[00:01] eod as well === gatox_tv is now known as gatox === gatox is now known as gatox_away [04:08] EODs, bye bye o/ === dduffey is now known as dduffey_afk === yofel_ is now known as yofel === zumbi_ is now known as Guest94149 === statik is now known as 64MAAQOA7 [09:24] morniing! [09:25] scarleo, I can give you a hand, but let me reboot first.. some updates need it :( [09:25] mandel, ok thanks [09:25] scarleo, np [09:25] * mandel reboots [09:33] scarleo, I'm back! [09:33] great! [09:33] scarleo, so which os are you running? [09:33] Ubuntu 11.10 [09:34] scarleo, and the probelm is that there are some images that are not syncing, is that correct? [09:34] yes, they never get the green tick in Nautilus [09:35] most sync but I have maybe 5 that doesn't [09:35] scarleo, hm.. interesting, are there any app that are using the images? [09:35] mandel, nope [09:35] scarleo, on linux we wait until you close the file so that we update the correct version.. [09:35] and they've been on the pc for quite some time [09:36] scarleo, ok, are they in ~/Ubuntu One or in a UDF, for example ~/Images [09:36] thay are in ~/Pictures/webbilder and Pictures is the folder I've set to sync === gord_ is now known as gord [09:37] scarleo, lets first see if it is indeed there, just in case :) [09:37] scarleo, run u1sdtool --list-folders [09:38] scarleo, and u1sdtool --status will also be useful [09:39] mandel, http://pastie.org/3353566 [09:39] * mandel looks [09:40] scarleo, if you do u1sdtool --info=PATH where PATH is one of the images, what do you get? [09:41] mandel, http://pastie.org/3353584 [09:43] scarleo, and u1sdtool --waiting-content ? [09:43] scarleo, I'm trying to find out in what state are those files.. [09:44] --waiting-content gives Deprecated, use --waiting. --waiting gives no output [09:45] scarleo, hm.. funny, let me see if one of the QA is there, but nothing seems to be bad with those files [09:45] mandel, I can see them when I log in to one.ubuntu.com [09:46] scarleo, have you check the web ui, are the files there? [09:46] :) [09:46] yes [09:46] scarleo, oh, so they where uploaded yest nautilues gives the wrong icon [09:46] mandel, yep [09:46] scarleo, I think that is a bug in the nautilus plugin then... :( [09:47] ok [09:47] scarleo, the developer that takes care of that is based in the USA, do you find filing a bug with all what we have looked at and explain that the files are ok in the web [09:47] also every time I login to my desktop U1 tells me it is syncing one image but it's already synced [09:47] scarleo, I'm sure we will take a look at that asap [09:47] ok, I will [09:47] scarleo, I think there was a bug with the notifications where it had 'memory' [09:48] ok [09:48] scarleo, AFAIK we are working on that, but we have been looking at porting to other platforms lately and have not had the time, sorry for the PITA [09:48] mandel, no problem [09:49] scarleo, also, add the logs and set the bug to private, we will make it public if there is nothing that could compromise your account or details about you [09:49] scarleo, better safe than sorry [09:50] mandel, which log and where is it? === zumbi is now known as Guest47393 [09:50] scarleo, let me check [09:51] ok [09:53] scarleo, should be here: .cache/ubuntuone/log [09:54] ah, mornings [09:54] mandel, ok, filing a bug report then [09:54] scarleo, thx, we really appreciate that! [09:55] np [09:55] scarleo, also, if you are here a little longer we can ask rye, he may know :) [09:55] rye, do you have back logs? [09:55] yeah I've got time [09:56] mandel, yes, i have them, looking [09:56] rye, thx! [09:57] scarleo, I leave you with rye he is great at finding problems like this [09:57] mandel, is it the syncdaemon log I should provide from .cache/ubuntuone/log [09:57] mandel, ok [09:57] Hi rye [09:57] * mandel goes back to be a code monkey [09:58] scarleo, re: already synced images - it is a bug in client display code [09:59] scarleo, reading further... [09:59] scarleo, what system are you now on? [09:59] Hmmm, the file always telling me it's uploaded every time I log in is in syncdaemon-exceptions.log [09:59] rye, Ubuntu 11.10 [10:00] h, launchpad down? [10:03] scarleo, the first is bug #807005 - what ubuntuone-client do you have - apt-cache policy ubuntuone-client and see the Installed version [10:03] Launchpad bug 807005 in ubuntuone-client (Ubuntu Precise) (and 7 other projects) "Filename in notification does not reset. (affects: 9) (dups: 3) (heat: 38)" [Undecided,Fix released] https://launchpad.net/bugs/807005 [10:04] rye, 2.0.0-0ubuntu2.3 [10:04] interesting [10:05] scarleo, ok, could you please download the script form http://people.canonical.com/~roman.yepishev/us/u1filestatus.py and run it in the terminal [10:05] rye, can I paste the syndaemon-exceptions.log ? The file that is noticing is in there but it contains some id's and stuff. Is it safe? It's to a shared folder [10:05] scarleo, it will show what files are not synced [10:06] ok, I'll download it [10:06] scarleo, download and run it as "python u1filestatus.py" [10:07] rye, http://pastie.org/3353677 [10:08] scarleo, hm, where are the files that are not synced located? [10:08] rye, they are synced, I can see them in web UI, they do not get the green tick in Nautilus. The files affected are in ~/Pictures/webbilder [10:09] scarleo, ok, this looks like the emblem issue, are other files marked as synced? [10:10] rye, yes, it's about 5 .png and one .gif that doesnt get the emblem [10:17] rye, so emblem issue is known? [10:23] scarleo, well, this part does not have extensive test coverage, I am booting my virtual machine with oneiric to see how it works now [10:27] Oh, and good morning, world! Happy Friday! :) [10:28] meh, you people changed API for u1sdtool 3 times, the 3rd time breaks my u1filestatus in precise. boo, [10:28] oh, i know, i just propose it for merging :) [10:53] hi! [11:10] gatox, morning! [11:11] mandel, buenas [11:11] gatox, in a few mins I'll have lots of reviews to be done trying to land a feature today, mind if I bother you with the later? [11:11] mandel, no problem!! [11:12] gatox, great, is an good amount of code, but is just lots of branches very small [11:17] buenos días! [11:18] nessita, buenas [11:19] hola gatox, how is it going? [11:19] will review now! [11:19] nessita, fine..... finishing with u1-client-network-detect [11:20] nessita, good morning, I did a code review for https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/move-ping/+merge/92378 and looks great [11:20] nessita, let me run the tests and I'll approave, I completely forgot to do that [11:22] mandel: awesome!!! [11:26] gatox: oh, a branch of mine yesterday landed and your branch has one conflict (should be easy to fix: [11:26] Text conflict in ubuntu_sso/utils/ui.py [11:26] 1 conflicts encountered.) [11:26] nessita, ok..... fixing that right now [11:29] nessita, fixed [11:29] gatox: thanks! [11:40] gatox, what is orange and sounds like a parrot? [11:40] mandel, carrot [11:40] i read your twits [11:40] jejeje [11:40] gatox, heheh [11:40] mandel, i abstain to do any comment about that REALLY BAD JOKE! [11:40] jeejej [11:41] is not THAT bad hehe [11:43] gatox: APPROVED [11:43] nessita, just like that?! [11:43] jejeje [11:43] YEAH [11:43] AWESOME! [11:43] :-) [11:44] ralsina_: hola! would you be around? [11:46] nessita, FYI +1 on https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/move-ping/+merge/92378 [11:46] yey! [11:46] mandel: thanks [11:59] nessita: yes, here I am [11:59] mandel: what's brown and sticky? [11:59] ralsina_, I should know... [11:59] mandel: A STICK! [12:00] ralsina_, and they say my jokes are bad.. [12:00] mandel: it's from a movie. Hollywood says it's funny! [12:02] ralsina_, he.. === ralsina_ is now known as ralsina [12:03] ralsina, che, for you wife in a few days: sqrt(cos(x))*cos(300x)+sqrt(abs(x))-0.7)*(4-x*x)^0.01, sqrt(6-x^2), -sqrt(6-x^2) from -4.5 to 4.5 [12:03] ralsina, super super gay.. [12:04] I refuse to figure out how to graoh that ;-) [12:04] graph [12:04] graficar [12:04] blah [12:04] ralsina, is good, trust manuel.. [12:10] Ok, got it. :-) [12:12] ralsina, see, is nice :) [12:12] ralsina, and geeky, so double points [12:12] mandel: wife is a lawyer. geeky is -.25 points [12:13] If I were to give her a poster of Mazinger-Z and Aphrodite-A on the beach, for instance, she would hate it. Can you believe it? [12:13] pufff... [12:15] ralsina: hey there! not sure if you saw my messages yesterday evening about a couple of reviews needed [12:15] ralsina: one from me and one from gatox [12:15] nessita: I did yours, gatox is taking a bit longer [12:15] ralsina: ah, great! [12:16] ralsina, we can do the 1-1 whenever you want today [12:16] did you vote in it? (this morning had no vote) [12:16] nessita: in fact, your branch has 2 now, so you can approve it [12:16] nice [12:16] ralsina: thanks! [12:16] gatox: I am going to take a short break to walk while P updates, let's do it in 30' [12:16] * ralsina walks himself [12:16] ralsina, roger that [12:18] ralsina: and when you have a free slot, I desperately need help with the qt spawner. Somehow, the instance of the QProcess is garbage collected so is not working [12:19] ralsina: I've tried holding to that instance adding classes and all, but there is no way [12:19] nessita: point me at the code, I will take a look immediately [12:19] ralsina: so, current version (without adding classes) is lp:~nataliabidart/ubuntu-sso-client/execute-ui [12:20] ralsina: once you branch that (tested on linux only for now) [12:20] please remove your U1 token and run this command in the branch: [12:20] USE_QT_MAINLOOP=True U1_DEBUG=True PATH=bin/:$PATH PYTHONPATH=. bin/ubuntu-sso-login [12:20] nessita: ack [12:23] ralsina: once the sso service is running (you will get output in the terminal), you can open the gtk controlpanel and click on "I already have an account". Do not login, just click on "I've forgotten my password" and put this email in the field: a@example.com [12:23] when you click the default button, you should get this trace in the terminal: [12:24] http://pastebin.ubuntu.com/836470/ [12:24] ralsina: bah, a little different since I have some local changes where I put some stuff in classes to try to hold to the objects a little longer (does not work) [12:25] nessita: got it [12:25] the trace? [12:25] nessita: yes [12:25] ralsina: awesome! [12:25] nessita: ok, I'll try to see if something comes to mind [12:25] ralsina: I'm up for any chanchada to resolve that. Is driving (more) crazy [12:25] nessita: MY SPECIALTY! ;-) [12:25] hehe [12:26] ralsina: ok, I will move to some reviews to put some air in my mind and give you time to "fix" it :-D [12:26] nessita: cool [12:27] holas! [12:27] hola alecu [12:27] nessita: the version of spawn_program I have from bzr will never work because indeed process will be GCd right after start() [12:28] ralsina: yes, I can push you what I have now [12:28] nessita: let me try something first [12:28] ralsina: I have a couple of classes and hold to the references, but they disappear the same :-/ [12:28] ack! [12:28] I will make some mate for now [12:33] nessita: working proof of concept chanchada: https://pastebin.canonical.com/59923/ [12:34] nessita: if you push what you did with classes I can take a look to see if I can do something less nasty [12:34] nessita: I would even rather leak the QProcesses (since SSO dies a little later anyway) [12:36] ralsina, nessita: what if we have a "processmanager" singleton, that holds a dict of active processes, and has a "remove_process" method attached to process.finished? [12:36] alecu: same idea, except I am using a list ;-) [12:37] alecu: why a dict, what would be the key? The qprocess itself? [12:37] ralsina, right: I'm using something similar for QTCPsockets [12:37] (or whatever they are called) [12:37] This (GC of some objects) is the area where Qt and Python have some impedance mismatch [12:38] ralsina, definitely [12:38] ralsina, well, looking at your code it seems that a set() would do better than either a dict or a list. [12:39] alecu: and just connect lambda: thing.remove(process) [12:39] ralsina: I like your solution, is much less chanchada than I imagined [12:40] nessita: use a set instead of a list, remove remove_process and it's not a chanchada at all [12:40] ralsina: will push what I have so you can please explain to me why this last thing I tried is also not wokring [12:40] nessita: happy to [12:41] * ralsina orders a "Not bad for an old ex-nerd" t-shirt [12:41] ;-) [12:41] ralsina: Pushed up to revision 865 [12:42] ralsina: yes, I know that the class Runner in qt would be buggy if spwan process is called more than once, but was just trying stuff [12:43] nessita: my guess is that the Runner from __init__ was also being GCd [12:43] nessita: where were you instantiating that from? [12:44] ralsina: credentials.py:148 [12:44] ralsina: I'm holding the ref there [12:44] since _show_ui will not finish until runner.spawn_program finishes [12:47] nessita: problem with my solution: there is no guarantee about the order in which things connected to finished will be called. So probably better to just leak the QProcesses [12:48] ralsina: what's the problem with the order? is I call set.remove... order does not matter, no? [12:48] nessita: you are not connecting anything else to finished? In that case: OK! [12:48] ralsina: ah, yes, I am. I *now* understand [12:49] I can remove the process from the set in the callback I already connected to finished [12:49] you can remove it from processes in your call, though [12:49] right [12:49] exactly [12:49] ok, will revert all this classess that add noise [12:49] ralsina: thanks!!! [12:50] nessita: maybe it's getting GCd because the calls to finished will only be processed after it reaches the qt mainloop. Perhaps your deferred is triggered before that, and then _show_ui ends, runner gets GCd => crash [12:50] nessita: but it's just a guess [12:50] nessita: you're welcome! [12:50] ralsina: well, the deferred is fired when the finished callback is executed [12:51] nessita: it's a bit too complicated for me to guess quickly, I suppose [12:51] no problem [12:51] I will use your solution [12:52] nessita: cool. I will be back in 30' then :-) [12:52] ack [12:55] IT WORKS [12:55] I CAN'T BELIEVE IT [12:55] * nessita cries, a lot [12:56] ralsina, Why was 6 scared of 7? [12:56] mandel: because seven ate nine [12:56] ralsina, hehehe [12:56] mandel: what did 0 say to 8? [12:57] what? [12:57] mandel: "nice belt!" [12:57] ouch [13:05] nessita, where was the spawn method in sso? I cannot remember.. [13:05] mandel: ubuntu_sso/utils/runner. The API is about to change wiuth a branch of mine, but is working now! [13:05] nessita, ah, runner ok :) [13:06] Morning #ubuntuone ! [13:06] hola karni! [13:06] hola nessita ^_^ [13:07] nessita, if I have a command line such as --app_name name --domain domain it should be passed as ('--app_name', 'name', '--domain', 'domain') is that correct? [13:07] mandel: yes sir [13:07] mandel: and the runner will escape those accordingly [13:07] nessita, superb! I think we are going to get there on time :) [13:08] :-) [13:33] gatox, nessita, ralsina simple review of the collection of reviews coming today: https://code.launchpad.net/~mandel/ubuntu-sso-client/support-user-name-url/+merge/92467 [13:34] mandel, on it [13:34] mandel: ack [13:34] mandel: I would advice having one alecu's review on each one of this [13:34] nessita, ack [13:34] mandel: I would also advice one mine, and one alecu's [13:34] mandel: you think that may complicate the schedule? [13:35] nessita, no, if we do it correctly :) [13:35] nessita, we can always work around that, not a big deal [13:35] mandel: ok, I will try to be as responsive as possible with this [13:35] nessita, ralsina I think I have one more extra hour to finish the connection between the webclient implementations and the dialog (that means using spawn_process and writing tests) [13:36] mandel: so, is it ok if gatox keeps working on his huge refactor and me and alecu do the review? [13:36] nessita, ralsina I'll do after lunch (going now) and then will have the entire afternoon to fix the code according to the reviews, that gives me 3/4 hours just for this [13:36] nessita, certainly, we can keep gatox busy with other stuff, as long as I have the reviewers I'm happy [13:37] mandel: ack [13:37] nessita, FYI the pipeline is as follows: support-user-name-url [13:37] creds-dialog [13:37] keyring-integration [13:37] creds-dialog-script [13:37] load-creds-retry [13:37] * webclient-use-dialog [13:37] mandel: where is support-user-name-url there? [13:38] nessita, paste fail, is the first :) [13:38] nessita, after follow: [13:39] nessita, each of them is small, so it should not burn you brain out, all merges add appropiate tests so they can be merged without screwing up trunk [13:39] nice [13:39] ok, super quick lunch for me so that we land this guy today :) [13:39] * mandel lunch [13:54] wow. twinview on linux is apparently horribly broken :-/ [13:58] howdy! [14:00] hola ralsina [14:09] gotta reboot, brb [14:19] so [14:20] i am running a python script under gdb and it's crashing inside an assembly file in libc; anyone have any clues how to get some deeper debug info about the assembly? :) [14:22] dobey, strace? [14:24] hrmm, maybe [14:24] but probably not === dduffey_afk is now known as dduffey [14:31] alecu: can you send me a direct message on twitter real quick to test gwibber with? [14:40] crowd, ralsina says he computer will not boot [14:40] he's fixing [14:41] :S [14:44] dobey, sent. [14:44] * mandel back [14:45] * alecu is otp [14:47] alecu, ping [14:48] mandel, pong [14:48] dobey: is it working? :-) [14:49] alecu, got a question in the _on_authenticate method on the libsoup implementation, what does the try mean? [14:49] nessita: not quite yet. seems to not be refreshing [14:49] boo :-/ [14:49] And back, it was just a 40 minutes fsck [14:50] ralsina: yeey! k=just in time for a lovely branch [14:50] nessita: hit me! [14:50] https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/dont-let-it-go/ [14:51] ralsina: ^ [14:51] nessita: on it! [14:51] mandel, hhmmmm [14:51] mandel, looking [14:52] alecu, got it: http://developer.gnome.org/libsoup/stable/SoupSession.html#SoupSession-authenticate [14:52] alecu, and it means what I want it to mean :) [14:52] mandel, oh, you mean "retrying", not "try" [14:53] alecu, sorry.. [14:53] mandel, as per your question, I was looking for a "try... except" [14:53] alecu, jesus.. I'm stupid [14:53] :-) [14:53] alecu, understandable [14:53] mandel, keep praying, he may fix that. [14:54] alecu, so, I'm close to be done with the dialog, I just have an issue with it, the proxy creds are being saved as '%(appname)s porxy creds for %(domain)s' do you think that is correct? [14:55] alecu, or should I just go for the domain and ignore the app_name that was used to access sso? [14:55] alecu, I'm considering just having the domain since the app name adds nothing useful [14:55] mandel: you can use the app_name if you want proxy credentials by app [14:55] mandel, how do other applications name them? [14:56] if you want them to be "global", we should drop the app_name [14:56] mandel, chrome at least saves them in the keyring, right? [14:56] nessita, alecu I think global seems reasonable [14:56] @ping [14:56] pong [14:56] mandel, both "global" and "compatible with other apps" is reasonable :-) [14:56] alecu, chrome does not, empiphany browser does, but in a very ugly way [14:57] launchpad is painfully slow today. I suppose it's to be expected [14:57] mandel, also: this sounds like something fixable on a bug, so don't worry too much right now. [14:57] alecu, nessita yeah, I was going to say exactly the same, we send the feature, add a bug if the creds name is wrong [14:57] great [14:57] makes sense, right? [14:58] right [15:00] me [15:00] me [15:00] me [15:01] mandel, dobey, ralsina, gatox: stdup? [15:01] me [15:01] me [15:02] me (no notes) [15:02] oh meh [15:02] dobey: meh? [15:02] hehe [15:02] DONE: weekly call, micro sprint with gatox, debugged execute-ui branch, proposed a couple of priori fixes for that, FF talk with dobey and joshua, used some alecu's brian cycle that were excellent [15:02] TODO: Have execuet-ui working (IS ALMOST THERE), propose the branch and have some rest [15:02] BLOCKED: nopes [15:02] NEXT: alecu [15:02] DONE: worked on a branch for bug #929207 [15:02] TODO: reviews as needed, more work on bug above [15:02] BLOCKED: no [15:02] NEXT: briancurtin [15:02] Launchpad bug 929207 in ubuntuone-client "Proxy "tunnel" for syncdaemon (affects: 1) (heat: 6)" [High,Confirmed] https://launchpad.net/bugs/929207 [15:02] i was hacking [15:02] DONE: fighting through tests, updated merge proposal, more fighting with tests to be sure i gave roberto the right branches [15:02] TODO: fight through more tests, try to isolate why this is killing everything, try to come to some determination on bug #820350 [15:02] BLOCKED: none [15:02] NEXT: mandel [15:02] Launchpad bug 820350 in ubuntuone-client "WIndows: either set_dir_readwrite and/or set_dir_readonly are not doing what they should (affects: 1) (heat: 7)" [High,In progress] https://launchpad.net/bugs/820350 [15:02] DONE: Gone through the entire code of the dialog for creds, got to the point where the only last bit necesary is the connection with the webclients [15:02] TODO: connect the webclients, propose all the diff parts of the pipeline and fix accordingly. [15:03] BLOCKED: no [15:03] next, gatox [15:03] DONE: [15:03] A huge refactoring in SSO, fixed u1-client-network-detect branch [15:03] TODO: [15:03] Finish with more refactoring in SSO, adding signals to the pages to avoid the pages to know about the wizard [15:03] BLOCKED: [15:03] No [15:03] ralsina, go [15:03] DONE: weekly call, thought a lot (hey, it takes time), reviews, misc. TODO: reviews, think a bit more, finish setting up better Windows VM, P VM, BLOCKED: no [15:03] briancurtin: wanna pause what you're working on and move to another branch, until next week where I can put some time in helping you? [15:03] briancurtin: to another bug*, I meant [15:03] dobey [15:04] nessita: yeah i can do that [15:04] λ DONE: team call, nightlies fixes, gwibber poking [15:04] λ TODO: finish gwibber pokery, twisted docs, move more code around [15:04] λ BLCK: none. [15:04] briancurtin: I know the frustration feeling and is not nice :-) [15:04] briancurtin: and sometimes syncdaemon tests have some dark corners [15:04] any comments anyone? [15:04] comments: nessita, mandel: do you guys need reviews? [15:05] alecu, I have noticed that we are using http://developer.qt.nokia.com/doc/qt-4.8/qnetworkaccessmanager.html#authenticationRequired instead of http://developer.qt.nokia.com/doc/qt-4.8/qnetworkaccessmanager.html#proxyAuthenticationRequired in the qt implementation, that looks like a bug, right? [15:05] alecu, yes, let me get you one [15:05] alecu: I will before going to lunch. THE LAST BRANCH to complete the spawnning thing! [15:05] alecu, https://code.launchpad.net/~mandel/ubuntu-sso-client/support-user-name-url/+merge/92467 [15:06] all: in theory, I will not be working on Monday, but due to FF, I will pass by to see how everything is going. Or perhaps I just swap that day to another day, need to change some scheduls [15:06] mandel, it's not a bug. It's you looking at your belly button! [15:06] alecu, why? [15:07] alecu, nada, olvidate, soy tonto [15:07] alecu, I'm just going to fast.. [15:08] mandel, I think we'll end up using both. [15:08] alecu, by what i read very very quickly, you get the first the proxy and from that point on the other [15:08] mandel, since we may need the first one to authenticate against the destination server, and the second one to authenticate to the local proxy. [15:09] mandel, right, in the order you just said. [15:20] nessita: not important, don't change it, but I so much prefer if things like failed_to_start = process_error == process.FailedToStart have () arount the right-side of the assignment [15:21] ralsina: can change that :-) [15:21] I agree [15:21] nessita: cool, thx [15:22] nessita: +1 anyway [15:22] great! [15:23]