/srv/irclogs.ubuntu.com/2011/10/21/#ubuntuone.txt

=== eu is now known as Guest73896
=== Guest73896 is now known as adorilson
notKlaatuanybody know what this kind of error might mean:  http://ubuntuone.com/0fUscV451225ulVILAFYqs02:49
notKlaatu"Could not locate object"02:50
notKlaatuafter trying to publish a picture to the web on a brand new trial ubuntu one account.02:50
=== zyga is now known as zyga-afk
mandelmorning all07:31
JamesTaitHappy Friday, world!08:29
mattiJamesTait: ;)08:41
JamesTaitmatti: o/08:42
karniHi everyone!08:53
matti:)08:56
gatoxhi10:04
mandelgatox, buenas!10:05
gatoxmandel, how aree you?10:05
mandelgatox, what are you planning to do in your freaky friday?10:06
mandelgatox, good, in pain from rugby, but good :P10:06
gatoxmandel, in the morning i'm going to keep working in the expand_user path actually.... and then after lunch, i'm going to work in a branch to use pyqt api210:07
mandelgatox, cool, I'm trying to fix that same issue in python directyl :P10:07
gatoxmandel, that would be even better! :P10:08
mandelgatox, I'll be sending a patch in a few mins to python that uses the correct C API to expanduser, and I already sent one to fix listdir when using \\?\10:08
mandel:D10:08
gatoxgroso!10:08
mandelgatox, who said that canonical did not contribute upstream ;)10:09
gatoxmandel, :D10:09
mandelgatox, is amazing how slow is vs2008 compiling big pojects :(10:10
mandelgatox, I don't know how you would work with that..10:10
gatoxmandel, which language?10:10
gatoxmandel, i used to work with vs2010 in c++ in my previouus job..... and it literally disconnect me from internet and hung up my machine when it was compiling  because it use all the resouurces it seems :P (i wasn't the only one)10:12
mandelgatox, is C, I'm compiling python with all the extra libs to see if my changes work as expected..10:13
mandelgatox, so far vs2008 has chrashed twice10:13
gatoxmandel, ahhh yep.... same situation10:13
gatoxmandel, vs works fine..... if you use it for .net...... everything else is horrible!!10:14
mandelgatox, I do have good memories with it on .Net, but is ofcourse due to the way in which Intellisense works10:15
gatoxmandel, the debugger is quite good too....... and i hate vs, so...... it's good jjeje10:15
mandelhaha10:16
nessitahello crowd!11:31
mandelnessita, good morning and happy freaky friday ;)11:32
nessitamandel: hey there! I will do the reactor-import review soon11:32
gatoxnessita, hi11:33
mandelnessita, no worries, atm for the freaky friday I'm trying to send patches to python itself for os.listdir (so it works with \\?\) and for the expanduser issue (unicode)11:33
nessitagatox: hi there!11:33
mandelnessita, so hopefully we can remove those hacks from our code :)11:34
mandelok, the dog needs me, bbl11:35
gatoxnessita, i'm finishing expand_user branch in the morning.... and then work on my freaky friday project :D11:35
nessitagatox: what would be your freaky friday project?11:35
gatoxnessita, look how to migrate u1 to use pyqt api2.... so it will use python string instead of qstring, python object instead of qvariant and some other stuffs.... also it will be compatible with pyside api.... so with 4 lines of code we could change from pyqt to pyside and use any of that frameworks11:37
nessitagatox: nice!11:37
gatoxnessita, it seems really interesting.... solve some problems11:38
gatoxnessita, because the api2 is more compatible with python311:38
gatoxtoo11:38
nessitamandel: broken tests in reactor-import11:52
nessitamandel: plus a tiny change in a fake11:52
nessitagatox: added a comment on sso's network-detect12:13
gatoxnessita, ok12:13
nessitagatox: also, tests are broken on windows, but I guess it was alecu's branch. I file bug #879407 assigned to him12:15
ubot4Launchpad bug 879407 in ubuntu-sso-client "Test failure after token name change (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/87940712:15
nessitabrb, I need to run some quick errands12:16
gatoxnessita, ok, i'll be check that  in a while12:16
ralsinamorning!12:25
gatoxralsina, hi12:29
mandelnessita, ok, I'll take a look after lunch12:40
mandelralsina, https://answers.launchpad.net/ubuntuone-client/+question/17553012:41
* mandel lunch12:43
=== txwikinger2 is now known as txwikinger
nessitagatox: ping12:54
ralsinamandel: yes, so? :-)12:56
* mandel back for stand up12:59
ralsinamandel, alecu, nessita, dobey, gatox: standup in 1'12:59
ralsinaDon't say I don't give previous notice ;-)12:59
nessitagatox: in https://code.launchpad.net/~diegosarmentero/ubuntuone-windows-installer/connect-files/+merge/79611, did you fix the last lint warning I pasted? ( 41:  [E0611] No name 'tools' in module 'ubuntuone.platform' )?13:00
mandelme13:00
nessitame13:00
nessitaralsina: you forgot to give notice to yourelf :-P13:01
nessitagatox, alecu, dobey, ralsina?13:01
ralsiname13:01
dobeymeh13:02
nessitamandel: go!13:02
alecume13:02
gatoxme13:02
gatoxnessita, pong13:03
nessitamandel: can you go please?13:03
mandelhttp://bugs.python.org/issue1320713:03
mandelone sec13:03
nessitamandel: you can catch up at the end13:04
nessitaDONE: 1-1 with ralsina, reviews (still lots pending!), monkey coder13:04
nessitaTODO: magiciada, try to wrap up the epic branch. Almos there!13:04
nessitaBLOCKED: nopes13:04
nessitaNEXT: ralsina13:04
ralsinaDONE: nessita 1-1, administrative stuff, started expense report, preparing for trip TODO: finish starting the resourcing, arrange missing 1-1s, finish expense report, more trip preparations, BLOCKED: feels like it, but no, NOTE: I will not work next week's thursday and friday (Disneyworld!)13:04
ralsinadobey!13:04
dobeyλ DONE: commited banshee fix upstream, process e-mail13:04
dobeyλ TODO: banshee/libu1 SRUs, hack day?13:04
dobeyλ BLCK: No.13:04
dobeyalecu13:04
alecuDONE: moved the office13:04
alecuTODO: work on proxy support design document13:04
alecuBLOCKED: no.13:04
alecuNOTE: all my body aches now..! I should exercise more often. Or move once a week.13:04
alecuNEXT: gatox13:04
gatoxDONE:13:04
gatoxSome branch fixing. Improves in expand user.13:04
gatoxTODO:13:04
gatoxFinish expand user, work on freaky friday project13:04
gatoxBLOCKED:13:04
gatoxNo13:04
gatoxmandel, go13:04
mandelDONE: sent patch for http://bugs.python.org/issue13234 and written patch for http://bugs.python.org/issue13207.  TODO: 1 on 1 with ralsina. Fix mp for fix-reactor-import branch following nessitas review. Propose. Edge of delete password to pykeyring upstream. BLOCK: no13:05
nessitagatox: got my message for the pong?13:05
gatoxnessita, yep13:05
nessitagatox: great!13:05
nessitagatox: have an answer? lint is complaining about that for me13:05
nessitacomments? eom?13:06
gatoxnessita, ahhh i forgot to disabled to the lint issue about tools....13:06
gatoxnessita, doing it right now13:06
nessitagatox: ack13:06
mandelDONE: sent patch for http://bugs.python.org/issue13234 and written patch for http://bugs.python.org/issue13207.  TODO: 1 on 1 with ralsina. Fix mp for fix-reactor-import branch following nessitas review. Propose. Edge of delete password to pykeyring upstream. BLOCK: no13:07
nessitamandel: we got it the first time....13:07
mandeldid you guys get my last message?13:07
ralsinamandel: both messages today :-)13:07
nessitamandel: yeap, twice :-)13:07
* mandel hates irc at times..13:08
nessitamandel: you mean IRC in your mobile?13:08
nessita:-P13:08
mandelralsina, can we have the 1on1 in 30 mins?13:08
ralsinamandel: sure!13:08
gatoxmandel, cool about the proposal to python! :D13:09
=== m_conley_away is now known as m_conley
ralsinamandel, if you are trying to tell me something somewhere in IRC, I am *not* seing it. Just ping me after lunch ok ;-)13:15
gatoxnessita, done13:15
mandelralsina, ok..13:19
mandelI'll ping ASAP... mierda de irc13:19
nessitaalecu: ping13:29
alecunessita, pong13:30
nessitaalecu: hi there! I think this new test failure from sso in windows has been added from your latest branch: bug #87940713:30
ubot4Launchpad bug 879407 in ubuntu-sso-client "Test failure after token name change (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/87940713:30
mandelnessita, e13:37
mandelnessita, I meant to say yes.. stupid finger13:37
nessitamandel: yes to what? :-) I'm lost13:38
mandelnessita, nothing, I think that I'm not a human being between 2 and 3 pm..13:38
mandelnessita, ralsina, dobey, alecu, gatox would it be a huge PITA if we moved the standup 30 mins?13:39
alecumandel, earlier?13:39
mandellater, since it would be better for you13:39
alecumandel, \o/13:39
gatoxmandel, no problem for  me....13:39
nessita+1 for me, but may be too early for dobey13:39
mandelI'm trying to have lunch for free (mum cooking) and I'm failing to do the stand up at the same time13:40
ralsina+113:40
mandelnessita, I meant later, so it should be better for dobey, right?13:40
ralsinaif it's to may your mommy happy, agreed13:40
dobeylater13:40
mandelhurray!!!13:41
mandelralsina, is mainly for my wallet and my stomach, my mother does not care :P13:41
mandelralsina, 1 on 1, mumble?13:42
ralsinamandel: in 1'13:42
dobeynative american flute lullabies is probably not the best hacking music13:44
nessitadobey: would you have a couple of mins to see a backtrace and see if you know what's going on? is gi related13:45
dobeysure13:45
nessitagenerating...13:45
nessitadobey: I'm think is related to ssl... if that is even possible13:46
dobeyis it coming from u1trial?13:46
nessitahttp://pastebin.ubuntu.com/715127/13:46
nessitadobey: nopes, when running syncdaemon in a branch of mine13:46
nessitadobey: I made some refactoring, and the syncdaemon is working perfectly on windows. The whole test suite is fully green on both OSs13:47
nessitadobey: but syncdaemon gets that trace when transitioning to CHECK_VERSION13:47
nessitadobey: if I never connect the syncdaemon, I don't get the trace13:47
dobeyit has nothing to do with ssl13:47
nessitadobey: ok... any clues?13:48
dobeyit seems Notify is being imported from gi, after pynotify has already been imported13:49
nessitadobey: the trace is making syncdaemon completely die, it does not even respond to CTRL-C.... does that concur with your theory?13:50
dobeyis there some way in python to see exactly when things get imported?13:50
nessitadobey: not that I know of, but I never really looked13:51
dobeynessita: well, gi + static bindings can cause lots of problems, but the freeze could be unrelated13:51
nessitadobey: so, in ubuntuone/platform/linux/notification.py (code I never touched in this branch) has the conditional import for Notify13:52
dobeyyes13:52
nessitaI can't see how that can be importing the same from 2 different sources13:53
nessitaand grepping gives no other Notify import13:53
dobeynessita: something is probably importing notification, then importing launcher, then notification gets imported again13:53
nessitadobey: so, a quick test would be to comment out one of the branches of the import. IN natty I can try using only the Nofity from pynotify, right?13:54
nessitadobey: indeed, that solved the freeze13:55
dobeynessita: well, the gi import will never work, until twisted is fixed and we switch reactors13:55
nessitadobey: right13:55
nessitait works now!!!13:56
nessitanow I have to see how I can solve this13:56
dobeyand this is also why we don't have a progress bar on launcher in oneiric13:58
nessitadobey: any suggestion, before I start doing wild guesses and tests? :-)13:58
dobeytrace the import of ubuntuone.platform.notification up until you find some reason it would be imported multiple times (probably something getting imported within a function)14:00
nessitadobey: multiple imports is "easy".... if you import from  ubuntuone.platform.notification import Notify you get one import, and import ubuntuone.platform.notification gets you another14:01
nessitadobey: I mean, importing the same module but using different "ways" ("paths") will make the module be imported several times (and not reused)14:02
nessitadobey: I'll track this down, thanks!!!14:02
dobeynessita: right grep notification|grep import14:02
nessitayeap14:02
nessitaalecu: great news!14:02
nessitaalecu: btw, did you see my ping question?14:02
alecunessita, great news?14:05
nessitaalecu: my branch is 30 minutes away to be ready :-)14:05
nessitaalecu: is working almost perfectly on both OSs...14:05
* alecu books the afternoon for a review.14:05
nessitaalecu: I saw you disconnected, did you happen to see my comment re: a test failure in windows sso?14:06
alecunessita, yes, I saw it, and I've been trying to resurrect my windows vm, because it seems to be damaged14:06
nessitaalecu: I can help debugging if you need14:07
alecunessita, so now I'm setting up the desktop computer that's disassembled in the middle of the living room14:07
nessitaheh14:07
nessitagood luck14:07
alecu(with the other thousand bags of crap from the move)14:07
alecunessita, so, as soon as I have a working vm I'll test it, to see what I broke.14:08
nessitaalecu: no rush, just wanted the ack14:09
nessitaalecu: quick question, does a call to bus.request_name() starts the service?14:31
nessitathe service under the name being requested14:31
alecunessita, I usually do bus.get_object(), and that surely starts the service if not started.14:33
nessitaalecu: right, I'm trying to *not* start the service14:33
alecunessita, why?14:33
nessitaalecu: so I was wondering if request_name starts it or not.... I'm finding 2 implementations for is_running()14:33
nessitaalecu: in linux, we have 2 different implementations (is_already_running and is_running)14:34
nessitaalecu: in windows, we have 2 definitions of those methods, the former is implemented, the latter is not!14:34
nessitaalecu: so, SDTool is using the latter, and ergo, in windows we certainly start serveral instances of syncdaemon14:35
nessitaseveral14:35
ralsinanessita: this user is complaining ob things "not uploading" https://launchpadlibrarian.net/83394886/syncdaemon.log.report My current guess is "he added a whole lot of files at once (queue at 70K) and they are hidden files (.cache.noindex/whatever) and that's why he's not seeing uploads yet". OTOH, he says it keeps changing from "syncing" to "disconnected"14:35
dobeynessita: request_name() will either tell you the name is already taken, or give you ownership of it, iirc14:35
nessitaalecu: I'm trying to have only one implementation for each OS, and ideally, the correct one :-)14:35
nessitadobey: hum...14:36
nessitadobey: in u1client we have these 2 ways of checking is u1client is running http://pastebin.ubuntu.com/715154/14:38
nessitabin/ubuntuone-syncdaemon uses the first one14:38
nessitaSyncDaemonTool uses the second14:38
nessitawe're clearly failing on windows to check, from SDTool, if sd is running, so we're starting it multiple times14:38
dobeynessita: is_running seems wrong. should just change everything to use the is_already_running() method14:40
nessitadobey: I concur. But we need to be certain that is_already_running will not start SD if not running14:41
dobeynessita: it shouldn't, but is easy to test. shut everything down, and call it from python shell14:42
nessitadobey: you're so clever! and always making *me* work :-D14:42
dobeynessita: but i'm pretty certain it doesn't14:42
nessitadobey: I guess so, otherwise we would be starting u1 in current releases14:43
dobeynessita: dbus.bus.NAME_FLAG_DO_NOT_QUEUE means "don't give the name to me if it doesn't exist" iirc14:43
nessitanice, testing IRL now14:43
nessitadobey: confirmed what you said14:44
nessitaok, sticking to is_already_running14:44
nessitaalecu: just FYI for when you review the branch ^14:45
nessitaralsina: sorry, I think I missed a ping from you14:56
nessitaralsina: what bug report is that?14:57
dobeynessita, ralsina: as well he shouldn't be syncing files, being offline and all14:58
dobey2011-10-21 15:14:37,435 - ubuntuone.SyncDaemon.InteractionInterfaces - DEBUG - emitting state changed: SET_CAPABILITIES (error=False connected=True online=False)  Queue: WORKING  Connection: With User With Network14:58
ralsinahow can he be offline with network?14:59
dobeyralsina: ^^ he's probably constantly getting connection dropped from server as some people apparently still are14:59
ralsinadobey: makes sense14:59
alecunessita, so, are you adding the clean up for the activation functions to your megabranch too?14:59
nessitaalecu: yeah! is definitely part of the refactor (quit and start are unified now between OSs)15:00
alecunessita, so, are you keeping is_already_running and ditching is_running?15:01
nessitaalecu: yes sir15:01
alecugreat15:01
nessitaalecu: we should no longer have 2 syncdqaemon instances on windows (which we could easily have now)15:01
alecunessita, cool15:03
* mandel wonders wich OS is most used in japan.. can't be windows15:03
nessitadobey: question, why did pitti added the conditional imports for loading from gi if none of our twisted code works with gi?15:03
dobeynessita: because before the switch to the new gobject which broke the gi+static stuff, our stuff did work15:05
dobeynessita: the conditional isn't the problem you're hitting though. it just exposes the problem :)15:06
nessitadobey: gah15:07
dobeynessita: since it happens only in your branch, and not in oneiric proper, perhaps something you change, altered the timing of some imports, causing the issue?15:08
nessitadobey: yes, for sure15:10
nessitadobey: I've cleaned up a lot of code, among those, import clauses15:10
nessitadobey: I'm trying to find the cleaner solution for this, now15:10
nessitato do so I was trying to understand why we need the conditional import (besides looking for the "double" import, which so far I can't find)15:11
dobeynessita: we'll probably need the conditional to support both old and new Ubuntus, going forward15:11
nessitadobey: right, but right now we support nothing "new" since twisted will not work15:12
dobeynessita: but we have to fix that. and we do support it (on 11.04 at least, where it still works right)15:14
dobeynessita: and we don't want to have it break when we do get twisted fixed15:18
nessitadobey: right15:18
dobeynessita: did you move some imports into the functions that use stuff from them?15:19
nessitadobey: not really... but I think I found it15:20
dobeyok. i'm going to get lunch :)15:21
dobeybbiab15:21
nessitadobey: thanks!!!15:21
mandelnessita, https://code.launchpad.net/~mandel/ubuntuone-windows-installer/fix-reactor-import/+merge/79842 is ready for re-review, I update the tests in main since they did not make sense any longer and wrote tests for15:52
mandelsucces and error_cb15:52
mandelgatox, ping15:52
gatoxmandel, pong15:53
mandelgatox, did you mentioned the bug about your evil username?15:53
gatoxmandel, yep15:53
mandelgatox, cool15:53
gatoxmandel, no..... not so cool15:54
gatoxjejej15:54
mandelgatox, well, you know, cool that you found it15:55
gatoxjeje15:55
* mandel hates haveing to write ctypes code..16:00
alecunessita, a fix for the broken test: https://code.launchpad.net/~alecu/ubuntu-sso-client/fix-test-appname/+merge/8008416:09
nessitaalecu: looking!16:10
alecuand everybody else: trivialish review above16:10
mandelalecu, I hate when people use '' and "" in the same line ;)16:13
Chipacamandel: what about people who use " and """ on the same line?16:17
Chipacamandel: "like this"""16:17
mandelChipaca, that I don't mind as much, is always the same char :)16:19
mandelnessita, FYI I have just updated https://bitbucket.org/mandel/pykeyring-delete-password/overview to use the same code as the trunk from pykeyring, I'll be proposing it for merge so that we can depend on trunk and not that fork16:23
nessitamandel: great news!16:25
gatoxlunch..... i'm starving....... brb!16:25
nessitaalecu: approved16:41
mandelok, EOD for me have a great weekend16:41
mandelnessita, if you give a +1 to https://code.launchpad.net/~mandel/ubuntuone-windows-installer/fix-reactor-import/+merge/79842 feel free to add the aproval since is a critical bug16:42
nessitamandel: ack16:42
mandeleveryone, adios!16:43
nessitabye mandel16:43
nessitaalecu: branch proposed! https://code.launchpad.net/~nataliabidart/ubuntuone-client/u1sdtool-multiplatform/+merge/8008817:18
nessitano alecu? oh17:18
nessitalunchtime then17:18
ryeduanedesign, i have an idea17:34
ryeduanedesign, that's not gnome 3 issue17:34
duanedesignoh really17:34
ryeduanedesign, that's gnome-keyring-daemon erroring-out with unrelated error and sso or anything else simply fallbacks to the code that queries the old interface and therefore failing17:35
ryewe need to get the original exception17:35
ryeduanedesign, i will be back in 30 minutes and will start poking the gnome-keyring-daemon17:35
ryeduanedesign, we may get some info out of user's .xsession-errors, searching for e.g. keyring17:36
duanedesigngreat thank again sir17:37
ryeduanedesign, i will create updated sso package and we'll ask the users to test17:38
nessitaalecu: hey there!17:40
nessitaalecu: the branch is now proposed at https://code.launchpad.net/~nataliabidart/ubuntuone-client/u1sdtool-multiplatform/+merge/8008817:40
alecunessita, nice!17:40
nessitaalecu: I reported most of the bug things I fixed, and linked the bugs to it17:41
nessitaalecu: you can start looking at it, I will write some guide notes in the MP itself17:41
alecunessita, 22498 lines (+6824/-12688) 32 files modified!!!!!17:42
nessitaalecu: that's a lit17:42
nessitalie*17:42
alecuI love that it removes a lot more than it adds :-)17:42
nessitaalecu: you will *love* how stdool implementation turned out17:43
nessitaalecu: 90% of the code is shared multiplatform17:43
nessitaalecu: and the ipc interfaces are modified so is trivial replacing them with something better (a generator maybe)17:44
nessitaand I removed tons of mocker tests, adding regular tests instead17:44
alecuultra-mega-cool17:44
ralsinacongratulations on the branch nessita17:51
nessitaralsina: thanks! some weird issues will be fixed from now on17:52
ralsinayeah, quick scan, I like much more the green parts than the red ones! ;-)17:52
dobeythe red parts are more fun17:53
nessitaralsina, alecu: I made some heavy testing on both OSes, but anyways, once landed, we need to test it a lot before releasing17:53
nessitaalways17:53
* ralsina thinks dobey is thinking of budapest or something like that, and backs away slowly17:54
dobeylol17:55
nessitaalecu: commit message added, and some detailed description as well (mainly to run u1sdtool)18:01
nessitadobey: I will make the windows stable branch next week, I want to play a bit with magicicada now :-)18:14
dobeyheh18:16
alecunessita, I did a first part of the review, and added some notes.19:12
* alecu needs to run some errands now19:12
nessitaalecu: looking!19:13
gatoxEOD for me.... see you on monday19:41
=== yofel_ is now known as yofel
dobeyhave a good weekend all!21:14
=== gatox is now known as gatox_away
=== m_conley is now known as m_conley_away
=== caravone_ is now known as caravone
nessitaok, I'm done21:27
nessita(for the day and the week)21:27
=== gatox_away is now known as gatox
=== edvm is now known as x-ip

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!