=== mmcc_ is now known as mmcc === mmcc_ is now known as mmcc [08:58] morning all! [09:38] mandel: aloha [09:38] czajkowski, hello :) [11:20] good morning [11:23] gatox, morning, how is the tooth? [11:23] gatox, if your face look like a balloon take a pict! [11:24] mandel, painnnnn, but manageable....... but the doctor recommend me not to get surgery before my trip because soomething can go wrong... so i will need to extract it when i get back : [11:24] :S [11:25] not, my face looks normal....... well, normal......... as usual [11:25] jejee [11:25] gatox, ouch! well, I hope it does not give you any problems in NZ [11:26] mandel, i hope that too..... my major concern is not being able to speak the next sunday.... but the doctor gave a lot of painkillers just in case for my trip jeje [11:26] gatox, if not, speak via the pc :) [11:27] gatox, that and a wheelchair and you will be more than impressive hehehe [11:27] jejejjejejjejeje [11:27] mandel, espeak "hello, i'm going to talk about ninja-ide" [11:28] just like that hehehe [11:28] jejejje [11:28] gatox, if I where you I would record it just in case, at least you can be there and show a video in person :) [11:28] jeejeje [11:29] gatox, I saw that ralsina fixed lots of the ui problems in mac, right? [11:29] na..... i'm going to talk..... even with pain.... i prepare a really funny presentation http://ubuntuone.com/7fMIOKbq3xfPjwEOVghGMD [11:29] mandel, i heard something about it, but didn't see any branch [11:30] gatox, awesome, I'm in the process of fixing the dead references once and for all :) [11:30] I hate that ipc code.. [11:30] gatox, and turns out that the macfsevents guy used to work for canonical, so the changes should land in upstream :) [11:31] jeje yes, i saw your twit [11:31] small world [12:07] hello, all! [12:08] alecu, morning! [12:09] I need to get lunch or I'll die, catch you in 30/45 min [12:13] alecu, hi [12:25] good morning! [12:25] ralsina, hi! [12:25] mandel, gatox: yes I did and they are merged already [12:25] cool [12:25] mandel: see why you didn't have to worry about those? ;-) [12:43] ralsina, alecu i've made some updates in my branch [12:45] gatox: ack [12:45] alecu, sorry for the long branch..... it was too late when i realize..... but a lot of parts of the branch are really trivial [12:46] alsooooo... please remember that today is my last day before the vacation.... so if you could do the reviews today, so i fix anything i need to fix, IT WOULD BE MUCH APPRECIATED! [12:46] gatox: yup, I remember! [12:47] thx! :D [12:47] * alecu is not happy reviewing long branches. [12:47] i know.... sorry.. [12:47] * alecu reviews anyways, and makes a small note to play saboteur tonight, and kill a few ninjas. [12:51] gatox: in ubuntuone/controlpanel/backend.py, in def get_public_files()... [12:51] gatox: I think there's a returnValue missing [12:52] alecu, mmmm that method doesn't return anything....... just trigger the action, and the informatoin is retrieved in the signal PublicFilesList [12:53] gatox: oh, right. [12:58] Hello, this is me on the backup IRC client, so if you said anything I need to know since friday, please repeat ;-) [12:59] gatox: why do we set qss inside the code??? I mean btn_icon_style='''....''' for instance. [13:00] gatox: isn't there a way to just change the style by qss id name or something similar? [13:00] alecu, i did it that way..... because we use the same enhancer..... to build different line edits with buttons, and in one of them we need a special style but not in the other [13:01] gatox: I understand why we need to change the style. [13:01] alecu, i could do 2 qss styles, and let the user specify just the name [13:01] gatox: right! [13:01] do you prefer that? [13:01] gatox: that's what I was asking about. [13:01] gatox: yes: I would prefer if we keep all the styling together. [13:02] gatox: anyway: not a blocker for this branch, so please add it as a new bug. [13:02] alecu, as you wish..... i can change that in a few secs [13:03] alecu, do you mind if i do it now? [13:03] gatox: no, go ahead! [13:03] alecu, roger that [13:03] gatox: I've added a couple of comments to the branch, but I'm still reviewing it. [13:03] alecu, ack [13:10] gatox: same question regarding the inside the QLabels.... can't they get the style from the QSS? [13:11] alecu, nop [13:11] alecu, we have label styling in several places [13:11] gatox: :-( [13:11] gatox: yes, I see that it's done all over our code. [13:11] alecu, that's why you can customize different parts of the same string in different ways.... as we do there [13:13] gatox: right, but perhaps QLabels have a way to do that like in html. Like , and in the qss you define the right style. [13:13] alecu: no, you would have to set a CSS stylesheet for the QLabels to do that [13:14] alecu: and then set the same stylesheet on each QLabel [13:15] gatox, ralsina_: if there's no option to do this, then let's move all these custom styles (colors and font sizes) into just one file. I think we should not have styles scattered everywhere. [13:15] gatox: anyway: not blocking for this branch, don't worry about fixing it here. [13:15] alecu: yes, a few constants is a better idea [13:15] gatox, alecu: file a bug [13:15] alecu, ralsina_ i'll file that bug [13:19] ralsina_, alecu bug created... [13:19] gatox: awesome [13:23] ralsina_, yes, I saw :) [13:23] * mandel back [13:29] ralsina_, alecu, mmcc, do you know how to reproduce the 'Stale Broker' exception in any of the platforms (Windows or Mac) ? [13:30] gatox, ^ [13:30] mandel: not intentionally [13:30] mandel, nop..... [13:30] mandel: on mac, apparently it's enough to just switch tabs a few times [13:31] ralsina_, ok, good to know, I wonder why it happens because that means that the object ref in the server side is not valid anymore, I want to try if re-requesting the ref is ok or the root is also funky [13:31] mandel: my guess is the TCP socket for IPC reconnects and you get a new ref but that's just guessing [13:31] mandel: and then the old ref is stale [13:33] ralsina_, which can only happen if the root object is gc, so in theory it should also be out [13:34] ralsina_, expecting the exception in the tools code and reconnecting if needed should fix the problem, got a solution like that I'll test it and will propose [13:34] ralsina_, should also fix the problem on windows :) [13:35] mandel: awwwwwwwesome [13:35] this is the "fix the stupid bugs we are dragging for a year" week! [13:36] lol [13:37] gatox: a few more comments on the branch (still reviewing) [13:37] alecu, ack [13:39] Oh, sqlite, I love you but I hate you so much. [13:49] alecu, I should not be using iteritems in new code, right? what is the new method to be used called? [13:50] mandel: just use items [13:50] ralsina_, but that is not an iterator, right? [13:51] mandel: it's one in py3, not in py2 [13:51] ah... [13:51] mandel: and yes, that means it's less efficient in py2. Unless you expect a bazillion items... [13:52] ralsina_, should not be a problem, is just the signals that has already been connected so that on reconnection we do still work correctly [13:52] mandel: ok [13:53] mmcc, mandel: I've been getting an error quite a bit recently from the mac app, something about "calling stale broker" [13:53] any ideas? [13:53] chaselivingston, yes, is my top bug atm :) [13:53] mandel: awesome, good to know :) [13:53] chaselivingston, seems that the ipc reconnects, there is a gc call and the remote objects are gone [13:53] in case you wanted to know the possible reason [13:54] mandel: haha thanks === m_conley_away is now known as m_conley [13:57] chaselivingston: we probably are going to get a new build today that fixes a few f the reported issues [13:57] ralsina_: awesome, looking forward to checking it out [13:57] chaselivingston: may even have something useful in the "share" tab :-) [13:58] ralsina_: haha, nice! [14:10] * mandel feels dirty than usual when patching __getattr__ [14:13] mandel: that doesn't feel like something that would have to be done in any sane code [14:13] ralsina_, is for a testcase because the perspective_broker implementation of sdtools uses __getattr_ to wrap any attribute call with a possible deferred call [14:14] mandel: ok, if you are testing crazy code I suppose it's ok [14:14] ralsina_, yes, it is crazy code [14:16] alecu, i'm a little confuse about this: There's an extra blank line after "class ShareLinksPanel" and its docstring. ...... i was told that was the way to do it [14:17] is it wrong? [14:17] gatox: Leaving an extra line between the class and the docstring if the class has more than one "block" in it is ok according to the PEP [14:18] ralsina_, more than one "block"? [14:18] multiline docstring? [14:18] gatox: but there's two blanks BELOW the docstring [14:19] gatox: no, if a class has, for example, two methods, which are separated by blanks, then the docstring should be separated by blanks too [14:19] gatox: oh, wait. that's only one below, sorry [14:19] gatox: that docstring is ok, except it's missing the final "." [14:19] ralsina_, so..... my docstrings are ok? [14:21] gatox: ok, according to pep257, one-line docstrings have no blaks before or after [14:21] ok..... so deleting that.... grrrrr contradictory peps! [14:21] gatox: the extra blank lines before and/or after is only for multiline docstrings [14:21] gatox: and don't forget the "." :- [14:21] ) === zyga is now known as zyga-afk [14:35] oh bugger. ui freeze *is* on thursday :( [14:35] dobey: I thought it was next week! [14:36] nope. next week is b1 release; ui and b1 freeze are thursday (aug 30) [14:36] dobey: just checked, you are right :-/ [14:37] dobey: so we need to do releases again, right? [14:37] 3.99.90 release is scheduled for today, yes :) [14:38] dobey: hmmm late toay I hope ;-) [14:38] gatox: sorry, I'm feeling not quite well today, so I was afk. [14:38] alecu, no problem [14:38] dobey: at least u1cp [14:38] alecu, get better! :) [14:38] gatox: re: the extra blank lines, I mean the empty lines between the class definition and the docstring [14:39] alecu, yes, ralsina already explain it to me [14:39] gatox: ah, ok. [14:39] time dilation bubble would be awesome right now [14:42] dobey: yeah [14:43] dobey: in any case, we don't have any UI changes pending AFAIR except for diego's and the rb stuff that was going to need exceptions anyway [14:44] dobey: and all the dash stuff is in libunity and the exceptions are njpatel's to request [14:47] icons [14:47] dobey: right, missed that one [14:49] good morning everyone! [14:49] meh [14:49] good morning mmcc! [14:49] dobey: so, average morning? ;-) [14:50] good news - this morning I got a response from an Apple dev relations person about the blocking issue I had on Friday :) -> no time spent waiting for help! [14:50] reminder of all the things you wanted to get done and don't have time to do, morning [14:50] mmcc: awesome [14:50] also, OS security-enforcement code that crashes when you send it a string instead of a dictionary -> yikes! [14:51] mmcc: so, feeling good about another build today? [14:51] ralsina_: right now I am. ask again in 20 minutes ;) [14:51] mmcc: will do :-) [14:53] mmcc, before you do the build, let me know there is a branch I'd like you to include for testing [14:53] mmcc, dobey, alecu, gatox, mandel, briancurtin, thisfred_: standup starts in 7, ends in 9 ;-) [14:53] ack [14:53] mandel: sure. [14:55] yipyip [14:55] ralsina_: do we have techleads call today_ [14:55] ? [14:56] alecu: apparently not [15:00] me [15:00] me [15:00] me [15:00] me [15:01] me [15:01] dobey, thisfred_, mmcc, briancurtin: tick tock [15:01] oops, not you briancurtin :-) [15:01] meh [15:01] yessss, i'm out of the twitter shaming [15:01] me [15:02] mandel: go [15:02] DONE: Could not work on friday due to stupid telefonica changing all the network. Looked at the stale broker problem, we have reconnection problems which seem to be common on mac. Made code that will reconnect and get new remote objects with the correct python references. Talked with macfsevents to land our changes in, should be ok, also developer is an ex-canonical :) [15:02] TODO: 1-1 ralsina. Find what import is screwing up my tests. Test the code irl. Move to the next ugly bug (to be defined). [15:02] BLOCKED: No [15:02] gatox, your turn! [15:02] DONE: [15:02] Propose Share Links tab branch, with all the functionality. Start fixing the branch according some comments. [15:02] TODO: [15:02] Finish fixing the branch. Prepare for my vacations. Enjoy vacations. [15:02] BLOCKED: [15:02] No [15:02] COMMENT: [15:02] I'll be back on September 10th. [15:02] briancurtin, go [15:02] DONE: subprocess, read up on python-defer [15:02] TODO: figure out cross-version subprocess stdout/stderr once and for all... [15:02] NEXT: alecu [15:02] DONE: got vala doing oauth and calling webservices with libsoup (thanks rye!). reviews, more reviews. [15:02] TODO: more reviews, more vala, learn nux [15:02] BLOCKED: no [15:02] NEXT: ralsina_ [15:02] DONE: redid mac env, fixed a couple of mac UI issues (yay), helped around, reviews, canonicaladmin, etc. TODO: check schedules, check exceptions, hopefully fix some other stuff, coordinate with other teams BLOCKED: the UK bankholiday is not helping NEXT dobey [15:02] DONE: reviews, py3 rb plug-in [15:02] TODO: releases, icon generating magic, music store work [15:02] BLCK: None. [15:02] thisfred_: go [15:02] DONE: refactored v1 and v2 music web api to use common base.py TODO: introduce u1db to base.py BLO(CKED: no NEXT [15:02] NO SHAMING TODAY! [15:03] ralsina_, yes! mmcc [15:03] oops! [15:03] or he is not around? [15:03] mmcc? ;-) [15:03] gatox: he was a little while ago [15:03] * ralsina_ goes work on the shaming [15:03] alecu, branch has been updated [15:03] gatox: great [15:04] https://twitter.com/ralsina/status/240102457026437122 [15:04] jejejjej [15:05] sorry, toddler emergency, back now. perfect timing… [15:05] DONE: launchd, updated dev instructions, discussed daemon [15:05] upgrade and versioning issues [15:05] TODO: launchd [15:05] BLCK: no [15:06] cool, comments? [15:06] mmcc: sorry, had I known it was because of toddler issues I woul have held it ;-) [15:07] ralsina_: no prob :) [15:08] I wouldn't have expected this to work, but we are having much cleaner standups lately :-) [15:09] EOM [15:09] twitter shaming is a powerful weapon [15:09] jeje [15:10] gatox: double dot here: """Get the information of the shares..""" [15:11] alecu, fixed [15:12] gatox: do you have the number for the bug I asked you to create earlier today? [15:13] ah yes..... adding to the comment [15:14] alecu, done [15:16] * gatox lunch === gatox is now known as gatox_lunch [15:22] ok; lunch and then a lot of work. bbiab [15:23] mmcc, could you build the app bundle using lp:~mandel/ubuntuone-client/stale-broker ?? [15:24] ok mandel - you mean right now or just when we do the next build to share? [15:24] btw, have you tried building the app yourself? I still don't know if anyone else has done it recently [15:24] mmcc, next time :) [15:24] mandel: k [15:24] mmcc, no I have not, if you point me to instructions I'll do it right now :) [15:25] ubuntuone-windows-installer/scripts/README-mac.txt [15:25] alecu, i forgot to fix something.... doing it now [15:26] mandel, probably best to be using the new buildout [15:26] mmcc, ok, I'll give it a try [15:26] great [15:29] mandel, one thing I just noticed in there. py2app might already be installed on your system but it's probably out of date. you will need to get py2app and its dependencies from bitbucket [15:30] mmcc, ok, so no easy_install, right? [15:30] mandel: I don't think he's released the changes you need yet… [15:30] ack [15:32] dobey: you about ? [15:33] mandel: here are the deps you need to install -- py2app last: altgraph modulegraph macholib py2app [15:33] all from bitbucket.org/ronaldoussoren [15:33] czajkowski: he just went to lunch a few minutes ago [15:33] briancurtin: thanks [15:34] I just did 'python setup.py develop' from each, but you could install instead of develop, if you don't need to tweak it [15:34] and you shouldn't [15:34] dobey: https://launchpad.net/ubuntuone-servers-libs was registered but licience left empty can you get someone to change it please, it's in my review queue [15:34] thanks [15:34] mmcc, ok, I'll just be like a machine and follow each step [15:35] * mmcc is one step closer to a botnet [15:40] ah, back with the good IRC client :-) [15:41] ralsina, on trunk I see a funny mirroed loading widget, is this known? [15:41] mandel: no idea what you mea [15:41] n [15:41] ralsina, let me post a pict, I'll do it as soon as I'm done testing :) [15:42] mandel: ok [15:42] ralsina, just happens at the very beginning of the app [15:42] mandel: oh, wait, you see two overlays? [15:42] ralsina, yes [15:42] mandel: one is global the other one is on the tab, no idea why that started happening [15:42] ralsina, oh, ok [15:43] ralsina, 1-1? or are you to busy right now? [15:43] mandel: it clears up in a couple of seconds though [15:43] mandel: in 17'? [15:43] ralsina, yes, I noticed [15:43] ralsina, 17' mins then :) [15:53] ralsina, if you use --with-icon the overlay seems to stay there [15:54] mandel: hmmm you shoud be getting an exception somewhere in that case [15:55] ralsina, ok, I'm looking into solving the json decode issues we have which are usually error pages coming from the server ad will take a look closer to that, the stale brokers problems seems to be fixed or hard to reproduce when running from comandline [15:56] mandel: ok === salgado is now known as salgado-lunch [16:00] alecu, mmcc, can I have a review for the following: https://code.launchpad.net/~mandel/ubuntuone-client/stale-broker/+merge/121456 [16:00] mandel: mumblez [16:00] ralsina, sure, launching the thing [16:10] gatox_lunch: I added one final needsfixing to the branch. I might agree to move it to a new bug, so let's discuss it when you get back. [16:10] gatox_lunch: otherwise the branch is in good shape to me. [16:12] mandel: do you know why the PB is getting disconnected? just reconnecting seems like a band-aid to me… [16:13] mmcc, no idea, the issue does happen on widows again. Twisted will reconnect automatically, what the code does is to re-request the references of the server side objects so that we do not ask for the wrong id [16:14] mmcc, the way pb works is that it uses the python object id to point to the object that should perform the requested method, and when that objects is gc you get a stale broker error [16:14] mmcc, the main reason would be that the factory is release the root object which is the only guy that has the reference to all remote objects which then are also gc [16:14] mandel: understood, but why is the connection getting broken in the first place? [16:15] mmcc, no idea [16:15] mmcc, will need to look deeper in twisted to know exactly why, alecu any idea? [16:15] someone mentioned TCP timeouts earlier, but we're not using TCP, right? are we just mis-configuring the initial connection or something? [16:16] I mean, auto-reconnecting might be useful to have anyway, and your code is logging when it happens, that's good, but I'd also like to try to fix the root problem too [16:17] mmcc, I'll look closer during the tests to catch the reason, I also want to know why :) [16:17] at least we have a way to fix the issue or better said, not expose the issue to the user [16:19] ralsina, an example of the json decode issue is shown in the image of bug 1040910 [16:19] Launchpad bug 1040910 in Ubuntu One Control Panel "U1 Darwin Client Crash on OSX " [Undecided,New] https://launchpad.net/bugs/1040910 === gatox_lunch is now known as gatox [16:19] alecu, reviewing your comments [16:19] gatox: one second, I'm adding one more thing. [16:20] mandel: ack [16:20] alecu, i've just fixed the FakeDesktopService thing that i forgot [16:21] gatox: done. Two things of notice in the comments: some tests are missing, and the UI gets unresponsive for 15 seconds on startup. [16:23] mmcc, mandel: the only time I noticed that the PB connection is dropped is when one of the two process speaking dies or finishes, or is killed. [16:23] alecu, mandel - what was triggering the stale broker errors in the bug? I never saw them… were they with syncdaemon or some other backend process? [16:24] alecu, any idea on where to start looking on why twisted reconnects? the issues is definitely related to the refs being removed [16:24] mmcc, I have seen it via changing the tabs in control panel, chaselivingston has seen them too that way [16:25] alecu, i'll split the keyPress code..... and try to see how to avoid the blocking thing [16:25] gatox: I think we can split keypress and add tests to it on a different branch. [16:25] gatox: I'm more worried about the blocking thing [16:26] gatox: my guess is that it might get worse for people with much more files. [16:26] gatox: how many files do you have in the account you tested with? [16:26] alecu, probably...... because here is less than 15 sec [16:26] alecu, i only have 854mb [16:28] gatox: I've got 17000 files (mostly photos) [16:29] alecu, i don't have so much files [16:29] gatox: 30Gb [16:29] mandel: so if changing tabs makes it happen, does that mean it's the control panel backend and not the syncdaemon? [16:29] gatox: so, this will get worse for people with more files. [16:29] alecu, yes [16:29] mandel: I'm trying to find the connection between the CP backend and the code you changed in u1-client on your branch… [16:29] mandel, mmcc: it seems as if uploading has stopped as a result of that stale broker issue as well [16:30] alecu, i tried to do something about it.... but it seems didn't work.... will take a deeper look at that now [16:30] chaselivingston: ok, noted… :\ [16:30] chaselivingston: are you observing that via the UI, or looking at the web interface? [16:31] mmcc: that is assuming that the size of the u1 folder locally should match its size on the website [16:31] mmcc, so the back end uses ubuntuone/platform/tools to make the calls to the server side, I need to double check how the control panel backend and the control panel interact but AFAIK they do not use ipc [16:31] ralsina, gatox, how do the control panel and the control panel backend interact on mac, do they use IPC? [16:32] mmcc: guess that answered your question :) [16:32] mmcc, if the answer to that ^ is a yes I have to add similar code for them just in case [16:32] chaselivingston: yes… [16:32] mandel, yes [16:32] chaselivingston, I have made the changes and have not seen a problem yet [16:32] gatox, yes.. bummer or 'mierda' as you please.. [16:32] mandel: meaning you're running a new build? [16:32] gatox: it uses code in sso, though, right? [16:32] chaselivingston: he's running from source [16:33] chaselivingston, I'm doing irl tests from source :) [16:33] mmcc, not for that.... or i'm missing something [16:33] mmcc, mandel: gotcha, well good to know something is coming that should fix what I'm seeing [16:33] gatox: and here's a bug: if you scroll not using the down key (for instance with the scrollwheel), no new items are added to the list. [16:33] gatox: well, grepping for 'ipc' gets nothing in ubuntuone-control-panel/ [16:33] gatox: also, pgdown pgup do not work. [16:33] chaselivingston, mmcc know the branch with the fix, we want to put it there for irl testing before landing [16:33] so I figured it was getting it somewhere else === salgado-lunch is now known as salgado [16:34] mmcc, I know that ubuntuone/platform/tools uses ipc, which is what is used to talk with sd [16:34] mmcc, I'll keep looking there to see if there is more ipc to fix [16:35] alecu, ok..... i'll try to add all of that today...... but i suggest that something like that is missing a t the end of the day, we should open bugs for that..... because that is not blocking... and i'm not going to be here for 2 weeks [16:35] mandel: right, I see the sd tool being used in cp, but nothing else. but how is changing tabs related to our connection to SD? and SD isn't dying, right? do we have more than one connection to SD? (one per tab?!) [16:35] gatox: of course! do not waste time with pg/pgdown nor scroll: those can wait. [16:36] gatox: those are just bugs [16:36] gatox: but the delay that is added is more serious, because nightlies will have that delay for every user with many files. [16:36] mmcc, we have one sdtool per tab AFAIK, when there is a reconnect all those sdtools have old references but the update is just done when you move to the tab, we can say is lazy [16:37] mmcc, but I'd need to double check that by looking at the code, it is not an area I've spent a long time with [16:38] mandel: ok. so I wonder if we're causing problems by reusing the same socket for multiple connections between the same two processes [16:38] mmcc, should not be a problem, twisted should deal with that [16:38] but yet again, god only knows.. [16:39] It's not inherently wrong to connect over the same socket multiple times between two processes, but I bet twisted doesn't expect you to do that [16:40] mmcc, it does, twisted in that matter is really good, but is worth looking at the internals [16:44] gatox: I think I found inklings of the problem [16:44] alecu, shoot [16:44] gatox: trunk uses 0.8% of my 8Gb of ram. Your branch uses 5.8%!!!! [16:45] gatox: that's only second to firefox with a lots of tabs opened. [16:46] gatox: my guess is that the scanning of the files is using too much cpu on start, plus a lot of ram. [16:46] alecu, mmmmmm..... that's probably because we have the information of the files in memory.... the other solution will be to store the info with sqlite and do queries while the user type...... not so sure about this [16:46] mandel: in my syncdaemon log just now, it looks like something's using TCP: ubuntuone.SyncDaemon.ActionQueue - INFO - Connection started to host 127.0.0.1, port 53108. [16:47] gatox: that sounds like an awful solution, using sqlite. [16:47] i agree [16:47] gatox: I would guess that there are a bunch of more elegant, and less resource intensive solutions. [16:47] mmcc, yes, but that is action q, nothing to worry about regarding this bug, that is too low within sd [16:48] mandel: ok, but aren't we supposed to be using unix sockets for the IPC? just looking to see where we might be missing something here [16:48] gatox: I'll try to come up with one while I have lunch. [16:48] mmcc, all ipc was moved to sockets yes, we might have left out something not related with ipc but I would be surprised since we did not notice it on the windows port [16:49] mandel btw, does the monitor code log which monitor it is using anywhere? [16:49] alecu, ok..... i'm improving the keyPressEvent method [16:49] mandel, I'm trying to verify that i"ve actually connected to the root daemon [16:49] mmcc, lets ask in #chicharra (internal) about action q [16:49] mmcc, hm.. i might not do it, let me check [16:49] gatox: let's mumble with ralsina after I get back from lunch, and we can decide if we land it like this, or if we wait after you get back and come up with a different approach. [16:49] ok [16:50] mmcc, it does not log it [16:51] mandel: ok, well at least that means I might be using the right one :) [16:51] mmcc, please file a bug and I'll have that fix tom morning [16:52] mandel: ok [16:52] thx [16:55] gatox: this line may have part of the blame: [16:55] files_list += [os.path.join(root, f) for f in files] [16:55] mandel: bug #1042336 [16:55] Launchpad bug 1042336 in Ubuntu One Client "platform filesystem monitor code does not log which monitor it used" [High,Confirmed] https://launchpad.net/bugs/1042336 [16:55] gatox: try it with files_list.extend(...) [16:55] mmcc, awesome, I'll do that asap [16:56] mandel: thanks [16:56]