[04:09] i guess a web ui for uploading multiple files is filed under 'nice to have'? [08:07] rye: good morning [08:32] morning everyone! [08:55] morning# [08:57] mandel: what do I have to do to setup storageprotocol [08:57] I was trying to setup the client last night on my windows VM and ralsina_ didnt have the link handy [08:58] fagan: is in lp, let me get the exact project name for you [08:58] cool [08:58] fagan: there you go : https://launchpad.net/ubuntuone-storage-protocol [08:58] ah ok so just grab the branch and setup install thats [08:58] ok [08:58] fagan: I need you to test a branch on a windows machine for sd, so let me know when your system is ready :) [08:59] sure ill ping after I get that branch [09:02] good morning! [09:03] morning karni [09:03] \o === rodrigo__ is now known as rodrigo_ [09:10] mandel: im missing protobuf-compiler could you give me a link for it [09:11] fagan: do a google search for protobuf and wil return the exact page you need [09:11] I found a code.google.com page [09:11] you have to download the binary compiled version for windows [09:12] but it gave me out a .exe for doing c++ programs there was no installer [09:12] well there is no installer, you have to place it in a location in your path where the setup.py can find it [09:13] like the Programs folder [09:13] ah ok [09:13] I'm guessing you found: http://code.google.com/p/protobuf/downloads/detail?name=protoc-2.4.0-win32.zip&can=2&q= [09:13] right? [09:13] yeah [09:15] mandel: I put it in program files but its still not finding it [09:15] :/ [09:17] fagan: try adding it to the PATH var in the env variables of your system [09:17] if not, place it next to the setup.py, is should find it there [09:17] mandel: ok ill give it a go [09:21] worked after I set the path [09:25] mandel: ok so now where do I find contrib.testing.testcase [09:26] what is the proble? [09:26] fagan: ^ [09:26] its aparently not there [09:26] 'cause it is in the ubuntuone-client trunk [09:26] contrib/testing/testcase.py [09:26] u1trail should be loading the packages from there [09:26] mandel: im using ubuntuone-client but its the branch that was proposed yesterday [09:27] fagan: it is there, in trunk 100% sure [09:27] well the test cant find it for some reason [09:28] how are you running the tests? [09:28] 'ello 'ello 'ello! [09:28] mandel: im running the file system notifications test [09:29] oh I forgot to add the u1trial part my bad [09:30] didn't I add instructions in the merge proposal? [09:30] mandel: you did [09:30] ok im still getting it [09:30] :/ [09:30] fagan: can you pastebin how you run the tests and the output? [09:31] mandel: http://paste.ubuntu.com/596440 [09:32] fagan: read the instructions please [09:33] what is the diff between your command and mine? [09:34] fagan: ? [09:35] mandel: well you were running it from the project root dir and I was running it in the tests\platform\windows and I should have been doing it in tests\platform [09:35] Irc from my phone ftl. [09:35] oh and its giving out about google protobuf again :) [09:36] aquariuso: ha its not that bad [09:36] fagan: execute the tests from the root of the dir, if the ubuntuone-storage-protocol was correctly installed it should be working [09:36] fagan: it really is. :) [09:37] mandel: well I ran setup install it should be there [09:37] :) [09:37] easiest way, either go to site-packages and check, or do an import from the command line to test [09:40] mandel: yeah its in there [09:41] fagan: execute the tests like I wrote in the instrcutions, if they do not work, pastebin the command and the result again [09:41] mandel: ok ill pastebin [09:42] mandel: http://paste.ubuntu.com/596444 [09:43] (and ubuntuone-storage-protocol is in the site-packages) [09:44] and the path is setup [09:44] fagan: how did you install the google protobuf package? [09:44] did you use easy_install ? [09:44] mandel: downloaded the binary [09:44] fagan: then you did not install protobuf, you just installed the compiler [09:45] did you try to do the import that error uses? [09:45] ah ok [09:47] mandel: well I just tried to run it and see what was missing [09:47] I didnt do the import [09:47] anyway now I did an easy install protobuf and it cant find the setup script [09:48] fagan: that is the people at google did a poor job, go to the downloaded data by easy_install and navigate to the python dir [09:48] ah ok [09:48] they download everything C++ etc.. and setup.py gets confused [09:54] hi [10:01] mandel: http://paste.ubuntu.com/596461 [10:02] protobuf is installed now [10:03] fagan: oh, that is my fault, copy the clientdefs fro the windows dir to the ubuntuone we dont use autogen.sh on widnows, sorry for that [10:04] ah ok i was wondering because I could see the folder [10:06] mandel: so is 2 skips and 15 errors expected in the tests? [10:06] no errors are expected [10:06] may I see the errors? [10:07] mandel: sure [10:09] mandel: http://paste.ubuntu.com/596468 [10:10] it seems to be something about Level trace [10:10] fagan: oh, that, annoying… turns out that if we just run a single set of tests they brake because ubuntuone.logger is not imported [10:10] very annoying.... [10:10] ah ok [10:10] hm let me think a nice way to fix that [10:10] give me a min [10:10] cool [10:11] I already did a code review last night so its just the tests that I wanted to run [10:14] fagan: try to pull it now, it should work [10:15] kk [10:17] mandel: a little better but still errors http://paste.ubuntu.com/596473 [10:23] fagan: hm I suspected that might happen according to the system performance... [10:23] fagan: can you let me know the settings of the machine? [10:23] mandel: its a vm running on 1 2.6 ghz core [10:25] fagan: can you let me know the RAM, type of hd of the machine, exact processor type and Windows OS too? [10:25] I was expecting issues in vms… I don have some in the vm I have [10:26] mandel: 320mb of ram xp and an amd athlon ii x4 620 [10:26] I could increase the ram if that is too low [10:26] but shouldnt be [10:27] * fagan does have 4gb to spare but wanted to replicate a bad ish computer [10:27] (testing on bad computers makes the testing better) [10:28] fagan: hm try to improve the vm to see if they pass under 2 seconds, if not we can change the timeout for the results [10:29] mandel: cool [10:29] fagan: could you create a similar vm on your machine with M or N and run the tests on linux? I'd like to see if they pass [10:29] and would be great to have it in the exact same env [10:30] mandel: im on natty atm so i can just run it on this machine [10:30] or do you want it on a slower machine [10:30] fagan: I want the slower machine, so an exact same vm but with a diff os [10:31] ah ok will take 15 mins [10:31] with vms it can be a 1000 things so I want to make sure is my code and not the env [10:31] oki doke [10:31] fagan: cool, let me know what you find out [10:32] mandel: sure [10:32] I actually didnt know my computer had a load of vm optimisation in the cpu already [10:33] its pretty handy [10:34] * fagan takes a quick break while the iso downloads [10:36] morning all [11:00] Hmm it's odd, yesterday my files suddenly appeared with proper Artist / Album and now they are back to the broken single "Unknown" entry [11:13] hi duanedesign [11:14] rye: I've let myself assign one question from ubuntuone-servers to you, for that undelete file case. I hope that's OK. [11:14] karni, yes, that's ok [11:14] rye: cool [11:15] karni, well, for natty we won't need to do this since it deletes to trash [11:15] * rye <3 syncdaemon in natty [11:15] rye: right. I just saw someone loose something important, to I thought I'd call rye's powers [11:16] /s/loose/lose [11:16] almost ready with that vm for mandel [11:17] my internet is supprisingly slow today :/ [11:18] fagan: ok, no worries, just make sure that both vms have the exact same settings and that you run the test in the same env, so all apps closed etc... [11:19] mandel: yep already setup 350mb or ram and 2.6ghz single core processor [11:19] just waiting on the ubuntu install now [11:53] hehe one of my housemates is trying to do something in debian and he used ssh into a server and then tried to ssh in to the same server again from the ssh session [11:53] I lold [11:54] Id love to see an ssh to an ssh to an ssh.... on the same machine [12:00] mandel: it passes on the linux vm with the same spec [12:01] hmmm annoying [12:02] mandel: so it might be something specific to windows with low spec [12:02] fagan: can you edit the test files and change the timeout to one more second and let me now [12:02] mandel: sure [12:02] it could be that windows takes 1 sec more, which is a pain [12:02] mandel: or it could be the ram usage of both oses [12:03] windows xp is using 100 mb without anything running and ubuntu is running on about half [12:04] yeah, I'm not surprise, also the arch of the way we get the events is diff [12:04] fagan: in crease the timeout one sec at a time and let me know when it does start working [12:05] I'll also like to know how well it works in a real machine like the one that ralsina has [12:05] mandel: sure thats cool [12:05] im doing it right now so shouldnt take 5 mins [12:06] fagan: we want to keep it as low as possible, just to test how bad is windows or my implementation on windows... [12:07] mandel: 3 secs didnt work [12:07] fagan: did you change it in both places? [12:07] there are 2 of them [12:07] oh I just did it in the py I was running where is the other one [12:08] fagan: in that same py, there are two of them [12:08] (test_filesystem_notifications.py) [12:08] oh ok sec [12:08] mandel: fails at 5 secs [12:09] but passes a few more [12:10] no my bad it is the same as 2 and 3 secs [12:10] fagan: may I see the output of the 5 secs one? [12:10] oh actually 1 more pass at 5 secs than 3 secs [12:11] move up to 10s lets see what happens, then we move down [12:11] and upperbound would be nice to know [12:11] still is failing [12:12] :/ [12:12] 8 max pass [12:12] with 10 seconds? [12:12] really? [12:12] yep [12:12] can I see the outut? [12:12] output* [12:12] give it a sec ill get the full output the scrollback is long [12:12] ok [12:13] also, can you check the processes that are running when you execute the tests? [12:13] use process explorer gives more info [12:13] can you dump the output? [12:13] of the process list? [12:14] fagan: no, or I at least dont know how [12:15] mandel: 9 passes at 10 secs http://paste.ubuntu.com/596494 [12:16] Ill give 20 secs a go [12:17] fagan: ok, but that is waaaay to high we might have a situation here… :( [12:17] mandel: yeah [12:17] id like to know the acutal wait time anyway just to see [12:18] if it fails at 20 secs it would be pretty dumb [12:18] mandel: I just checked and there is only ms processes for windows itself [12:18] no other ones going [12:18] (like I dont have an anti virus or anything doing a scan or anything) [12:18] fagan: yes, I'm wondering how is XP interacting with the vm hd at the kernel level... [12:19] do you have a real machine around? [12:19] mandel: I do but id have to do the setup of the deps again [12:19] and its a really fast machine [12:19] so it wouldnt dup the results if your machine doesnt [12:20] mandel: 20 secs has 7 successes so id say its the vm [12:20] fagan: yes, but it will narrow it down to slow machines :) [12:20] it doesnt seem to have consistant results [12:20] fagan: it does look like it... [12:21] hm, annoying that means that we have to be careful with what is used for testing [12:21] bullocks! [12:21] I hate windows! [12:21] hahah [12:21] and doors too [12:21] http://paste.ubuntu.com/596497 [12:22] puta windows [12:22] fagan: maybe is the os too, I tried with a windows 7 and no xp... [12:22] :) [12:22] ah ok [12:22] this is xp I can do it on a vm for 7 if you want [12:22] (with the same constraints) [12:22] fagan: if you do that it would be very welcome [12:22] * mandel hopes is xp... [12:23] I hope its the VM [12:24] because then its just my tests that are going to have a problem [12:24] ok ill go on break for an hour and get back to it when I get back [12:25] ill have to install the deps again so it will take a few minutes [12:28] ok [12:31] * mandel walks dog [13:25] * mandel back [13:27] ralsina_: ping [13:27] im back too === alecu-away is now known as alecu [13:29] hello [13:29] alecu: morning :) [13:32] alecu: do you have a windows vm? [13:37] mandel, I may have one [13:37] mandel, how can I help you? [13:38] alecu: you just made me a happy man :) [13:38] hehe [13:38] alecu: I have a very strange issue with file system notifications, and I want to test if it is related to the vm that faga set up or just my code [13:38] alecu: I'm going to guess you do not have the env setup for developing on sd on windows, right? [13:39] mandel, right, I don't have the env setup, but that's something I should do [13:39] mandel, I'll start with a fresh vm anyway [13:39] mandel, should I ask fagan help to get the dev env set up? [13:39] alecu: im not the greatest windows guide :P [13:40] alecu: r fell free to ask me, either of us should be able to answer :) [13:40] yeah [13:40] even ralsina_ has a vm ready :P [13:40] python is always a good start [13:40] fagan, mandel: ok. so.. is there a wiki with steps to follow to set up the vm, or something? [13:40] alecu: nope [13:41] alecu: no, and there should be, let me right it :) [13:41] you need python, pycrypto, twisted [13:41] alecu: internal wiki, right? [13:41] python setup tools [13:41] mandel, for a start the internal wiki sounds fine [13:42] mandel, should I setup it on 7 or XP? [13:42] xp to see if its a problem with my setup id say [13:42] im going to try on 7 now [13:42] alecu: having both would be nice, to see the diff issues we might find [13:43] mandel, fagan: are you using VirtualBox, or qemu or what? [13:43] * alecu is very used to VBox [13:43] im using vbox [13:45] alecu: I'm using vmware, us what ever i better for you [13:45] I dont know if vbox does use the hardware vm features... [13:46] yup, it does. kvm does as well [13:46] mandel: well its taking advantage of my cpu features [13:47] so its cool [13:47] oh, cool then :) [13:49] another thing you need is coverage [13:49] aloha u1! [13:50] good morning ralsina_ [13:50] mandel: here I am, what should I review? [13:51] ralsina_: we found an issue with last night branch [13:51] it fails bad on my vm with everything fine ish [13:52] ralsina_: hello :) [13:52] hi fagan, mandel === ralsina_ is now known as ralsina [13:53] ralsina_: can you look at https://code.launchpad.net/~mandel/ubuntuone-client/close_files_in_tests/+merge/58343 [13:53] ralsina: if you can run it in a real machine and in a vm would be great, [13:53] I can run it on a real machine [13:53] in my vm everything works find under 2 seconds for fagan it fails even when the tests are set to 20 seconds [13:53] and I need to set up the missing pieces for the VM [13:54] which is weird and scary at the same time [13:54] more scary that weird [13:54] yeah it is weird [13:54] ok, let me try it [13:55] however, if it works for me, that means nothing, right? [13:56] ralsina: whats your memory like on the computer [13:56] fagan: 4GN [13:56] im on 350 mb ram [13:56] 4GB [13:56] fagan: 350mb ram won't be enough [13:57] ralsina: well it was enough for the linux client [13:57] * alecu reboots [13:57] fagan: well, this ain't linux ;-) [13:57] hah [13:57] fagan: my XP vm crashes with anything below 768 [13:57] oh I didnt check if I increased the ram what would happen [13:57] ralsina: well my computer is special :) [13:58] fagan: well, you can always TRY assigning more RAM. [13:58] ill give it a go [13:58] fagan: and if it works, it means u1 doesn't work on 1994 computers. [13:58] ralsina: well, I wanna see if with a dcent m it will work [13:58] mandel: ok [13:59] fagan: what vm software are you using? [13:59] ralsina: so if it works with your real machine, my 3 vms and another I might get from alecu I'll be happy to say that the vm is the issue [13:59] vbox [13:59] fagan: how much ram do you really have? [13:59] 4GB [14:00] fagan: then why limit the VM to a ridiculous 350? The default vbox offers you should be 768 already. [14:00] ralsina: well I was doing the oh it would be good to test for older machines since we are supporting xp [14:00] noone has xp on less than 512 [14:00] it's the suggested minimum [14:00] Ive used a computer with less [14:01] or well 350 mb [14:01] you can't put 350 on a computer, can you? :-) [14:01] its one of those ones before 500 mb ram chips came out [14:01] there is all sorts of sizes back in the day [14:01] :) [14:02] you can put 256 + 64, or 256 +128 only. [14:02] and you need an antiquarian ;-) [14:02] * fagan has a bad memory :P [14:02] argh, freaking bzr and limbo permissions and windows [14:03] fagan: it's easy, memory chips always come as powers of 2 [14:04] and 350 is not a reasonably short enough sum of powers of 2 :-) [14:04] binary :) [14:04] passes with 1.4gb of ram [14:04] oh nooooo [14:05] crap some of it did fail [14:05] ralsina, you can have a system with 350 main memory if it's one of those with cheap video chipsets that use the same ram chips, and that let you select the amount of memory in the bios. [14:05] yes, 350 in binary is 101011110. So how many memory banks do you need to get 350MB of RAM? ;-) [14:05] alecu: ohhhhh. Ok, you win :-) [14:05] hahah [14:05] ralsina, so, 384 - 32 video = 352 [14:05] mandel: interesting news [14:05] alecu: close enough! [14:05] ~~~ [14:06] ot fails with more ram [14:06] *it [14:06] and the timeout at 20 [14:06] fagan: may I see the error, can you run it a second time? [14:06] fagan: also, do tell me you restarted the vm [14:06] running [14:06] mandel: did [14:07] ralsina, yesterday I remembered to add tomorrow and friday to canonicaladmin.... sorry for being so late. [14:07] me? [14:08] alecu: np, I will approve it [14:08] mandel: ImportError: cannot import name clientdefs [14:08] me [14:08] alecu: 1 hour early [14:08] me [14:09] alecu: oops, no, right time,sry [14:09] me [14:09] ralsina: you have to move the clientdefs file to ubuntuone [14:09] hmm [14:09] ralsina: oh sorry I forgot to mention that, copy clientdefs from the windows folder in the root of ubuntuone-client [14:09] it gets genrated by runtests and you are not using it, mea culpa [14:09] me [14:09] mandel: its not the root its the ubuntuone folder I think [14:10] ralsina, fagan started meeing. I just followed the pack [14:10] ralsina: fagan is right, copy it from root/windows to root/ubuntuone/ [14:10] mandel: I get some timeouts :-( [14:10] FAILED (skips=2, errors=4, successes=13) [14:10] fagan: start [14:11] ralsina: can you increase the timeout of the tests to 3 seconds, there are two vars in the test that sets it [14:11] DONE [14:11] * went to test mandel's branch and broke it [14:11] TODO [14:11] * figure out what the problem is [14:11] Blocked [14:11] mandel: ok [14:11] * Nope [14:11] * mandel wonders why windows is reporting changes so slow... [14:11] alecu: Gogog [14:11] DONE: got the fix for the blocker issue in banshee merged in trunk and stable; did a fix for broken tests on u1-client on maverick and before [14:11] TODO: help mandel with windows and vms [14:11] BLOCKED: no sir [14:11] NOTE: national holiday tomorrow and friday. [14:11] mandel: my disk is not terribly fast [14:11] NEXT: thisfred [14:11] * DONE Investigated bug #746276, bug #762722, bug #762717, bug #764057 [14:11] * TODO fix bug #762722 [14:11] NEXT: ralsina [14:11] Launchpad bug 746276 in desktopcouch (Ubuntu) "desktopcouch-service crashed with RuntimeError in start_couchdb(): Can not start couchdb. (affects: 13) (dups: 5) (heat: 82)" [Undecided,Incomplete] https://launchpad.net/bugs/746276 [14:11] Launchpad bug 762722 in ubuntuone-control-panel "Opening the Control Panel during sync removes progress indicator (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/762722 [14:11] Launchpad bug 762717 in ubuntuone-control-panel "Launcher icon shows syncing progress when not connected (affects: 1) (heat: 6)" [Wishlist,Confirmed] https://launchpad.net/bugs/762717 [14:11] Launchpad bug 764057 in ubuntuone-client (Ubuntu) (and 1 other project) ""program is still running: ubuntu one file synchronization in progress" at logout (affects: 2) (heat: 10)" [High,Triaged] https://launchpad.net/bugs/764057 [14:11] DONE: calls calls calls, some designer work [14:11] TODO: way too much stuff [14:11] BLOCKED: by life [14:12] hah [14:12] alecu, please [14:12] ralsina, seriously boss, get more sleep! [14:12] alecu: I slept a lot last night, like 9 hours. I am feeling fine nw :-) [14:13] ralsina: anyway your one isnt failing as bad as my machine [14:13] oh and if you need help with any of the designer stuff im around to pick up some slack [14:13] then mandel: go [14:13] DONE: looked at the sd code I left blocked by sso. Done some work with fagan trying to find why windows is slower than linux. Some work on updating the wiki to allow people to set windows vms for testing. [14:13] oh well [14:13] TODO: more wiki editing, find WTF is going on with windows and the bloody fs events speed [14:13] λ DONE: beyond-last-minute releases, scanned sprint receipts [14:13] λ TODO: maybe urgent beyond-beyond-last-minute uploads, SRUs [14:13] λ BLCK: None. [14:13] BLOCKED: no, but windows is a PITA [14:14] who is next? [14:14] ralsina, no nessita today? [14:14] the others are off [14:14] I wasgoing to say dobey but he did it already :P [14:14] alecu: she is teaching will be here later [14:15] oh, it's wed, and she teaches today. [14:15] right. [14:15] ralsina: increase the timeout and we will see, also if you can show me where are the error it would be a nice indication, reporting an slow move mught be ok, reporting a slow delete is a problem [14:15] mandel: setting timeout to 5, I get two errors [14:15] ralsina: may I see them? [14:15] I'll pastebin [14:15] cheers [14:16] mandel: https://pastebin.canonical.com/46422/ [14:16] http://paste.ubuntu.com/596542 [14:16] one is dir_create! [14:17] I don't know which other one is failing [14:17] ralsina: do a second run to see if there is a race condition, even better incrase the timout to 10 [14:18] ralsina: is trial that reports always 2 instead of one… ubber lame [14:18] running at 10 [14:18] mandel: I increased it to 20 with the extra ram and its still failing [14:18] oh, so it's only 1? [14:18] dir_create passes at 10 [14:18] dir create passes for me at the same one [14:18] fagan: that pastebin shows 2 seconds not 20 [14:19] test_file_create_close_write failed [14:19] mandel: yeah I reran it at 2 [14:19] ill pastebin a 20 sec now [14:19] fagan: move it up to 10 and run [14:20] ralsina: yea, is a bloody race condition, the events are not returned fast enough and the buffer in the async IO is no firing the vent early enough…. [14:20] mandel: more tests fail on 10 than on 5: https://pastebin.canonical.com/46424/ [14:20] I'd not be surprise if you rerun it and get all of them green [14:20] mandel: so basically the tests are useless on windows [14:20] hah [14:20] puta windows [14:21] ralsina: yes, I have no way to control when ReadDirecotryChangesW witll return the events, which are later processed and sent to the muted filter, this means that there are several env situations that will make the test not to ran [14:22] ralsina: remove the timeout var from the class (which will default it to 120) I'm courious to see what happens [14:22] mandel: figured out as much. Suckitude :-( [14:22] ok [14:22] mandel: http://paste.ubuntu.com/596545 [14:22] I get errors about there not being a timeout member ;-) [14:23] running at 120 [14:23] ralsina: oh, I forgot I changed the callLater part… ups [14:23] funny enough the code is tune to always pass in my machine hehe [14:24] mandel: that's the magic machine effect in action :-) [14:24] mandel: nice machine [14:24] the developer always has the magic machine that works. Or he would have noticed ;-) [14:24] it reminds me of something [14:25] ralsina: I thin is that a touch the frequency of the pooling of the data in the reactor main thread in probably the right spot [14:25] ralsina: which is utterly useless for anyone else... [14:25] exactly [14:25] "jiggle this until it passes" is not a valid test guide. [14:26] test_dir_create fails at 120 [14:26] ralsina: can you set the logging so that you see its output o stdout, I wonder if any event is given to the watch thread [14:26] just enable DEBUG? [14:26] ralsina: yeah, the thread is being scheduled by something in the wrong place [14:27] ralsina: hmm Ilet me check [14:28] ralsina: looks like enabling debug should be enough [14:28] there is no env variable that is used for that... [14:28] weird I have it set to 1 [14:29] oh... set logging level. Ok [14:29] ralsina: yes :) [14:29] sorry, I assumed :P [14:30] mandel: level is set to DEBUG in the code already [14:30] and it does not go to the stoudt, lame... [14:30] yep, let me change the handler [14:32] ralsina: ok [14:32] ralsina: we are looking for something of the form Added %s to raw events queue. [14:32] mandel: ok [14:32] ralsina: where raw events are the events sent by the OS to be processed and converted to pyinotify events [14:33] ralsina: if those are there, it means that the processing thread isdoing something funny, if we dont get them, I blame kernel32.dll [14:33] weird, even if I use logging.StreamHandler() it doesn't go to the terminal. [14:34] ralsina: I think the tests should be written the logs in the hidden AppData dir within xdg [14:34] * mandel can remember the exact dir... [14:34] * ralsina looks [14:36] ralsina: here: C:\Users\Mandel\AppData\Local\ubuntuone [14:36] do chante the username ;) [14:36] yes $APPDATA :-) [14:36] exactly :P [14:36] ja, I don't have an AppData [14:37] ok, creating and re-running [14:37] ralsina: sorry I'm stupid, it is in _trial_temp [14:37] ralsina: like this C:\Users\Mandel\Documents\ubuntuone-client\close_files_in_tests\_trial_temp\xdg_cache\ubuntuone\log [14:37] ok [14:38] ralsina: also, you do have AppData, but is hidden [14:38] should be in ~/AppData you have to tel the shell to show hidden dirs [14:38] mandel: well, I should be able to cd into a hidden folder, and mkdir succeeded :-) [14:38] ralsina: is windows… [14:38] But everything is possible :-) [14:40] locked up my computer [14:40] having to do uploads today makes me a sad panda [14:40] dobey: well it is getting close to release the sooner the better [14:41] mandel: nothing of the sort in the logs [14:41] mandel: in fact, nothing interesting in the logs at all [14:41] fagan: yes, generally these things should be done before freeze [14:41] which we are already past [14:41] dobey: you are the least panda-like human. [14:42] ralsina: sad ninja then [14:42] ralsina: now, that is weird, I get logs [14:43] mandel: let me run on 2sec and do a full trial. On 120 it takes forever. [14:43] ralsina: sure :) [14:43] ralsina: I just ran mine, and the logs are: _trial_temp\xdg_cache\ubuntuone\log [14:44] yes, test.log [14:44] test.log? I get syncdaemon.log [14:45] hmmm [14:45] oh, much more interesting :- [14:45] :-) [14:45] test.log is pretty much what you get on terminal [14:46] mandel: https://pastebin.canonical.com/46429/ [14:48] ralsina: look at the logs, the events are there.. I mean: Error encountered while handling: FS_FILE_CLOSE_WRITE in [14:49] mandel: then why does it fail? [14:49] ralsina: It smells to be due to raise AlreadyCalledError [14:50] dobey: are you going to be creating a new u1-client package soon - i'm thinking of bug #764646 in particular [14:50] Launchpad bug 764646 in ubuntuone-client (Ubuntu Natty) (and 4 other projects) "music store widget dies with «TypeError: find_credentials() takes exactly 3 arguments (2 given)» (affects: 3) (dups: 2) (heat: 26)" [Critical,Triaged] https://launchpad.net/bugs/764646 [14:50] ralsina: I\m jumping to #chicharra to ask the smart people :P [14:50] mandel: always a good idea :-) [14:51] joshuahoover: i will be uploading with patches, but not making a new tarball relaese/upload to ubuntu. hopefully we can get the patches in [14:51] ralsina: could explain why it fails with long timeouts :) [14:52] dobey: ok, cool ...pitti said we were fine on that bug in particular and any "bug fixes" right now [14:52] mandel: right, because the timeout is not the problem :-) [14:52] ralsina: clearly, t has plenty of time, the eq mich be wondering wtf is going on under me [14:53] mandel: yes, the event comes quickly [14:53] mandel: but it fails anyway [14:54] ralsina: doyou have the output of the test for those logs you gave me? [14:55] mandel: sure, it's on another file, let me pastebin it [14:55] https://pastebin.canonical.com/46431/ [14:56] hrmm, i wonder how well sqlite does with storing "large" images inside it [14:56] mandel: keep in mind that the first log seems to have more than one run in it fr some reason [14:57] mandel: you want from 10:43 onwards [14:57] ralsina: can you delete both logs and reproduce it again? [14:57] sure [14:57] we really dont care about which error, it looks the same to me [14:59] mandel: syncdaemon.log --> https://pastebin.canonical.com/46433/ [14:59] ralsina: and stdout? [14:59] mandel: test.log -> https://pastebin.canonical.com/46434/ [15:00] mandel: stdout --> https://pastebin.canonical.com/46436/ [15:06] ralsina: can you go to the tests and modify the HitMe class so that you see how many times you get an event, also print the event to see what is happening [15:06] mandel: sure [15:06] ralsina: verterok had a brilliant idea of what could be the error [15:08] Ha! All successes [15:09] mandel: https://pastebin.canonical.com/46441/ but it's all green, I'll rerun [15:14] mandel: the thing decided to start passing the tests. Give me 2' to check why ;-) [15:15] ok, getting failures again... === clarita_ is now known as clarita [15:23] mandel: when the tests fail, HitMe doesn't seem to get the events [15:24] mandel: stdout: https://pastebin.canonical.com/46447/ [15:25] ralsina: let em update the branch with some more logging, I'll need a couple of mins [15:26] mandel: ok [15:35] mandel, windows 7 has just finished installing.... I couldn't find my XP around. [15:35] mandel, were you able to write the wiki page on how to install the dev-env? [15:36] alecu: I'm got part of it, let me save it so that you can start with it [15:36] alecu: give me a min [15:36] ralsina: can you pull the branch with the changes? [15:36] mandel, cool. Anyway, I need to run some errands right now, I'll be back in an hour or so. [15:36] alecu: ok, even better then :) [15:36] ralsina: the logs should have way more useful info === alecu is now known as alecu-errands [15:54] alecu-errands: what was that zombie movie? [15:57] ralsina: looks like the twisted main thread in your machine is very gready and that is why the others dont get it… let me try and find a way to fix it [15:57] alecu-errands: ah nevermind, found it :) [15:58] mandel: ok [16:00] alecu-errands: and netflix has it on dvd! :) [16:01] mandel: did you push another branch? [16:02] some bastard bough a bird and place it next to my window and it keep whistling like I was a good looking lady... [16:02] hahahha [16:02] most annoying thing ever!! [16:02] evil taunting bird [16:02] I have dogs next door that just bark all day long [16:02] * fagan wants to move [16:03] (or find a way to silence them dogs [16:03] ) [16:04] gun? [16:04] knife? [16:04] baseball bat? [16:05] or cricket, if that's your fancy [16:05] dobey: cricket if I wanted to bore them to death [16:05] baseball is similar [16:05] :) [16:05] oh that would do [16:07] ok, well, client uploaded, and hopefully will get accepted soon [16:07] ralsina: I've made the task that checks for events in the twisted main thread to be a better friend, can you pull and let me know, if this does not work I'll need some more time 'cause I wont have a clue of why is not workiing [16:07] mandel: ok, doing it [16:08] mandel: fails in pretty much the same way [16:08] and now, i think i will go get some lunch, and then fix up nightlies and see if i can do some SRU work [16:08] bbiab [16:12] ralsina: may I see the logs... [16:12] mandel: sure, 1' [16:12] looks like some race condition… but I really dont know how to tell the task not to execute when there are no events... [16:13] mandel: I'll re run to get clean nes [16:13] ok [16:15] mandel: log https://pastebin.canonical.com/46460/ [16:16] mandel: stdout: https://pastebin.canonical.com/46461/ [16:17] ralsina: how is your one passing 12 when mine does 7-9 [16:17] fagan: it varies from run to run. I even had some 100% success ones [16:17] I dont think ive gottn one [16:18] (100% I mean) [16:22] fagan: is a multithreading issue, there is a greedy thread and one of the filosofers is dying, in this case the events one [16:23] ah ok [16:23] so its a race [16:23] that's why making the timeouts longer doesn't make a difference [16:23] because what matters is the order of the events, not how much they take [16:24] * fagan remembers tales from the lecturer in 2nd year about the dangers of not controlling execution timings of threads [16:24] and it sucks :-( [16:24] I'm off to lunch === ralsina is now known as ralsina_lunch [16:25] ill be off in a half hour [16:25] ill just keep playing about with my merge tester while I wait to test some more [16:41] ralsina: the problem I have is the way in which the processing of the events is hooked into the main thread, atm is a task that is repeated every second, but that does not ensure that the other threads are allows to run, so even if you could signal the thread that you have an event [16:41] dobey: ping [16:41] it does not mean that the thread that gets the events is ever executed, which seems to be the issue... [16:42] with the select reactor and pyinotify using a file it works, because you know that the changes in the file will work... [16:42] dobey's on vacation this week, right? [16:42] mandel: so, it works IRL, fails on testing? [16:43] rodrigo_: dobey is working today, he just went out for lunch a few minutes ago [16:43] ah ok [16:43] I guess he's answer tyo my ping when he's back [16:43] he'll [16:43] ralsina_lunch: yes, which is a PITA [17:02] ralsina_lunch: ping [17:02] mandel: pong [17:04] ralsina_lunch: there is one last trick I want to try, but it will only work because I know the tests… which is cheating [17:04] hahaha [17:04] mandel: feel free :) [17:05] ralsina_lunch: I'll give you the diff to apply I dont wanna push it to the branch ;) [17:06] mandel: cool [17:17] rodrigo_: hey [17:19] hi dobey [17:19] rodrigo_: que pasa mi amigo [17:19] dobey: so, is it important to keep evo-couchdb compiling for lucid, karmic, etc? [17:21] dobey: I ask because for evolution 3.1.1 there are more API changes coming, and the code is becoming ugly with all those #if EDS_CHECK_VERSION [17:21] rodrigo_: trunk compiling on karmic is definitely something we don't need to keep happening. and it already doesn't compile on lucid afaict [17:21] rodrigo_: yeah, the changes for GTK3 will make that an issue [17:22] dobey: ok, so supporting >= 2.32 is enough, right? [17:22] dobey: well, for gtk3 I don't see much problem, since we can detect what gtk version to use based on the evo-plugins version [17:22] we use very basic GTK, so it should compile ok on both [17:23] the problem is all those EDS_CHECK_VERSION in the backend code [17:23] rodrigo_: well, what is in maverick? [17:23] have a nice around-Easter-week [17:23] 2.30, I think? [17:23] rodrigo_: i think we should probably support that. [17:23] well, no, 2.32, like natty [17:23] can someone with maverick check? [17:23] rodrigo_: 2.30 [17:24] ugh [17:24] ok then [17:24] https://launchpad.net/ubuntu/+source/evolution [17:25] rodrigo_: although, in our Buenos Aires sprint a couple weeks ago, we talked a little about possibly making some rather big and invasive changes for desktopcouch. and i think evolution-couchdb would fit into that scope, so maybe we should just wait a few weeks and you, thisfred, and i can chat in person in Budapest about it [17:25] ok [17:26] dobey, thanks for including the fix for dbus order into the release! [17:26] rodrigo_: big and invasive, as in rewriting desktopcouch in vala, and just providing the GIR interface and libdesktopcouch-glib/libcouchdb-glib directly in desktopcouch [17:27] rye: sure. i think "infinite restart loop" is a pretty serious issue :) [17:27] dobey: oh really? [17:27] rodrigo_: yeah, it should not be hard to do, and would make a lot of stuff work much more smoothly, and be simpler [17:27] dobey: we are discussing in d-d-l about couchdb syncing for gnome, and some replacement for desktopcouch is in discussion [17:27] so maybe that fits that [17:28] rodrigo_: yes, we were also discussing similar things [17:28] I am going to think about how to do without having to call dbus to get the port [17:29] not sure how to do it though [17:29] dobey: are we really considering that? [17:29] +1 from me [17:29] mandel: yep :) [17:29] rodrigo_: why? === alecu-errands is now known as alecu [17:29] dobey: since when? is we, you rodrigo and I? [17:29] z/I/me [17:29] dobey: because it doesn't work well, if couchdb crashes, for instance [17:30] dobey, plaga zombie 2? [17:30] alecu: yeah. zona mutante? [17:30] dobey, yes!!!! :-) [17:30] rodrigo_: yes, that's one of the issues we want to fix as well [17:31] mandel: well i think thisfred and chad are more convinced now. we discussed it very shortly in buenos aires [17:31] alecu: it's in my dvd queue, so i'll probably get it next week :) [17:31] dobey: ah, you bastards! and I was no there... [17:32] dobey: another thing is just having a daemon (in *cough* python *cough*) just for that [17:32] I hate windows even more now… take the time I should be allowed to do cool stuff [17:32] mandel: well we will discuss more in Budapest i'm sure :) [17:32] rodrigo_: the daemon isn't just for that, and right now it's 3 daemons :) [17:32] 3? [17:32] dobey, one thing to note: all the characters in the movie speak "neutral spanish", mocking the lousy dubbing of hollywood action movies that are shown on tv here. [17:33] alecu, zona mutante? [17:33] rodrigo_: yes, desktopcouch-service forks itself into 2 additional processes. one is dbus, one is replication, and the other is some other stuff [17:33] rodrigo_, "plaga zombie 2: zona mutante" [17:33] and if you see a zombie extra hanging from a little bridge, that's me. [17:33] dobey: ah, right, thought there were other 2 different ones [17:33] rodrigo_: part of that is because of the way stuff works in python [17:34] * thisfred is committed to learning vala and doing stuff with it [17:34] go thisfred go! :) [17:34] rodrigo_: rewriting in vala, we could make them be 1 process again, probably, and use threads [17:34] rodrigo_: I will not join you in hating python though ;) [17:34] thisfred, I don't hate it [17:34] i do [17:34] oh just dobey then :) [17:34] hate hate hate [17:34] thisfred, I just don't like it being used for stuff it shouldn't :) [17:34] but I like it [17:35] * alecu is willing to join the vala team as well. As awful as a "vala programmer" sounds in es_AR [17:35] I use it for lots of scripts and small apps [17:35] dobey: well, threads would make more sense [17:35] pretty much everything python is used for, is stuff it shouldn't be used for :) [17:35] alecu: lol [17:35] alecu, this one -> http://www.taringa.net/posts/tv-peliculas-series/2290477/Plaga-Zombie-2-_-zona-mutante_.html ? [17:36] alecu: i just always get confused by the pronounciation of 'll' in .ar when i go there :) [17:36] rodrigo_, that's it! [17:36] dobey: Good! Your hate has made you powerful. [17:36] alecu: are you one of the zombies in it? [17:37] thisfred: now i just need an underground lair [17:37] dobey, yes a zombie extra, but I show up maybe on 10 seconds of the film... :P [17:38] alecu: i think that's more air time than miguel got in Antitrust :) [17:38] jajaaja [17:38] and 5 of those seconds are on the credit screen. [17:38] hehe [17:38] :) [17:38] if my theory is right we should no longer see the ubuntuone-syncdaemon crashed in DBus __new__ [17:39] ok, time to go out and for 5 days of national holidays!! [17:39] mandel, call me if you go toi donosti [17:39] dobey: I just have one complain about sing vala [17:39] rodrigo_: I will certainly do [17:39] mandel, do you have any wiki page for me? [17:40] dobey: I dont want to have to deal with compaling that on windows... [17:40] alecu: partial one, got cauth in stupid crap, but is enought to get you busy for 4 hours at least [17:41] mandel: heh === ralsina_lunch is now known as ralsina [17:44] alecu: https://wiki.canonical.com/UbuntuOne/WindowsPort [17:44] cool [17:44] alecu: getting vs2008 takes ages, so you can get it running til I finish the rest [17:44] uh [17:45] sorry there is no much, but I have been trying to get multithreading right and I'm failing like a little girl in a rugby game… [17:46] alecu: you as a quick note, using python for x86 make everything a lot easier [17:46] x64 is a major pain in the ass [17:47] cool [17:48] mandel, I thought all rugby players were little girls! [17:48] (my father in-law would kill me) [17:49] alecu: wait? your father in-law is a little girl?! please do invite me to the next family dinner :D [17:50] mandel, he's a rugby trainer [17:50] mandel, and he has played rugby like... forever [17:51] alecu: oh.. I preferred to imaging him as a little girl hehehe but an argentinian rugby coach is also interesting [17:51] hahaha [17:51] mandel, ok, next time we sprint around here we'll have asado at his place. [17:52] alecu: that is a deal I wont forget ;) [17:54] * mandel walks the dog now that he thinks he knows how to fix the issue he has... [17:54] ralsina: I'll ping you in a couple of hours, I need to try it or I wont be able to sleep during the night [17:54] * mandel walks dog [17:55] mandel: ok [18:13] thisfred: haha, there is a python users group meetup down here tonight. i could go and say "you're doing it all wrong!" :) [18:15] hola mundo! [18:16] hola prof nessita [18:16] hi dobey. Do you want to take the same exan we just gave our students? :-) [18:16] holanessita [18:16] hola thisfred [18:16] dobey: do it! [18:17] any news? [18:17] apparently iphones have a databse of locations [18:17] *gasp* [18:19] nessita: also, you don't need to nominate any bugs for 11.10 yet. [18:19] dobey: by 11.10 you mean ubuntu later? [18:20] nessita: i mean Oneiric. not the milestones but the "Nomiate for series" link on bugs [18:21] nessita: until 11.04 is actually released next week, anything fixed there will end up in 11.10 automatically [18:22] nessita: i saw you nominated one of the u1client bugs from yesterday for Oneiric, which is why i'm mentioning it. doesn't hurt anything, but it's not necessary [18:28] dobey: right, I now remember. Sorry for that then. [18:31] dobey: I will be doind the stable-1-0 branch for u1cp today, I'll reply to your email in ubunet when is done [18:31] now, I'll quickly grab something to eat [18:31] nessita: 1.0 or 0.10? [18:34] 1.0 [18:34] dobey: I released 1.0.0 yesterday [18:36] oh [18:41] nessita: congrats ;-) [18:44] oooh, a nessita! [18:57] congrats? [18:58] hi Chipaca, ralsina [18:58] * nessita is catching up with email [18:58] nessita: i know you know this, want to answer/explain? http://askubuntu.com/questions/36062/folder-shared-from-unknown-user [18:59] * nessita looks [18:59] agrh, complicated answer without saying that our shares infrastructure is poor :-( [19:00] nessita: a 1.0.0 always deserves a congratulation. [19:00] Chipaca: do you know that there is no way to revert the unknown name in shares, right? [19:00] * ralsina never released an actual 1.x of any of his programs [19:00] ralsina: ah! :-D /me agrees [19:01] Chipaca: I'll reply explaining the situation, saying that we're planning to improve that [19:02] nessita: yes, i know there's no way to revert it [19:34] dobey, alecu: hey there, a report of gsd crashing becasue U1 just came in: bug #757786 [19:35] Launchpad bug 757786 in ubuntuone-client (Ubuntu) "gnome-settings-daemon crashed with SIGSEGV in g_return_if_fail_warning() (affects: 1) (heat: 10)" [Medium,New] https://launchpad.net/bugs/757786 [19:35] nessita: dup [19:35] dobey: ah, I pinged your becasue apport didn't mark it as such [19:35] dobey: can you please update it? [19:38] yeah; don't know why apport didn't dup it. fix is in maverick-proposed though [19:40] ah [19:43] mandel: ping [19:44] nessita: pong [19:44] mandel: could you please answer the question in https://code.launchpad.net/~mandel/ubuntuone-client/close_files_in_tests/+merge/58343 ? [19:44] so I can finish the review [19:45] nessita: is not the answer good enough? [19:45] mandel: is there an answer? /me curses firefox [19:46] nessita: according to lp from 10 hours ago :) [19:46] mandel: I opened FF and all the tabs were restored, I thought that they were reloaded but they were not :-( [19:46] sorry! [19:46] nessita: no worries, it can happen to anyone :) [20:13] I am taking a bit early break, will be back later tonight. See you all later! [20:20] bye ralsina [22:52] * nessita -> eods and eows [22:52] bye all!