: i get the sorry an error has occurred dialog before the email/pass would pop up [19:23] dobey, i don't think so… i'm going to take a look after i finish with the branch i'm working on right now on mac… but i'm having some problems with mac and ubuntu, so i don't know when is going to be…. i hope soon [19:24] briancurtin: i'd look at the sso logs first then. seems like something is going wrong in there [19:29] dobey: yeah thats where im looking now. its raising that CredentialsError for not finding creds (which sort of makes sense), but it doesnt allow you to then provide any, so something's not hooked up correctly [19:30] briancurtin: I did change the code for finding binaries (eg, to launch the SSO gui). maybe that's not working right. IIRC it was working for me when running from trunk on windows, but not super well tested… [19:31] mmcc: could be it. running windows IRL from source tends to lag behind. i really need to get that "dev channel" updater setup some day...would catch this stuff right away [19:49] so if I add an action to a menu and add that menu to a menubar, that action doesn't show up in menubar.actions(), and I can't see a way to get a list of menus from the menubar… [19:49] mmcc: is actions() empty? [19:49] ralsina: yes [19:50] mmcc: but is menu.actions()? [19:50] ralsina: in this context I don't have menu [19:50] mmcc: actions() probably has the actions that were added directly to the menubar [19:50] mmcc: so, you may have to iterate something. let me check docs [19:50] yeah, that makes sense. I was just expecting to be able to get a list of menus from the menubar [19:51] mmcc: when you call addMenu() you should get a QAction [19:52] mmcc: which should be in menubar.actions() unless the gc is doing some trick [19:52] mmcc: also, you can try menubar.children() but I am guessing that's empty [19:53] ralsina: I do get the qaction, but I wasn't exposing that from the function I'm testing. I wanted to call the function then test that it put the right action in the menubar [19:53] mmcc: try children() and let's see what's there [19:54] but from the test function, I can't seem to do step 2. does that make sense? [19:54] ok [19:54] mmcc: actions() is returning actions that are added directly, you are adding a menu which has an action in it, so it's not in menubar. You may want to keep a reference to that menu :-( [19:55] hmm. children is this: [] [19:55] not a menu [19:56] keeping a reference so I can test it seems like it's defeating the purpose of the test - I want to check that it's in the menu bar, but if I keep the menu ref around for testing all I can really check is that the ref is what I expect it to be, which is less useful IMO [19:57] mmcc: that's ... strange [19:57] here's what I'm talking about. I have this function: http://paste.ubuntu.com/1256753/ [19:58] I wanted to write a test that finds the 'quit' action in the menu bar and checks that it has the same method hooked up to the triggered signal [19:58] mmcc: that QToolButton.menu() is probably the menu you are creating. [19:58] hmm. ok, checking [19:58] mmcc: so, menubar.children()[0].menu().actions() should contain your action [19:59] ugh. menubar.children()[0].menu() crashes… [20:00] mmcc: grmbl [20:00] mmcc: if you can put a name to the QAction, you can call menubar.findChildren() [20:01] aha, ok. I will try that when I get back from lunch… I've been telling them 'one mnute' for about ten minutes [20:03] ok… eod for me… i'll keep fighting with twisted tomorrow... [20:03] bye people [20:18] ugh, 41ms ping [20:20] ralsina: what did we decide to do with the menu separator branch from gatox for the sync menu? jsut try to get it in an SRU? [20:20] err, just even [20:23] dobey: hmmm [20:23] dobey: I wouldjust do a 4.0.1 and abuse that they let us get microreleases in [20:23] dobey: as a SRU [20:23] rye_, second (and last) branch to solve the syncdaemon problems we talked about: https://code.launchpad.net/~facundo/ubuntuone-client/fix-path-retrieval/+merge/127583 [20:23] ralsina, alecu, pfibiger, ^ [20:24] facundobatista: looking... [20:24] ok [20:28] ralsina: https://code.launchpad.net/~dobey/ubuntuone-client/update-4-0/+merge/127586 [20:28] dobey: got it [20:34] dobey: +1 [20:36] facundobatista: I don't think I can finish reviewing that one today [20:39] EOD for me [20:39] bye! === mmcc_ is now known as mmcc === m_conley is now known as m_conley_away [21:55] later all === eu is now known as Guest51072 === Guest51072 is now known as adorilson [23:18] ugh, need to keep a reference to the menubar around or it reverts to old default menubar. wtf [23:18] will update my branch and add a test for that tonight. as-is, my branch doesn't work [23:19] because I moved things around to clean it up, I broke it. whee