[00:17] mmcc: nice branch [08:03] Good morning all! :D === eu is now known as Guest16925 [11:44] hello, all! [12:21] hello alecu [12:21] hello mandel as well I assume! [12:22] hola ralsina :-) [12:25] alecu: good news mandel is going to the nux sprint [12:31] ralsina: that's great! has he already left for the sprint? ;-) [12:32] alecu: it starts monday so I hope not! [12:32] * alecu is green with envy... [12:32] I just love London on summer :-) [12:35] ralsina, alecu, hello, yes I was out with the dog for a few mins, he need to pee :P [12:36] and I'll go there if the people from the agency do answer my emails.. [12:37] alecu, I've been using your branch to test the generic preview idea, I think is great to start with, but what about encapsulating that in a vala class that extends generic view adds the info and exposes properties for the info we have to pass? [12:40] mandel: I lost connection for a bit, so I don't know if this got through: if the agency takes too long, just do orbitz and ask for a refund [12:40] ralsina, yes, that is my idea.. they are not very good [12:40] mandel: I would take the trains on saturday with my kindle, but that's just because I love trains :-) [12:41] ralsina, I'm done the deal of manchester-madrid by train, not that great.. [12:49] lunch time here [12:51] hrmm [13:13] ralsina: I'm looking at your branch that needs review. I see that most of the code being added touches code that already has tests, so I was wondering if you are able to add some? [13:13] ralsina: in the past we had no issues with submiting huge branches to FFe, if the "huge" in the branch was tests. [13:14] ralsina: also, since most of the branch is .ui changes, I think the tests won't make it grow so much. [13:22] alecu: ok, let me think it through a bit. [13:44] hi guys. I'm unable to move a folder inside my Ubuntu One dir, after "mv" the folder apears again in the old place [13:44] some ideas? [13:47] if I dosconnect the daemon everithing is ok, but on connect it moves the folders on the original position [13:49] looks like my internet is really flaky today [13:49] alecu: doing the tests, somehow I missed where I had to add stuff. There is one bit that still won't have them (the change in the starting page logic) but that's much smaller and I will do a bug for it (mostly because I don'thave much coding tie for this today) [13:52] brb, need to reboot. X has gone a bit wonky [13:54] hey all, ralsina says his internet connection has gone down completely. [13:55] DNX: that's really weird. [13:55] rye: have you seen behaviour like DNX says above? [13:58] alecu, also I noticed that it generates some conflicts between these folders [14:00] DNX: could you please check whether that folder already exists at that place @ https://files.one.ubuntu.com ? [14:01] alecu: not really, but I was not testing for this [14:01] rye, yes, the folder has been correctly syncronized [14:02] DNX: but it is missing at that location locally, right? [14:02] but something is wrong with these folders, I can't reproduce this situation again [14:03] rye, yes [14:06] alecu: ^ looks like remote and local roots diverged [14:07] alecu: so the folder is in fact there somehow at the server... [14:07] DNX: what's the last update date for that folder on the server? [14:08] rye, 31 minutes ago [14:11] and my changes doesn't apear on the server :| [14:12] I'll try to move this folder outside Ubuntu One dir and then paste it here again [14:13] can I proceed or you need some other details? [14:14] hm\ [14:14] alecu: so ralsina has no internet, and we're waiting for him to add more tests to that branch? :( [14:15] alecu: i just found a folder on the server that was not present locally. Created the folder and renamed it locally and the one at the server has replaced the server-side folder with the empty local one. I feel that's not how it was supposed to happen [14:16] rye: that indeed sounds like a bug [14:16] alecu: can you please check Move() checks that the destination does not exist already? [14:17] alecu: because this bug #987767 suggests it does not for rest api and if it is using the same code path then for SD too [14:17] Launchpad bug 987767 in Ubuntu One Servers "[rest] Renaming a file into a folder overwrites the folder" [Undecided,Triaged] https://launchpad.net/bugs/987767 [14:25] rye: in the client, once the Move() command is created and scheduled in the action queue, no such checks are done. [14:25] rye: it's up to the sync state machine to decide that. [14:25] rye: so, I'll check that too. [14:35] alecu: well, I am more worried about the server part [14:35] rye: in that case we should ask facundobatista or verterok [14:38] facundobatista or verterok, is Move() operation checking that the destination does not already exist? [14:40] rye, isn't Move a "overwritting operation"? [14:40] facundobatista: for files, it is. [14:41] rye, locally you do "touch a; touch b; mv b a", this overwrites, in the server it should overwrite too [14:41] facundobatista: for directory [14:41] rye, ah! don't know [14:41] facundobatista: one directory was moved over another one w/o any error here [14:41] facundobatista: for directories, there are no unix semantics; most file managers ask you if you want to merge. [14:42] rye, alecu, how did you move one dir over other in the client? [14:42] "mv file directory" typically moves the file in the directory; the problem is that the dir doesn't exist locally on u1 [14:42] rye, alecu, need to pickup my kid from kinder, brb [14:43] and "mv directory file" will fail [14:45] also "mv directory directory2" moves it *inside* the second dir, not overwritting it. [14:45] yes [14:46] and "mv file directory file2 file3 directory2 directory3" will move all the previous arguments into directory3 [14:46] the problem here is that we aren't dealing with local files (because it didn't get synced for some reason) [14:47] dobey: alecu / $ move blah.txt targetDirectory [14:48] dobey: alecu in cmd_client makes blah.txt inside targetDirectory... [14:48] rye: yes; locally, if targetDirectory exists and is a directory, mv will move the file into it; if it doesn't exist, or is a file, mv will place it there [14:49] either way, there is clearly a bug in u1 if it overwrites a directory with a file [14:50] dobey: it seems to be overwriting a directory with a different directory [14:50] dobey, alecu: pushing branch before the internet goes away again [14:50] ralsina: great, thanks! [14:50] great [14:50] ralsina: what about team call in 10 min? [14:51] ralsina: mumble, hangouts, or postpone? [14:51] alecu: there are a couple of missing tests I would like to have for sso_wizard.py and a comment about a couple of flaky asserts that are commented [14:51] alecu: I will file bugs for both [14:51] dobey: postpone until tomorrow [14:51] ralsina: great, thanks [14:51] ok [14:51] because I don't know how long I will have connection :-( [14:51] alecu: pushed [14:52] aha! [14:52] well if alecu can approve your branch, perhaps i can at least get that FFe done today; and possibly get the icons FFe done as well [14:52] dobey: that would be awesome [14:55] ralsina: can you move the commented lines after "There is a problem here, these oscilate 0/1" to a different test method, and use that string as the skip reason? [14:55] alecu: facundobatista dobey reproduced with cmd_client - if i prevent it from getting the target folder id the directory will freely overwrite anothe directory and SURPISE move blah.txt targetDirectory will make "targetDirectory" into file [14:55] i mean that targetDirectory becomes a file with blah.txt contents [14:55] So client can do weird things... [14:56] but that's still not something DNX experienced, I was not getting any conflicts [14:56] alecu: I can try if arnet lets me ;-) [14:57] ralsina, standup in mumble, right? [14:59] mandel, mmcc, briancurtin, thisfred: no mumble today: it's been moved for tomorrow [14:59] (since ralsina's connection is going haywire) [14:59] alecu, oh, thx fort the info [15:01] ralsina: if you prefer to fix that skip bit in a different branch I can approve like it is. [15:01] me [15:02] me [15:02] me [15:02] me [15:03] me [15:04] me [15:04] mandel: go! [15:04] DONE: Worked on the dash preview factory to detect that is that GenericPreview has an specific title then return a u1paymentpreview. Got to have the daemon running showing always the payment dummy preview for not present songs in the system. Played a little with nux and layouts. Sorted out flights etc.. for the DX sprint in london. [15:04] TODO: More on the preview UI work. Understand better nux to do some interesting stuff. Should be able to have a example preview working for tom. [15:04] BLOCKED: no [15:04] thisfred, please [15:05] hola? [15:05] thisfred [15:05] DONE: u1db server noodling TODO: turn noodling into code BLOCKED: no NEXT: alecu [15:05] DONE: discussed search api details with server guys, and lens-dash interfacing with mandel. Coded a sample lens branch so mandel can move forward with dash ui. forgot about reviews, doing some now. [15:05] TODO: make lens use real server apis [15:05] BLOCKED: no [15:05] NEXT: briancurtin [15:05] DONE: some small windows testing, working on HTTPSServer implementation [15:05] TODO: fix squid test on 3, HTTPSServer [15:05] NEXT: dobey [15:06] DONE: icon magic, music store poking, file sso FFe, review sso FFe fix [15:06] TODO: release/package icon magic [15:06] BLCK: None. [15:06] DONE: fixed sync indicator, fought with windows [15:06] TODO: fseventsdaemon [15:06] BLCK: no [15:06] NEXT: no [15:06] NEXT: ralsina! [15:06] (as long as he is still connected) [15:07] Ok, time for comments [15:07] * mandel finds cpp ugly [15:07] and I always forget the bloody ; [15:07] mandel: and how long does it take to compile? [15:08] mandel: a little less than "forever", right? [15:08] alecu, hehe yes, is kinda slow but the big tower can deal with it.. in london with the laptop is going to be a diff story [15:09] apparently I am connected but I have no notes [15:09] alecu: I will be pushing with references to the missing test bugs & the skip in a minute [15:09] do you need to twitter shame yourself? [15:09] mandel: distcc ftw! [15:09] briancurtin: probably! [15:10] mandel: kde invented a whole thing called icecream to help build large C++ projects on sprints [15:10] ralsina: I thought the romans invented icecream! [15:11] alecu: well, the name is because at first it took so long they went out for icecream [15:11] http://en.opensuse.org/Icecream [15:12] it's like a communal distcc [15:19] smells like socialism to me [15:21] thisfred: typical germans [15:21] communicable distcc [15:24] stdcc [15:24] briancurtin: unless alecu is already reviewing this, can you take a look at: https://code.launchpad.net/~mikemc/ubuntuone-control-panel/fix-sync-status/+merge/122974 [15:24] mmcc: sure, will look shortly [15:24] thanks [15:25] mmcc: I had that tab opened, was about to review it shortly... [15:26] briancurtin: don't worry, I'll do it. [15:26] alecu: thanks [15:33] lunch time; bbiab [15:33] mmcc: did you see mfoord's mail about the u1 mac client? [15:33] mmcc: I think having no dock icon makes sense in general for us [15:33] ralsina: yes. if we want to do it that way, it's a very minor change [15:34] mmcc: just putting it out there for you to think about it. If you can reply to it, awesome+ [15:34] ralsina: ok. [15:39] ralsina, mmcc, I think the dock icon should be seen when the control panel is opened, if it is not don't show it [15:40] also our interaction with the dock is quite bad [15:47] mandel: yes, there's a bug about not re-opening the main window when the dock icon is clicked. we could solve that bug by not having a dock icon :) [15:48] mmcc, yes, I wonder what happens if a user has several spaces and does not have a way to know the cp is open [15:48] however, we can either have the dock icon always or not at all. either the dock icon is always shown for as long as the controlpanel is running (even if the window is closed) or never… [15:48] mandel, the status menu is shown in all spaces [15:49] mfoord's suggestion was to be just like dropbox and skydrive, ie - just have the menu item [15:49] mmcc, yes, is a matter of sending the user to the right space if the cp is already opened [15:51] mandel: sending them to the right space when? I'll have to test this, but I assume that when you do something that results in a window.raise_() , it'll do something sane so you can use the window… [15:51] mmcc: on your branch, the status_changed_handlers property can probably be removed, since it does not seem to be used anymore. [15:52] mmcc: when activating by clicking on the menu thingie I guess [15:52] alecu: it's used in tests… I couldn't convince myself that those tests were unnecessary, so I left it in. let me find where they are [15:52] mmcc, you have cp in space 3 and the go to the icon menu to launch it you should be sent there [15:52] ralsina, exactly that [15:52] mmcc: and raise_ is half-buggy on windows and ubuntu, I would not expect it to be perfect on mac :-( [15:52] ralsina: yes, on osx it's like "raise_() then show()" or something [15:53] mmcc: exactly. And sometimes activate() [15:53] mmcc: I see that most tests directly access the underlying list (_status_changed_handlers). Only a few use the property... [15:54] dobey: I've approved ralsina's branch, so it should be landing soon. [15:54] alecu: thanks! [15:54] ralsina: thank you for the tests :-) [15:55] alecu: np, they had to be there, and arnet gave me time ;) [15:55] alecu, yes in fact now that I look at it, the actual backend property is never used, and the tests that use the property are using the version from the MockBackend class, which is annoyingly near complete [15:55] mmcc: right! [15:56] mmcc: so the mock is a copy of the thing? [15:56] alecu - so I'm going to remove it from the real backend and tweak those tests that use the mock [15:56] * ralsina remembers why mock sucks [15:56] ralsina: well, at least for this particular bit, the real function is so small that the mock is basically the same (but missing an IPC call, so it's still useful) [15:56] ralsina: I think that's a custom mock, not the "mocker" module. [15:56] ok [15:57] yes, mocker has been nearly killed in our code [15:57] ralsina: anyway, that was my first reaction too :-) [15:57] we have a bunch of places where the fake is >= than the original [15:57] mandel: I nearly don't trust that statement. [15:57] but better than not having a fake I guess [15:57] mandel: :-) [15:58] alecu, hehe at least the test we have been doing recently, I do no some places that are dominated by mocker [15:58] mandel: I would pay actual money to get rid of it. Everytime I had to fix a mock test it was easier to rewrite it. [15:59] ralsina, maybe in another life when we have the time :) [16:00] mandel: maybe there is no other life, but a hell forever fixing mocker tests! [16:00] THE HORROR [16:00] where all the bad kids go.. hehe [16:00] "wrong number of calls to method flogMeWithAPineApple" for eternity [16:02] better than C++ template type errors [16:08] mmcc: when I coded for KDE I once got a two-page template error. It meant "there is one space too many to the left of your third >" or something similar. [16:09] ralsina: nice! (wow) [16:12] anyone know why this hasn't merged: https://code.launchpad.net/~mikemc/ubuntuone-fsevents-daemon/commserver-tests-2/+merge/122741 [16:13] let me check [16:13] mmcc: looking... [16:13] mmcc, fixed :) [16:13] missing comment [16:13] mandel, aha thanks. why doesn't it complain? [16:14] mmcc, the tarmac gods... dobey ? [16:14] and do we need to toggle the status again to make it notice? [16:14] mmcc, AFAIK we don't but it would not do any harm [16:15] mmcc, did it plus an other +1 :P [16:15] mandel: good. it's important that you approve of your own code :) [16:16] mmcc, I know hehehe [16:16] mmcc, I'v got a big ego :) [16:20] I clearly don't know what to say to that [16:23] mmcc, is more in the latin meaning than the eng one, although they are close [16:37] ping alecu, just pushed your suggestion: https://code.launchpad.net/~mikemc/ubuntuone-control-panel/fix-sync-status/+merge/122974 [16:37] mmcc: thanks [16:42] mmcc: done [16:42] alecu, thanks [16:55] mandel: i presume you meant "commit message" rather than "comment" there? [16:55] dobey, yes, sorry [16:55] you didn't need to vote for that :) [16:59] dobey, having fun :) [17:01] well, rugby time, I'll catch you all tom! [17:05] bye mandel! [17:39] * briancurtin lunch [17:39] dobey: want a rubberstam on https://code.launchpad.net/~dobey/ubuntuone-client-data/more-updates/+merge/123131 or are we supposed to do real code reviews? ;-) [17:40] ralsina: not yet [17:40] ralsina: found some annoyance i'm trying to fix [17:41] dobey: ack [17:41] what's the right regex to compress multiple blank lines into a single blank line? [17:42] dobey, is ".art" in that COPYING.art a standard naming convention? why not COPYING-ART.txt so e.g. OS X knows how to open it? [17:42] just curious, really [17:45] mmcc: it's follows the common convention of using COPYING.lesser or COPYING.lib for the LGPL exceptions to the GPL, and such [17:45] dobey: ah, ok [17:45] it's COPYING.icons in ubuntuone-client currently for example; but i figured .art would be better since this project may include more art :) [17:46] oh [17:47] mmcc: that reminds me. what was the icns creation tool you mentioned previously for osx? and is it open source? [17:48] dobey: the one that apple ships is called iconutil, and a quick google scan says it is not open source. I wouldn't have expected it to be [17:48] damn [17:49] why? [17:49] i was hoping it was; and written in a suitable language for running on other platforms [17:49] because png2icns doesn't support the @2x retina stuff [17:49] oh, definitely not other platforms :\ It uses CoreFoundation stuff [17:51] ralsina: ok, now you can 'review' it :) [17:51] dobey: ha [17:52] dobey: there [18:10] ralsina: btw, looks like the new shares tab is in the omgubuntu 12.10 features preview video :) [18:13] whoa [18:13] nice :-) [18:14] hope we had had another month to polish that one [18:14] heh [18:15] hey look at that. a green ball next to the ubuntuone-windows-installer build in jenkins [18:15] yay me! [18:16] :) [18:17] * dobey waits for the beta1 release announcement so he can upload new sso [18:18] * briancurtin back [18:19] wasn't much of a lunch, more of a try to buy two boxes of sudafed (different types) and get denied because you could cook meth with it [18:19] blame iowa [18:20] briancurtin: that's funny, last time I did that they just took my drivers license info. just like when you get codeine… [18:20] also, bummer [18:20] mmcc: they did that too, but after the first scan it locks you out for X amount of days (which the cashier couldnt tell me X) [18:20] briancurtin: bad idea buying sudafed in a porkpie hat. [18:21] at least i had a good comeback: i dont want that second box anyway, i'd rather use methlamine [18:21] (i watch too much Breaking Bad) [18:21] bad timing - today was lab coat day at briancurtin's [18:21] hahaha [18:23] man; waiting for the battery to die on my sansa fuze is very annoying [18:29] dobey: yay green balls! They beat blue any time! === yofel_ is now known as yofel [18:36] brb [18:50] hmm.... should I get the new kindle fire hd, or a nexus 7? [18:53] alecu: you hate kindles! [18:53] * ralsina is so getting the paper white [19:11] * dobey is so not buying a tablet [19:11] i kind of want a win8 tablet (seriously) [19:15] briancurtin: well i can have one for $40 apparently (just upgrade win7 to win8 on my dell duo) [19:19] mmcc, mandel: there are no merge proposals for https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/jenkins-daemon-main, yet https://code.launchpad.net/~mikemc/ubuntuone-fsevents-daemon/mungeproj/+merge/122314 depends on it [19:20] dobey: yes, we're aware of that. I talked with mandel yesterday and he says he needs to add some more tests to it before he proposes [19:20] ok [19:21] dobey: I don't see the kindle as a tablet. I see it as a book. [19:21] which is why I want the paper white, not the fire [19:25] anyone have a minute for a four line diff (+2/-2)? https://code.launchpad.net/~brian.curtin/ubuntuone-dev-tools/py3-exception-message/+merge/123151 [19:25] briancurtin: sure [19:27] briancurtin: global +1 [19:27] thanks ralsina [19:40] man, it's not easy to tell what these IPC tests are actually testing [19:40] I keep commenting out things that I think should break the tests, and they resist breakage! [19:41] mmcc: delete! [19:42] ralsina: :| - I think it'd be even harder to prove that they're useless… They're just not clear [19:42] I was expecting to find a test case that would send a signal and check that it gets received without error, or at least a place to add one [19:44] there's a test_remote_signals, but it patches emit_signal, so we only test that we call emit. Nothing is testing that the round trip is OK [19:45] so I removed the patch to emit_signal, just to see what happened, and only a few of the tests failed :o [19:45] meh, and the launchpad ppa builders are being really slow right now :( [19:48] dobey: happens every thursday [19:50] well, every thursday when there's a beta release, anyway [19:51] oh duh, only the first test in the loop has to fail for the whole test case to fail [19:55] alecu: locked himself out of his house, so he will only be back later tonight [19:55] sorry, that explains alecu locked himself out :-) [19:56] tab complete strikes again! [20:12] meh [20:36] bah and now youtube videos are not playing on the youtube site itself, but the embedded videos on g+ play fine. wtf [20:38] dobey: I get all youtube videos at 2x speed for a couple of weeks already [20:39] saves a lot of time [20:39] thisfred: kinda ruins the music [20:39] ralsina, maybe it's the RIAA's new plot [20:40] ralsina, you just have to think of everything the alvin and the chipmunks version [20:42] thisfred: allow me to retort: yeeeech [21:27] have a good evening all! [22:53] OK, time for me to go. If anyone feels like reviewing - this fixes the on_download_finished exception, and adds a test that'll hopefully help avoid that kind of thing in the future [22:53] https://code.launchpad.net/~mikemc/ubuntuone-client/fix-signal-exceptions/+merge/123189 [22:53] TIL more than I wanted to know about u1client's twisted IPC and tests [22:53] that was a warmup, tomorrow: fix the IPC with the root daemon [22:54] good night