[09:41] Morning all! [10:23] hey all. a few of us in the uk have machines which have files waiting to sync but no transferrs [10:24] popey: hi [10:25] hullo [10:25] popey: is SD status connected? [10:26] popey: and how big are the files && how much data was already transferred? - u1sdtool --current [10:26] rye: TheOpenSourcerer has a similar issue [10:26] hello. [10:26] as does AlanBell [10:26] yes, me too [10:26] Adding a dir and nothing seems to be happening [10:26] is_connected: True [10:26] no syncing or even getting the dir listed in the U1 client or the web [10:27] * AlanBell has added a directory of stuff and it synced for a bit but appears to have stalled [10:27] --waiting shows stuff waiting [10:27] --current-transfers shows 0 [10:28] nice. [10:28] by the way, i guess I can recommend installing the indicator until design comes up with a better alternative, to see whether anything is actually happening [10:28] http://paste.ubuntu.com/992121/ last 200 lines of my syncdaemon log including a little python traceback [10:29] AlanBell: https://bugs.launchpad.net/ubuntuone-client/+bug/988534 Hmm [10:29] i cant use the indicator right now, machine is at home and I am in millbank [10:29] Ubuntu bug 988534 in Ubuntu One Client "KeyError in the Upload" [High,In progress] [10:29] anybody else having a KeyError with marker? [10:30] rye daily or from your extras ppa? [10:30] AlanBell: what version of ubuntuone client are you running? [10:30] rye: fully updated precise [10:30] TheOpenSourcerer: extras will do [10:30] alan@wopr:~/.cache/ubuntuone/log$ grep KeyError syncdaemon.log | wc -l [10:30] 63 [10:31] http://paste.ubuntu.com/992122/ [10:32] popey: ok, you have upload failure from the server [10:33] AlanBell: could you please archive the whole log dir and send it to ubuntuone-support@canonical.com? [10:33] rye: is that good or bad? ☺ [10:34] can I start the indicator app without logging out? [10:34] TheOpenSourcerer: yep, it is in /usr/lib/indicator-ubuntuone/indicator-ubuntuone [10:34] ty [10:35] seriously? bug #978903 ???? [10:35] Launchpad bug 978903 in ubuntuone-client (Ubuntu Precise) "[precise] Client is stuck due to Upload executing before MakeFile" [High,Triaged] https://launchpad.net/bugs/978903 [10:35] ping joshuahoover, I can has SRU? [10:36] AlanBell: in your case that's bug #978903 which is supposed to be SRU'd ASAP but it is not for some reason, will follow up on this today [10:36] Launchpad bug 978903 in ubuntuone-client (Ubuntu Precise) "[precise] Client is stuck due to Upload executing before MakeFile" [High,Triaged] https://launchpad.net/bugs/978903 [10:38] popey: looking at your logs. something is not right there too, could you please grep -C 20 -E KeyError.*marker ~/.cache/ubuntuone/log/syncdaemon.log* ? [10:39] rye: no hits [10:39] last entry in my syncdaemon-exception.log: http://paste.ubuntu.com/992138/ [10:39] rye: ok, logs on the way [10:40] No obvious errors in syncdaemon.log [10:40] rye: SRU is being uploaded today [10:40] ralsina: oh, that's good to know, I was about to start alarming [10:41] rye: we delayed it bit because of the SSL problems, didn't want to lock users out :-( [10:41] so mine will sync tomorrow or something? [10:41] AlanBell: yes, maybe in two days. Earlier if you install the update manually frm proposed [10:42] will do [10:42] AlanBell: you may want to restart the client - it is a recoverable error [10:43] TheOpenSourcerer: your log message is quite spectacular, do you mind archiving the ~/.cache/ubuntuone/log/syncdaemon* files and sending them to ubuntuone-support@canonical.com too? There is a state mismatch and metadata does not correspond to the expected state [10:44] TheOpenSourcerer: talend/.metadata/.plugins/org.eclipse.core.resources/.root/.indexes/properties.index - is it inside a UDF ? [10:44] can do. I've been trying to add a new dir to U1 all morning and it just doesn't seem to be doing anything at all... [10:44] * rye wants to reproduce [10:44] TheOpenSourcerer: the directory was added but in a weird way [10:44] rye: I have removed that dir from U1 via the web interface. [10:44] TheOpenSourcerer: could you please try restarting syncdaemon to get fresh logs after you send current ones? [10:44] OK. [10:46] 1st batch of logs in their way. [10:47] what's the right way to restart the syncdaemon? [10:47] not a "service" by the looks of it [10:53] rye - talend dir and files are in ~/Development/Workspaces/talend... (plain old ext4 spinning rust file system) so not sure what you mean by UDF. These are created by Eclipse (Talend is an eclipse framework product). [10:59] TheOpenSourcerer: u1sdtoool --quit [10:59] TheOpenSourcerer: and then you can start it with u1sdtool --start [11:00] ah - I killed the process id then restarted it with /usr/lib/ubuntuone-client/ubuntuone-syncdaemon & [11:00] TheOpenSourcerer: then see whether the talend dir is still being synchronized - u1sdtool --list-folders - https://wiki.ubuntu.com/UbuntuOne/u1sdtool [11:00] TheOpenSourcerer: well, that also workd [11:00] works [11:01] taled is not in the folder list. [11:01] talend [11:02] I tried to add a new dir earlier called talend-5.1 - this is not in the list either. [11:03] OK - have added it agin and now it seems to be working. [11:03] do you want the new syncdaemon log files? [11:07] TheOpenSourcerer: ok, the old ones will do, since there is no file which causes the issue, there is no error [11:07] np. [11:08] thanks [11:09] ahh - it has stalled again. [11:09] http://paste.ubuntu.com/992174/ [11:11] TheOpenSourcerer: is there an Upload entry before that traceback ? [11:12] hang on I'll pastebin the log file. [11:13] http://paste.ubuntu.com/992178/ [11:15] TheOpenSourcerer: great, now you are definitely experiencing bug #978903 and unfortunately for now you have two options - 1. restart u1sdtool upon failure or 2 - patch the code with the diff http://bazaar.launchpad.net/~dobey/ubuntuone-client/queue-limit-3-0/revision/1184#ubuntuone/syncdaemon/action_queue.py [11:15] Launchpad bug 978903 in ubuntuone-client (Ubuntu Precise) "[precise] Client is stuck due to Upload executing before MakeFile" [High,Triaged] https://launchpad.net/bugs/978903 [11:15] TheOpenSourcerer: it is a ">" to ">=" change [11:15] :-) [11:15] thanks. [11:15] TheOpenSourcerer: yes, i know :-/ [11:18] patched. [11:19] restarted syncdaemon [11:19] cd - [11:19] opps [11:20] transferring files :-D hoorah! Thanks rye [11:22] AlanBell: you might want to do this one-line patch too in the mean time ^ [11:23] TheOpenSourcerer: thank you for confirmation. I really hope we can get this SRUd as soon as possible [11:24] np - thanks for the "quick-fix" :-) [11:25] where is the file that needs patching? [11:25] in /usr/share/pyshared/u1-client/... [11:26] in the syncdaemon dir [11:27] line 1098 add a "=" [11:27] :-) [11:27] good morning! [11:38] gatox, morning! I did the review you asked me to [11:38] gatox, let me know if there is a re-review needed [11:38] mandel, yes...... i'm fixing the branch === yofel_ is now known as yofel [11:49] AlanBell: /usr/share/pyshared/ubuntuone-client/ubuntuone/syncdaemon/action_queue.py [12:08] saluton ĉiuj! [12:08] alecu, morning! [12:09] alecu, hi [12:24] mandel, if you can, the branch is ready for re-review [12:25] gatox, I'll have to do it after lunch, is that ok? [12:25] mandel, of course! [12:35] http://people.ubuntu.com/~alanbell/Screenshot%20from%202012-05-17%2013:33:46.png is this normal? [12:36] loads of CPU and occasional bursts of network stuff [12:36] I am adding a folder of a gig or so of assorted stuff [12:36] ok, I'm off to have lunch [13:07] alecu, facundobatista: sent you email with what I understood yesterday, feel free to make corrections/additions before I send it upstairs :-) [13:08] ralsina, awesome, looking. [13:12] hmm [13:13] dobey: can you point me to the stable branch merges? I will do them starting in a bit. [13:15] ralsina, you should use my canonical mail account for this kind of stuff [13:15] facundobatista: sorry, I get autocompletion by name and don't notice [13:16] facundobatista: hapens to me about once a week [13:16] facundobatista: will try! [13:16] ralsina, no problemo [13:23] mandel: what do you think of doing it so when syncdaemon doesn't start we give an error that says something like "Cannot connect to syncdaemon" instead of "None has no attribute freaking get_rootdir?" [13:23] * ralsina starts branch [13:24] ralsina: https://code.launchpad.net/~dobey/ubuntuone-client/fix-test-fails-3-0/+merge/106062 [13:24] ralsina: taht one's pretty trivial, and should land first :) [13:24] dobey: on it! [13:25] dobey: global aprove [13:28] https://code.launchpad.net/~dobey/ubuntuone-client/fix-984407-3-0/+merge/106063 [13:28] https://code.launchpad.net/~dobey/ubuntuone-client/fix-getdelta-3-0/+merge/106064 [13:28] ^ and those are the other 2 [13:30] dobey: +1 on the 1st one, the second is going to take a bit [13:35] dobey: fix-getdelta works, passes tests, and everything but what's a change that adds a function that's not called supposed to do? [13:35] Oh, got it [13:35] nevermind [13:35] heh [13:35] damn context [13:36] dobey: There, +1 and global approved on all those branches [13:36] cool [13:37] * mandel back [13:37] ralsina: you didn't actually *vote* on https://code.launchpad.net/~dobey/ubuntuone-client/fix-test-fails-3-0/+merge/106062 but only set the global status to approved [13:37] ralsina, I do think is a much nicer error to be propagated [13:37] mandel: isn't it? [13:37] mandel: yeah, I will do it [13:37] dobey: oops [13:38] dobey: probably did it too quickly [13:38] ralsina, that is not the first user I see with that issue.. I hav been following @ubuntuone on twitter because some users report bugs that one.. [13:38] * mandel wonders what other uses do people find for twitter.. [13:39] briancurtin, I did a review for https://code.launchpad.net/~brian.curtin/ubuntuone-windows-installer/build_installer/+merge/105403 let me know what you think about the comment [13:41] mandel: good comments - yeah the bitrock stuff could be a bit cleaner. i'll fix it up [13:41] dobey: there [13:42] briancurtin, great, let me know when I can do another review :) === duanedes1gn is now known as duanedesign [13:46] ralsina: been seeing a lot of "NoneType 'object has no attribute 'get_rootdir'" . Trying to find one with logs for you [13:47] duanedesign: thanks [13:48] ralsina: I had someone come in IRC last night with the get_rootdir problem and as you suspected yesterday afternoon, it was in fact that SD wasn't running because of the "run as root" confusion [13:48] briancurtin: well, the good news is we can do a release to fix that [13:48] briancurtin: after today [13:51] alecu: we talked yesterday, so no 1-1 [13:51] gatox: we can do yours instead [13:51] ralsina, ack [13:51] ralsina, right. But we still need to do the 1-1 for the performance review. [13:51] ralsina, whenever you want [13:51] alecu: ohhhhh really [13:52] alecu: late today? [13:52] whee, new terminal is open at ATL now it seems [13:52] ralsina, ok, let me know. [13:52] * alecu hates ATL. And every other template library coming out of MS. [13:53] * alecu in fact hates old C++ compilers. [13:53] gatox: to the mumble! [13:53] okkkkkkkk [13:53] briancurtin: you are next because I missed yours too [13:54] alecu: uh, ok :) [13:54] ralsina: i'm ready whenever [13:54] ralsina: also the 1-1 review thing [13:54] dobey: yes, you too [13:54] gatox: 1 sec [13:54] ralsina, ack [13:58] dobey, I'm trying to push the code I have so far to lp:~ubuntuone-control-tower/ubuntuone-fsevents-daemon/trunk so that I have the initial code but apparently I'm not in the right group, can you do that for me? [13:58] mandel: where is the code at now? [13:59] dobey, I'd like to push the code form lp:~mandel/+junk/fsevents [13:59] dobey, also, I think I need to change the group to ubuntuone-control-tower, is that correct? [13:59] change the group? [14:00] briancurtin: to the mumbles! [14:00] dobey, right now this is what I've got setup: https://launchpad.net/ubuntuone-fsevents-daemon/trunk [14:00] dobey, I think is correct... [14:07] now it's more correct [14:07] :) [14:08] dobey, thx, and the branch, how can that be done? [14:10] mandel: i will do it [14:10] alecu: i am surprised that qtnetwork doesn't support proxy autoconfig. that's quite odd [14:10] dobey, great, let me know when do so I start adding bugs and proposing fixes :) [14:11] mandel: though, how will merges be handled? [14:12] mandel: can we compile it and run tests on linux? or what? [14:12] dobey, is a good question.. is mac only code (there are some tests to be run etc..) does tarmac work on mac? [14:13] i imagine it could if there was a way to legally run mac in a vm :) [14:13] though it would need to compile without the xcode gui and such as well [14:14] dobey, yes, that is some work I need to get done.. is one of the bugs I wanted to fix [14:14] dobey, ralsina, alecu, can we talk about that ^ after the team meeting? [14:14] mandel: sure [14:15] mandel: so let's figure this out first, before pushing it to a place where only tarmac and a couple people can push :) [14:15] mandel: we will probably do like we do with windows, and get a mac jenkins [14:15] dobey, qtnetwork supports proxy autoconfig only on windows and mac. "On other systems, there is no standardised method of obtaining the system proxy configuration" says the docs. http://doc.qt.nokia.com/4.7-snapshot/qnetworkproxyfactory.html [14:16] ralsina, and for tarmac? [14:16] dobey, ok, we talk about it and we try to find a solution [14:16] mandel: like for windows, tarmac doesn't care [14:16] alecu: but if we know the url, we can't feed it to qtnetwork? [14:17] dobey, the url of the proxy? [14:17] ralsina: well, we do merge code with tarmac for windows-installer. though it doesn't run any tests, just does the other stuff [14:17] alecu: the url to the proxy.pac file [14:17] mandel: we have tarmac and jenkins backwrds. Jenkins should pick up the branches, check that it passes on all platforms, then send to tarmac, which runs no tests and merge [14:17] dobey: ^ [14:17] dobey, oh. Qt does not support .pacs afaik [14:17] ralsina: i disagree :) [14:17] lol [14:18] alecu: no, we need to parse it ourselves [14:18] alecu: surely it does, if it does autoconfig [14:18] ralsina, dobey lets talk about it after the team meeting :) [14:18] alecu: there are several parsing libraries though [14:18] dobey, I think that windows itself resolves the pac files, and lets qt know about the proxy server to use for a given request. [14:18] dobey: I kinda expected you to disagree ;-) [14:19] ralsina, there's also this: http://blog.ivandemarino.me/2011/03/21/How-to-handle-Proxy-PAC-configuration-with-Qt [14:19] alecu: i'd guess not, since qt doesn't use IE on windows. but does include the webkit bits :) [14:21] alecu: that means adding QtWebkit to the installer (yeech) [14:21] alecu: OTOH http://code.google.com/p/pacparser/ [14:21] dobey, it does parse the pacs on windows anyway, according to: http://qt-project.org/doc/qt-4.8/qnetworkproxyfactory.html#systemProxyForQuery [14:22] dobey, it does not parse them on osx. [14:22] alecu: well, we don't support osx yet either :) [14:22] dobey, right, not just yet. [14:24] alecu: so, on windows we suport .pac already? [14:24] ralsina, it seems that pacparser includes the whole of spidermonkey. [14:24] alecu: haha [14:24] ralsina, we probably support it on windows. But I have not tried it at all! [14:25] ok, so .pacs can contain arbitrary JS. How screwed up is that? [14:25] we probably don't if we're pulling settings and shoving them into qt [14:25] dobey, we are pulling settings *only* on ubuntu. [14:25] alecu: OTOH, that's 400KB, webkit is like 10MB [14:25] alecu: are we using the systemProxyForQuery on win? [14:25] dobey, yes. [14:25] I think we are :-) [14:26] alecu: ok, so it should just work there then; and on linux we should be able to tell qt to read a pac file [14:26] so, yes we are: ./ubuntuone/proxy/tunnel_server.py: proxies = QNetworkProxyFactory.systemProxyForQuery(query) [14:26] cool [14:27] dobey, but qt is not (yet) able to read a pac file [14:28] ralsina, so, since pac files are so screwed up that they include javascript, I decided not to explicitly support them in this first iteration of proxies. [14:28] ralsina, if many people request them, then we can consider doing them. [14:29] ralsina, but right now only one person asked for them, and even then she was able to extract the proxy settings by reading the pac. [14:30] alecu: how can qt not be able to read a pac file? [14:30] alecu: surely, pac files work in qtwebkit browsers [14:31] alecu: pac files aren't so secrewed up that they include javascript; they are specified as being javascript. that is how PAC is defined [14:33] dobey: webkit is one thing, but we are not using webkit to access the network [14:33] ralsina: eh? why would anyone use webkit to access the network? qtwebkit uses qtnetwork to access the network [14:33] dobey, yes, I know that they are fully js. I've already used them. [14:34] dobey: yes, and webkit has a JS interpreter, which means it can parse .pac files and tell QtNetwork which proxy to use [14:34] ralsina: *qt* has a js interpreter. [14:35] also, libsoup can handle pac, or at least, it used to; not sure if it does with the new glib proxy bits [14:35] dobey: yes, it does, and it's inside QtWebkit: http://qt-project.org/doc/qt-4.8/modules.html [14:35] or is it QtScrit? [14:36] i think QtScript is what QtNetwork uses for the pac files [14:36] dobey, no [14:36] dobey, on windows, it uses the WinHTTP DLL. And that does the pac parsing. [14:37] dobey, that's why there's no pac support in qt on mac nor on linux. [14:37] alecu: interesting [14:37] ugh. what nonsense [14:37] dobey, ralsina: please read this bit of qt docs that I pasted several times already today: http://qt-project.org/doc/qt-4.8/qnetworkproxyfactory.html#systemProxyForQuery [14:37] alecu: reading :-) [14:38] aquarius: would be great if https://one.ubuntu.com/developer/music/stream_music/cloud contained the actual URL for the service [14:38] alecu: doesn't say no pac support on linux, though [14:38] alecu: but we aren't using that function on linux [14:38] alecu: and it says it supports the environment variable, which I thought we did not support? [14:39] ralsina: ^^ we are not using that method on linux :) [14:39] ralsina, it seems to be supported on 4.8. It surely was not on 4.7. [14:39] alecu: free feature! ;-) [14:39] ralsina: we're explicitly setting the settings from gsettings by hand [14:39] dobey: right, because this doesn't take them from gsettings [14:40] so maybe we should use this as a fallback [14:40] or something. My head hurts now. [14:40] ralsina, fallback sounds good. [14:40] well, we can also pull the url to the .pac file from gsettings, and pass it in to qt on linux [14:40] which should generally 'just work' [14:41] dobey: citation needed for that one [14:41] dobey: but maybe [14:41] * ralsina goes have some ibuprofen before team call [14:41] dobey, I don't think there's a way to "pass the .pac url to qt". At least I can't find it. [14:42] dobey, that's maybe because qt does not parse pacs. [14:43] morning folks. [14:43] so proxy support in qt is not as great as we originally thought [14:43] mandel thanks for pointing out the lint complaints - I think I need to fix my lint setup, I thought it was running and not complaining, but I guess it wasn't running... ? [14:43] dobey: not on ubuntu, at least [14:44] dobey, we never thought it would be great. Only that it's much better than any other alternative. [14:45] dobey, we always assumed that there's plenty of proxy setups and configurations, and that with our limited resources we could only support some. [14:46] mmcc, no worries i know gatox had some issues with it, he might be able to give you a hand [14:46] * dobey still thinks we should have just told people to fix their broken networks instead :) [14:46] mmcc, what is happening? [14:47] alecu: although, would have been good to know this last week, so we could point out all these flaws, in the proxy session :P [14:47] gatox, it looks like I wasn't actually running pylint when I thought I was. I'll look into it and ask you if I need help [14:48] mmcc: lint in what project? [14:48] mmcc, ok, let me know... in mac, for sso, pylint should run with the run-mac-tests script [14:48] dobey, in ubuntu-sso-client. [14:49] mmcc, did you update your environment with the latest changes in the doc? [14:49] mmcc: they should be run with ./run-tests; but maybe it has a check for pep8 and you don't have it installed, and that's what is failing? [14:50] gatox - the changes from this week? I thought those were for protobufs [14:50] dobey: pep8 is there on my mac but maybe not on my linux vm... I'm poking around now [14:51] mmcc, this week was just protobuf... but when i was getting the buildout to work on mac, i made a couple of changes, i don't know if you follow all of them..... it might be a reason [14:52] i hope the meeting will be quick today [14:52] gatox, ah OK. I'll check again but I'm pretty sure I have everything up to this week's changes. [14:53] mmcc, for example, you shouldn't have u1lint and u1trial in the bin folder as in some moment the document says [14:57] mmcc, mandel please remember to review this: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-os-helper/+merge/106014 when you have a moment :D thx [14:58] dobey: 30' as usual [14:59] 30 minutes or its free [15:00] so if you guys order a pizza now, we all have lunch after the call [15:00] OTOH, team call! [15:00] alecu, mandel: call [15:01] joshuahoover: desktop call if you are available [15:01] gatox: call! [15:01] joining... [15:01] oh, already there [15:01] ralsina, i'm already there [15:27] mandel: whenever you have time, https://code.launchpad.net/~brian.curtin/ubuntuone-windows-installer/build_installer/+merge/105403 is now updated with your comments [15:27] it works on python 2 and 3, which is the reason for that _winreg/winreg import dance [15:31] briancurtin, I know rename, much nicer without the _ [15:41] ugh python3 [15:41] ralsina, http://virtualmacosx.com/index.php/server-plans [15:41] ConfigParser -> configparser :-/ [15:42] alecu: it's something [15:43] ralsina, right: we should try to get our mini running as a server, but that looks like a nice fallback. [15:43] alecu: yeah. I suspect jenkins will destroy the "lite" one :-) [15:44] why is qt network api so hard to follow [15:47] dobey: Norwegian cunning. [15:47] mandel: lp:ubuntuone-fsevents-daemon pushed [15:47] * gatox lunch === gatox is now known as gatox_lunch [15:49] dobey, thx, really appreciated! [15:53] mandel: and proposed a branch to add config for it to tarmac [16:00] dobey, superb, let me finish a little bit of code and will take a look [16:04] ok, well i really need to get some lunch now. so bbiab :) [16:17] And lunch for me to [16:17] o [16:22] ralsina, joshuahoover: i emailed some details about the installer, but here it is http://ubuntuone.com/7cnc3EWLWTSblGPPPOV7DK [16:22] briancurtin: ah, nice, thanks! [16:25] briancurtin, ralsina: i filed the rt about doing a release on monday...in the mean time, i'll test this new installer and see if rmcbride can help test today/tomorrow as well [16:34] joshuahoover: briancurtin ralsina I can certainly get some installer testing in today/tomorrow [16:35] rmcbride: awesome [16:35] team, turns out I have to skip lunch and go to the bank. Depending on hunger I may take a bit longer to come back [16:36] ralsina: I learned as a child that coins are not very filling. Bear that in mind as you schedule [16:36] rmcbride: you have never seen a hungry bank, I guess. [16:36] hehe === gatox_lunch is now known as gatox [16:50] briancurtin, branch approved [16:50] dobey, I don't see the branch with the config for tarmac [16:50] mandel: thanks! === IAmNotThatGuy is now known as C0nfus3d [16:58] mmcc, gatox, ralsina: thinks will start happening from now on here: https://launchpad.net/ubuntuone-fsevents-daemon (related to the root daemon) [16:59] mandel, cool [17:01] ralsina, briancurtin: I've taken a look at device insertion/removal on windows. It seems that there are nice events for USB drives (and I assume esata too), but they don't work right for sd-like card readers. [17:01] I've found lots of pointers to the same, but the most relevant is: http://timgolden.me.uk/python/win32_how_do_i/detect-device-insertion.html [17:01] mmcc, added you to bug 1000875 [17:01] Launchpad bug 1000875 in Ubuntu One FsEvents daemon "Daemon does not run on 10.6" [Medium,Triaged] https://launchpad.net/bugs/1000875 [17:02] ok mandel, thanks [17:02] ralsina, briancurtin: for card readers the recommended solution is to poll using wmi. [17:02] mmcc, I'll be adding the script to run the tests from the command line tom morning to make things as simple as possible [17:03] now, is my time to go and get a bone broken [17:03] * mandel runs to rugby [17:03] alecu: is this for a new feature? i'm not familiar with where this stuff came from [17:03] bye mandel, enjoy [17:04] briancurtin, we are pondering how hard it would be to do "UDFs on removable volumes" [17:05] * alecu considers we should not worry too much about other kind of devices. Ie: floppies and cd/rw :-) [17:09] alecu: i'm about to leave for my half day, but i will take a look at that stuff. i'm familiar with using WMI for a lot of things and its usually easy to work with [17:11] briancurtin, no hurry, we can take a better look tomorrow AM [17:13] gatox, mmcc, before I go, please take a look at https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/add-run-tests/+merge/106227 whenever you have time [17:13] all, laters! [17:13] mandel, ok..... bye! [17:14] i'm also leaving for my half day. see everyone tomorrow [17:14] mandel, ok will do [17:15] gatox, thanks for the extra instructions and the new env-mac, it fixed my lint problem [17:15] mmcc, great! [17:15] gatox - also, I'm trying to run tests on darwin on your darwin-os-helper branch and it looks like it's trying to run some linux tests: http://paste.ubuntu.com/992712/ [17:16] mmcc, are you running the tests as i said in the merge proposal? [17:16] ./run-mac-tests tests/platform/os_helper/ [17:16] gatox: er... :) [17:16] :P [17:16] mmcc, the other tests (for the other packages) are not working yet [17:29] mandel: you were probably looking in the wrong place. but it's already merged [17:32] ralsina, alecu: do you know what CA bundle qt is using by default for ssl? [17:32] dobey, no [17:42] dobey, I'm taking a look at qt source. It's using openssl, and it does not seem to include any certificates. [17:43] dobey, so I assume it's using the system certificates. [17:45] hrmm === alecu is now known as alecu_lunch [18:08] gatox, +1 on your os-helper branch. sorry that took so long, had to tweak my test setup on two platforms :) [18:09] mmcc, no problem! thanks! [18:11] hey folks, I picked up a stomach flu or something yesterday, and it looks like it's not going away - so I'm going to try to rest and come back in a couple hours. [18:12] mmcc, oops... hope you get better [18:17] * dobey makes some 3.0.1 releases === alecu_lunch is now known as alecu [18:40] ralsina, rmcbride: i'm currently running the win release through the windows manual tests [18:40] as am I actually [18:44] what win release? [18:46] 3.0.1 candidate [18:48] joshuahoover, rmcbride: awesome *2? ;_) [18:48] dobey: the one we discussed on the call this morning :-) [18:48] ralsina: so far it's going quite well [18:51] ok, candidate [19:14] ralsina, rmcbride: i've run into a problem w/ this test: http://testcases.qa.ubuntu.com/Applications/UbuntuOne/Windows#Select_the_folders_to_synchronize [19:14] joshuahoover: ouch [19:14] joshuahoover: what happened? [19:14] ralsina, rmcbride: it won't setup once i click the next button, if i don't select a folder to sync, all works fine [19:15] ralsina: let me get some logs for you [19:15] Hmm [19:15] joshuahoover: is it a very big folder? [19:15] ralsina: it's empty [19:15] joshuahoover: that's bad [19:16] yeah, let me try once more (3rd time is always a charm) and then get you logs...one moment [19:18] school run [19:18] joshuahoover: if it fails, send me the logs I will try to do a quick fix since brian is gone for today [19:20] rmcbride: can you try that test case too? http://testcases.qa.ubuntu.com/Applications/UbuntuOne/Windows#Select_the_folders_to_synchronize [19:21] as a sanity check [19:21] joshuahoover: yea, setting up to do that now. didn't get presented with that withthe upgrade path I tested first [19:29] ralsina: syncdaemon.log from my failed test - http://paste.ubuntu.com/992924/ [19:35] whups. WebClientError u'SSL handshake fail' u'' [19:35] while removing device [19:36] rmcbride: really? can you see what your sso.log has in it? [19:36] joshuahoover: checkking [19:37] joshuahoover: not seeing one on my VM. what's the path again? [19:37] stupid windows and its stupid different paths for diferent versions [19:37] joshuahoover: nm found it [19:38] rmcbride: in your home folder: AppData\Local\xdg\cache [19:38] heh [19:38] 2012-05-17 15:35:09,404:404.000043869 - ubuntu_sso.credentials - INFO - find_credentials: self.app_name u'Ubuntu One', result is {}? False [19:38] 2012-05-17 15:35:09,404:404.000043869 - ubuntu_sso.main - INFO - CredentialsManagement: emitting CredentialsFound with app_name u'Ubuntu One'. [19:39] 2012-05-17 15:35:09,457:457.000017166 - ubuntu_sso.utils.ipc - INFO - Emitting remote signal for CredentialsFound with callback . [19:39] joshuahoover: ^ [19:40] rmcbride: ooh, you got ssl error? awesome [19:41] dobey: heh...isn't it sad that we both got excited about that? ;) [19:41] it's windows though [19:41] yeah [19:41] yea [19:41] we've seen it on both, windows and ubuntu [19:41] first time I've been able to repro that one I think [19:41] on windows anyhow [19:42] though maybe qt isn't loading the system certs on windows [19:42] just like it isn't on ubuntu, apparently [19:42] whee [19:42] reverting image to last september and retrying [19:43] dobey: eh? It is loading them in ubuntu, or else everyone would see the failure [19:45] ralsina: not according to strace it isn't [19:45] ugh. vm image fail probably [19:45] dobey: weird. Are you tracing all threads? [19:45] as far as the revert goes. THis might take a while [19:46] ralsina: i did strace -Ff yeah [19:46] FWIW I do not get the SSL fail on my physical machine [19:46] dobey: it has EMBEDDED CERTIFICATES [19:47] altough [19:47] dobey: run strings on libQtNetwork.so [19:47] ralsina: right [19:47] after removing and trying to sing in, I get ZipImportError [19:47] dobey: which is nuts [19:47] ralsina: and it's loading 1 certificate file out of /etc/ssl/certs/ [19:47] ralsina: did that log tell you anything? http://paste.ubuntu.com/992924/ [19:47] ralsina: which on my system happens to be a file named ValiCert...pem [19:47] joshuahoover: looking... [19:47] dobey: whoa [19:47] ralsina: so, basically, magic. [19:48] yea. can't sign back in after removing the device [19:48] lrwxrwxrwx 1 root root 23 Feb 8 09:35 /etc/ssl/certs/55a10908.0 -> ValiCert_Class_2_VA.pem [19:48] well, it's loading that symlink, which points to the ValiCert [19:48] dobey: was about to ask you if it was a symlink like the others :-) [19:48] clicking on signin gives me a ZipImportError on ubuntuone\dist\library.zip [19:48] joshuahoover: nothing special there, it seems [19:49] dobey: may be noticing it's newer than its copy? [19:49] i guess i should tell is to close the rt [19:49] ralsina: i think qt itself might not include that one? [19:49] dobey: yes, but the ubuntu users should have it (and they do have it) and it loads it in your system [19:50] joshuahoover: second look, that seems to say you are not authenticated? [19:50] ralsina: yes. well, haven't explained why exactly it's failing for some [19:50] joshuahoover: around line 174 [19:51] ralsina: but that's after i log in...i'm following this test case: http://testcases.qa.ubuntu.com/Applications/UbuntuOne/Windows#Select_the_folders_to_synchronize and i never get past the part where i click "next" after selecting the folder [19:52] joshuahoover: which is why "No user With Network" is surprising [19:54] ralsina: HAH! [19:54] The issuer certificate of a locally looked up certificate could not be found (11) [19:54] Yea I can't do a fresh install at all with this installer, only upgrade [19:54] rmcbride: yikes [19:55] I get teh ZipImportError [19:55] ralsina: yikes indeed [19:55] especialyl yikes that the upgrade "worked" [19:55] ok, looks like I have to use windows tomorrow the [19:55] n [19:57] where did this file come from [19:59] joshuahoover: i think i have a workaround for the ssl cert issue! [19:59] dobey: share? [19:59] dobey: really? excellent! [20:00] sudo ln -s /etc/ssl/certs/ValiCert_Class_2_VA.pem /etc/ssl/certs/55a10908.0 [20:01] also, i have no idea how language would matter [20:01] dobey: yeah, me either...i just passed along what the user told us he did that worked for him :) [20:02] and i manage to get an AuthenticationError as well as a WebClientError [20:03] but maybe the former is a result of the latter [20:05] creating the symlink seems to work though [20:05] but i have no idea where it originally came from [20:05] dobey: hmmm... maybe it's doing an uppercasing of the filenames to find VALICERT in it? [20:05] dobey: that would fail in turkish :-) [20:06] ralsina: no, it seems to be looking for a file named by the cert's id [20:06] dobey: I have both [20:06] ralsina: yes, but qt is loading the /etc/ssl/certs/55a10908.0 one [20:06] dobey: interesting [20:06] eod for me people..... tomorrow i'll keep trying to kill the filesystem :P see you! [20:06] ralsina: if you rm it, and try to log in, the error happens [20:06] dobey: awesome, great job [20:07] * ralsina tabulates... about 40 dev-hours spent on this crap [20:07] would be awesome if i could tell where that file was coming from, though [20:07] that's a couple grand [20:07] ca-certificates doesn't own it, or create it during postinst, nor does update-ca-certificates create it when run [20:08] ralsina: if i'm filing a bug against the win wizard, which project do i file it against? [20:08] joshuahoover: ubuntuone-control-panel [20:08] dobey: firefox? [20:08] thanks [20:09] ralsina: no, firefox has its own db, it doesn't use system certs [20:10] dobey: /etc/ssl/certs id full of symlinks to firefox in my system [20:10] ralsina: the word "mozilla" does not mean they are *in* firefox [20:10] dobey: ok [20:10] ralsina: it means those certs were extracted from the mozilla db [20:11] however, mozilla/firefox still has its own db [20:30] ralsina: bug #1000970 from testing win 3.0.1 [20:30] Launchpad bug 1000970 in Ubuntu One Control Panel "Windows 3.0.1 - AttributeError: "'NoneType' object has no attribute 'stop'"" [Medium,Triaged] https://launchpad.net/bugs/1000970 [20:30] rmcbride: maybe you can try to reproduce? ^^ [20:30] joshuahoover: oh, interesting [20:32] joshuahoover: I have some trouble here. One system won't complete an install, and the other is a VM image that has issues. give me a few to resolce [20:32] but that's an intersting/pythony error [20:32] rmcbride: ah, ok [20:38] joshuahoover: that bug you reported looks easy to fix, at least [20:38] ralsina: that's good :) [20:39] ralsina: i'll file another one for the problem i ran into (which may be somewhat related, seems to be something with udfs in the wizard) [20:41] man i love when people comment on bugs about some completely unrelated problem they're having, with absolutely no info [20:45] dobey: let me guess, the turkish bug's last comment? === salgado is now known as salgado-afk [20:46] of course [20:54] ralsina: tarball for u1-client 3.0.1 is uploaded to launchpad, and the branch is tagged. will try to get the other projects all done asap. shouldn't be too bad, just tedious, as there aren't any changes elsewhere really [20:54] ralsina, rmcbride: all but 2 tests passed here for me: http://testcases.qa.ubuntu.com/Applications/UbuntuOne/Windows after 2 test runs and 3+ on the 2 failed tests... [20:54] ralsina, rmcbride: failed tests were: http://testcases.qa.ubuntu.com/Applications/UbuntuOne/Windows#Select_the_folders_to_synchronize & http://testcases.qa.ubuntu.com/Applications/UbuntuOne/Windows#Sign_in_after_removing_the_device [20:55] ralsina, rmcbride: and the bugs i filed were: #1000983 and #1000970 [20:56] joshuahoover: awesome [20:56] bug #1000983 and bug #1000970 [20:56] Launchpad bug 1000983 in Ubuntu One Control Panel "Windows 3.0.1 - Adding folder to sync in Wizard prevents setup from completing" [High,Confirmed] https://launchpad.net/bugs/1000983 [20:56] Launchpad bug 1000970 in Ubuntu One Control Panel "Windows 3.0.1 - AttributeError: "'NoneType' object has no attribute 'stop'"" [Medium,Confirmed] https://launchpad.net/bugs/1000970 [20:56] dobey: I did some changes in u1cp [20:56] which changes? [20:57] dobey: simplified bandwidth dialog, fixed it for RTL locales [20:57] dobey: don't recall any others but maybe something small [20:57] there's the duplicate instances branch [20:58] i guess there are other things we need to backport? [20:58] duplicate instances? really, that was not in 3.0.0? [20:58] dobey: that was on the last SRU, IIRC [20:58] some of which probably need approvals [20:58] ralsina: nope. it was right after 3.0.0; we didn't do an sru for it yet [20:58] dobey: ack [20:58] it's not a critical fix [20:59] well, it was done by request [20:59] right, there is a bug for it [20:59] i guess we need to decide which other fixes we want in 3.0.1 for cp then [21:00] dobey: ok, we can do that first thing tomorrow morning, because I am close to EOD [21:00] ok [21:01] and looks like a few are mis-triaged [21:06] dobey: the LTR one may go in, it changes no strings, and only fixes an obvius bug [21:06] dobey: the cleanup of throttling changes strings [21:06] dobey: same about the one setting units for translation [21:06] so maybe just the duplicated instances, and whatever I fix that failed the manual tests today [21:07] but that just or the windows release [21:08] no it's not [21:09] there is no more "just for foo release" [21:13] joshuahoover: looks like my particular install problem is due to the system needing to reboot to complete previous install. I can't do that right now so that particular system is unavailable for a bit. checking on the image now [21:13] rmcbride: got ya...what vm software do you use? [21:14] vmware. the system that I can't reboot right now is a physical system though [21:14] rmcbride: i've been using vbox for testing for the last 3 years and have had good luck with it...which means it will start to crash and burn immediately now that i've typed that [21:14] joshuahoover: I reverted the vmware image but it's been an issue [21:14] weird [21:15] joshuahoover: yea I use KVM for my linux boxes. VMWare does very nice automated installs and such (which it looks like I'm going to have to do again) [21:15] joshuahoover: if I'm reading the bug reports you gave right, I expect we're looking at at least one more candidate build yes? [21:16] rmcbride: yeah, i haven't seen if that happens in the current 3.0 build but i think we need to fix at least the one that is marked high, the other is slightly less likely to happen but should be fixed as well (they seem related to udfs anyway) [21:17] joshuahoover: OK. I want to do another pass on install, but I'll wait on the new build. THat'll let me get a non broken vm image going for tomorrow [21:19] dobey: you are right, thanks for reminding me to stay in the straight and narrow [21:20] rmcbride: sounds good [21:30] bad ralsina [21:33] branch which fixes a bug, with the bug number in the branch name, and in the commit message, but no commit --fixes= :-/ [21:52] ok, have a good evening all [22:53] dobey: really? I tryto do the --fixes first thing when starting the branch :-/ Sorry