[08:37] Happy Friday, everyone! :D === zyga__ is now known as zyga === zyga is now known as zyga-food === zyga-food is now known as zyga === zyga_ is now known as zyga [11:06] good morning! [11:44] gatox, morning! [11:44] mandel, hi [11:44] gatox, how is everything going? [11:45] mandel, fine.... i'm finishing with a branch for sso.... yesterday i lost internet connection.... but i wanted to close 2 issues, one for cp and another for sso, before keep working on mac port, i almost finish with this issue...... [11:46] gatox, great! I have a process that runs as root, gets the events and I'm considering using http://en.wikipedia.org/wiki/Grand_Central_Dispatch for the dispatch of the events to the different users [11:46] mandel, you? [11:47] mandel, sounds good [11:48] gatox, should be very efficient when working on machines with more than one processors, which is all of them (at least all macs) :) [11:48] gatox, that should mean that the io operations could be dispatched to a diff process and we wont have to take care [11:49] mandel, even better...... as ralsina says: good programmers are lazy programmers :P [11:50] mandel, if we can use something already prepare for that, it's awesome! [11:51] gatox, well, it is not many lines of code right now, I don't think it is over 1000 [11:56] good morning guys, early day for me :) === mmcc_out is now known as mmcc [11:56] mmcc, i see! jeje good morning [11:59] mandel, AFAIK GCD is only for managing threads within a process - I don't think it has anything that helps you with IPC :( [12:01] it might simplify the daemon itself, though [12:13] mmcc, I was considering for the simplification of the async task [12:13] mmcc, no ipc related, but dispatching the dispatch of the events to the clients via that [12:14] mmcc, I might have an example daemon with nearly all the pieces by today [12:14] mandel, oh, ok - I misunderstood. yeah, that sounds great [12:14] mandel cool! [12:15] mmcc, now that I read that, I used dispatch too many times hehe [12:15] mmcc, is an objective-c proof of concept, we can move it to c very quickly :) [12:19] mandel: sounds good. I guess if you have something working in ObjC, maybe we should keep it that way [12:19] mandel - it'd make unit testing easier [12:20] mmcc, well, it has its pros and cons, I just used it because I started with it like that, no other reason [12:20] mmcc, porting to c (is very small) should be easy [12:21] mandel: sure. we'll have to think about it I guess. I did find a C unit testing framework that looked OK (check.sf.net) [12:22] mmcc, yeah, unittesting in c is a PITA, I know some of our projects do not have proper unittesting at all [12:27] ok, I'm off to have lunch :) [12:28] good morning! [12:29] ralsina, hi [12:30] gatox: I put a needsinfo in your branch === yofel_ is now known as yofel [12:31] ralsina, i answer it, and i'm submitting the fix now [12:31] gatox: ósom! [12:32] ralsina, the value we are receiving was dbus.UInt32(3).... but it make sense to check for mapping first [12:32] that was my comment [12:32] ah, ok [12:32] but yes, since we are expecting a mapping, we have to check that first. Or when someone sends a Uint32(0) ... DISABLED :-) [12:33] ralsina, done! [12:33] gatox: cool, re-viewing! [12:34] damn, thunderbird has gone all crashy the last two days [12:36] I hate github, but I don't miss launchpad's "updating diff" when I use it. [12:37] gatox: +1 [12:37] ralsina, great! [12:37] mmcc, did you have the chance to review the branch i sent you yesterday? [12:39] gatox: it's like 6am at mmcc's. [12:39] ralsina, but he is around [12:39] * ralsina looks at backlog... [12:39] whoa [12:40] * ralsina has to update his preventing burnout speech [12:42] jeje [12:46] ralsina, ah! again..... if you have a moment for the 1-1 today let me know!! internet connection works like crap yesterday (all the day!) [12:50] gatox: right after nessita's, so that would be about 10:30 [12:50] gatox, I looked at the branch but need to look again. will do now [12:50] ralsina, I'm always up this early - it's just a question of whether I'm watching IRC or a toddler [12:50] mmcc: hahaha [12:51] and today I have to cut out a little early in the afternoon, so here I am [13:04] gatox, still looking. sorry, I have to read more because I'm not familiar with the code yet [13:04] mmcc, no problem [13:25] gatox, the code looks good, but I was going to ask you to update docstrings -- how should i do that launchpad? "Approve" but make my suggestion in the comment, or "needs fixing"? [13:25] mmcc, need fixing in launchpad [13:26] gatox: ok done. let me know if my comments don't make sense [13:27] mmcc, reviewing and fixing [13:29] gatox also I'm curious why we get sent invalid arguments to _process_file_sync_status at all.. this code handles that case gracefully, but why was it happening? [13:29] mmcc, it seems to be a problem in u1-client, but we didn't debug that part yet, because everyone is really busy [13:30] gatox, ack [13:30] any reason why a branch which was set to approved on tuesday would still not be merged by today? [13:31] briancurtin, any kind of message was added to the mp? [13:31] from launchpad [13:31] mmcc, branch updated! [13:31] mandel: whenever you have time, https://code.launchpad.net/~brian.curtin/ubuntuone-dev-tools/jenkins-skipping/+merge/104950 and https://code.launchpad.net/~brian.curtin/ubuntuone-client/jenkins-fix/+merge/104962 [13:31] gatox: nothing [13:31] briancurtin, can you show me the branch? [13:31] gatox: this is the MP that was approved - https://code.launchpad.net/~brian.curtin/ubuntuone-client/fix984407-windows-is_root/+merge/104967 [13:31] briancurtin, usually the person to ask that is dobey , but he is in UDS right now :P [13:32] no messages on there, nothing came in email either [13:32] ah [13:32] it's not urgent, just weird [13:32] briancurtin, mmmmmm and you have a commit message..... i can't see why it's not merging [13:34] gatox, approved [13:34] mmcc, great! thanks [13:36] gatox: 1-1? [13:37] ralsina, yep.... connecting to mumble [13:37] ralsina, can you hear me? [13:38] gatox: can you hear me? [13:39] gatox: wait, mumble is updating :-/ [13:39] ralsina, kind of..... you seem to connect and disconnect a lot [13:39] ralsina, ok [13:39] dropped again [13:40] ralsina, let me know when you are ready [13:41] ralsina, do you know that u1sdtool --publish-file fails in not-ascii paths? (bah, the operation succeeds, but it fails to return the published url) [13:42] facundobatista: news to me! [13:43] ralsina, ok [13:43] facundobatista: file a bug, please, assign to Diego "Unicode" Sarmentero ;-) [13:44] briancurtin, looking! [13:45] ralsina: treeview and I do not seem to agree on anything [13:45] thisfred: that's a popular feeling [13:45] In particular, they seem almost impossib [13:45] le to style [13:45] thisfred: let's go with treewidget, it's easier for an example [13:45] ok :) [13:45] thisfred: it's completely stylable [13:46] thisfred: you just need to use a custom delegate (yeah, "just") [13:46] I'm sure it is, but I'm not getting it to work [13:48] mandel: the only thing that change doesn't include is that we'd need to set JENKINS=1 env var in the jenkins setup [13:49] briancurtin, ok, do you want me to make such a change in jenkins? [13:50] mandel: sure if you want to and know how. i was just told that would be the way to do it - haven't done it yet [13:50] briancurtin, yes, is very simple :) [13:53] facundobatista, ralsina jejeej yes..... i'm considering changing my middle name to Unicode [13:53] or Uñicode [13:53] gatox, oh... "Uñicode" is an awesome name for a superhero [13:53] Úñî©õdê [13:54] * ralsina needs something for the d [13:54] jejeje [13:55] facundobatista: leo Uñicode y me acuerdo una compañera de la facu con uñas de 5 cm tecleando en una vt100 [13:55] * ralsina goes back to english [13:55] Ok, time to pretend I live in Turkey and setup a VPN [13:55] ralsina ɗ [13:56] mmcc: nice! [13:56] Úñî©õɗè [13:56] There's your new nick gatox ;-) [13:56] jajajajajaaj [13:56] good luck explainging "I am Úñî©õɗè in freenode" ;-) [13:57] good luck for the people who try to talk to me jeje [13:58] gatox, jajaj [14:04] anyone know if there are issues with U1 account upgrades? mine failed. [14:15] hmm, neither setBackground or setBackgroundColor on TreeWidgetItems seems to have any effect. So maybe it was not TreeView's fault after all. [14:16] or maybe i'm overriding it in the stylesheet [14:16] thisfred: care to share? [14:17] ralsina: sure: lp:~thisfred/u1db/u1todo-design-1 [14:17] thisfred: looking... [14:19] I don't think it's the stylesheet. It seems the methods do set the property, but it's just ignored when rendering [14:19] thisfred: give me 1' [14:19] yeah, sry, just thinking "out loud" [14:20] thisfred: how do I run this nowadays? [14:20] ralsina: PYTHONPATH=. python cosas/ui.py [14:21] thisfred: thanks [14:21] ralsina, mandel, gatox: I'm runnning the client from a branch, with "PYTHONPATH=. bin/ubuntuone-syncdaemon", if I put a print in the Main inside the branch, I see it ok... however, I'm touching code from ubuntuone/platorm/tools/__init__.py and it's not being used, how could it be? [14:22] thisfred: http://qt-project.org/doc/qt-4.8/stylesheet-reference.html#qtreeview-widget [14:22] facundobatista, give me a sec...... looking [14:22] facundobatista, what exactly do you mean with 'touching code from'? [14:23] mandel, gvim, etc === m_conley_away is now known as m_conley [14:23] thisfred: you need to style treethingie::item [14:23] ralsina: yeah I found that, but it all assumes you want all the items to look the same [14:23] thisfred: ok, if you want to do custom styling for items, you need to do something somewhat more annoying, give me 1 sec [14:23] mandel, I added some prints, new lines... not only I don't see the new prints or log calls, but also when I do u1sdtool and it crashes, it shows me the "old" line number [14:23] I had it almost working with dynamic properties [14:24] but you can't set those on items either [14:24] thisfred: yes, items don't have any "visual" information [14:24] facundobatista, can you share the branch to test it here? [14:24] ralsina: they do have setBackground and setBackgroundColor methods though [14:24] yes, lets take a look ath the branch [14:25] anyone know if there are issues with U1 account upgrades? [14:25] listed in the API docs, not just by accident [14:25] thisfred: checking... [14:25] thisfred: view this http://qt-project.org/videos/watch/the_power_of_qt_model_view_delegates [14:25] zirpu, which kind of upgrade?? upgrade from 5gb to more storage..... or upgrade with Ubuntu to 12.04? [14:25] ok, so back to view I guess :) [14:25] thisfred: and remember that a QTreeWidget is just a QTreeView with a tree model in it :-) [14:26] gatox, simpler, this is the idea: http://pastebin.lugmen.org.ar/7444 [14:26] gatox: upgrade from 5gb to the music 20gb package for 1 year. [14:26] thisfred: I don't see those functions here http://qt-project.org/doc/qt-4.8/qtreewidgetitem-members.html [14:26] facundobatista, yes..... but i wanted to execute it by myself..... copying that [14:27] ralsina: http://www.riverbankcomputing.co.uk/static/Docs/PyQt4/html/qtreewidgetitem.html#setBackgroundColor [14:27] thisfred: oh, there is a setBackground [14:27] thisfred: the PyQt docs are a lie [14:27] zirpu, i don't think that there is a problem there...... what is happening? [14:27] thisfred: but setBackground *should* work [14:27] ralsina: both exist, and both change the background() property [14:28] facundobatista, can you upload a branch please?? it would be better [14:28] gatox: it just says failed. i'll try again to see if there's a specific error. [14:28] it just has no visual effect whatsoeve [14:28] r [14:28] thisfred: although it takes a QBrush, not a QColor [14:28] facundobatista, so i can test it here too [14:28] thisfred: there is supposed to be automatic casting, but... [14:28] ralsina: the docs say you can pass a color instead of a brush anywhere [14:28] gatox, too much work, don't worry [14:28] thisfred: yes, you know docs [14:29] yeah :)h [14:29] thisfred: I'll do some exploratory hacking [14:29] anyway, watching the video now. Why can't people write web pages. I HATE videos for information transfer [14:30] it makes me pause my music [14:30] for one thing ;) [14:30] thisfred: https://pastebin.canonical.com/65883/ [14:31] ralsina: oh, I probably need to push 1 sec [14:31] gatox: it just says: We were not able to process your payment. [14:32] ralsina: anyway, that's not important for the issue at hand, just don't click the side buttons ;) [14:32] thisfred: but I never see calls to set_color [14:32] ralsina, do you know who'll be the person for this? " gatox: it just says: We were not able to process your payment." [14:32] gatox: joshuahoover / duanedesign [14:32] ralsina: add an item with a hashtag [14:32] zirpu, please ask joshuahoover or duanedesign about that [14:33] ok. i'll just give up for now. got to go to work. [14:33] ralsina: also on startup it gets called once for each item that has a tag [14:33] just put a print in there, you'll see [14:33] zirpu: can you please contact support at https://one.ubuntu.com/help/contact ...it's likely being declined because your card has "verified by visa" or "securecode" activated on it [14:33] thisfred: now I see it [14:34] hello zirpu [14:34] good thing i replied...sigh [14:34] thisfred: do you have a stylesheet here somewhere? [14:34] * joshuahoover goes back to writing and running tests [14:34] ralsina: the weird thing is, setForeground does work [14:34] thisfred: could even be style-dependent [14:35] ralsina: I do, in qt designer I set it on the mainwindow [14:36] ralsina: yeah it was [14:36] ralsina: sry I'm an idiot [14:36] thisfred: what was it? I like to know how I fix things ;-) [14:36] Or else I am like Wilson in House MD [14:37] ralsina: not sure yet, but emptying the stylesheet makes the colors show. [14:37] "oh, right, this conversation ust gave you the clue to solve the case. And now you walk out without telling me what it is." [14:37] thisfred: ok, then :-) [14:37] It's weird though, I don't set the background for those [14:37] I don't think [14:37] thisfred: you are probably setting something very non-specific, like QWidget [14:38] ralsina: I set QTreeWidget::item, but not the background-color [14:38] nothing more general than that that should apply [14:38] thisfred: where's the stylesheet? [14:39] thisfred: and I am not going to ask what that verticalspacer on the left does ;-) [14:40] ralsina: it floats the tag buttons to the top [14:40] ralsina: the stylesheet is on the mainwindow [14:40] ralsina: setting this: [14:40] QTreeWidget::item{ [14:40] border: 1px dotted rgb(102, 102, 102); [14:40] border-top: 0px; [14:40] } [14:40] prevents setting the background color from code [14:40] take it out, and it works [14:41] I am revising my opinion of Qt by +1 micro-wtf [14:41] thisfred: there is a "thing" with stylesheets. In some specific cases, if you set one property, you have to set them all [14:41] thisfred: I suspect ::item is one of those [14:42] ping ralsina, syncdaemon.conf gets created in %LOCALAPPDATA%\ubuntuone\ unlike everything else living in %LOCALAPPDATA%\xdg\{cache,ubuntuone} [14:42] ralsina: is it correct? [14:42] rye: well, there are no rules in windows land [14:42] ralsina: I suspect so too, the problem is, you can't do ::item[foo="bar"] either [14:42] rye: we keep it out so people can delete xdg and not break stuff [14:42] since items take no dynamic properties [14:43] rye: or something similar. I suspect we are following some MSDN guideline for that [14:43] so that's me shit out of luck then, unless I set the borders from code too [14:43] if I can [14:43] thisfred: or you set the stylesheet on each item [14:43] thisfred: from code [14:43] ralsina: can't [14:43] thisfred: which you can't, right [14:43] thisfred: there is an alternative [14:43] maybe the video will help [14:44] thisfred: let me find you the docs ;-) [14:44] ralsina: ok, now at least i understand why there's ubuntuone folder after installation [14:44] thisfred: yes, it's supposed to be in the video, too "custom delegates" [14:44] right [14:44] thisfred: basically, you do all the rendering yourself [14:44] I'll watch it and google where appropriate [14:44] thx [14:44] thisfred: cool, don't hesitate to ask [14:44] I won't ;) [14:46] gatox, ralsina, https://bugs.launchpad.net/ubuntuone-client/+bug/998079 [14:46] Ubuntu bug 998079 in Ubuntu One Client "--publish-file doesn't end correctly for non-ascii paths" [High,Confirmed] [14:46] facundobatista, ack [14:47] gatox, the problem is evident... but I didn't want to go deeper to find where any of both should be converted to the other datatype and how [14:52] just when I thought I was out… they pull me back in. I just had to create an Oracle Single-sign-on accoun to report a crashing bug in VirtualBox [14:52] lol [14:53] 3 times this week it crashes my ubuntu guest, apparently while trying to update the dock icon (from QT, so this might be something to remember) [14:54] mmcc: crashes how? [14:54] * rye wants to try virtualbox [14:54] the guest VM just disappears [14:54] rye, I've used VBox a lot on an OS X host, and it's always been pretty reliable and acceptably fast [14:55] I've never tried anything else so I can't draw comparisons [14:55] but I was always using it on OS X 10.5 (which I was using up until I left oracle) [14:55] * rye used vmware workstation, vmware gsx server, then after network issues jumped to qemu and now using qemu-kvm only [14:56] vbox seems a little more unstable on osx 10.6, I've had this problem and one kernel panic (yow) [14:57] I think the kernel panic was related to USB audio support in VBox, though. So now I pause my VMs when I mumble. And knock on wood, wave a dead chicken, etc [14:57] mmcc: you could just disable audio for the VM, unless you need it or something [14:59] ralsina: yeah, I might do that. So far it's only panicked once, and there are other USB audio problems anyway - I can't use my headset to listen to music because when CPU load gets even medium-high the audio wets the bed [14:59] er, starts to skip [15:00] rye: I don't know what to answer here http://askubuntu.com/questions/20343/what-does-this-icon-mean-in-the-ubuntuone-sign-up-screen [15:00] rye: ignore, chipaca got it [15:01] me [15:01] me [15:01] mmcc, mandel thisfred briancurtin standup? [15:01] oops, typing now [15:02] * mmcc also oopsed [15:02] this "standup" guy is *always* late for the meeting [15:03] jejee [15:03] me [15:03] me [15:03] me [15:04] thisfred! [15:04] ok, thisfred is last. Now known as THAT fred who was always late for standup. [15:04] go gatox [15:04] DONE: [15:04] Fixed issue in control panel about sync status, finishing bug in sso about networkstate (couple of issues yet). Found a couple of issues with u1-client in the mac port, create the bugs for this issues, started working on them. [15:04] TODO: [15:04] Propose the sso branch, keep working in the mac port until the end of the world, just kidding, i hope to finish before that. [15:04] BLOCKED: [15:04] No [15:04] ralsina, go [15:04] DONE: 1-1s, weekly team call, reviews, helped around, misc stuff, askubuntu answers, canonicaladmin cleanup, all clear on allhands... TODO: code a little, more reviews, help around if needed BLOCKED: no NEXT: briancurtin [15:04] DONE: finished windows installer automation, tested and it seems to work fine, proposed the branch [15:04] TODO: pushing on with windows bugs [15:04] BLOCKED: no [15:04] NEXT: mandel [15:04] DONE: Got a root process that gets events from the system, filters them per user and writes the event details to a domain socket. [15:04] TODO: Get the twisted code to get the events. Clean the code. Integrate root process with launchd so that it is the one that creates the socket. Get the client uid and check passed settings. [15:04] BLOCKED: no [15:04] mmcc, please! [15:04] DONE: figured out problems with network detection and ctypes, did a [15:04] review. think I'm forgetting something [15:04] TODO: expand tests for network detection, land that branch [15:04] BLCK: no [15:05] [15:05] that empty line brought to you by that fred [15:05] me [15:05] yes YOU [15:06] ;-) [15:06] DONE: Started applying lisettte's design to cosas (formerly known as u1todo), changed treewidget to treeview and back, broke the world. TODO: Fix world, add due dates to cosas, pack and prepare laptop for u1db sprint BLOCKED: no NEXT: [15:06] sry! :) [15:06] ok, then! Comments? [15:06] nop [15:07] EOM then [15:07] ralsina, 1-1 whenever you are free [15:07] mandel: now! [15:09] ralsina, ok, getting there! [15:14] ralsina: what is the current version of u1 for windows? [15:14] rye: 3.0.0 [15:14] ubuntuone-3.0.0-windows-installer.exe 09-May-2012 13:51 22M ? [15:15] rye: yes [15:15] yay, fresh exes [15:15] * rye updates [15:15] * gatox lunch === gatox is now known as gatox_lunch [15:21] mandel: have you tests a condition with proxy when only squid was having DNS connectivity and client was not able to get to resolver? [15:24] rye, no, I've never seen that, why? [15:25] mandel, rye: I think alecu did try that, because it's a (not too) common configuration [15:36] mandel: user's log claims our twisted client is scared to death when it can't find the way to contact the resolver enve though it is using proxy [15:38] rye, uh, weird.. is this on windows, or linux.. [15:42] mandel: looks like linux [15:42] mandel: but nevermind, asked user to clarify whether resolver is actually working [15:43] lunchtime! [15:53] empathy... i love it so much when a video call causes a segfault in logger === gatox_lunch is now known as gatox [16:28] mandel: what happens when a hasher in U1 can't hash a file ? [16:28] Hasher: hash error [Error 2] Le fichier spécifié est introuvable: u'\\\\?\\C:\\Users\ ... [16:28] rye, uhm, you should as facundobatista he should know better === salgado is now known as salgado-lunch [16:29] facundobatista: ^ file not found during hashing (i.e. file was there during rescan, but no longer there when the queue finally got to it) [16:31] mandel, rye, it sends HQ_HASH_ERROR (or similar), that is handled by sync... if node still exists in the system, it pushes a new hash request with the new path... if node is not there anymore, it's discarded [16:35] gatox, I have a twisted question about some test code you wrote: in ubuntu_sso/networkstate/tests/test_linux.py, TestConnection.setUp() has the inlineCallbacks decorator and does 'yield super(TestConnection,self).setUp() -- why is setUp calling its super.setUp with a deferred? [16:35] I hope I've said that right so the question makes sense. Still getting used to twisted terminology [16:36] mmcc, for each TestCase, when you need to add something to the setUp method, you need to add the @inlineCallbacks decorator and do: yield super(TestConnection, self).setUp() because of the way ubuntuone-dev-tools is workign [16:36] facundobatista: https://pastebin.canonical.com/65891/ [16:39] gatox, ok so it's something about how u1trial runs the tests. got it [16:39] gatox, thanks [16:39] mmcc, yes [16:40] rye, was that file removed? or is there but not accesible somehow? [16:41] mmcc, gatox, the reason for that is that setUps and tearDowns should return a deferred and ofcourse call their parents setUp/tearDown methods, there was some code check added by MentalGuy (not longer with us) to ensure that we write the code like that [16:41] facundobatista: that's what can't be found, because all 5 1Mb logs show that hasher is very interested in that file [16:42] facundobatista: and they already rotated several times [16:42] rye, ok... if the file was removed, it should have been deleted from the system [16:42] 2012-05-07 21:25:37,377 - ubuntuone.SyncDaemon.sync - INFO - T:NONE:F 46a65855-d4d2-42e2-b4e5-6928faa97b83 ['a6a4b9c9-566d-46a8-af2a-ab15b2761ee9'::'e0b67d95-4cd5-46d6-b391-a9d34c3b97e8'] ''Org_scol\\Planifications\\12-13\\Untis\\12-13_054.gpn~RF4a8d23.TMP'' | Calling calculate_hash (got HQ_HASH_ERROR:{}) [16:42] 2012-05-07 21:25:37,378 - ubuntuone.SyncDaemon.HQ - DEBUG - HashQueue: inserting path 'C:\\Users\\ft\\Google Drive\\Org_scol\\Planifications\\12-13\\Untis\\12-13_054.gpn~RF4a8d23.TMP' mdid 46a65855-d4d2-42e2-b4e5-6928faa97b83 [16:43] rye, those lines (and the zillion like those) ^ show that the node is still alive in the internal MD [16:43] and with the same path! [16:43] so, it seems it wasn't removed, and it wasn't renamed [16:43] facundobatista: yes, i agree, but it looks like it is a temporary file from the name [16:43] mmcc, gatox, no adding the super call makes the tests less reliable and not returning the deferred too since the tests are not deterministic (got to love async code!!) [16:43] rye, yeap, but it seems that it is a temporary file nobody else can use? [16:43] facundobatista: i am installing google drive to see how it does downloads [16:44] mandel, thanks - it makes sense now, that the test harness uses twisted reactors and expects deferreds from the test cases... [16:44] or that it's removal didn't send a filesystem notification? [16:44] facundobatista: aha! exclusive lock? [16:44] rye, no idea, that enters mandel's domain [16:44] mandel, yep..... async code is making me suffer right now with sso [16:44] facundobatista, uh, me pasaste la bola! let me see.. [16:44] gatox, what are you working on in sso right now? [16:45] gatox: is it the network detection wizard page? [16:45] mmcc, the issue about network detection page....... it's trivial..... if it wasn't for twisted :P [16:45] gatox: aha. ok [16:46] mandel, is it possible to know from the logs if a file is there but can not be really accessed? [16:46] rye, ^ [16:47] facundobatista, rye, we are getting errono 2 which means not found: http://msdn.microsoft.com/en-us/library/windows/desktop/ms681382%28v=vs.85%29.aspx [16:48] mandel, but the file was there at some point, could it be that it was removed but SD didn't get to know it? [16:48] facundobatista, rye, yet I know that some of the api from windows says that when there are security issues, like you want to change the rights and the ACLs do not allow you [16:49] briancurtin, we do not get errno 2 when the file is locked, right? /cc facundobatista [16:49] facundobatista, I wonder if google drive is doing something 'fun' with the file.. [16:49] mandel: i don't remember the exact code you'd get, but 2 is not it [16:50] hello again! [16:50] mandel: it would be 5, ERROR_ACCESS_DENIED, if the file is locked [16:51] ralsina: so I set up a delegate, and have overridden paint, set its pen to a 2 px dotted grey QPen, but for the life of me, I cannot figure out how to tell it to draw borders on the items. (Ideally only on 3 sides, but for now on all sides would be a huge step forward) [16:51] facundobatista, rye, we would need to know if the file is really there, but certainly 2 usually means is not, could we have lost an event.. maybe but would be the first bug report we have like this [16:52] thisfred: well, you draw them [16:52] thisfred: so probably I am not understanding the question ;-) [16:52] ralsina, possibly neither am I [16:52] mandel: could it be that there is a permission problem with the parent folder? [16:53] ralsina, I don't think so.. but is worth to test it, also we might as well install google drive an see if we can reproduce it [16:53] mandel: usually having two sync programs running on the same flder is not going to work [16:53] ralsina, you type faster than I swear ;) [16:54] thisfred: so, push? [16:54] * ralsina has fast fingaz [16:54] mandel: so, some of the other test cases don't have an @inlineCallbacks - decorated setUp. What's the rule for when that's necessary [16:54] ? [16:54] mmcc, ok, simple, if the test inherits form twisted.trial.unittest.TestCase you need it [16:55] ralsina 278 pushed. I assume I must call something around line 88 [16:55] mmcc, otherwise you don't, if you forget the tests will complain [16:55] thisfred: ok, will look, report in 5' [16:55] kk [16:55] facundobatista, rye, we have to test to see how bad we interact with dropbox, we might have issues in that case [16:57] you know what [16:57] these google docs files are actually links to the documents in google, not real files [16:57] 167 bytes of pure links [16:57] mmcc, there are some tests that just use the 'normal' TestCase because they do no test async code, but most of them do unless they use self.patch which comes from from twisted [16:58] syncing them to U1 is completely useless [16:58] rye, ok, time for spanish lesson: son unos pedazo de hijos de puta!!! [16:58] rye, what time of links? and ofcourse the do not end with .lnk which is what we use to test if the path is a link... [16:59] mandel, hmm. ok, so if it uses self.patch it has to have an inlineCallbacks setUp, even if it doesn't inherit from Twisted.t.u.TestCase - right? [16:59] mandel: a file with a json document in it :) [17:00] mmcc, I'd be surprised if we have any self.patch test case that does not inherit from the twisted test case, so as an easy rule of thumb, yes [17:00] mandel: but that's google docs files, not real pdfs or anything [17:00] rye, oh well.. you always learn something.. [17:03] mandel yeah it looks like everything does descend from twisted eventually. I was getting thrown by ubuntu_sso/networkstate/tests/test_linux.py , TestConnection(TestCase), where TestCase is ubuntu_sso.tests.TestCase, which is declared as TestCase(unittest.testcase) --> but that's actually twisted.trial.unittest, thanks to an import I didn't see. [17:03] if you read all that you get a cookie [17:04] mmcc, yep, sounds correct and evil from our part to do those imports like that :) [17:04] mmcc, found I nice article that describes @inlineCallbacks if you want more info of what they do: http://blog.mekk.waw.pl/archives/14-Twisted-inlineCallbacks-and-deferredGenerator.html [17:05] mandel luckily I can follow import paths up to about 7 +- 2 separate files. after that I get confused and angry [17:05] mandel, thanks. I'll look. [17:05] dobey: re: installers and ubuntuone. Landscape has an installer that installs an installer [17:05] mmcc, I'd probably get angry earlier than that :) [17:05] ok, EOD for me [17:05] dobey: when you are aroun... https://code.launchpad.net/~brian.curtin/ubuntuone-client/fix984407-windows-is_root/+merge/104967 [17:06] yeah wtf is that [17:06] rye, facundobatista, I'll install google drive in my machine and will see what happens, maybe it screws things up [17:06] briancurtin: that, my friend, is some docbook nonsense, AFAIK [17:06] ah, no [17:06] just one installer [17:06] briancurtin: did you know our linux development environment includes LaTeX? [17:06] Or used to ;-) [17:06] bye mandel. if you look at google drive on osx see if you can tell if the daemon is running pyobjc [17:07] * briancurtin backs away, hugs windows machine [17:07] mmcc, good idea :) [17:08] ralsina, briancurtin, one day in the distant future an smarter species will look at our code and will say something like.. ok, so this LaTeX thing means the world will end in 2012 [17:08] ralsina: see other channel [17:09] * mmcc would share how much he loves LaTeX but can't remember how to format it properly [17:09] oh it failed now [17:09] now, really going or dog will pee everywhere! [17:09] all, have a great weekend!!! [17:09] mmcc: I wrote 15KLOC of python whose main defining feature is "and you don't need LaTeX to do that anymore" [17:09] bye mandel! [17:09] ./tests/platform/os_helper/test_windows.py: 35: 'platform' imported but unused [17:09] bye mandel, have a good weekend [17:09] ./ubuntuone/platform/os_helper/windows.py: 38: 'version' imported but unused [17:09] ah [17:10] ralsina: interesting, what was it? [17:10] mmcc: rst2pdf.googlecode.com [17:10] holas! [17:11] rye: what about the installer? [17:11] ralsina: hey that looks useful. [17:11] mmcc: it is! [17:11] mmcc: it is also kinda broken and abandoned [17:12] dobey: nevermind, landscape employs the same thing as we are - you install the installer from software center, then it installs landscape-client [17:12] ralsina: bummer. [17:12] * mmcc *shakes head* remembers he isn't writing research papers anymore, and gets back to code [17:13] (I was about to ask about reStructuredText and BibTex databases) [17:13] rye: right. doesn't mean we should though :) [17:13] mmcc: well, it's one of the most often requested things in the docutils mailing list [17:13] mmcc: I must confess I know very little about it [17:14] ralsina: well, if you ever need to know too much about bibtex, I can help you with that. [17:15] mmcc: that's the thing, I have not written a paper since they were actually typed for presentation quality :-) [17:15] mmcc: and my reference management was a piece of paper [17:17] ralsina: heh. thank goodness for computers... and incompatible file formats, character encoding, etc etc [17:17] mmcc: luckily noone ever asked my about my typewriter's character set. I think it was not unicode. [17:20] ralsina, in typewriter character sets, do "1" and "l" share the same bit representation? [17:21] alecu: I don't recall if there was a | [17:21] alecu: it has been a while [17:21] ralsina, no, not "|" (pipe char). I'm asking about "1" (one) and "l" (lowercase L) [17:22] alecu: no, not on this typewriter [17:22] alecu: that was only in really old mechanical ones, this was a selectric [17:22] selectric? are those the cousin typewriters to scalectric? [17:22] alecu: more noisy, less fun! [17:23] * dobey wonders how to test python-oauth in python3 [17:23] argh, setting up this VPN is *painful* [17:24] ralsina: which? [17:24] rye: I bought a turkish VPN [17:24] ralsina: type? [17:24] rye: I ordered PPtP but I got L2TP [17:24] ralsina: oh, openl2tp is your friend [17:25] ralsina, a turkish VPN! awesome! [17:25] rye: right [17:25] alecu: yeah, ... awesome... right [17:26] ok, going to windows, it's going to be easier than making this crap work [17:26] brb === salgado-lunch is now known as salgado [17:27] OpenVPN FTW! [17:27] ralsina: get the refund for that one too [17:27] dobey: he [17:27] dobey: if it helps us fix it, they can keep the 20 euros [17:28] the other thing does show the valicert thing here too, and i filed the RT to ask IS to look at it asap [17:35] dobey they were looking last night and nothing suggested itself [17:35] mandel: interesting. Indeed gdrive files initial hashing attempt results in error 2 [17:35] mandel: and i guess this was somehow fixed in 3.0 version with introducing the delay before hashing === daker is now known as daker__ [17:37] well i was asked to file rt, and i did :) [17:40] hm - http://paste.ubuntu.com/982073/ [17:42] Hello from Turkey! [17:43] uh [17:43] dobey: yes, my IP to you will appear to be in Jersey or something [17:43] dobey: my LAN spans continents [17:44] i have ukrainiang vpn and US vpn. i can say the same [17:44] :-P [17:44] ralsina: so, how is it there, in Turkey? [17:44] rye: censory! [17:44] plus, it would have been a better idea to download 3.0 BEFORE switching [17:44] ralsina: you didn't go anywhere. :) your irc is still in the same ip it was before [17:45] dobey: because I am using a decentralized IRC proxy [17:45] yeah [17:45] dobey: but whatismyip says I am in Ankara [17:45] right [17:46] chrome gets me the exact same cert as from .ar [17:46] right [17:50] * mmcc lunches [17:55] and u1 works great [17:56] awesome [17:56] what does openssl -connect do? [17:57] but this is from windows, not precise [17:57] so, vlunteers to setup l2tp+ipsec on ubuntu? ;-) [17:57] damn [17:59] heh [17:59] oh crap i am not censored [17:59] dobey i would've to rtfm [18:00] oh, yes I am. But they are cesoring rollingstone.com but not (ahem) xvideos.com [18:03] No error whatsoever, everything Just Works ® [18:03] going back to ubuntulandia [18:04] haha [18:05] i think elmo is going to fix that cert though not sure when exactly, but hopefully soon [18:20] it expires in a month anyway [18:26] dobey: apparently openssl -connect does a CONNECT then quits. Surprise! [18:26] heh [18:26] doesn't spew the ssl stuff? [18:28] dobey: http://www.madboa.com/geek/openssl/#cert-retrieve [18:28] dobey: only if you use the s_client thing. openssl's CLI is just weird [18:28] but it printed stuff last night [18:30] dobey: yes, that was using echo | openssl s_client -connect login.ubuntu.com:443 [18:30] dobey: without the echo, press enter, of course [18:31] is there a really simple script which uses python-oauth somewhere? [18:31] right, what does s_client say? [18:32] dobey: if there is one, I don't have it. I was on windows, I am now trying again to connect from Linux [18:33] hmm ok [18:34] or I could go back to windows and get openssl or windows from somewhere, I guess. But the cert validated, so I expect to get the exact same result as in Ubuntu from .ar [18:39] ralsina, after all the twisted problems....... when you have a moment please: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/network-page/+merge/105517 :D [18:39] gatox: queued! [18:39] ralsina, thx [18:40] * ralsina is not going to ever miss a mention again, because his freaking android phone buzzes and blinks whenever soeone says ralsina on IRC now thanks to quasseldroid [18:43] I give up, the VPN didn't connect but killed my defaultroute anyway [18:46] ralsina: I'll be sure to not install quasseldroid then, [18:46] ralsina [18:47] because I hate being pinged needlessly, ralsina :) [18:47] rmcbride: I see [18:47] rmcbride: I have your home phone number and *free* skype calls, dude. [18:47] hehheh [18:47] rmcbride: so don't start things you can't win ;-) [18:47] I have a blackhole utility for my android phone, but point taken [18:48] hehe [18:48] heh [19:03] I wonder if our jenkins config should be in bzr [19:04] ralsina: isn't it part of the big puppet config in bzr? [19:04] Chipaca: I know it's in puppet somewhere, but I suspect we are editing jobs through jenkins [19:05] Chipaca: thus I have this nagging feeling of "doing it wrong" [19:32] ok folks, early EOD for me - will probably be back later this evening after everyone else leaves. have a great weekend [19:32] thisfred: https://pastebin.canonical.com/65903/ [19:32] thisfred: green thin rectangles for items [19:34] gatox: +1 [19:35] ralsina, great thx! [19:35] mmcc, you too! [19:37] ralsina awesome! [19:40] and it works, even with other colors ;) [19:41] thisfred: hehe [19:41] thisfred: I suspect there are nicer ways to do it, but it would involve redoing all the drawing instead of using the parent class [19:42] this is fine, I have dotted lines, and background colors based on tags. Spent way too much time on this already [19:43] thisfred: happens with this kind of things. "oh, is dot-dot or dot-dash better?" [19:44] I know, but I'd prefer to get the functionality done before I descend into that kind of polishing [19:45] the dotline is more like a dashline for instance [20:04] ralsina: moving the mouse over the items does remove the borders, however. Sometimes... [20:05] eod for me people!! enjoy your weekend! [20:05] you too gatox [20:05] thisfred_, bye [20:10] ralsina: I'm almost thinking it might be easier to do t [20:10] his in actual html? === m_conley is now known as m_conley_away [20:11] just kidding, mostly [20:12] probably [20:12] lunch [22:22] back for a bit on "take your sleeping toddler to work (in a stroller next to the desk) day" [22:30] dobey: http://www.catbirdseat.org/catbirdseat/bingo.jpg [22:31] hahahahah [22:38] ralsina: there's an issue with the qt sso in P [22:38] ralsina: or perhaps it's with how usc uses it [22:39] ralsina: it doesn't let you go from sign up to sign in (if you already have an account) [22:39] ralsina: can you look into it? [22:40] Chipaca: i filed a bug about that before p release :) [22:40] dobey: ouch [22:40] dobey: is it in usc, or is it in ussoc? [22:40] well, couple days [22:40] ussoc [22:41] and only happens for upgrades [22:41] dobey: bug#? [22:41] installing gtk package if missing should fix [22:41] let me find it [22:43] bug #974637 [22:43] Launchpad bug 974637 in ubuntu-sso-client (Ubuntu) "Qt Registration and Log-in dialogs have no way to perform the other action" [High,Triaged] https://launchpad.net/bugs/974637 [22:44] dobey: thanks [22:45] uhm [22:46] i am in the USA, and when i try to buy a song from u1ms, it prompts me to make an account [22:46] but i already have one [22:46] (a UK one) [22:46] popey: sign in from control panel first [22:47] * popey signs in and tries again [22:56] gone again on "dog woke up the baby" day [23:05] Chipaca: you should come to the proxy session [23:06] dobey: i am going to the proxy session [23:06] dobey: it is at 16:45 [23:06] no [23:06] 16:15 [23:06] cock [23:06] heh [23:06] apparently, one is allowed to say "cock" [23:06] i can't fathom why [23:06] anyway [23:06] session! [23:12] hi moch [23:12] hi