[08:32] Happy Friday, everyone! :) === zyga is now known as zyga-afk [08:47] morning! [10:33] * mandel reboots due to updates [11:03] good morning! [11:19] gatox, buenas! [11:19] mandel, buenas === zyga-afk is now known as zyga [12:02] hello everyone [12:02] nessita, hi [12:05] gatox: would you be able to do a review? https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/validate-in-sd/+merge/100041 [12:06] nessita, yes [12:06] gatox: depends on latest nigthlies of u1client [12:06] upgrading like a baus [12:06] baus? :-) [12:07] nessita, didn't you see the video? [12:07] apparently not :-P [12:07] nessita, http://www.youtube.com/watch?v=NisCkxU544c [12:07] nessita, morning [12:07] is really famous :P [12:07] * mandel super quick errands! [12:08] hola mandel [12:08] nessita, before I go, for the autoupdate we need a UFe :( [12:08] nessita, missing strings.. [12:08] mandel: when you come back, we will talk about it [12:08] mandel: we may not mark for translations those [12:09] or we may try to use other strings already there [12:09] nessita, ok, I'll take very few mins, I need to pick up a pair of keys downstairs [12:09] mandel: ack [12:17] holas! [12:18] alecu, buenas [12:19] ralsina, I've got branches to fix the issues on windows. I did a lot of IRL on them and auth proxy, non auth proxy and non proxy cases seem all to be working ok. [12:19] I still need to create the bugs and do some cleaning up. [12:19] nessita, +1 [12:21] nessita, I'm back [12:21] mandel: would you have a list of the new strings? [12:21] nessita, looking for them right now, give me a few seconds [12:21] sure [12:21] will grab some milk [12:22] nessita, this are the two new ones: http://paste.ubuntu.com/907123/ [12:28] mandel: are those the only new needed ones? [12:28] nessita, yes [12:28] nessita, those are the title on the balloon shown for the update and the actual message [12:29] mandel: ok, what about making the titte "Ubuntu One"? [12:29] nessita, I have not double check with design or ux (I just realized) but there should be just two [12:29] nessita, sure, using the app name sounds reasonable :) [12:29] mandel: let me grep a little for the second [12:31] sure [12:32] hey! the ubuntu font issue in Qt apps is now fixed! [12:32] seems like the update made it to ubuntu [12:32] nessita, gatox, mandel, ralsina: finding this was an awesome moment: http://stackoverflow.com/questions/1736015/debugging-a-pyqt4-app [12:33] alecu: looking [12:33] it's basically calling QtCore.pyqtRemoveInputHook before doing pdb.set_trace() [12:33] and it allows you to debug qt apps in a sane way. [12:33] alecu, i think there is a package in pypi that does that [12:33] alecu: wow! [12:34] the other day we had that problem, and martin gaitan upload it i think [12:35] gatox, awesome. What is it called? and... Why didn't you tell us about it?!!!! [12:35] nessita, wait wait wait, why didn't we know about this?! [12:35] mandel: about the pdb thing? [12:35] alecu, gatox ^ :P [12:35] alecu: oh, nice :-) [12:35] alecu, :P don't remember.... let me check [12:35] nessita, yes, it was a joke :P [12:36] nessita, with no context I just realize is a shit joke hehe [12:36] and good morning! [12:38] alecu, can't find it... i'll ask later to martin which is the name [12:39] mandel: regarding the new string, I would say let's ask ralsina and joshuahoover [12:39] ralsina: how late is it to ask a UIFe for a new string? [12:39] nessita, I agree [12:39] nessita: incredibly? [12:39] ralsina: mandel needs one new string for the autoupdate dialog [12:39] nessita: that's windows only [12:40] nessita: so we may not set it up for translations, since we are not translating windows until after P release [12:40] ralsina: yes, but the string will in the multiplatform project and will be listed for translation (if we mark it) [12:40] ralsina: right, that was my suggestion [12:40] nessita: so let's do that :-) [12:40] but wanted to confirm is too late first [12:40] too late and not needed [12:40] at least not strictly needed [12:41] mandel: ack? ^ (do change the title to "Ubuntu One" please) [12:41] nessita, ack! [12:43] hi u1 team. any update on the authentication issue with u1, updated to 12.04 beta and still unable to log in, been weeks now [12:44] psypher246: hi there. There is no issue as far as I know, was someone in particular helping you when you first reported this? [12:44] ralsina: would you please review https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/validate-in-sd/+merge/100041 ? [12:45] nessita: on it! [12:46] nessita: when it first bombed out i got an erro which lead me to this bug (https://bugs.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/711162) i have asked about it on the channel a few times but got no response [12:46] Ubuntu bug 711162 in ubuntuone-client (Ubuntu) "ubuntuone-login crashed with ValueError in call_async(): Unable to guess signature from an empty dict" [High,Confirmed] [12:46] since updating about 3 weks agho u1 will not log in [12:46] psypher246: that error means one of the U1 dbus services is not working for you. Let's debug [12:47] nessita: ok [12:47] psypher246: can you please pastebin the output of running the following in a terminal "ps aux | grep ubuntu"? [12:48] http://paste.ubuntu.com/907154/ [12:48] psypher246: great, can you please close the controlpanel and kill ubuntuone-syncdaemon? [12:49] done [12:49] psypher246: also, could you please close/kill the ubuntuone-indicator as well? (just to be sure what we have running) [12:49] nessita: +1 [12:49] psypher246: so, after confirming that no ubuntuone nor ubuntu-sso-login process are running, please pastebin the output of: [12:50] U1_DEBUG=True python /usr/lib/ubuntu-sso-client/ubuntu-sso-login [12:50] http://paste.ubuntu.com/907160/ [12:51] psypher246: perfect, do not kill that. Can you now paste the output of (run in another terminal): [12:51] U1_DEBUG=True python /usr/lib/ubuntuone-client/ubuntuone-login [12:51] 2012-03-30 14:51:43,295 - ubuntuone.credentials - INFO - Starting Ubuntu One login manager for bus u'com.ubuntuone.Credentials'. [12:52] psypher246: great, now in a third terminal, would you please run u1sdtool -c? [12:52] psypher246: a UI for login should pop up for you [12:53] psypher246: let me know how that goes [12:53] don't think it's coming, both other terminlas just spat out a bunch of output and stating it's being shutdown, will pastebin [12:53] psypher246: please, thanks [12:53] http://paste.ubuntu.com/907162/ [12:54] and http://paste.ubuntu.com/907164/ [12:54] psypher246: ok, so, what does u1sdtool -s says? [12:55] should be in READY state [12:55] lunch time for me!!! I'm starving! [12:55] * mandel lunch [12:55] State: AUTH_FAILED [12:55] connection: With User With Network [12:55] description: auth failed [12:56] is_connected: False [12:56] is_error: True [12:56] is_online: False [12:56] queues: IDLE [12:57] fyi never changed pw and not able to see how you entre a new one [12:58] psypher246: so, seems like you do have credentials in the system... would you please open seahorse and search for any entry with UBuntu One in the name? [12:58] psypher246: also, please let me know if you see your current device (computer) listed in https://one.ubuntu.com/account/machines/ [12:58] yes the machine is listed [12:59] don't see u1 saved in keys, which tab am i looking for? [12:59] passwords [12:59] psypher246: passwords [13:00] gatox, nessita, ralsina, thisfred, urbanape: can I ask for reviews? [13:00] https://code.launchpad.net/~alecu/ubuntu-sso-client/qt-defer-to-thread/+merge/100132 [13:00] nessita, can you check if this sounds good for you? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/unicode-info/+merge/100007 [13:00] https://code.launchpad.net/~alecu/ubuntuone-client/fix-broken-tunnel/+merge/100133 [13:00] alecu: indíd! [13:00] alecu: reviewing the first [13:00] psypher246: could you please also confirm that you see your current device (computer) listed in https://login.ubuntu.com/+applications [13:01] ralsina, can you check this branch please: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/unicode-info/+merge/100007 ? [13:01] passwords only has one item, login, can't do anything with it [13:01] gatox: sure! [13:01] used to be able to expand i think? [13:01] nessita, ralsina: sorry, was wiping tears from my eyes seeing that there was a suggestion that we may have needed a uife ;) [13:01] ralsina, nessita i tested that on windows and linux and it works [13:01] joshuahoover: sorry! ;-) [13:02] return unicode in both cases [13:02] nessita: yes i see my device listed twice, once under software center and then again under u1 [13:02] joshuahoover: no worries, we will not mark the new string for translation, so no new UIFe needed [13:02] alecu: we missed 1:1 yesterday, quick mumble? [13:02] nessita: cool :) [13:02] psypher246: ok... did you find your u1 entry in seahorse? [13:03] psypher246: right-click on login and choose "unlock" perhaps if it is locked? [13:03] nessita: i only see" paswords: login" [13:03] nope only lock [13:03] psypher246: right click will allow you to "unlock" it? [13:03] oh [13:03] no [13:03] dobey: no? [13:03] even if i lock it and then unlock again still no subfolders [13:03] dobey: ping [13:04] dobey: we need to have a video lens package ready for upload on short notice if/when we get a different URL for queries [13:04] dobey: sorry about the rush job [13:04] psypher246: i think there is a bug in gtk+ which seahorse is hitting :( [13:05] ralsina: uh ok [13:05] * thisfred reviews https://code.launchpad.net/~alecu/ubuntuone-client/fix-broken-tunnel/+merge/100133 [13:05] ralsina, dobey, just ping me when you have it if you want. [13:05] psypher246: ok, we can workaround this :-). If you open controlpanel, from the u1 icon in the messaging menu, what do you get/ [13:05] ? [13:05] davidcalle: oh, awesome :-) [13:05] ralsina: i suppose i need to set up nightlies for it too :P [13:06] dobey: probably but not urgent, since we are in hell-freeze already and nothing is going to change except this [13:06] right [13:07] nessita: the u1 client opens [13:07] psypher246: does the info loads? in particular in the last tab (Account) [13:07] gatox: just to be on the safe side, since options.path_info comes from an external library, I would check that it is a string before trying to decode it [13:08] nessita: yes [13:08] gatox: since u'á'.decode('utf8') fails miserably [13:08] nessita: personal detail has my username and email, your services is blankl [13:08] ralsina, sound reasonable.... i'll do that [13:08] gatox: just to futureproof it against whatever [13:08] psypher246: would you please share a screenshot of that? [13:09] psypher246: ah! the bug with the expanding in seahorse appears to be an issue in the theme :-/ [13:09] at least for seahorse it's the theme. evolution seems to still be broken [13:10] https://www.dropbox.com/sh/71su31e0ac2hz9b/735GElJ-cA [13:10] dobey: do you perhaps have the bug nr for that? [13:10] psypher246: i don't know if there is a bug filed for it yet. i just confirmed the problem though :) [13:10] psypher246: one second, need to check something [13:11] interesting [13:12] maybe it is a gtk+ bug [13:15] psypher246: I'm doing some queries for your account, will get back to you asap [13:15] nessita: ok thx\ [13:17] gatox: did you notice that the overlay in ussoc is no longer a transparent overlay? :-/ [13:17] nessita, mmmmmm nop [13:17] gatox: when you have some slot, would you check if that's the case for you? [13:17] nessita, i can do that now, before getting innto the seg.... [13:17] sigsegv [13:19] gatox: please [13:21] gatox: aaahhh I think that was a fix that ralsina added... [13:21] dobey: logged the issue: https://bugs.launchpad.net/ubuntu/+source/seahorse/+bug/969203 [13:21] Ubuntu bug 969203 in seahorse (Ubuntu) "12.04 - seahorse (passwords and keys) does not allow me to unlock or expand passwords" [Undecided,New] [13:21] gatox: perhaps, when he set that every widget has to have white background, the overlay broke [13:21] nessita, ok..... i'll check that now..... is there an issue for this?? or just do it! [13:21] gatox: please open a bug if you can reproduce and fix :-) [13:21] nessita: I did not, but maybe I did one widget too many [13:21] nessita, roger that [13:22] gatox: should be a simple fix though, sorry :-/ [13:22] ralsina, yes [13:22] no problem [13:28] wow, i totally just hit a not so fun issue in launchpad [13:29] psypher246: i've marked your bug as a dup of mine, which is bug #969209 [13:29] Launchpad bug 969209 in gtk+3.0 (Ubuntu) "Missing expander arrows in some tree views" [Undecided,Confirmed] https://launchpad.net/bugs/969209 [13:29] ok u have way more info :) [13:29] dobey: thx [13:34] ralsina, sorry, didn't see your request to mumble. [13:34] ralsina, shall we do it now? [13:37] alecu: I'm about to mumble with beuno, maybe in a bit? [13:37] ralsina, sure. [13:37] alecu: also, I posted a needsinfo in oe of your branches [13:38] ralsina, awesome. Yes, I added that bit to both branches, and I forgot to add the tests to the -sso one. [13:39] alecu: cool, with the test, +1 from me [13:39] alecu: pre-approving :-) [13:39] thanks! [13:39] psypher246: still waiting on some queries (just FYI) [13:41] psypher246: appears there's already fix which should get into precise soon [13:41] ok cool [13:44] ralsina, question: "why did you add this: background-color: white; for the qwidgets in the qss?" [13:44] just trying to understand which was the problem [13:46] ralsina, to see how to change that..... and don't break what that was fixing [13:47] psypher246: hi [13:47] hi rye [13:47] * rye reads the backlog [13:47] gatox: to make it work with dark color schemes [13:47] gatox: see the attached bug for my branch [13:48] ralsina, ok..... i'll look that [13:48] gatox: just put an objectname on the overlay and give it a style [13:48] gatox: it will win by higher specificity [13:49] psypher246: so, the syncdaemon status is AUTH_FAILED, from u1sdtool --status, right? [13:49] rye: and from controlpanel status [13:49] rye: yes [13:50] s/work with/ignore/ :) [13:50] ralsina, gatox: we already have a section for the overlay in the qss where we modify it by name, I think [13:50] yes [13:50] nessita, gatox: so let's just set the bg color [13:50] ralsina, i'm playing with that [13:51] psypher246: do you have old-style tokens in the seahorse? I suppose based on the bug report you _might_ have them but they are not accessible, right? [13:51] most likely [13:51] psypher246: is it precise? [13:51] yup [13:52] psypher246: does "login" still show there if you search for "UbuntuOne" in seahorse, or does it disappear? [13:54] dobey: no nothing happens [13:54] briancurtin: hola! any update on the uninstall issue? was wondering if I should seek more reviews or fix anything [13:54] nessita: nothing yet, still trying to get it working [13:55] psypher246: so "Passwords: login" still appears in the list, with "ubuntuone" in the search box? [13:55] yes [13:55] you have an old style token in your keyring then :) [13:55] the searchbox has a "missing icon" picture in the box itself as well [13:55] nessita: im just going to start creating new accounts since i cant figure out where non-admin user creds are being stored in order to get the fresh install to run [13:56] briancurtin: sorry :-( [13:56] psypher246: ok, that's probably a theme issue [13:56] nessita: no worries, it's all a part of the game :) [13:58] psypher246: if you go to https://one.ubuntu.com/account/ and remove the device entry for your computer, then quit the u1 control panel if it's running, and start it back up, what happens? [14:01] dobey: ok cool, no gonna have to re-upload though? [14:01] ralsina, are you very busy atm? [14:01] psypher246: nopes [14:01] mandel: kinda, how can I help you? [14:01] ralsina, I was wondering if we could move or 1-1 a little earlier [14:02] psypher246: no, it's just like logging out and back in. if you have files already uploaded, you should be fine [14:02] dobey: the local entry will not be removed, since sso does not manage it, only reads. we need to remove the token programmatically [14:02] dobey: control panel still hasthis devive in the list [14:03] mandel: let's do it right now [14:03] ralsina, superb! [14:03] dobey: [14:03] wheni try remove the device on the web it does not remove it, just takes me to the service screen [14:04] worked from the control panel though [14:04] ok === urbanape` is now known as urbanape [14:05] dobey: ojk re-added this device still getting auth failed [14:06] but my name is there and all the data usage stats [14:06] right [14:09] dobey: i suppose the filesync servers use somehow different way of detecting the correct token to use === yofel_ is now known as yofel [14:11] rye: it seems like syncdaemon is getting the old token, and control panel is maybe getting the new one [14:11] but not sure [14:12] psypher246: could you please run this script - http://paste.ubuntu.com/907263/ - this will remove the token from gnomekeyring [14:12] (if i correctly specified the arguments) [14:13] psypher246: wait, no [14:13] psypher246: http://paste.ubuntu.com/907265/ - this one [14:15] u1: 5: u1: Syntax error: "(" unexpected [14:17] names script u1 fyi [14:17] named* [14:17] psypher246: are you running it as "python scripname" ? [14:17] oooh sorry i thought it was bash [14:18] psypher246: sorry, should have specified that, if it errors out with NoMatchError, then no, you did not have the token, otherwise if it prints nothing, then it removed the token [14:18] psypher246: if you re added the computer, you need to restart syncdaemon [14:18] http://paste.ubuntu.com/907273/ [14:18] psypher246: since it will not pickup automatically the new token [14:19] psypher246: well, this means there is nothing to delete... at least nothing i am searching for [14:19] should i just wait till that other seahorse bug is fixed? [14:19] wait [14:20] i just restarted like nessita said, connected now [14:20] psypher246: YEY [14:20] yay [14:20] :) [14:20] i would have tried removing it and adding, i am just always so worried about having to re-upload [14:21] I have lost my storage that i bought? [14:22] psypher246: tokens don't have direct connection to your data, you can create new ones and they will link to the same account. The only thing that matters is that you should not authorize your computer with another ubuntu one account, in this case you will get ROOT_MISMATCH [14:22] how can i check when my 40gb expires? [14:22] gotta run for a doctor's appointment, will be back in 80 minutes, and have no lunch. See you all soon! [14:22] psypher246: Mar-05-2012 Expired [14:23] ah! [14:23] ok [14:23] don't recall getting a warning\ [14:23] should I? [14:24] gatox: is this ready https://code.launchpad.net/~diegosarmentero/ubuntuone-client/unicode-info/+merge/100007 ? [14:25] nessita, yes [14:25] gatox: who suggested adding the isinstance? [14:25] webm0nk3y: ^ psypher246 [14:25] nessita, ralsina told to be nice to check that options.path.... is in fact a str [14:26] nessita: alrighty. i *finally* got my Popen uninstaller to work successfully on a XP non-admin account (still has to be executed from "run as"). it works fine on an admin account as well. i can test it on Win7 now if you'd like [14:26] gatox: I would definitely not use isinstance, that means we have no idea what we're getting and we're just guessing :-). I would definitely recommend a try-except UnicodeDecodeError [14:26] rye: ? [14:26] nessita: however, i still can't explain why the old way, shellexecute, does not work. i'm looking into that as well [14:26] nessita, ooooook [14:26] rye: is there no system which notifies you that your storage is going to expire, looking at email I don't see anything [14:27] briancurtin: could you please check with gatox and mandel about that? they built the code the first tme [14:27] briancurtin: and they may defend it, or help debugging it [14:27] briancurtin, may I have some context? :) [14:28] mandel: the "uninstall_application" at the bottom of this MP: https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/uninstall/+merge/100012 [14:28] * mandel looks [14:28] when i run IRL, i get a traceback (i'll have to dig it up, don't have it handy) that the ShellExecute call does not work. however, if i just change it to use subprocess.Popen, everything is fine [14:28] i'll dig up that traceback [14:29] mandel: with ShellExecute being used, I get https://pastebin.canonical.com/63408/ (Access is denied.) [14:29] briancurtin, let me see.. [14:30] so the path being given to it is correct, and just passing it to Popen works...but somehow ShellExecute can't do it. this was taken from somewhere in ubuntuone-windows-installer [14:31] i switched it with """Popen([uninstall_path, "--mode", "win32"])""" and it works fine [14:32] briancurtin, without runas? [14:32] briancurtin, what user do you have? have you tried with a non admin user? [14:32] briancurtin, I would be very happy if we can remove runas because it will allow users with an AD account :) [14:33] mandel: i've tried the Popen way with both admin and non-admin on XP. i'm about to start trying on Win 7 [14:35] mandel: the installer itself needs runas, at least on XP. since this uninstall is called during the installer (when you disagree with the license), the process is already elevated and doesn't need explicit elevation to run uninstall [14:35] i'm going to build an installer that uses ShellExecute without runas and see what happens [14:36] briancurtin, please do, using runas is a dirty horrible trick [14:36] briancurtin, and is possibly related to the installer working in a funny way regarding the privileges [14:39] nessita, i've just push the changes into u1-client-unicode-info [14:39] gatox: thanks! [14:39] FWIW, if we don't need runas, then we also don't need ShellExecute. it's effectively just doing a bare CreateProcess with no special "verb" (the runas) [14:42] briancurtin, and it would be great! [14:42] thanks for all the help guys in getting me back up and running, gotta go, have a great weekend! [14:42] briancurtin, also fixes the AD users bug which I had no idea how to solve and I suspected that we would have because people in corporate networks use it [14:44] mandel: do you have a reference to that? The installer will still need to be started by right clicking and selecting "Run as..." (at least on XP). in my tests i have Administrator, brian (an admin), and limited (a non-admin) - Admin and brian can just click on the installer and it goes. limited can't -it fails to even begin copying the files [14:44] mandel, gatox: can I ask for one review on this? https://code.launchpad.net/~alecu/ubuntuone-client/fix-broken-tunnel/+merge/100133 [14:45] briancurtin, to the AD bug? [14:45] alecu, sure [14:45] alecu, ok! [14:45] mandel: if there's a LP # [14:45] alecu, read the u1-internal we have a very stupid bug related to proxy, I'll fix asap [14:45] briancurtin, sure there is, let me find it :) [14:45] mandel: we won't be able to install to "C:\Program Files" without admin elevation. this is part of the reason why python installs to C:\PythonXY, so everyone can have it even if their sysadmin doesn't give admin access :) [14:48] briancurtin, yeah, that is why we used the dirty runas, but it certainly makes no sense.. [14:50] briancurtin, if we remove runas from everywhere you are fixing bug 894471 [14:50] Launchpad bug 894471 in Ubuntu One Control Panel "Autoupdate is broken on machines that have joined a domain with AD" [Undecided,Confirmed] https://launchpad.net/bugs/894471 [15:00] me [15:00] me [15:01] meeeeeeeeeeeeeeeeeeeeeeeee [15:02] me [15:03] me [15:04] alecu, nessita ralsina dobey ? [15:05] * gatox listen the sound of the wind...... [15:05] zzzzzz [15:05] meh [15:05] me should write some notes [15:05] the wind sounds like that in spain.. [15:06] mandel, lazy wind! [15:06] * dobey needs to plan some holiday [15:06] dobey, to listen the wind in spain? [15:06] :P [15:06] no [15:07] but to go to spain [15:07] dobey, you know you are welcome :) [15:07] i would say gatox go..... the rest can be last....... [15:07] * thisfred recommends Portugal: the people are much nicer there :P [15:08] gatox: go [15:08] DONE: [15:08] Propose a branch for the encoding problem in u1sdtool --info. Fixing the overlay issue in sso (almost done). [15:08] TODO: [15:08] Start working on: #940669 ubuntu-sso-login crashed with SIGSEGV in QSocketNotifier::setEnabled() [15:08] BLOCKED: [15:08] No [15:08] urbanape, go [15:08] DONE: got the new buildout dev environment up, interviewed someone for the iOS position [15:08] TODO: get ubuntu-sso-client branch back into play and finished, diagnose a crasher in the iOS Music app [15:08] BLOCK: None [15:08] mandel: por favor [15:08] oops [15:08] me [15:08] DONE: All hands, if anyone else asks me for a review I will hunt them down and make them suffer as much as I do in that web page. Got autoupdate branch up to date and working. [15:08] TODO: IRL for the autoupdate branch. Propose and move to bug 969280 [15:08] BLOCKED: no [15:08] briancurtin, go! [15:08] Launchpad bug 969280 in Ubuntu One Client trunk "If http_proxy is enabled then an incorrect assumption about https_proxy is made" [High,In progress] https://launchpad.net/bugs/969280 [15:08] DONE: testing windows branches almost every second of the day. heard the XP startup sound more than i ever needed to [15:08] TODO: figure out this windows uninstall/autoupdate situation, eventually move back into windows read/write for testing/jenkins improvements [15:08] BLOCKED: no [15:08] NEXT BATTER: thisfred [15:08] DONE: reviews / u1db mapping indexes progress TODO mapping indexes BLOCKED no NEXT dobey [15:08] λ DONE: team meeting, bug #961342, bug #968555 (protocol) [15:08] λ TODO: more bug #968555, bug #969262 [15:08] λ BLCK: none. [15:08] Launchpad bug 961342 in ubuntuone-client (Ubuntu Precise) "GI error while trying to show notifications" [High,Triaged] https://launchpad.net/bugs/961342 [15:08] Error: Launchpad bug 968555 could not be found [15:08] Launchpad bug 969262 in libubuntuone trunk "Error while getting credentials: The '/org/freedesktop/secrets/collection/login/24' object does not exist" [High,In progress] https://launchpad.net/bugs/969262 [15:08] alecu [15:08] DONE: IRL testing of sd on windows, a few fixes for the tunnel: bugs #969150, #969157, #969159 [15:08] TODO: catch up with SD bug backlog [15:08] BLOCKED: no [15:08] COMING UP: nessita [15:08] Launchpad bug 969150 in Ubuntu One Client "Proxy tunnel is trying to use QtDbus on Windows" [High,Triaged] https://launchpad.net/bugs/969150 [15:08] Launchpad bug 969159 in Ubuntu Single Sign On Client "Windows keyring uses twisted deferToThreads even when no reactor is installed" [High,In progress] https://launchpad.net/bugs/969159 [15:08] Launchpad bug 969157 in Ubuntu One Client "Tunnel is not using the system proxy on Windows" [High,In progress] https://launchpad.net/bugs/969157 [15:09] DONE: a lot (landed like 5 last-minute branches yesterday, for controlpanel mostly) [15:09] TODO: some more, add missing tests to controlpanel [15:09] BLOCKED: no [15:09] TODO.append: reviews [15:09] any comments anyone? [15:09] nop [15:09] alecu, he, bug 969159 makes sense, the keyring was never updated not to work without a reactor [15:09] Launchpad bug 969159 in Ubuntu Single Sign On Client "Windows keyring uses twisted deferToThreads even when no reactor is installed" [High,In progress] https://launchpad.net/bugs/969159 [15:10] everyone: reminder that next week is national holiday in argentina on Mon, Thru and Fri [15:10] mandel, already fixed and about to land :-) [15:10] I will be working Tue, Wed and Thru [15:10] alecu, sweet! [15:10] nessita, not monday *thru* friday! [15:10] alecu, gatox: could you please specify which days will you be working next week? [15:10] nessita, just mon, thu, fri. [15:10] alecu: LOL [15:11] nessita, tue, wed, thu, fri [15:11] nessita, all: I'll only be working tue and wed. [15:11] thanks [15:11] eom then? [15:11] * alecu needs to check canonicaladmin to see if it's in order. [15:11] nessita, I'm updating my holiday, I'm working on monday and tues for what is worth it, the rest of the week I'm out [15:12] mandel: thanks for the update [15:13] the US needs to get some of these many-day-a-week holidays going [15:17] briancurtin: we do. it's called March Madness. [15:20] ah this is true. it became second spring break in college [15:21] alecu, +1 [15:21] * gatox lunch+bank [15:21] briancurtin: So, I got the buildout complete, and while the ubuntuone-dev-tools egg is built, I don't see u1trial anywhere in the devsetup. [15:22] briancurtin, mandel: did you guys reach any conclusion in the uninstall issue? [15:22] nessita: i'm running on win7 right now a few cases, will report the final findings shortly === gatox is now known as gatox_away [15:23] nessita, if we can really get rid of the runas and shell execute we will, is cleaner code [15:23] urbanape: ah, i haven't written the equivalent of what on windows is the "env.bat" script (in devsetup/) - it copies u1trial and u1lint to devsetup/bin, then sets up the PATH and whatnot [15:23] nessita, briancurtin, is doing IRL for that [15:23] mandel, briancurtin: nice! we should apply the same solution to both: the autoupdater and the uninstaller, agreed? [15:24] briancurtin, I just realize, for the autoupdate, we need the runas because the execution is done as the user that executes the control panel [15:24] nessita, ^ [15:24] mandel: i haven't looked explicitly at autoupdate, just uninstall, but i see it there as needing the same investigation [15:24] mandel: the uninstall as well, no? [15:24] briancurtin: ^ [15:24] the uninstall does not work with runas [15:24] (on XP, testing on 7 now) [15:25] briancurtin, nessita, if we do a release with that we should make sure that rmcbride and elopio know about it so that we don't brake things [15:25] briancurtin: excuse my ignorance, but not sure what that means [15:26] nessita: whether you are an admin or a limited user, uninstall does not work when "runas" is an option to ShellExecute [15:26] it gives that "access is denied" error [15:27] nessita: your uninstall branch either requires the "runas" option to ShellExecute to be removed, or to use Popen [15:27] nessita: but this is only what i've seen on XP within a workgroup. apparently it may be different in a domain. it may also be different in Win7, which is what im looking at now [15:28] but what i'd expect is the majority of users, people using Windows on their home computer which is just a workgroup, your uninstall branch won't work for them [15:31] ah-ha...looks like there's a difference between XP and 7 here. still running tests, but i think i know how to fix this [15:36] briancurtin: just FYI, officially we don't support XP (we always try to), but when in doubt, prioritize win 7 [15:36] mandel: so you said you'd work Mon and Tue only? [15:37] nessita, yes, let me update my calendar and add it to the online services one [15:37] mandel: according to the canonicaladmin, you're in holiday Mon and Tue [15:37] and on swap day on Wed [15:37] mandel: and you work Thu and Fri [15:38] nessita, yes, I talk with ralsina, the house I was going had an accident and had to change mon and tues [15:38] mandel: you broke a house without even being there? [15:38] lunch, bbiab [15:38] nessita, hm.. weird, I though I ask for 5/4 and 6/4 let me check [15:38] mandel: your breaking powers amazed me [15:39] nessita, frozen pipes.. water everywhere :( [15:49] nessita, mandel: here's what we have...https://pastebin.canonical.com/63477/ i think we should not use "runas" for the uninstall [15:50] briancurtin: from that, I would say +1 [15:50] briancurtin: will change that after having lunch [15:50] briancurtin: thanks a lot for testing this! [15:50] brb [15:50] :) [15:51] * mandel looks [15:54] briancurtin, so in all the systems * ShellExecute (without runas) - success, right? [15:54] mandel: yep, that seems like the best route. i don't currently have a Vista VM, though [15:55] briancurtin, do you have the script you use for testing? I want to try it on Vista, is an evil OS and will like to know what it does [15:55] briancurtin, I have one, usually never boot it but got it just in case :) [15:55] mandel: i just created three different installers with the code changes, then ran them manually [15:55] mandel: i can upload them somewhere if you want to run them [15:55] briancurtin, sure, put them in a u1 folder and share it with me :) [15:59] nessita, FYI I'll do the same tests for Vista [16:03] I'm back! [16:16] ralsina, all the fixes for the tunnel issues on windows have landed on trunk. [16:16] ralsina, should we make branches to merge them against stable? [16:16] briancurtin: the removal of the runas parameter has been pushed to revno 309 [16:17] briancurtin: please let me know if I blindly fixed the windows tests correctly [16:17] briancurtin, FYI I might take a little to test the installer, waiting for vista to do updates.. [16:18] nessita: will look [16:20] alecu, nessita: good question. If we do that, we can try to make a windows release that is synced to 2.99.92 [16:21] nessita: are we still merging everythig on this release, or cherripicking? [16:21] ralsina: I will update stable branch with what we have in trunk on tuesday [16:21] ralsina: we're merging everything [16:21] nessita: cool then. Alecu, no need. [16:22] awesome then. [16:42] ralsina: shall I propose a branch that wipes everything outside 'scritps' from the windows-installer project? [16:42] mandel: how's the autoupdate branch going? [16:43] ralsina: if we ever need something from there, will be in trunk's history [16:43] nessita, doing IRL tests [16:43] nessita: sure [16:43] ralsina: do you guys use anything outside scritps/ now? [16:44] nessita: no [16:45] * briancurtin stepping away for a few minutes [16:45] meh i don't want to look at this bug :-/ [16:47] dobey: just close your eyes [16:48] mandel: I would advice merging https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/uninstall/+merge/100012 in your autoupdate branch [16:48] i'm still working on that as a means to fix bugs [16:48] mandel: since it may generate conflicts for you [16:49] nessita, ok, will do now and see what goes wrong [16:50] mandel: let me know, I can help [16:50] will have lunch now [16:50] but I will be back [16:58] * alecu says it's time for lunch, errands and whatnot. === alecu is now known as alecu_errands [17:09] finally back! [17:09] the transit is a MESS! and the streets are full of people (that scares me :P jjee) === gatox_away is now known as gatox [17:12] gatox: and is hot! [17:12] nessita, also! [17:12] ralsina, after your unicode post..... do you want to review this? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/unicode-info/+merge/100007 [17:13] gatox: indeed :-) [17:13] gatox: like manuel says, this project is an endless source of blog posts ;-) [17:13] jejeej [17:13] gatox: that would actually be UnicodeDecodeError, UnicodeEncodeError :-) [17:14] if path_info is unicode, it will give a DECODE error. If it's invalid bytes, it will be ENCODEERROR [17:14] ralsina, ahhhhhh true [17:14] or viceversa [17:14] if it was unicode es Encode [17:15] nessita, FYI there are some conflicts fixing and lets see if tests pass :) [17:15] mandel: ack [17:16] ralsina, done [17:16] gatox: checking [17:16] ralsina: you can't have an encode error there [17:16] ralsina: since we're never encoding, just decoding [17:16] nessita: yes you can [17:16] ralsina: how? [17:16] nessita: try u'á'.decode('utf-8') [17:16] nessita: SURPRISE! :-) [17:17] hum, why is raising a UnicodeEncodeError? [17:17] unicode magic [17:17] nessita: http://t.co/wChRKHzF [17:18] ah, ok, thanks [17:18] I learnt something new [17:18] nessita, add_to_autostart is gone, right? [17:18] mandel: no, is moved so is alphabetically ordered [17:19] nessita, ah, ok I did not see it up there, thx [17:20] ralsina: are unicode objects actually guaranteed to be utf-8 though? [17:20] dobey: they are guaranteed to be ENCODABLE as utf-8 [17:20] asking if a unicode object "is utf8" is meaningless. [17:20] ralsina, briancurtin: when you can, https://code.launchpad.net/~nataliabidart/ubuntuone-windows-installer/wipe-it/+merge/100204 [17:20] except as internal representation which is unimportant [17:21] mandel: also, if you could review the uninstall branch, that would rock [17:21] nessita, certainly, let me fix merge issues and take a look at the the mp [17:21] -9056! nice! [17:21] mandel: sure [17:21] nessita: will look shortly [17:21] ralsina: and +0! [17:21] briancurtin: no rush [17:22] nessita: that branch should close a bunch of bugs, too! ;-) [17:22] ja [17:22] nessita: +9056 on that branch [17:29] nessita, in https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/uninstall/+merge/100012 the perform update does need the runas AFAIK since it is a process that will require the admin right as is spawned by a process that does not have them [17:29] briancurtin, do you think so? ^ [17:29] mandel: then that would break, no? (as per briancurtin's tests) [17:29] i don't know specifically for the update one, ive only looked into uninstall. they may operate differently [17:30] briancurtin: ah, ok, shall I restore the runas there then? [17:30] nessita, I would do so and I'll make sure we use the correct one in the autoupdate branch [17:30] ok, will revert that change there [17:30] nessita, that way we reduce the chance to get it wrong :) [17:31] mandel: true [17:31] nessita, I'll do IRL for this specific case [17:31] nessita: i'm not sure yet. also i think the way you have it written, removing hte "runas" parameter entirely, will not work. what i meant was to make that paramter ""...running the installer in 1' to double check this [17:31] briancurtin: ah, ok [17:35] nessita: yeah you will need to add back in a "" where the "runas" used to be [17:35] briancurtin: added, committed and pushed to revno 310 [17:35] mandel: ^ [17:37] mandel: also restored the runas for updater [17:37] nessita, sweet I'll remerge with mine :) [17:54] nessita: uninstall MP approved [17:54] briancurtin: awesome, thanks! [17:56] nessita, in here: https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/uninstall/+merge/100012 in FrozenTestCase is there a way to get the cleanup calls closer to what the clean? [17:56] nessita, if it is no possible is not an issue [17:58] nessita, https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/uninstall/+merge/100012 +! [17:58] +1 :P [17:58] mandel: we can have them closer, but then we'd need duplicate some calls, because the same cleanup may be needed by any guard [17:59] nessita, then is fine, the branch is approve I trust your decision [18:00] thanks! [18:02] dobey: +1 [18:07] overlay done..... it was more difficult than i thought [18:07] it wasn't just a style thing [18:09] gatox: really? [18:10] ralsina, yap..... i tried setting a specific style for the overlay, with background-color: none/transparent... but didn't work [18:10] i needed to set the stylesheet directly to the widget, to delete any previous style...... and change a bit the paintEvent for the animation [18:11] not the paint event, but who paint event is being listened [18:11] i'll propose now [18:19] ralsina: any chance you review this? https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/unicode-syncing-error/+merge/99990 [18:19] nessita: sure! [18:20] nessita: I had that queued yesterday, must have skipped it [18:20] ralsina: np [18:21] nessita, ralsina FYI I have I did the IRL of the update and works ok but I have tests failing and I ran out of time, I' work on thins on monday and will ping briancurtin to help me fully test it [18:21] dobey: lp:~diegosarmentero/ubuntuone-client/unicode-info just bounced very weirdly on tarmac [18:21] dobey: giving an error about a PDF of all things [18:21] mandel: ack, let me know whatever you need [18:21] mandel, ralsina: we need this landed today if we want to have it in stable, no? [18:21] mandel: awesome [18:22] mandel: what if briancurtin follows up on that? ralsina, is that possible [18:22] ? [18:22] in stable next Tues, I mean [18:22] mandel: oh, I thought you meant you ran out of time in the test fixes you were working yesterday! [18:22] ralsina, nessita review please: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/overlay-transparent/+merge/100216 (and with pretty pictures :P) [18:22] nessita, I can work a little longer later today, just give me an hour to take the dog out a little [18:22] mandel, nessita, briancurtin: yes, please, brian, take over mandel's branch, let's try to merge that today [18:22] ralsina, nop.. on the autoupdate [18:23] mandel: no need to over work, perhaps briancurtin can follow up [18:23] mandel: can you send me the branch and some info about what's going on with it? [18:23] mandel: what nessita said, we are not THAT hurried [18:24] gatox: I really really really not like putting the qss like that, in the code. You 100% there is no other way? [18:24] gatox: have you asked ralsina for more options, using qss only? [18:24] ralsina, ack [18:24] briancurtin, let me push the branch for you, give me a sec [18:25] nessita, i'll try another option and let you know [18:25] nessita, gatox: nothing comes to mind, really, I don't understand the problem yet [18:26] nessita, ralsina let me try another option........ but, something positive (at least for me)..... is that if you reuse this component (the loading overlay) the other projects won't need to know the qss for this component to look properly [18:28] gatox: and that's a con... each project should be able to customize the overlay as the project needs [18:28] gatox: another project may need the background orange a the bullets white [18:31] nessita, ok.... the other option didn't work either..... i can keep with this if you want.... trying more options.... but i'm not sure that is going to be doable just with the qss [18:31] gatox: I don't understand why the QSS is not applied. You know? [18:32] ralsina, the qss is being apply...... but for some reason, i can't remove the background........ mmmmmmm another idea! [18:32] gatox, ralsina: so, looking at gatox's branch, I realized that the indentifier frm_box is just the "small" box with the bullets in it [18:32] gatox, ralsina: I think we should set a name for the "parent" of that box, no? [18:32] and set the style to it [18:32] nessita: the parent is the big one? [18:33] gatox: did you try that ^? [18:33] ralsina: looking [18:33] ralsina: "should be" [18:33] nessita, yep [18:33] gatox: and that did not work? [18:33] gatox: there's a lint error in your branch [18:34] ralsina: ^^ [18:34] nop..... i did a setObjectName to the LoadingOverlay class..... and didn't work [18:34] setting the style for that name in the qss [18:34] ralsina: the pdf message isn't a fatal error [18:34] dobey, really?? a TODO? [18:34] gatox: no a unicode error in bin/u1sdtool [18:35] ahhh the other branch [18:35] in unicode-info yeah [18:35] 160: local variable 'UnicodeEncodeError' is assigned to but never used [18:35] ahhhhhhhhhh silly me...... [18:36] gatox: oh your code is wrong [18:36] who approved that! [18:36] you need parens! [18:36] argh, my fault [18:36] * ralsina goes read the python tutorial as penance [18:37] my bad, i forgot that [18:37] briancurtin, the branch is here: bzr push lp:~mandel/ubuntuone-control-panel/auto-update-looping-call [18:38] briancurtin, the idea is to add a looping call that will check for updates every x minutes, will pop a message in the app tray icon and will perform the update when the user clicks on it [18:38] briancurtin, there are two tests failing due to the crazy merge I had to fix [18:39] gatox: I'm testing a branch where the styled is applied, but something is not letting the 'transparency' work [18:39] mandel: ack, thanks, i'll start working on it [18:39] gatox: so, I added this to the style: [18:39] 93 LoadingOverlay { [18:39] 94 background: red; [18:39] 95 border: 3px solid green; [18:39] 96 } [18:39] ok, all have a great weekend! [18:39] and ar, have a great easter :P [18:39] mandel: easter is next sunday? [18:39] gatox: and I have a redish background with green border, the problem is that the redish is not full red, but like it had something applied to it that make sit softer [18:40] anyway have good weekend you too! [18:40] ralsina, gatox: can this be explained with your knowledge? http://ubuntuone.com/5trjjPDlwqzRWFRO7hkGha [18:41] ralsina, gatox: that's the result of applying the style I pasted above ^ [18:41] nessita: since i'm doing a similar thing with manuel's branch that you did with my volume_manager notify branch...is there some formal way of "taking it over", or do I just branch from his code and push to my own? [18:41] nessita, why is not fully red? [18:41] nessita: two widgets, one transparent the other red, one on top of the other? [18:41] because the painter is adding a transparency [18:41] briancurtin: no other way other than hijacking :-) [18:41] gatox: you tell me why is not fully red [18:41] nessita, yes [18:42] nessita: LoadingOverlay has an opacity less than 1 [18:42] nessita, which was the questio? [18:42] gatox: what is the background red not fully red? [18:42] why* sorry [18:42] nessita, ah... so the answer was correct [18:42] gatox: what's the answer? [18:42] nessita: LoadingOverlay has an opacity less than 1 [18:42] dobey: "where"? :-) [18:43] nessita: i don't know, i'm not looking at the code. but that is *why* :) [18:43] nessita, the painter that "paint" :P the widget has this to fill the area: QtGui.QColor(255, 255, 255, 135) [18:43] where the last number is the opacity [18:43] 255 will be not transparent [18:43] gatox: let me play with that then [18:43] already did [18:44] I made it work [18:44] I'm getting the transparent background [18:44] let me share the branch [18:45] gatox: this is working in my env http://pastebin.ubuntu.com/907682/ [18:46] ralsina, gatox: that diff gives me http://ubuntuone.com/5trjjPDlwqzRWFRO7hkGha [18:46] yes..... :S [18:46] nessita: looks good to me! [18:46] crap.... [18:47] nessita, are you going to propose that? [18:47] gatox: all I've learn for qss I've learnt from you, so you can consider this your win [18:47] gatox: no no, you do it [18:51] nessita, ralsina changes pushed.... [18:51] gatox: thanks! [18:52] dobey, https://code.launchpad.net/~diegosarmentero/ubuntuone-client/unicode-info/+merge/100007 updated [18:53] gatox: approved! [18:54] nessita, thanks...... and sorry about that.... i was too focused try to make the thing transparent! that i miss to do that completely :S [18:54] man, I love the ubuntu font not being bold anymore [18:54] and with objetName wasn't working [18:54] nessita: the overbolding is fixed now? [18:55] gatox: yes, I tried objectName... I don't like saetObjectName, since I think is buggy when more than one widget is created [18:55] dobey: YES YES YES [18:56] gatox: anyways, the only thing I would like you to take from this is that if you find doing some nasty Qt things, ping ralsina for help... because if we reach a place where Qt make us do nasty things, we need to rethink higher level stuff [18:56] gatox: or not apply the fix altogether [18:56] still bold to me in control panel [18:56] nessita, gatox +1 [18:56] dobey: did you update today and rebooted? [18:57] gatox: same thing goes to nessita and I, nasty stuff needs more eyes [18:57] roger that [18:57] nessita: why reboot? [18:57] dobey: font cache [18:57] gatox: anyways, thanks a lot for working on this [18:57] dobey: with unity that's the only thing that works for me to apply changes [18:58] dobey: but I haven't tried that hard [18:58] it looks SO much better now [18:58] ralsina: nonsense :) [18:58] ralsina: are bold font gone for you as well? [18:58] dobey: hahaha [18:58] nessita: ah, i do see a font update now, that wasn't there earlier [18:58] nessita: never had it because I am not using ubuntu font [18:58] dobey: I have non-bold fonts since this morning 9am ART [18:59] nessita: I only had it on my test-fresh-user accounts [18:59] ah [18:59] maybe i didn't do an update this morning then, or forgot to do dist-upgrade [18:59] it's all just a blur to me [19:01] dobey: hum, perhaps you should quit drinking scotch when you wake up :-P [19:01] ralsina: any nice way to access the last row in a treewidgetview? [19:01] nessita: hmmm... is it a firt-level item? [19:01] nice would be without grabbing an iterator and going to the end [19:01] ralsina: yes [19:01] nessita: give me 1' [19:02] nessita: know the text? [19:02] ralsina: the text that is inside the row? [19:02] nessita: topLevelItem(topLevelItemCount()) [19:02] nice! [19:02] thanks [19:03] nessita: what would be a maximum acceptable time for the control panel to connect [19:03] with necessary object names inserted [19:03] nessita: I am not 100% sure that works right if items are sorted on insertion though [19:03] ralsina: is ok, we do not sort [19:03] are NOT sorted [19:03] thanks! [19:03] nessita: yes we do [19:03] elopio: hum... to connect to the web? [19:03] ralsina: heh? [19:03] nessita: what else would i drink with the gummy bears? [19:03] dobey: milk!!! [19:04] with gummy bears? crazy [19:04] nessita: I recall it not being the same order as the volume list but I may be remembering wrong [19:04] ralsina: we order the volume list and then fill in the folder treewidget, so strictly speaking we do not re-order the treeview [19:04] nessita: then it's ok [19:04] ralsina: thanks [19:05] elopio: sorry, I'm not sure what to answer since I'm not sure what you're specifically asking :-) [19:08] nessita, I'm making a script that clicks the connect button and waits for the control panel to say it is connected. [19:08] nessita: but I need to set a maximum time for the script to wait until connected. [19:09] elopio: you mean the 'file sync connect' button? [19:09] nessita: yes. [19:10] elopio: even if we assume you have netowrk, and the speed of the connection is within average, the connection can take a very long time, since it depends on our U1 servers answering to syncdaemon [19:10] elopio: if we have issues in the server, you can "never" connect [19:10] elopio: under normal circumstances, I would say 30 seconds [19:10] (as a maximun) [19:11] or a minute [19:11] nessita: here it's taking around 1:30. For now I'm going to set a really big maximum so doesn't fail because of the load in the server. [19:11] elopio: ack [19:12] but one good usability goal in the future would be to make that a maximum of, i don't know, 30 seconds. [19:12] elopio: main issue is that the proces of 'connect' involves, among others, a server rescan [19:12] I'm sure we have a lot of other issues to take care :) [19:12] elopio: and network, which is impossible to restrict [19:13] nessita: but if my network is overloaded, we should say: "it's taking a lot of time, check your connection or patiently wait". Like gmail does. [19:14] just wait for ever doesn't sound nice. [19:14] elopio: right [19:15] nessita, ralsina: good news are that the qt-testability driver works great for the automation. [19:16] elopio: awesome! [19:23] yay the bold is gone [19:23] also, ubuntu is a really compact font [19:29] need to reboot..... brb === alecu_errands is now known as alecu [19:33] ralsina: just FYI, the last one is folders.topLevelItem(folders.topLevelItemCount() - 1) [19:34] nessita: hahaha [19:34] you know what they say, there are 2 difficult things, naming stuff, cache invalidation, and off-by-one errors [19:34] heh [19:54] nessita: yeah i dont know what is going on with this mandel branch. i fixed one thing to get tests further along, but i'd say there's a 1% chance of this being figured out by the end of the day :/ [19:55] briancurtin: may I help somehow? have traces I can look at? [19:56] nessita: here's the branch https://code.launchpad.net/~brian.curtin/ubuntuone-control-panel/mandels-updater and here's the trace https://pastebin.canonical.com/63500/ [19:57] briancurtin: did you merge latest trunk into it? [19:57] yes [19:57] that part should not have changed at all [19:57] ralsina: i started from trunk, merged mandel's branch [19:57] briancurtin: branching and looking [19:58] briancurtin: he probably reverted a fix I did in main [19:58] briancurtin: anyways, I'm seeing a LoopingCall there that worries me, since in the controlpanel we no longer have a reactor (on linux, we do on windows) [19:58] nessita: the autoupdater stuff should be windows-only [19:59] nessita: but it doesn't look well factored out [19:59] ralsina: yes, we could use a qtimer the same, no? [19:59] nessita: indeed [19:59] nessita: harder to test, maybe [19:59] ralsina: also, we agreed with mandel that the strings will not be marked for translations, and I see this [19:59] UPDATES_TITLE = _('Updates') [19:59] UPDATES_MESSAGE = _('There is a new update available') [19:59] :-/ [19:59] perhaps mandel did not push everything? [20:00] nessita: in which case we are out of luck [20:00] ralsina: and the branch is creating a LoopingCall in multiplatform code! [20:00] nessita: right [20:00] :-( [20:01] on linux the LoopingCall is in ubuntuone\controlpanel\gui\qt\task\linux.py, on Windows it uses twisted (if that's what you were talking about earlier) [20:01] ralsina: and there's new code in the systray test file? what does have to do autoupdate with systray? [20:01] nessita: when called with some options, u1cp's window is not instantiated [20:02] nessita: also, that's where notifications should come from [20:02] nessita: as in "balloons" [20:02] ralsina: but this is not the code and MO we had in the former windows-installer, right? [20:02] ralsina: I thought that we were just moving what we had in installer to controlpanel [20:02] nessita: no, but that was for checking updates n startup only [20:03] ralsina: or did I misunderstood? [20:03] nessita: this is about having regular checks for updates. I don't know what you talked with mandel, but that's what he told me [20:03] ralsina: on the call yesterday... what we agreed about this? [20:04] nessita: that there was no autoupdate code, and that we would add it [20:04] nessita: the specifics of which one of the two ways to autoupdate were not mentioned [20:04] nessita: and I guess everyone assumed what he wated [20:04] ralsina: ...moving what we had in the windows-installer to controlpanel, that's what I recall. Perhaps I just thought that? [20:05] nessita: maybe [20:05] nessita: we could do the "just on startup" for this release [20:05] nessita: and kick the looping check for the next one [20:05] ralsina: that was my idea all this tume [20:05] ralsina: given the time constraints [20:05] nessita: ok, so miscommunication [20:05] nessita: right [20:05] apparently [20:05] nessita: when mandel spoke about his branch he was talking about this, he never had a branch for the "on startup" checl [20:06] ralsina: I can see 2 options: [20:06] ralsina, nessita ein? [20:07] * ask briancurtin to make a brand new branch with the code that checks only at startup (moving from installer to controlpanel) [20:07] mandel: one sec [20:07] ralsina, if the looping call is run as on launch, it does the check on startup which is when the app is started, later it will be done once every 24 hours [20:07] * wait for this branch to land next week, and to fully test it, and make another stable-3-0 update after the update for releases, making that an update for 2.99.92.1 [20:08] no, let's go with the first option [20:08] ralsina: the last option means we don't ship this code in the tags and updates we're doing on Tuesday, but I can make another stable-3-0 update after Tuesday [20:09] since the previous release had only on-startup checks, it's not a regression if this doesn't go in [20:09] mandel: apparently there was some miscommunication about what each one of us meant with 'auto update' feature [20:09] nessita, ah.. [20:09] mandel: I thought you were moving *just* the (working) code from installer to controlpanel [20:09] mandel: and I guess you understood you should finish the autoupdate "cool" branch [20:09] nessita, yes, that is what I though.. [20:10] mandel: right, sorry for not being more explicit [20:10] nessita, well, sorry for not understading.. my brain 'subcsociente' surely wanted to do the 'cool' one [20:10] mandel: anyways, now brian can't make the test suite pass, and I see the diff and is very very large to land in a rush (it may break Ubuntu "very easily") [20:11] nessita, ok, makes sense, I can do more testing on monday with IRL etc.. or I can get a branch with a single one, as you wish [20:11] mandel: so, the safest path I guess is just adding the windows isolated code (from installer). But we can ask briancurtin to do that :-) [20:12] mandel: so, enjoy the weekend, you can keep fixing this branch next weel [20:12] nessita: so what should i be adding? [20:12] nessita, indeed, is saver to just move the working code [20:12] mandel: but please do not land it until we agree "is safe" to do so [20:12] briancurtin, nessita , sorry for 'el marron' [20:12] mandel: is ok [20:12] nessita, I'll add you as a reviewer :) [20:13] briancurtin: in the windows-installer project, you will find some code that runs the autoupdater at controlpanel startup, we should move that to the controlpanel [20:13] ok, I'm off to watch tv, laters! [20:13] briancurtin: let me point you to it [20:15] briancurtin: ubuntuone_installer/gui/qt/utils/windows.py, methods are_updates_present, check_updates (you will need to do some magic to not need the 'gui' parameter), and perform_update [20:15] ralsina: that code needs to interact with the UI ^ :-/ [20:16] ralsina: to confirm with the user he wants to upgrade [20:16] ralsina: I'm not sure that's doable today... what do you think? [20:16] briancurtin: _get_update_path is already in controlpanel code, is the get_exe_path function [20:17] nessita: ack [20:17] briancurtin: anyways, I'm not that sure that's doable in a couple of hours [20:18] nessita: i'm also the wrong person to be rushing GUI-specific stuff :/ [20:18] briancurtin: indeed [20:18] (at least for now ;-)) [20:18] i'll get better :) [20:19] ralsina: what were the plans about the windows release for next week? regarding you being on vacations, was any plan for anyone else to move that forward? [20:21] briancurtin: perhaps we killed ralsina with this talk? [20:21] ralsina: you alive? :-D [20:21] * alecu wishes we have at least *one* windows release for internal testing. [20:21] before we start spreading it around. [20:22] we don't need signing for that, right? [20:22] alecu: I'm pretty sure we can have one without autoupdate code [20:22] briancurtin: am I lying? [20:22] nessita, that would be awesome. But if we get to test autoupdate, it would be even better. [20:22] agreed [20:22] nessita: we have one that i sent out last week or so [20:23] alecu: have any suggestion to solve this autoupdater issue? [20:23] nessita: oh actually we said don't use that one [20:23] i can create an installer from trunk in a few minutes if we want to send it out [20:23] alecu: ^ [20:24] briancurtin, that would be awesome. [20:24] nessita, no ideas on the autoupdater, no. [20:24] nessita: sorry, was afk [20:24] nessita, I've not seen that code in depth. Nor in surface. [20:25] let's do a test build on monday, brian knows how [20:25] alecu: I have in surface, rough estimate to move it and having working, I'd say one day [20:25] ralsina: and what about the other things I asked? :-) [20:25] then I can coordinate the RT moving forward on tuesday if the tet build is ok [20:25] ralsina: about autoupdater [20:26] ralsina: we have no autoupdater, and apparently we will no have one for MOn [20:26] about autoupdater? I said let's do the on-startup now and move the other until next release [20:26] ralsina: right, and I said: [20:26] nessita: the autoupdater on startup is not that hard. I can do that and propose tonight [20:26] (05:15:57 PM) nessita: ralsina: that code needs to interact with the UI ^ :-/ [20:26] (05:16:11 PM) nessita: ralsina: to confirm with the user he wants to upgrade [20:26] (05:16:26 PM) nessita: ralsina: I'm not sure that's doable today... what do you think? [20:27] ralsina: is complex as in the code is outside the qt/ dir, and the autoupdater requires an UI [20:27] ralsina: how would you solve that? [20:27] nessita: what code is outside qt/ dir? sorry if I am dense, but I don't remember [20:27] ralsina: the whole utils/ dir [20:27] where we have all the windows specific [20:27] hmmm [20:28] let me think 1' [20:28] ralsina: so the UI needs to check for updates, if there are some, present the user a confirmation dialog, if the user agreeds, the UI should call the thing that actually do the autoupdate [20:28] we could just launch the autoupdater on interactive mode [20:28] but that needs admin access. Sigh [20:29] ralsina: and the other options does not need it? [20:29] option* [20:29] nessita: no, just checking not-interactively doesn't [20:29] ralsina: but doing the actual install it does, right? [20:29] you are only asked for creds after you agree to update [20:29] which is not annoying [20:30] why not take that from utils and put it in qt/ with a noop if platform is linux? [20:30] we want no platform-specific code in qt at all? [20:30] ralsina: only that or the whole utils dir? [20:30] nessita: only the autoupdater [20:31] and then, when we want to do the looping call it [20:31] 's just a qtimer triggering that same thing [20:31] ralsina: bah, not sure why I asked, I'm -1 to either... we're risking breaking things [20:31] a new python module will break freezes [20:31] nessita: grmbl, but that means we never get to do a windows release again [20:31] you understand that? [20:31] even if in linux is a no-op (but the module has to be installed, etc) [20:32] ralsina: what about the logic I proposed? [20:32] "the UI needs to check for updates, if there are some, present the user a confirmation dialog, if the user agrees, the UI should call the thing that actually do the autoupdate" [20:32] nessita: the only way to check for updates is calling autoupdate.exe (or however it's called) and checking exit code [20:32] nessita: that *is* how it works [20:33] ralsina: and what's the problem with that? [20:33] sorry if it's obvious, I just don t see it [20:33] nessita: I don't follow you. mumble? [20:33] sure [20:33] briancurtin: wanna join us? [20:34] nessita: need a few minutes, waiting 2' for the microwave [20:34] sure, let me know [20:38] nessita: logging in now [20:44] eod here..... have a nice weekend everyone! [20:58] ok, guys, gal: have a great weekend. [20:59] see you some on tuesday. [20:59] ralsina, have a great vacations. [20:59] this is an early EOW for me, since I stayed up late yesterday fixing a few bugs... :-) [21:01] alecu: same for you! [21:06] have a good weekend/holiday/whatever all [21:34] ok, I'm of [21:34] f [21:34] bye all! [21:34] me too. BYe! [21:39] installer ready - http://ubuntuone.com/77HUgqLg7CbrISFAgclG1F - email sent [21:40] and im out of here as well. i need a beer.