[00:19] ok, I'm gone [00:19] bye all! === txwikinger2 is now known as txwikinger [06:49] dobey: There was always a KDE thing called Neon. Its the project which gets daily builds of kde. [06:49] but yeah, lack of imagination. We should have an app which spits out random project names :) [06:52] hello nigelb [06:52] hey duanedesign :) [07:08] morning [07:47] o/ [08:44] morning all [08:44] o/ [08:56] morning mandel [08:56] ralsina: buenos dias :) [09:10] morning [09:11] morning fagan [09:12] ralsina: want me to look into that nm bug [09:12] fagan: good idea [09:12] ralsina: cool I have a good idea about how and what to do [09:13] fagan: great! [09:13] ralsina: ill ask if I dont know anyway [09:21] fagan: ok, go ahead [12:01] * duanedesign wonders if mr rye is around? [12:27] * fagan break [12:35] hello everyone! [12:35] o/ [12:35] morning nessita [12:36] looking forward to london nessita? [12:37] fagan: I have different feelings, sometimes yes, sometimes I would like to stay at home (too much traveling lately). How about you? [12:37] nessita: well im just interested in whats going on more as a learning experience and all that [12:38] nessita: its always get being in those kind of meetings [12:39] nessita: like I learned a lot at UDS when I was there, I learned a lot at the windows sprint so any more experiences to take in is a plus [12:41] right [12:45] ralsina: hello there, you around? [12:51] anyone aware of any Tomboy-1.6.0 dataloss bugs that would cause you to lose notes? [12:57] I'm not, maybe rye knows? === teknico is now known as teknico_away [13:11] morning nessita [13:12] was having lunch [13:12] mmmm, lunch [13:12] duanedesign: there is no data losage reported in the last month or so, AFAIK [13:17] ralsina: hello! was it a good lunch? [13:17] nessita: pretty good yes! [13:17] ralsina: QT question, when you have a moment [13:17] right now is good [13:17] ok, so [13:18] I need to "walk" a QTreeWidget in order to do some asserts over its content. Google suggest using a QTreeWidgetItemIterator, which kinda sounds right, but when I constructs a QTreeWidgetItemIterator from my treewidget, the result is None [13:18] this is my code: [13:18] tree_iter = gui.QtGui.QTreeWidgetItemIterator(self.ui.ui.folders) [13:19] and tree_iter is None [13:19] let me check... [13:20] well, according to the C++ docs that should work [13:20] now let me see if it's broken on python... [13:21] nessita: workaround: http://www.mail-archive.com/pyqt@riverbankcomputing.com/msg11348.html [13:22] * nessita reds [13:22] reads* [13:22] but there are examples of people using the iterator, so can you point me to the code? [13:24] in fact, I don't know how tree_iter could be None, since that's a class instantiation :-D [13:25] let me re confirm [13:25] it may be PICNIC [13:25] * nessita checks for the third time [13:34] ralsina: so, you're right (as usual in this area) :-) [13:34] what I can't do, like you said, is iterate over that result [13:34] haha, you have found my area of expertise ;-) [13:34] I knew I had one! [13:36] :-) [13:39] * fagan thinks he found the bit of code he needs [13:40] nessita: Bug #790717 is in the cp not the sso [13:40] Launchpad bug 790717 in ubuntu-sso-client "Changes to the network manager break network detection (affects: 1) (heat: 6)" [Undecided,Confirmed] https://launchpad.net/bugs/790717 [13:41] nessita: its just in the detection of the network state [13:41] fagan: nopes [13:41] fagan: the main code is located in ussoc [13:41] see the imports [13:41] nessita: oh [13:41] 34 from ubuntu_sso import networkstate [13:41] ahhhhh [13:41] ;-) [13:42] nessita: well that shows why I was confused all morning [13:42] hehe [13:43] ralsina: if I make a branch and it gets merged into trunk to fix the network state thing wouldnt that break the nightlies for everyone no on 11.10 [13:43] *not [13:43] fagan: you tell me! You should make sure it doesn't ;-) [13:44] ralsina: good point [13:45] ralsina: pong [13:47] fagan: if you try to fix it, please take into account the following: === m_conley_away is now known as m_conley [13:48] fagan: the code needs to be test driven or at least be accompanied with tests covering all the changes. And it has to keep working with natty, maverick and lucid, since we provide PPA for those :-) [13:49] in fact, this is a neat case where a "small" fix is going to keep you busy for... a week or so ;-) [13:49] nessita: yeah ill make sure its working on at least natty and 11.10 [13:49] fagan: ping me for reviews! [13:49] nessita: yeah harsh harsh reviews :D [13:50] nooooooo, lies! [13:50] * fagan will break everything [13:50] I'm a good girl [13:50] nessita harsh? naaaaaaaah [13:50] exactly [13:50] * nessita is a sweet dolly doll-like girl [13:50] oh one question [13:51] * nessita thinks this f****ng trees are making my life miserable [13:51] nessita, wasn't chucky a doll as well? [13:51] lol [13:51] beuno: but that was a male doll :-D. Though hard to guess, I'm a girl doll [13:51] actually nope false alarm no question [13:51] or did chucky have a girlfriend? [13:52] nessita, he did in Chuky 2! [13:52] a bride [13:52] beuno: the one and only time I ever hitchhicked the person had a Chucky doll in their car. o.O [13:52] LOL [13:53] the bride of Chucky :) [13:53] duanedesign, does this story end in a tub full of ice? [13:53] Is it wrong that I dont know what you guys (and gal) are talking about [13:53] fagan: yes, it's unforgivable [13:53] the unforgiven! [13:53] hahah [13:53] specially about "Bride of Chucky" [13:53] * nessita requests metallica [13:54] beuno: :D fortunately no. [13:54] * fagan googles [13:54] fagan: careful, that movie contains puppet intercourse [13:55] ralsina: oh no then its not my kind of film [13:55] fagan: it was necessary for the plot! [13:55] * fagan is still a little sick from a flash of two girls one cup [13:57] ralsina: names of vars [13:57] fagan: it's not THAT kind of movie, you sicko [13:57] ralsina: I need to do a second state number for 11.10+ so what should I call it [13:57] ralsina: hahah [13:57] fagan: I don't follow you [13:58] ralsina: well at the moment they are like NM_STATE_DISCONNECTED = 4 [13:58] thats a different number for the new spec [13:59] so id need two vars 1 for the old style one for the new style [13:59] ralsina: Chipaca: ping [13:59] if there is a way to see the NM version, just check and set the right numbers [13:59] what should I call the var [13:59] voidspace: pong, saw the CV [13:59] ralsina: cool, after I sent it William realised he forgot to include details of his education in the cv [13:59] oooh that gives me an interesting idea on how to take on this [13:59] voidspace: I am thoroughly impressed and want to have his children, or something like that ;-) [14:00] ralsina: I have an updated version if you want it - but he already put the updated version into the system [14:00] ralsina: hah :-) [14:00] me [14:00] besides, I *have* tried to write a PLY thing to parse excel into python and am amazed someone succeeded ;-) [14:00] me [14:00] voidspace: I will mail him in like 15 minutes [14:00] ralsina: anyway, snap him up soon or we'll hire him for our team - Gustavo has already emailed him to arrange a voice chat about working on ensemble [14:00] me [14:00] ralsina: so the race is on... [14:01] voidspace: :-) [14:01] ralsina: yeah - the whole excel formula syntax -> python translate pipeline was great fun [14:01] ralsina: I worked on it quite a bit too [14:01] alecu, thisfred, dobey? [14:01] me [14:02] ralsina: but most of it was already implemented (by Will) when I joined Resolver - although we continually added new features [14:02] voidspace: you know the original excel parser is not actually a parser, right? It's just a stack of handmade if chains [14:02] me [14:02] ralsina: I didn't know that [14:02] ralsina: mandel too no? [14:02] ralsina: we did it properly :-) [14:02] DONE: more Cloud folders tab, started adding tests for that, got stucked and unstucked several times (but I keep moving forward). [14:02] TODO: please, finish the cloud folders! [14:02] BLOCKED: not ATM [14:02] NEXT: fagan [14:02] fagan: nopes [14:02] DONE [14:02] * finding where to do the work for the nm bug [14:02] * payroll [14:02] TODO [14:02] * fix the bug [14:02] * investigate if there is a way to query the nm version to set the var nicely rather than have some extras [14:02] Blocked [14:02] * nope [14:02] ralsina: anyway, I'll leave you in peace - but really you'll need to move quickly if he's interesting... [14:02] ralsina: I think is next [14:03] yep gogo [14:03] DONE: meetings, reviews, pinged design to confirm we have signoffs for everyhting, administrivia (yes, the same thing every day) [14:03] TODO: same as every day. Try to conquer the world pinky! [14:03] BLOCKED: no [14:03] that only happens at nights! [14:03] thisfred? [14:03] DONE: cry hot couchdb tears of sadness. Also was unhelpful with at twisted trial issues. TODO: more couchdb debugging BLOCKED: no NEXT: dobey [14:03] λ DONE: bug #789299, lots of debugging on 789300 [14:03] λ TODO: bug #789300, bug #771488 [14:03] λ BLCK: Twisted. [14:03] Launchpad bug 789299 in ubuntuone-dev-tools "DBusTestCase sometimes connects to real session bus (affects: 1) (heat: 5)" [Critical,In progress] https://launchpad.net/bugs/789299 [14:03] Launchpad bug 789300 in ubuntuone-dev-tools "DBusTestCase needs to work with Qt main loop as well (affects: 1) (heat: 6)" [Critical,In progress] https://launchpad.net/bugs/789300 [14:03] Launchpad bug 771488 in ubuntuone-dev-tools "u1trial should unset GTK_MODULES (affects: 1) (heat: 4)" [Medium,In progress] https://launchpad.net/bugs/771488 [14:04] voidspace: I will go really quick now! [14:04] dobey: any idea why your branch is not being merged? https://code.launchpad.net/~dobey/ubuntuone-dev-tools/dbus-priv/+merge/63022 I approved it this morning [14:04] no, i'll look [14:04] thanks [14:04] ralsina: is alecu coming today? [14:04] nessita: AFAIK he should [14:05] * fagan hopes tarmac doesnt blow up before the sprint [14:05] well it's not a huge deal if it does next week. not like there will be a massive amount of branches being landed [14:05] fagan: mashalla! [14:06] ok then I thought there would be at least some development going on next week [14:06] dobey: actually from what I heard there will be more coding and less talking than last yeat [14:06] year [14:06] wtf cron [14:07] I suppose I wasnt exactly fully aware of the entire content of the week [14:07] thisfred: i thought there were just going to be more cage matches (team building) [14:07] dobey: 4 of the 5 days are "mostly" hands-on [14:07] dobey: remember the first rule of sprint club [14:07] Chipaca: those are the best kinds of cage matches [14:09] thisfred: ooh I know "dont talk about sprint club" [14:09] Hola gente! [14:09] holalecu! [14:09] alecu: say me! [14:09] sorry I'm late, got delayed getting to kinder. [14:09] me === teknico_away is now known as teknico [14:10] alecu, go ahead ;-) [14:11] * fagan hugs dfeet [14:11] make sure you wash dfeet before you hug them [14:12] dobey: I like it dirty [14:12] : [14:12] :D [14:12] DONE: worked on the "devices" tab; worked on a hard merge with nessita's latest branch [14:12] TODO: finish merge and devices branch [14:12] BLOCKED: some weird dbus initialization issues are blocking the merge [14:12] * fagan grew up in the christina agulara generation [14:12] yuk [14:13] shen me niao! [14:14] ok, so why would cron just stop working all of a sudden? [14:14] 'cron' appears to be running, as does 'atd' [14:14] but it hasn't processed any jobs for 2.5 hours [14:16] dobey: you know noone will know. You can give me shell access and I can probably guess using strace ;-) [14:17] well i restarted cron and atd, but i think i did it just after the time had passed for it to run again, so will wait until 9:30 and see [14:18] * fagan is terrible with dbus [14:24] ralsina: ah, apparently there was a pam security update that broke it [14:24] ralsina: i guess it got installed automatically on my lucid server just before 07:00 today [14:24] which is why the last run of tarmac was at 06:45 on that box [14:24] dobey: fun! [14:25] ok [14:25] 5 minutes and my branch should land then [14:32] huzzah, it is back [14:32] dobey: hmm, I wonder if that explains why my mail suddenly stopped being fetched automatically yesterday. What did you do to fix cron? [14:33] or does it fix itself? [14:34] there is another update today for pam that fixed it [14:34] so just upgrade again [14:34] ah ok [14:37] ok, back to this awful twisted mess [14:40] * fagan sees the irony of the name twisted when dobey talks that way [14:42] i'm not entirely sure what to do here [14:43] * ralsina records the moment for posterity [14:44] ok, well i'm sure of what to do, but you won't let me do it :) [14:44] * fagan break [14:44] dobey: you are probably right ;) [14:44] and it still wouldn't solve the problem in the case where we do have to use twisted [14:45] dobey: can [14:45] dobey: care to explain a little bit what the problem is? Sometimes it helps [14:45] twisted's TestCase does not have a setUp or tearDown defined itself [14:46] * fagan was going to ask a question but then realised he read it in a book already anyway :) [14:46] and it expects all subclasses to return deferreds in setUp/tearDown [14:46] and also in test_foo() [14:47] dobey: well, not exactly true: If *any* of the subclasses return a deferred in setUp or tearDown, they will all have to and take care of properly chaining those deferreds [14:47] but, it seems, since it doesn't have those methods, when we do super() to call them, the ultimate setUp which gets called, is from unittest.TestCase, which returns None (because it's just an empty function) [14:47] thisfred: i am just trying to make it work correctly in devtools at least first; and it's not complex enough to hit that issue :) [14:47] right, that's the real problem [14:48] and so even when we check for the None result, and set up our own deferred, the method i added as a callback, never gets called, so the setUp pieces never actually get done [14:48] dobey: well if there wasn't the dbus test case, I'm pretty sure the super would just work [14:49] thisfred: well it does work, if we don't try to deal with deferreds [14:49] hmmmm [14:49] * thisfred stops confusing the issue [14:49] right [14:49] except when we try to be good citizens, and deal with deferreds, it breaks [14:49] we *could* monkeypatch TestCase [14:50] * ralsina ducks [14:50] i don't think it matters [14:50] the problem isn't necessarily that the parent TestCase has no setUp/tearDown [14:50] the real problem is the deferreds [14:50] to fix it properly, i think we'd need to rewrite much of the twisted TestCase [14:51] right [14:51] because, basically, there's no way to do synchronous code with it [14:51] and we need to make synchronous calls [14:51] I'm surprised noone ever ran into this [14:51] which is why all of our tests always end up failing completely randomly with nonsensical errors [14:52] well we're doing synchronous code with asynchronous superclasses [14:53] having a completely synchronous TestCase hierarchy works. [14:53] we aren't the first ;-) [14:53] http://twistedmatrix.com/pipermail/twisted-python/2008-July/018120.html [14:55] yeah the inlineCallbacks rubbish is what breaks :) [14:55] dobey: see? Everything has been tried before ;-) [14:56] which is ironic [14:56] because "do what twisted wants" is when everything breaks [14:56] but "do what seems right to me" is when everything works [14:56] so i can only take this to mean that twisted is obviously just completely wrong [14:57] but alas [15:01] dobey: have we tried using returnValue() [15:01] I think that may be the missing part [15:01] what is that? [15:02] with inlineDeferred, the method cannot return a value, because it's a generator, so it always yields None by default [15:02] but twisted has a returnValue() that can return the deferred instead [15:02] http://twistedmatrix.com/documents/8.1.0/api/twisted.internet.defer.html#inlineCallbacks [15:03] and http://twistedmatrix.com/documents/8.1.0/api/twisted.internet.defer.html#returnValue [15:03] \o/ ralsina I got the version of network manager [15:03] wtf [15:03] yeah [15:04] twisted is a mess [15:04] fagan: cool, you are 5% there ;-) [15:04] ralsina: well now all I need to do is sub in the other values and it should work [15:05] ralsina: (when the version is greater than the version im on since that was when the sate values changed [15:05] ( [15:05] ) [15:05] fagan: making it work is 10% of the work. You need to create a test case that fails without your fix, and works with the fix, too [15:06] ralsina: yeah I know ill have it tested to hell [15:06] well, ideally you'd start with that test [15:06] fagan: what thisfred said ;-) [15:06] which is easier than writing it afterwards [15:06] now you have to *remove* the fix before writing the test [15:07] yeah I know im trying to make sure that it is just right [15:07] hrmm [15:07] returnValue seems to make things worse, not better [15:07] but that follows with my experience of using twisted :-/ [15:07] so not exactly unexpected [15:08] hrmm, i think i need to postprocess some of my mp3s, particularly the classical pieces; so that the average volume is higher [15:09] can barely hear this symphony 1 in d minor recording [15:09] mp3gain/vorbisgain are your firends [15:09] yeah [15:10] but "having to do stuff" is not my friend [15:10] ok, coffee break for me. [15:10] says the guy who is designing a language [15:10] dobey: find -L . -iname '*.mp3' -print0 | xargs -0 mp3gain -k -r -f [15:10] done [15:11] although that does not do album gain, which you may want for classical [15:11] right [15:12] actually, what i *want* is for the music player to be smart [15:12] but i suppose that's asking too much in the world of foss [15:12] :( the version numbers in nm are dumb [15:12] uh, no they aren't [15:13] well the version number is 0.8.9997 in 11.10 [15:14] It means ill have to do a str comparison and id prefer to do a float [15:14] yes [15:14] uhm [15:15] im up for suggestions on how to do it though I may be just seeing it wrong [15:16] I was thinking of dropping the leading 0. and doing a float and having it going >= the version that state changed [15:16] It's not that hard if you don't have to handle alphas, betas, rcs etc [15:17] fagan: then what happens when it hits 1.0? [15:18] seriously, version comparisons is like the one thing that python can actually do well [15:18] dobey: could you explain it to me why that is [15:19] thisfred: well id say im just overcomplicating it because im overlooking something that other peole might know [15:19] because it is dynamic [15:20] also, floats are stupid [15:20] fagan: so write a test that covers a single case, and make that test pass, then think of a use case that is not covered by the existing tests, and repeat [15:21] this prevents analysis paralysis [15:21] thisfred: ok then thats cool so ill make it work first and work backwards [15:25] weird question, why are we doing the network state thing when we could use the bool that says if you are connected instead [15:25] id say that would be more of a question for alecu since he wrote it [15:26] meta-x pyshcoanalyze-zippy [15:26] and it would be backwards compatable too if we did it with the bool [15:27] i don't know what bool you're talking about, but you are wrong [15:27] dobey: there is a var in the dbus that we have access to that says if you are connected or not [15:28] NetworkingEnabled [15:28] that has nothing to do with connectivity [15:28] stop making assumptions [15:29] dobey: hmmm well I see what you mean now sorry [15:35] why oh why is twisted so insane [15:36] fagan, make a tuple out of the version string, and compare with a version tuples [15:39] alecu: cool will try that, just working it out with strings first and ill do that after [15:46] the only base type in python is object. any US-ASCII encoded number in a "string" can be an int [15:48] sigh [15:50] Exception dbus.exceptions.DBusException: DBusException('Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.',) in at 0xeb1768c>> ignored [15:50] i wonder where that came from. it is weird [15:51] but i get that in control-panel, and it has some very very weird usage of test cases [15:53] ralsina: ping? [15:54] ok this is messed up [15:58] tcole: ping [15:59] hrmm, does super() do the right thing with MI? [16:03] * dobey wonders [16:06] dobey: It does, though not always the thing you'd expect, depending on how well you know the MRO [16:07] dobey: if you're wondering MI is the problem, change the order of the baseclasses in the class that does MI [16:07] interesting [16:08] this *should* only matter if both of them implement the method you're supering to [16:09] but if they do, it does matter [16:09] ok [16:09] NOT interesting :'( [16:10] switching to super() with @inlineCallbacks in the u1client tests, caused memory usage to go to > 1.3G RSS, and caused lots of other errors in the tests directly [16:10] so uhm [16:10] time to rewrite i think [16:11] i think i'm going ot have to reboot now, just to fix up my memory usage [16:12] swap == slllllllllllllllllllloooooooooowwwwwwwwwwwwwww [16:21] so fixing ubuntuone-client seems to be not an option right now [16:23] let me know if there's anything I can pick up or look at [16:26] i am not sure there is [16:26] i think "do what actually works now" is better than "do what twisted seems to expect, but always fails to work correctly" [16:26] so i think i'll just propose what i have now [16:29] right. Fixing twistedtestcase is out of scope. We should contact jml to see if he knows a workaround that does work, I think, but barring that... [16:31] well, i am not sure it is as simple as "fixing twisted's testcase" [16:32] maybe not [16:38] https://code.launchpad.net/~dobey/ubuntuone-dev-tools/dbus-strict-priv/+merge/63134 [16:39] looking [16:40] wrong button... [16:40] don't click the (x) [16:41] actually I hit ^R to refresh my browser while xchat still had focus [16:41] I do that a lot [16:42] dobey: so what I don't understand is, if we're not using inline callbacks, why do setUp and tearDown still need to be generators? [16:43] well, it doesn't matter, but i'd rather have the change to go back to inlineCallbacks be the simplest possible [16:44] ok, looks good to me, +1ing [16:45] though tim should probably add his 2¢ [16:48] Think I have it [16:50] tcole: ^^ [16:50] i guess tcole's hilight in irc is acting up [16:51] or he might not be up yet, west coast slacker ;) [16:52] he's usually up early [16:52] he's usually up earlier in EST, than when he actually lived over here :) [16:52] I'm actually seriously considering moving to Portland next year, but getting up at 6 or so will take some getting used to [16:53] probably 5 if I have to walk the dog [16:53] also, it's portland [16:53] yeah, we covered that :) [16:53] i don't know how you'd survive without the weekly murders [16:53] Anyone on 10.10 or 10.04 that can test a branch for me [16:53] nessita: ^? [16:53] what branch? [16:54] fagan: I'm 11.04 [16:54] dobey: the one that fixes the nm thing [16:54] dobey: and hopefully doesnt break everything else [16:54] nm hasn't changed since before lucid, until O [16:55] dobey: he's proposing a change to work with O, and he can't break former clients :-) [16:55] dobey: well 11.10 the states have changed [16:55] nessita: yes, but 11.04 is old enough :) [16:55] oh you guys refer to the new release as 0? [16:56] fagan: it is Oneiric Ocelot [16:56] dobey: right, but he should (the same) test older clients, as a final test [16:56] or Ocular Oroborus [16:56] dobey: I cant spell it so I just go 11.10 [16:56] :D [16:57]