[09:32] Happy Hallowe'en! >:-D [09:32] JamesTait: and morning to you too! [09:34] rye_: Good weekend? [09:35] JamesTait: yep, guests, and Windows. And weird Ubuntu One bugs in Windows :) [09:36] rye_: I love the smell of frustration on a Monday morning. :-P [09:37] JamesTait: well, suddenly Ubuntu One found "My Pictures", "My Videos" in Documents folder, created them on the online storage, upon restart it ends up in infinite loop in local rescan by faililng to scan these directories which do not exist [09:38] So I am really frustrated [09:38] rye_: Interesting. So what are they then, virtual directories or something? [09:39] JamesTait: nothing, bug #883926 [09:39] Launchpad bug 883926 in ubuntuone-client "[windows] Adding watch to nonexistent directory (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/883926 [09:41] rye_: Nice tags on that bug. ;) [09:42] JamesTait: the u1-zomg-windows is being added by nessita to the bugs, so I guess they are tracking them [09:44] Also, a good thing to do before doing such jumps is to actually copy the SSH keys over to some external location. Now I need to swap the hard drives to get them [10:04] hi [10:04] mornings, gatox [10:04] rye, hi! [12:05] brb!! [12:16] hello! [12:17] alecu, hi [12:26] gatox, ping [12:26] mandel, pong [12:27] gatox, morning! [12:27] mandel, morning! how are you? [12:27] gatox, so, I just wanted to let you know that I'll be starting work a little later :) [12:27] mandel, ahhhhh okok.... so me too :P [12:27] gatox, 'cause I'mworking with alecu and dont want to make hime wake up to early :P [12:28] mandel, ack! [12:29] * mandel errands [13:10] gatox, does tarmac work? [13:15] mandel, no idea [13:15] gatox, I have a branch that has been accepted and not merged.. [13:16] mandel, i didn't try to merge anything today yet.... [13:17] mandel: url? [13:17] dobey, https://code.launchpad.net/~mandel/ubuntuone-control-panel/fix-tests/+merge/80587 [13:18] hmm [13:20] mandel: i think it's down [13:21] dobey, ok, no worries, I just wanted to know :) [13:22] dobey, how is the keynote going? [13:22] dobey, I've been told we have been mentioned :P [13:23] mandel: yeah, u1 was mentioned a few times [13:24] \o/ [13:28] mandel, alecu dobey standup in 2'? [13:29] no [13:30] gatox, should it me in an hour? [13:30] you guys can [13:30] dobey, true, uds for you [13:30] mandel, nok by me [13:30] ok [13:30] mandel, poke [13:30] mandel, bug #883926 [13:30] Launchpad bug 883926 in ubuntuone-client "[windows] Adding watch to nonexistent directory (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/883926 [13:30] me is a bit puzzled [13:31] rye, uh, nice title.. let me see [13:32] rye, ok, which version of the installer is that? [13:32] rye, sorry of the port [13:32] mandel, current latest [13:32] mandel, installed this saturday [13:32] * mandel thinks we should also provide the rev number of the client rather than a random number.. [13:33] rye, hm.. funny, 'cause those logs look that they do not folow the new logging, let me check [13:33] mandel: eh? [13:33] mandel: what do you mean about revno? [13:34] mandel, i don't have the logs at the point when it noticed these phantom directories, since the local rescan logs pushed these out, in order to get more info I would need to reboot into windows, as this is a real world installation [13:34] dobey, the windows client has a version, but has nothing to do to with the revno of the source code. When we have an issue there is not way to know in which revno we released the installer [13:35] dobey, we don't even have a spreadsheet that maps revno and releases.. and is a PITA and really bad practice [13:35] mandel: tag your releases! [13:35] dobey, we should, but our release process at the moment is crap, is from a branch from ralsina that has not yet been merge to the lp:ubuntone-windows-installer [13:36] mandel: but the alignment of windows/linux releases will help with that [13:36] dobey, hopefully it will [13:36] mandel: if it's not merged, it hasn't got a revno [13:36] :) [13:37] dobey, that would be for the installer, but the u1-client does :) [13:38] dobey, anyway, we are doing a shit job with the windows releases [13:38] mandel: and we will fix it [13:38] rye, I'd need more info from the start of sd to see why is thinking those paths exist.. [13:39] rye, can you add that to the bug report? I need to go for lunch now :) [13:39] mandel, https://bugs.launchpad.net/ubuntuone-client/+bug/883926/comments/2 ? [13:39] Launchpad bug 883926 in ubuntuone-client "[windows] Adding watch to nonexistent directory (affects: 1) (heat: 6)" [Undecided,New] [13:41] rye, ok, I think that bug is fixed and it was related to a bad implementation of a method on the windows file system watcher, look here: http://bazaar.launchpad.net/~ubuntuone-control-tower/ubuntuone-client/trunk/view/head:/ubuntuone/platform/windows/filesystem_notifications.py [13:41] rye, the logging in your machine does not match with the one present in trunk [13:41] mandel, it is the latest available .exe from the web site, so this version is in the wild [13:42] mandel, so, if that is fixed in trunk, it will still take some time to reach people. What I am worried is whether the next version will work properly with the data collected by broken one [13:42] rye, we shoould be making a release with the fixes this week that wills solve it [13:42] mandel, currently the sync is infinite loop for some reason [13:43] in local rescan i mean [13:43] rye, hm.. can you give me 30 min for lunch and we look into it? [13:44] mandel, sure! [13:45] rye, ok, I'll ping you when Im back :) [14:01] joshuahoover, bug #745540 - this is something wrong with keyring, not Ubuntu One. We are displaying the wrong bug [14:01] Launchpad bug 745540 in ubuntuone-client (Ubuntu) (and 5 other projects) "Method "CreateItem" with signature "a{sv}(oayay)b" on interface "org.freedesktop.Secret.Collection" doesn't exist (affects: 63) (dups: 2) (heat: 316)" [Undecided,Invalid] https://launchpad.net/bugs/745540 [14:01] i.e. we are not checking the exception and cause another one [14:02] rye: do we know what the cause is? because i can't get any users impacted by it past it [14:02] oh fun [14:03] rye: are you sure? [14:04] dobey, we switch to this old signature on ANY error from dbus, and cause the users to be notified about the signature error === jdobrien is now known as webm0nk3y [14:05] hmm [14:08] rye: sounds like maybe we need to file a new bug for this? [14:08] i just can't get it to fail to get the details of the error to trap it correctly [14:10] or can i... [14:10] let me break my keyring [14:14] rye, ping [14:17] mandel, pong [14:19] rye, so, getting back to the issue, I think the best thing would be to do an IRL tests with your account and trunk and see what happens, but I believe there should be no problems [14:19] rye, also, does the shell show that you have a My Music library? [14:19] no cmd, but the shell (explorer) [14:19] mandel, yes, i do, on a level upper than My Documents [14:20] which is also a view [14:20] joshuahoover, do users describe that Ubuntu One control panel simply grays out after they click that they have an account? [14:21] rye, hm.. I hope that we are not getting that in the os.listdir for somet stupid stupid reason.. [14:21] rye, let me know what you ddi with the folder and I'll do the same in my test vm.. I hope python is not THAT broken on windows 7 [14:22] rye: i've had one user mention that, yes [14:22] joshuahoover, I managed to get this now === m_conley_away is now known as m_conley [14:22] joshuahoover, I dismissed the keyring prompt after I moved my login keyring to another name [14:22] mandel, so, I had a real UDF folder called "Documents". I installed Ubuntu One on Windows, added it to my real world account. The files in Ubuntu One syncded perfectly [14:23] mandel, I wanted to test UDF so I subscribed to Documents UDF in control panel. It asked whether I'd like to merge the files, so I did that. [14:24] mandel, the directory at that time contained one text file and one directory, nothing was syncing that dir [14:25] mandel, the UDF started populating with files locally, as expected. I waited until it finshes syncing and was pleased with the performance. [14:25] rye, and then the My Music was left out? [14:26] rye, 'cause the issue is that we try to add a watch yo my music and then it creates it in the cloud, right? [14:26] mandel, then I decided to put a bunch of files to \Ubuntu One folder. It claimed that it got all files, but due to the bug with I/O on closed files I had to shut down syncdaemon and control panel and start it again [14:26] mandel, after that SD went into infinite loop [14:27] mandel, there were no "My Documents", "My Videos" or "My Pictures" directories or views in ~/Documents folder at any time during this [14:27] mandel, yet, SD managed to create the folders online, and now it iterates over them, failing, and iterating again in local rescan. [14:28] mandel, so there are 2 bugs as facundo noticed - one - the folder was added which does not exist [14:28] mandel, 2nd - local rescan breaks when it adds a watch on the folder tha does not exist with "Permission Denied" [14:30] rye, ok, the permission denied is because the windows api is plain stupid and returns that instead of file not found.. so that should be a simple fix, the second one is more complicated, I fear that at some point a shell function is used to get the dirs and returns library (which are not real paths) [14:30] rye, lets go step by step, can ou create a bug for the first one, and we will talk with nessita about how should fix it (but I'm sure is me :P ) [14:31] mandel: not necessarily :-) [14:31] rye, we create a second bug for the library issue and we talk with nessita again, mainly for the resources allocation, that is all [14:31] nessita, hehe [14:31] mandel, rye: with UDS going on this week, expect some delay on bug fixing. alecu and mandel need to move on on the proxy support feature if we want to make it in time for feature freeze [14:32] rye: but yes, bug reports and pinging next week is the best way to proceed :-) [14:32] nessita, well, I will not survive for that long in Windows [14:33] rye: why not? [14:33] nessita, it is scary over there [14:33] nessita, and I had no SSH keys with me [14:33] rye, there is a fix for the is_path_dir method which should aboid the issue nevertheless [14:34] rye, we should test that before we report anything since it has already landed in trunk [14:34] nessita, i'd like to reproduce the bugs but I haven't deployed the python environment yet [14:34] rye: right... is not so hard, the wiki page is pretty accurate [14:35] mandel: that fix is in trunk already isn't? [14:36] nessita, yes it is, so I think we should first test that that does not work as expected [14:36] mandel, well, I will grab all the keys I need and will swap the hard drives (poor-man bootloader switch)... === CardinalFang_ is now known as CardinalFang [14:39] joshuahoover, regarding bug #745540 [14:39] Launchpad bug 745540 in ubuntuone-client (Ubuntu) (and 5 other projects) "Method "CreateItem" with signature "a{sv}(oayay)b" on interface "org.freedesktop.Secret.Collection" doesn't exist (affects: 63) (dups: 2) (heat: 316)" [Undecided,Invalid] https://launchpad.net/bugs/745540 [14:41] joshuahoover, you mention trying numerous configurations. Have you tried setting the user to be the one that's used in autologin? [14:41] joshuahoover, probably setting it during configuration may make a difference too. [14:41] alecu: that is one i have not tried yet...though it sounds like rye thinks we're showing that error incorrectly [14:42] joshuahoover, oh, ok. [14:43] alecu: hi there! ralsina hasn't come yet, so as soon as he gets here, I will try to setup a proxy meeting time [14:43] alecu: were you able to build the poll? [14:44] do we have any stat from that? [14:45] * joshuahoover funny that proxy should be mentioned...just replied to a support request for a user behind a proxy server [14:45] nessita, no, I didn't do any poll. I was planning on doing it today [14:46] joshuahoover, do you have any specific details on the users requesting proxy support? [14:46] joshuahoover, like proxy types or something? [14:46] alecu: most have no idea [14:46] alecu: ack, please make that your priority so we can have the feedback as soon as possible, so we can plan the work for you and mandel to "start" [14:46] alecu: they're just behind "a proxy" that their company, university, etc. have [14:47] gatox: any ETA on having https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/tests-broken/+merge/80596 fixed? [14:47] gatox: that may block the rest of your ussoc branches since it fixes the tets suite :-) [14:48] nessita, mandel is helping me with an issue that i'm having with 2 tests suites, that are making twisted fail in an assert [14:48] joshuahoover, right. I guess most users would not know what type of proxy they are using; but perhaps they might be able to tell if it's a Microsoft Proxy or something else? [14:48] gatox: which test suite? [14:48] nessita, I think I know th fix, will try it in a sec :) [14:49] gatox: the error "SyntaxError: 'return' with argument inside generator" involves just changing the return for a yield, is that blocking you? [14:49] nessita, CredentialsManagementTestCase and SSOLoginTestCase, the previous tearDown has a return, and now with the yield i changed the return to defer.returnValue [14:49] gatox: show me the trace, please [14:49] alecu: i'm asking those users now as they come up to see if we can get more info on that [14:50] nessita, http://paste.ubuntu.com/724414/ [14:50] joshuahoover, I was planning on making a small survey to ask info about that... [14:51] gatox: and what did you change to have that error? [14:51] joshuahoover, can I send you the link in a few minutes, and you can tell me if it's too complex? [14:51] alecu: that would be good to, duanedesign could post that to our facebook page and twitter account [14:51] alecu: sure [14:51] joshuahoover, great, thanks. [14:52] https://blueprints.launchpad.net/ubuntu/+spec/desktop-p-u1-couchdb-sync [14:52] gatox: seems like you're returning the defer as a defer result instead of yielding in it [14:53] gatox: confirmed, you probably changed: [14:53] 577 return defer.gatherResults([self._tearDownServer(), [14:53] 578 self._tearDownClient()]) [14:53] with defer.returnValue, and that is worng [14:53] wrong* [14:53] nessita, yes, that's that mandel told me [14:53] mandel: that is wrong! :-) [14:54] gatox: but I need you to think this a little bit :-) returning a deferred is the same as yielding on the defer within inlineCallbacks [14:54] nessita, no, mandel told me to add the: gatherResults [14:54] gatox: nopes, just change the return by yield [14:54] ein? what did I say? [14:54] so, the final line of that tearDown should be: [14:55] gatox, nessita I don't recall saying using the gather resuls.. :P [14:55] 577 yield defer.gatherResults([self._tearDownServer(), [14:55] 578 self._tearDownClient()]) [14:55] I said this: http://paste.ubuntu.com/724415/ [14:55] mandel, maybe i got confuse [14:56] gatox, hehe no worries :) [14:56] gatox: the solution that mandel is showing is the correct one [14:56] mandel: sorry for boofing before [14:57] mandel, yes, that do the trick [14:57] THANKS [14:57] nessita, I was more worried about looking bad infront of you that anything else ;) [14:57] gatox, cool, I'm glad it worked :) [15:01] mandel: for a monet I got really scared that you were saying the wrong stuff :-) [15:01] moment* [15:01] nessita, there was a problem with the communication, that is all :) [15:01] nessita, no, was my bad, trying to explain what mandel was telling me [15:01] gatox: do you understand why defer.returnValue was worong and why yield works? [15:02] nessita, yes, mandel explains me and show me the doc [15:02] and the source code :) [15:03] yes, the source code actually :P [15:07] gatox: I think is very important that you see the difference, is key to understanding deferreds, and tetsing with [15:07] deferreds [15:07] nessita, yes, it's more clear now [15:08] gatox: great! I'm happy to explain more next week [15:08] nessita, are you going to be here next week? [15:08] gatox: yes! we can get together and do some IRL catchup [15:09] nessita, yes, i was thinking about that..... and do a "ping pong de preguntas y respuestas" :P to understand better some things [15:09] gatox: sure [15:09] gatox: wednesday may be a good day [15:09] gatox: or thrusday [15:10] nessita, no problem for me [15:16] ok, going offline for a while === eu is now known as Guest50440 [16:12] * mandel needs coffee to think === beuno is now known as beuno-lunch [16:35] alecu: GREAT work on the survey, thanks! loks really good and clear [16:38] nessita, we went live a few minutes ago when duanedesign posted the link on twitter, and we already have almost 40 respondents [16:38] alecu: that's great! :-) [16:38] alecu: I'm looking forward to see the summaries (I guess we need to wait at least until tomorrow) [16:44] ok, heading for lunch soon, going offline [17:43] alecu, ping [17:43] gatox, pong [17:45] alecu, i'm having an issue with a dirty reactor..... we were looking at the code with mandel, but we are pretty stuck [17:45] alecu, mandel suggestions help me to fix some of them, but i still have 1 dirty reactor error [17:45] tell me! [17:47] alecu, i'm working in a branch that fix the tests after the change in dev-tools, i added "yield" to the tearDown of CredentialsManagementTestCase and SSOLoginTestCase..... but those methods had a return, so we change that to gatherResults, and the errors with the dirty reactor appear [17:48] this is the trace i'm having now: http://paste.ubuntu.com/724537/ [17:48] and this is the branch: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/tests-broken/+merge/80596 === beuno-lunch is now known as beuno === lifeless_ is now known as lifeless === yofel_ is now known as yofel [18:31] alecu, everytime that i execute the test i have different failures :S} [18:33] gatox, ugly! [18:35] alecu, too much [18:36] gatox, alecu: can I help? [18:37] nessita, i'm testing different possibilities to check what's going on... but it's kind of tricky [18:38] gatox: can I have a little more context? is this windows? the unicode issue? are you writing tests or still testing IRL? [18:39] nessita, no, this is the branch that fix sso tests [18:39] nessita, and the error is only in windwos [18:39] gatox: still failing? have a race I can look at? [18:39] trace* [18:40] nessita, yes, let me give you [18:41] that [18:42] gatox: that? [18:43] the trace [18:43] nessita, i'm running the test again [18:43] ah [18:44] nessita, http://paste.ubuntu.com/724593/ [18:45] gatox, I got no errors on oneiric, even after enabling and updating nightlies [18:45] alecu: that's windows code... the one that has the "dummy" factory in it [18:45] alecu, running the sso tests? [18:45] (I found out that I was missing the nightlies when lint started throwing import errors for ubuntuone-dev-tools) [18:46] gatox, ./run-tests [18:46] gatox: is the branch updates in LP? I can branch and see [18:46] updated* [18:46] nessita, yes, it has some commented lines, i didn't clean it yet [18:46] ack [18:48] gatox: did you try just changing the minimal stuff? [18:48] gatox: I see a lot more code than expected [18:49] nessita, that branch contains the fix for all the tests, now only credentialsManagementTestCase and SSoLoginTestCase are failing [18:49] gatox, both "run-tests" and "run-test -qt" work fine on my fully-updated oneiric [18:49] gatox: yes, did you try adding the minimal fix to main/tests/test_windows.py? [18:50] gatox: on top of the other fixes [18:50] nessita, sorry..... what do you mean with "the minimal fix"? [18:51] gatox: well, in the branch you have much more that the minimal. You have some addCleanups, and some commented out code [18:53] nessita, yes the cleanup it's because, some cleaning was made in the tearDown, and that cause other failures.... i'll clean that code and look where to locate the cleaning code... [18:53] gatox: the addCleanup is not needed... as far as I can see [18:54] gatox: I have a diff that is working, let me pushit [18:55] nessita, the cleanup was to do the same that the tearDown was executing, but without executing that in the tearDown because that was presenting somme oother twisted issues [18:55] gatox: I guess you had some other issue [18:55] nessita, probably... [18:55] gatox: the tearDown as it was originally should stay the same, adding yields properly [18:55] gatox: this is all green for me on windows: https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/review_tests-broken [18:56] * gatox checking... [18:56] the tearDown with yield as we talk earlier was not working for mme and mandel [18:56] i'll try that [19:01] gatox: try the branch I emailed you [19:01] gatox: if it fails, change nothing and show me the trace [19:01] nessita, ok [19:01] i was trying the one you told me here..... [19:01] but fails [19:01] let me branch [19:03] nessita, where did you send it?? === ejat- is now known as ejat [19:03] (02:55:49 PM) nessita: gatox: this is all green for me on windows: https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/review_tests-broken [19:03] gatox: ^C [19:04] gatox: ^ [19:04] ahhhh...... that one..... i thought you send me another one to my e-mail [19:06] nessita, no, is not working..... maybe something is wrong here..... i'll show you the paste [19:06] gatox: ack [19:07] nessita, and now it work...... AHHHH..... every time i run the tests i have: PASSED, or different numbers of failures [19:07] gatox: well, make it fail once, and show me the trace [19:09] gatox: I need to go offline for a minutes [19:24] gatox: ok, shoot [19:26] nessita, http://paste.ubuntu.com/724619/ [19:26] nessita, i'm running them again.. to see what happend [19:26] gatox: are you 100% sure you're running my branch, with ni changes? [19:26] no* [19:27] nessita, yes, i just branch it when you tell me [19:28] nessita, i think that there is some timing issue.... i run the test again, and now i'm getting failures in tcpactivation only [19:29] gatox: what do you mean "tcpactivation' only? [19:30] nessita, that now, CredentialsManagementTestCase and SSoLoginTestCase seems to be fine, and tcpactivaction is failing... but in the paste that i gave you before, tcpactivation was fine, and Cred y SsoLogin were failing.... [19:31] gatox: what is "tcpactivaction is failing"? what kind of error you have? [19:31] nessita, and now i run it again, and show me 11 errors instead of 2..... the errors are always reactor unclean [19:35] nessita, are the tests working for you in windows always?? did you try to execute them several times to see if they fail? [19:35] gatox: yes, no failure [19:35] gatox: may be faster hardware in my case [19:36] nessita, probably... my machine is quite slow at this moment.... [19:36] (in the process of buying something new...) [19:37] nessita, now all the tests pass :P [19:37] without any change [19:39] gatox: so, stop working on this, and keep going with the unicode issue [19:39] gatox: I'll debug when have a moment [19:39] gatox: there are also some nasty code duplication that we should fix [19:39] nessita, ok... [19:39] gatox: so I'll propose a branch basing my work in yous [19:40] nessita, let me know if i can help in anything..... at least testing it in this hardware [19:40] gatox: sure! thanks [19:41] nessita, thank you [19:41] nessita, i'll go back to unicode issues now [19:41] ye,s good luck! [19:47] brb..... i'll take a little break === m_conley is now known as m_conley_away