[00:44] <diogobaeder> Anyone there could give me a hand with a branch for ~ubuntuone-hackers ?
[00:51] <ralsina> diogobaeder: sure
[00:51] <ralsina> diogobaeder: what do you need?
[00:52] <diogobaeder> ralsina, I created this branch to act as a sourcedep: https://code.launchpad.net/~diogobaeder/ubuntuone-servers/coverage
[00:53] <diogobaeder> ralsina, not sure if this is what I need to do, but I want to include it as a Python module for ubuntuone-servers, in one of my branches ( https://code.launchpad.net/~diogobaeder/ubuntuone-servers/coverage-optimization )
[00:53] <diogobaeder> ralsina, am I doing this right, so far?
[00:53] <ralsina> diogobaeder: I'm afraid I am not very familiar with that code
[00:53] <ralsina> diogobaeder: sorry :-(
[00:54] <diogobaeder> ralsina, it's just the "coverage.py" module (can be found at the cheeseshop)
[00:54] <ralsina> looking anyway, just in case :-)
[00:54] <diogobaeder> http://pypi.python.org/pypi/coverage
[00:54] <diogobaeder> I want it to be included as a dependency in the second branch I wrote here, so that I can use this external module
[00:54] <ralsina> ok
[00:55] <ralsina> let me check the second branch...
[00:56] <diogobaeder> ok
[00:56] <ralsina> is there a missing commit in the coverage-optimization branch? I don't see where you added the sourcedep
[00:56] <diogobaeder> ralsina, I didn't add it yet. How should I do it?
[00:56] <ralsina> diogobaeder: and that's where my ignorance means I can't help you :-/
[00:57] <diogobaeder> ralsina, I mean, shouldn't I push the "coverage" branch to be available from ~ubuntuone-hackers first?
[00:57] <ralsina> yes
[00:58] <ralsina> diogobaeder: also, server code needs to be discussed on u1-internal so let's switch ;)
[00:58] <diogobaeder> ralsina, ok
[01:46] <psychok7> hi there
[01:47] <psychok7> i am having trouble sync a bunch of large files to a ubuntu one shared folder
[01:48] <psychok7> i click on the disconnect and connect button and it says its synced but its not
[07:56] <shuerhaaken> hello all. I started rhythmbox with ubuntuone music store plugin from the terminal. I get the following output: http://pastebin.com/C03pL9ym ; Is there an issue with ubuntuone-client?
[07:57] <shuerhaaken> This is on precise
[08:08] <mandel> morning all!
[08:15] <shuerhaaken> morning!
[08:15] <shuerhaaken> I started rhythmbox with ubuntuone music store plugin from the terminal. I get the following output: http://pastebin.com/C03pL9ym ; Is there an issue with ubuntuone-client?
[08:16] <JamesTait> Good morning all! :)
[10:58] <mandel> psychok7, hm.. weird
[11:03] <gatox> good morning!! :D
[11:56] <alecu> good morning to everyone!
[11:58] <gatox> alecu, hi!
[11:58] <shuerhaaken> in this channel nobody says anything but good morning.
[12:05] <alecu> probably because everybody was sleeping...
[12:06] <gatox> alecu, he left
[12:06] <alecu> mandel: you've seen this branch that tarmac is bouncing, right? https://code.launchpad.net/~mandel/ubuntuone-client/correct-namespace/+merge/116264
[12:06] <alecu> mandel: any ideas?
[12:07] <mandel> alecu, let me look
[12:07] <gatox> mandel, hi
[12:07] <mandel> shuerhaaken: clearly depends on the time zone you are, maybe most people in this channel are in a +5 UTC and you have been asking questions when no-one was here ;)
[12:07] <mandel> gatox, hello hello
[12:08] <mandel> alecu, yes, I have seen that before in a previous branch, retrying the merge work, what worries me is a possible change in Q regarding the reader.. because tarmac is in Q, right?
[12:09] <alecu> mandel: probably not. But we should ask dobey.
[12:10] <mandel> alecu, also, that is happening  because I fixed the tests and ask the monitor to shutdown in a cleanup call.. so the problem might have been there all along but we have just started noticing
[12:10] <gatox> alecu, i fixed the merge problems in the tests-refactoring branch, i'm running the tests now in all the platforms to see if everything is ok
[12:10] <mandel> alecu, we can either ignore that on linux (ugly) or try to reproduce it, in my P machine it does not happen at all
[12:11] <alecu> mandel: do you have a Q vm? I need to install one.
[12:11] <alecu> gatox: great!
[12:12] <mandel> alecu, I have to set up one too.. I'll set up one after lunch and will test, lets try to set it to merge again and see if it works
[12:12] <mandel> last time it did
[12:13] <mandel> alecu, by the way, I updated the to MP which you reviewed, UNIX_PATH_MAX is not defined on Mac OS X AFAIK
[12:16] <ralsina> morning!
[12:16] <gatox> ralsina, hi!
[12:17] <ralsina> welcome back gatox!
[12:17] <gatox> ralsina, thx! :D
[12:19] <mandel> ralsina, morning ralsina!
[12:22] <ralsina> good morning mandel
[12:26] <mandel> well, I'm off to have lunch I'm starving and writing all that objective-c takes all my energy away.. god that bloody lang is verbose!
[12:27]  * mandel lunch
[12:27] <mandel> alecu, FYI the branch that was bouncing just got merged: https://code.launchpad.net/~mandel/ubuntuone-client/correct-namespace/+merge/116264
[12:28] <mandel> alecu, looks like our cleanups have a bug somewhere that might be async and might not be that easy to reproduce :(
[12:30] <alecu> mandel: :-/
[12:30] <mandel> alecu, yes.. that is the more children safe expression you can use
[12:31] <mandel> ok, really going now, stomach is making funny noises
[12:31] <alecu> mandel: buen provecho!
[13:02] <dobey> hrmm
[13:04] <dobey> alecu, mandel: q has twisted 12.0.0, so maybe there is a racing regression in it, or maybe the patch to add the gireactor got refreshed wrong
[13:05] <alecu> dobey: is tarmac running on Q?
[13:06] <dobey> trunk and stable-4-0 branches do, yes; though separate instances of q. trunk will always land on the development version of ubuntu
[13:07] <alecu> dobey: great, thanks.
[13:16] <ralsina> Oh, quantal VM, why do you hate me so.
[13:16]  * ralsina presses the virtual reboot button
[13:25] <lamalex> U1 got rid of notes????
[13:25] <dobey> only the web ui
[13:25] <dobey> you can still sync tomboy if that's what you are wondering
[13:27] <lamalex> no not really i really only used the web ui, but i guess i need to install tomboy
[13:39]  * mandel back
[13:39] <dobey> brb
[13:44] <ralsina> yay, I know what's wrong with our folders and devices list on Q!
[13:44] <ralsina> No idea how to *fix* it but I know what's wrong
[13:44] <ralsina> they are constantly getting resized to 100x30 for some reason
[13:46] <ralsina> And the items inside them probably have a similar problem
[13:47] <alecu> ralsina: likely the gtk theme renderer for qt. Or the scrollbar thingie.
[13:47] <alecu> ralsina: what if you switch qt to a standard theme?
[13:47] <ralsina> alecu: trying
[13:48] <alecu> (by scrollbar thingie I mean the custom unity scrollbars)
[13:48] <ralsina> alecu: works great with -style windows
[13:48] <alecu> ralsina: there you go. Let's ship with that syle!
[13:48] <alecu> *style!
[13:49] <alecu> ralsina: can you try with "-style Metro"?
[13:49] <ralsina> alecu: there isn't one AFAIK :-)
[13:52] <alecu> ralsina: guessed so. And there's no much love right now in the Qt trenches for Melop... I mean Metro.
[13:52] <ralsina> yeah
[13:53] <ralsina> so, I have to find a way to have the items not resized to 0x0 and we should be ok-ish
[13:53] <alecu> ralsina: shouldn't we be asking the qt on gtk guys to fix this?
[13:53] <ralsina> alecu: yes
[13:54] <ralsina> alecu: but I need to ask with a description of the problem
[13:54] <ralsina> alecu: also, we are doing strange things in those lists
[13:54] <alecu> oh, I see.
[13:54] <ralsina> alecu: we have custom delegates, which is unusual
[14:06] <ralsina> bummer, the items have the right sizes, they are just not drawn, totally a style bug
[14:07] <alecu> ralsina: awesome?
[14:07] <mmcc> hi people
[14:08] <ralsina> alecu: well... it depends. Means we depend on others fixing it :-/
[14:08] <alecu> hi mmcc!
[14:08] <ralsina> hi mmxx
[14:08] <mandel> alecu, ralsina, FYI I have the jenkins testing daemon working and copying the events to named pipes, what is left is to adapt the real daemon to take a new arg stating that  we are testing and which named pipe to use + a python test case
[14:08] <ralsina> mmcc I meant
[14:08] <gatox> mmcc, hi
[14:08] <ralsina> mandel: cool
[14:09] <alecu> mandel: that's great!
[14:09] <alecu> mandel, mmcc, ralsina: perhaps we should do a mumble to talk about the status of the macport and to discuss pending things.
[14:10] <mandel> alecu, +100 specially because I need to talk with mmcc on how to deploy the daemon
[14:10] <alecu> mandel, ralsina, mmcc: shall we do it after today's standup?
[14:11] <ralsina> alecu: yes
[14:11] <ralsina> and the bug is in overlay-scrollbar-gtk2
[14:11] <mandel> ok
[14:11] <ralsina> removing that makes everything work
[14:13] <ralsina> dobey: is there a way to disable a gtk module for a single app?
[14:13] <dobey> GLib (gthread-posix.c): Unexpected error from C library during 'malloc': Cannot allocate memory.  Aborting.
[14:13] <dobey> again
[14:14] <dobey> ralsina: run it with empty GTK_MODULES env var?
[14:14] <dobey> ralsina: there's no good way to disable a single module though
[14:14] <mmcc> mandel, alecu, mumble after standup sounds fine.
[14:15] <dobey> you'd have to parse the variable to strip the one you want to remove
[14:15] <ralsina> dobey: yep that works
[14:16] <ralsina> dobey: thanks
[14:16] <dobey> why the heck is this memory error happening on quantal
[14:16] <ralsina> dobey: I am getting tons of weird crashes on q today
[14:17] <ralsina> dobey: since overlay-scrollbar is the only gtk module enabled on Q by default, would it be awful if we cleaned that when launching u1cp?
[14:17] <alecu> mmcc: while running tests for your branch I get this: http://pastebin.ubuntu.com/1123448/
[14:17] <alecu> mmcc: any ideas?
[14:17] <mmcc> alecu: looking
[14:18] <dobey> ralsina: we should get cimi to fix it i guess
[14:18] <ralsina> dobey: yes, I am asking him
[14:18] <dobey> or whoever is working on that
[14:18] <ralsina> dobey: I just like having plan B in mind ;-)
[14:19] <dobey> i think plan B should be have them make the overlay scrollbar not get used in qt apps, because i doubt our app is the only one it breaks
[14:19]  * alecu just read "Plan 8", and thought we were doing a port to a previous version of Plan9
[14:20] <dobey> also, why the heck does tar complain about file name length in one instance of quantal, but not the other
[14:20] <ralsina> dobey: could be locale? One UTF8 one not?
[14:20] <dobey> both are utf-8
[14:21] <dobey> both are en_US.UTF-8 even
[14:21] <ralsina> dobey: then no
[14:26] <gatox> alecu, the conflicts has been resolved, and the tests has been updated using the new package structure for the imports, all the tests passing in all the platforms: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-tests-refactoring/+merge/113289
[14:27] <alecu> gatox: great. I'm branching it
[14:27] <mmcc> alecu - you need lp:testtools in your PYTHONPATH.
[14:27] <gatox> alecu, thx
[14:27] <mmcc> alecu, you need testools in your PYTHONPATH to run dirspec's tests
[14:27] <alecu> mmcc: thanks!
[14:36] <ralsina> oh great, playing with the overlay scrollbars broke my quantal :-(
[14:36] <dobey> ?
[14:36] <ralsina> and by "playing" I mean 'uninstalling and reinstalling' and now my session crashes
[14:38] <dobey> grr
[14:39] <dobey> this makes absolutely no sense at all
[14:39] <gatox> dobey, welcome to my world :P
[14:40]  * dobey puts that song on
[14:41] <dobey> i set ulimit for RSS to 384M but process used more than that when I ran it
[14:41] <dobey> and it's getting killed due to ulimit, when run in tarmac
[14:41] <dobey> but for some reason, trunk is landing fine
[14:42] <dobey> and it's also on quantal
[14:42] <dobey> also, trunk lander is not complaining about the file name lengths in tar during the make dist
[14:42] <dobey> so wtf.
[14:43] <ralsina> dobey: versions/implementations of tar?
[14:43] <dobey> ralsina: they would different between quantal and quantal?
[14:43] <ralsina> dobey: yes, there are like 4 tars in the repos :-)
[14:43] <dobey> eh?
[14:43] <ralsina> dobey: or maybe a missing update got them out of sync
[14:43] <dobey> there is only 1 tar
[14:44] <ralsina> dobey: gtar, star are the first two that come to mind
[14:44] <ralsina> then there is bsd tar but surely we are not using that ;-)
[14:44] <dobey> there is only 1 tar though, and they are the same version on both
[14:44] <dobey> and we don't depend on gtar or star so they wouldn't be installed (and they aren't)
[14:45] <ralsina> dobey: then the answer is midichlorians.
[14:45] <dobey> actually there is no gtar in the archives
[14:45] <ralsina> dobey: on ubuntu tar is gtar
[14:46] <dobey> nor is star in the archives
[14:46] <dobey> :)
[14:46] <dobey> starfighter and starplot are though
[14:46] <ralsina> dobey: use cpio, it's more standard!
[14:46] <ralsina> dobey: http://cdrecord.berlios.de/private/star.html
[14:46] <dobey> we'll us rar
[14:47] <dobey> yeah i know what star is
[14:47] <dobey> there's also "guitar"
[14:47]  * ralsina starts project patbena-tar
[14:47] <alecu> that's one of the best names ever
[14:47] <briancurtin> hahah
[14:47] <alecu> I mean "gui-tar".
[14:48] <dobey> but the tar issue isn't what's blocking the landing
[14:48] <dobey> it's the OOM issue that is
[14:48] <alecu> patbena-tar is good too ;-)
[14:48] <ralsina> dobey: could it be it's OOMing globally (the instance) and not via ulimit?
[14:49] <dobey> i'm pretty sure it's due to ulimit
[14:49] <dobey> i just ran the tests in the instance a few times, all successful
[14:50] <dobey> even when i set ulimit for myself to the same as for tarmac
[14:50] <dobey> which is weird; it should have OOMed :(
[14:51] <alecu> mandel: why did you land this?
[14:51] <alecu> mandel: https://code.launchpad.net/~mandel/ubuntuone-client/correct-namespace/+merge/116264
[14:52] <alecu> mandel: oh, crap, it's the branch from this morning.
[14:52] <alecu> mandel: it should not have landed, because it depends on this other branch: https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/correct-namespace/+merge/116263
[14:53] <alecu> hmmm....
[14:53] <dobey> should be fine, except osx is broken until the fsevents on lands i guess?
[14:53] <alecu> dobey: right.
[14:53] <alecu> dobey: but doing darwin reviews is even harder now.
[14:54] <alecu> s/harder/more annoying/
[14:54] <alecu> annoyingerester
[14:55] <dobey> well, you already +1ed the fsevents branch it looks like. maybe ralsina or someone can review it quick?
[14:55] <ralsina> looking...
[14:56] <ralsina> mandel: why are MPs private when trunk is public? Makes no sense! :-)
[14:56] <dobey> it's not
[14:56] <ralsina> oh, not in that project
[14:56] <ralsina> ack
[14:57] <ralsina> will +1 as soon as I run tests
[14:57] <dobey> we should get someone to make them public by default i guess
[15:00] <gatox> me
[15:00] <ralsina> me
[15:01] <mmcc> me
[15:01] <briancurtin> me
[15:01] <dobey> meh
[15:01] <alecu> me
[15:02] <dobey> thisfred: ?
[15:02] <thisfred> me
[15:02] <dobey> gatox: go
[15:02] <gatox> DONE:
[15:02] <gatox> Vacation and relax :D. Update tests-refactoring branch fixing some conflicts with another branches that landed. System Tray menu almost completely functional, investigating a little bit about progress bar in the menu that cause conflicts with some ubuntu menus.
[15:02] <gatox> TODO:
[15:02] <gatox> Add tests to the u1-cp and u1-client menu branches and propose them.
[15:02] <gatox> BLOCKED:
[15:02] <gatox> No
[15:02] <gatox> ralsina, go
[15:02] <ralsina> DONE: confirmed root cause of quantal issues, talking with cimi about it, drafted and negotiating twisted contract,  booooring stuff. TODO:  get an exception in scrollbar-overlay for u1cp (or something else) fix some buglet, finish negotiation NEXT: mmcc
[15:02] <mmcc> DONE: py3-unicode reviews, fixed PIL bug, py2app on 10.7 works now
[15:02] <mmcc> TODO: reviews for mandel, test mac certs in package, first-run code
[15:02] <mmcc> BLCK: none
[15:02] <mmcc> NEXT: briancurtin
[15:02] <briancurtin> DONE: walking back through py3-unicode-part-3 now that 1&2 are merged
[15:02] <briancurtin> TODO: debug a test failure in this part-3 before i propose it
[15:02] <briancurtin> NEXT: dobey
[15:03] <dobey> DONE: releases/uploads
[15:03] <dobey> TODO: client release/upload, fix review issues in refactor branch, reorg to drop ubuntuone-installer for good, SRUs for lucid/natty/oneiric
[15:03] <dobey> BLCK: OOM landing update-4-0 branch for client. :-/
[15:03] <dobey> alecu: go
[15:03] <alecu> DONE: pushed forward the review queue. We need more hands there! Started a 3k cleanup branch for storage protocol
[15:03] <alecu> TODO: more reviews, work on a bug list of py3k for u1c and u1sp, mumble re mac port
[15:03] <alecu> BLOCKED: no
[15:03] <alecu> NEXT: thisfred
[15:03] <thisfred> DONE: u1db example app, Bug #1030947 TODO: example app | package 0.0.5 | documentation BLOCKED: no NEXT: no
[15:04] <gatox> TODO: 1-1 with ralsina.... 14:30 is ok?
[15:06] <ralsina> gatox: sure
[15:08] <mandel> so, I got disconnected, sorry
[15:09] <mandel> my turn?
[15:09] <mandel> DONE: More work on the jenkins daemon. Got the daemon running an copying the events to a set of namedpipes.
[15:09] <mandel> TODO: mumble with ralsina, alecu and mmcc. Adapt real daemon to read from namedpipes in test mode. Canonical admin to get a few days off.
[15:09] <mandel> BLOCKED: no
[15:09] <mandel> ralsina, I think we can make the project public to be honest
[15:10] <mandel> @ping
[15:11] <alecu> mandel: @pong
[15:11] <alecu> so, before EOM: we need more eyes on the review queue.
[15:11] <mandel> alecu, thx.. I though i was still off
[15:12] <mandel> alecu, you guys saw my standup stuff?
[15:12] <dobey> yes
[15:12] <mandel> :)
[15:12] <dobey> alecu: today is you/ralsina, is that not enough?
[15:12] <mandel> alecu, are you in mumble already?
[15:12] <ralsina> dobey: I have been doing other stuff and neglecting reviews
[15:13] <ralsina> dobey: so I need some backup
[15:13] <alecu> dobey: there are a lot of darwin branches by mandel still, and it's not trivial to get everything up to speed to review them.
[15:13] <dobey> alecu: maybe gatox and mmcc can help out with those then?
[15:14] <alecu> dobey: right.
[15:14] <mmcc> I'll be doing some darwin reviews, today, yeah
[15:14] <ralsina> mandel: +1 on ubuntuone-fsevents-daemon/correct-namespace
[15:14] <mandel> ralsina, awesome
[15:14] <gatox> alecu, yes..... should i just start reviewing mandel's branches or do you have sometihng specific in mind?
[15:14] <alecu> mandel: btw: a few of those branches have a needs fixing, so before you move forward with other branches, please fix them and annoy people to have the current ones land.
[15:15] <mandel> alecu, they are all fixed from my morning work AFAIK
[15:15] <mandel> alecu, unless there are new comments
[15:15] <mmcc> no, you just need to annoy me to look at them again
[15:15] <alecu> mandel: ok
[15:16] <mandel> mmcc, feel annoyed :-)
[15:16] <mmcc> mandel: done!
[15:16]  * mmcc is downloading mumble
[15:17] <mmcc> I'm on 3g today, so cross your fingers about mumble quality
[15:18] <gatox> mandel, can you please tell me which branches do you need review? i see a lot..... or should i pick randomly? :P
[15:18] <dobey> ok, need to get lunch. bbiab; am starving
[15:18] <mmcc> tomorrow afternoon: fiber to the home :)
[15:18] <mandel> gatox, pick randomly :)
[15:18] <gatox> mandel, ack
[15:19] <alecu> mmcc: let us know if you prefer a hangout
[15:20] <alecu> ralsina: can you join us on the mumble?
[15:20] <ralsina> alecu: going
[15:20] <mmcc> alecu: let's try mumble and I'll let you know :)
[15:21] <alecu> gatox: I've some ideas regarding the progressbars on menues. Perhaps we can discuss about them after the osx mumble?
[15:22] <gatox> alecu, ok...... i've been looking at that.... and now i realize what is going on..... so i can tell you that too
[15:22] <mmcc> logging in to mumble, sorry this is taking so long
[15:27] <gatox> alecu, the mumble about the menu is going to be soon? or i can go to have lunch?
[15:28] <mandel> alecu, ralsina bug #1031815
[15:29] <alecu> gatox: go have, lunch.
[15:29] <gatox> alecu, ack
[16:11] <mmcc> brb
[16:23] <ralsina> lunchtime for me
[16:29] <gatox> alecu, let me know when you want to mumble
[16:35] <alecu> gatox: can we do it in an hour or so? I'm about to leave for lunch.
[16:35] <gatox> alecu, yes, no problem....... i'm reviewing mandel's branches right now
[16:38] <mandel> alecu, gatox I love you! a number of branches landed, thx!!! and also thx to mmcc who was the first!
[16:39] <alecu> mandel: this is not about loving... this is about using you as a mule for board games from spain :-)
[16:39] <gatox> jajajaja true story
[16:40] <dobey> GRRRRRRR
[16:41] <alecu> when dobey grrrrs, everybody whistles away in fear....
[16:41] <alecu> dobey: what happened?
[16:41] <mandel> alecu, lol
[16:42] <dobey> alecu: the OOM in the -quantal tarmac instance. i can't figure out why it's not also happening on the -trunk instance (which is also quantal)
[16:43] <dobey> running tests for both my branch, and trunk, in the -quantal instance, and u1trial goes up to about 428M RSS. which is well over the 384M RSS max set with u1limit :-/
[16:43] <alecu> dobey: do both VMs have the same amount of memory and swap?
[16:43] <dobey> appear to yes
[16:43] <dobey> both have 0 swap and 2G total mem shown in 'top'
[16:44] <dobey> of course doesn't tell me why it's using so much RSS to run the tests now either
[16:58] <mandel> well, EOD here, catch you all tom!
[16:58] <gatox> mandel, bye!
[16:58] <mmcc> bye mandel
[17:04] <dobey> hmm
[17:08] <dobey> maybe i should just land this by hand for now
[17:12] <dobey> too bad i can't tweak the summary output for the test runner easily
[17:26] <gatox> ralsina, 1-1 in 5min? irc or mumble?
[17:26] <dobey> not a hangout?
[17:26] <dobey> disappoint
[17:27] <briancurtin> argh, i still need to figure out that reocurring hangout meeting invite thing
[17:31] <gatox> ok, alecu and ralsina let me know when you have some time to mumble (i need a mumble with each of you :P)
[17:31]  * gatox goes back to the menu
[17:40] <ralsina> gatox: is now good?
[17:41] <gatox> ralsina, yap...... irc or mumble?
[17:41] <ralsina> gatox: mumbles
[18:04] <dobey> ralsina: should i just merge my update-4-0 branch into stable-4-0 by hand? the constant OOM from tarmac is getting annoying, especially since i can't seem to isolate why it's not happening in the other instance, or why u1client is using so much RSS to run the tests now
[18:05] <ralsina> dobey: YES
[18:05] <ralsina> oops, sorry for shouting
[18:05] <ralsina> dobey: maybe we can reset that instance later?
[18:05] <dobey> it was recently, afaik
[18:05] <ralsina> dobey: after you merge, I mean
[18:05] <dobey> i don't think the instance is the problem
[18:05] <ralsina> dobey: maybe even redeploy it
[18:06] <ralsina> dobey: there comes a time when fiddling to fix is just not worth it if deployment is automatic
[18:06] <dobey> it's clear why the OOM is happening. just not clear why the mem usage shot up in the tests, or why it's not happening on trunk
[18:06] <ralsina> dobey: yes, but we are having two apparently disconnected strange symptoms only on that instance (OOM and tar)
[18:07] <ralsina> dobey: Gregory House would raise an eyebrow
[18:07] <ralsina> dobey: and pop a pill or two
[18:07] <dobey> well, the tar thing is odd, but not fatal
[18:07] <dobey> and easy enough to fix
[18:08] <ralsina> dobey: ok, so yes to merging manually, trust you on what to do next, ping me if you don't :-)
[18:11] <mmcc> ralsina: are you doing reviews today? this one is short: https://code.launchpad.net/~mikemc/ubuntu-sso-client/fix-1031437-PIL-exception/+merge/117504
[18:11] <ralsina> mmcc: sure
[18:23] <ralsina> dobey, alecu: up to a very simple review that needs quantal? https://code.launchpad.net/~ralsina/ubuntuone-control-panel/no-scroll-overlay/+merge/117757
[18:25] <ralsina> dobey: that's a workaround to disable the overlay. Alternative is to find out what our  g_get_prgname is by patching the gtk module and asking Cimi to add an exception in the module, which sounds like a lot more work for the same thing.
[18:25] <alecu> ralsina: I don't like that we are removing this for all of nightlies on other releases other than Q
[18:26] <ralsina> alecu: there is no overlay-scrollbar module in any other ubuntu that's not Q
[18:26] <dobey> alecu: other versions of ubuntu don't have overlay scrollbars in qt anyway
[18:26] <alecu> ack then
[18:27] <ralsina> dobey: the bugno is there in the branch and in the commit message
[18:28] <ralsina> dobey: will add the comment though
[18:28] <dobey> ralsina: right, but that's only helpful if you know which commit added it :)
[18:28] <ralsina> dobey: don't understand
[18:28] <dobey> and figuring that out 500 commits later, can be troublesome
[18:28] <dobey> much easier to just see in the comment
[18:29] <ralsina> Oh, add the bugno in the comment! Gotcha
[18:29] <alecu> ralsina: also: TDD :-)
[18:29] <ralsina> alecu: there are no tests for main, because we can't have no tests for main :-(
[18:29] <dobey> ralsina: oh, yes. sorry if that wasn't clear :)
[18:29] <alecu> ralsina: please move that code inside the main module, then!
[18:29] <ralsina> alecu: ok, sounds logical. Will do.
[18:30] <dobey> i thought about mentioning the lack of a test, but you really can't test it
[18:30] <alecu> dobey: you can certainly test that whatever comes in GTK_MODULES does not break the parsing
[18:30] <ralsina> dobey: I can set GTK_MODULES and test that after calling main the overlay-scrolbar is gone from it
[18:30] <dobey> a unit test for this is basically a very expensive no-op
[18:31] <alecu> dobey: and that overlay-scrollbar is actually removed if it's there...
[18:31] <alecu> and, etc, etc.
[18:31] <ralsina> well, it's accidental-deletion prevention mostly
[18:32] <dobey> alecu: GTK_MDOULES is not set outside of the user session; and setting it in the test with a custom list of modules, means the test will always pass unless someone makes a really stupid mistake :)
[18:38] <alecu> gatox: let me know when you want to mumble
[18:38] <gatox> alecu, we can do it now if you want
[18:39] <dobey> well, i have a fix for the tar issue anyway
[18:39] <alecu> coming...
[18:40] <ralsina> grmbl trial should have a "run only the last test that failed" option
[18:40] <dobey> ?
[18:41] <ralsina> dobey: as you would say "sigh" ;-)
[18:42] <dobey> not in this case. it's certainly something that could be added to dev tools, but it's not entirely clear to me at the moment how you'd expect that to work
[18:42] <dobey> on the other hand
[18:42] <ralsina> dobey: when I am trying to ix a test, I would run that to know if the fix worked, and then run the whole suite
[18:43] <dobey> alecu, ralsina: trivial review please: https://code.launchpad.net/~dobey/ubuntuone-client/tar-ustar/+merge/117761
[18:44] <dobey> ralsina: right. i think the annoying bit there is all the ignoring of other files/paths we do, rather than having things just get skipped
[18:44] <ralsina> dobey: could be just reordering of the tests
[18:45] <dobey> at least, running all the tests in only a single file is easy
[18:45] <ralsina> dobey: although the change there looks much harder of course :-)
[18:45] <mmcc> is it a bug to have multiple linked devices listed for the same device? my devices list has two entries for my iPhone and two for my macbook...
[18:45] <dobey> what i'd really like is some memory summarization
[18:46] <dobey> mmcc: not necessarily, though it is annoying; depends on how you deleted the previous token, or how you managed to log in twice
[18:47] <dobey> brb
[18:49] <mmcc> ok, thanks dobey. I figured I was tweaking some edges, but wondered if I had broken anything in doing so
[18:49] <mmcc> sounds like not
[18:51] <ralsina> grrrr python's os.putenv doesn't change what os.getenv returns
[18:58] <mmcc> ralsina: and remember os.environ is not a real dict, if you need a copy, call environ.copy()
[18:58] <ralsina> mmcc: yeah, the thing is, I can't test it because I am modifying the "real" environment, but later in the test I am gettin the values from the fake never-updated one!
[18:58] <ralsina> mmcc: and the suggested solution is to use ctypes' gentev!!!!!
[18:59] <ralsina> Or, I can patch os.putenv to act like a decent function :-(
[18:59] <mmcc> ralsina: if you just set os.environ['foo'] that should work- getenv will read that correctly
[19:00] <ralsina> mmcc: but the real code uses a real putenv because it has to be seen from C
[19:00] <ralsina> mmcc: does setting os.environ work that way?
[19:01] <mmcc> the docs say os.environ calls putenv() on systems where it exists…
[19:01] <ralsina> mmcc: ok then, thanks!
[19:02]  * mmcc is curious to see if that works:)
[19:02] <ralsina> mmcc: will let you know ;-)
[19:03] <ralsina> dobey: global +1 on tar-ustar
[19:21] <ralsina> mmcc: +1 on the PIL branch
[19:21] <mmcc> thanks!
[19:23] <ralsina> alecu, dobey: re-review when you have a minute please https://code.launchpad.net/~ralsina/ubuntuone-control-panel/no-scroll-overlay/+merge/117757
[19:26] <dobey> ralsina: +1
[19:27] <ralsina> dobey: thanks
[19:27] <ralsina> mmcc: it did work
[19:27] <mmcc> ralsina: hooray
[19:27] <ralsina> mmcc: somehow the whole os.environ[]/putenv/getenv thing feels just wrong
[19:27] <mmcc> ralsina: only somehow?
[19:28] <ralsina> mmcc: ok, it feels really wrong
[19:29] <mmcc> yeah it's unfortunate. I saw a historical explanation about why it's weird somewhere but I forget where
[19:31] <alecu> breaking news: amelia loves "synergy" as much as I do.
[19:31] <alecu> she was just amazed moving the mouse between screens
[19:34] <ralsina> alecu: hahaha
[19:36] <ralsina> alecu: beware my son now expects all TVs everywhere to have netflix and all phones to have plants vs zombies
[19:37] <ralsina> alecu: so amelia will expecto to be able to move pointers to all screens in the house :-)
[19:38] <dobey> alecu: i'm just amazed synergy hasn't crashed for you :)
[19:38] <alecu> ralsina: thanks for adding tests to that branch. One tiny thing: since test customs suggest one assertion per test, would you mind making three tests out of it? Even one _test method that gets called with the content to put in GTK_MODULES, and the expected result.
[19:38] <dobey> ok, brb again
[19:38] <ralsina> alecu: sure, no problem, the only thing I don't like about that is triplicating the cleanup function
[19:39] <ralsina> So a _test it is
[19:40] <alecu> ralsina: right: a _test that receives two params, and that's called by three test_xxx methods.
[19:40] <alecu> ralsina: the cleanup can be in _test
[19:41] <ralsina> alecu: right thanks
[19:41] <alecu> dobey: I'm using the unstable release of synergy; the server is Precise, the client is osx 10.8
[19:43] <alecu> dobey: and if it crashes I don't notice, since I've got a script that restarts the client and its ssh tunnel whenever it breaks.
[19:43] <alecu> (for instance, after unsuspending the laptop)
[19:44] <mmcc> alecu: test style question - why use a _test helper function instead of making a separate TestCase subclass and using setUp (and tearDown for the cleanup)?
[19:46] <alecu> mmcc: we tend not to use tearDown: addCleanup is much nicer, since you can put the cleanup right next to the thing being setup
[19:46] <ralsina> dobey: did you already do the quantal release?
[19:47] <ralsina> dobey: because I am thinking this branch really should go in it
[19:47] <alecu> mmcc: even when doing setup stuff in setUp, it's much nicer to put the cleanup code in addCleanups right after setting something up.
[19:48] <alecu> mmcc: also, for these kind of tests, we would still need a _test helper if we want to parametrize the tests... right?
[19:48] <alecu> mmcc: I don't have anything against setUp though, and for this might make sense.
[19:49] <alecu> mmcc: anyway, I'm just a journeyer on these matters; we can find an expert at a ping notice.
[19:49] <mmcc> alecu: your point about addCleanup makes sense. and yeah, the helper is good if you want to avoid writing "main.main(blah) three times ;)
[19:50] <alecu> mmcc: in this case it would be setting the environ, calling main and doing the assert...
[19:51] <mmcc> alecu, thanks for explaining. I don't really have an opinion, just curious
[19:51] <mmcc> alecu: right.
[19:51] <alecu> mmcc: it's considered bad form to do asserts in the cleanup.
[19:51] <mmcc> alecu: yes, that makes good sense
[19:57] <ralsina> alecu: re-re-review please?
[19:57] <alecu> sure!
[20:01] <mmcc> ok, lunchtime here.
[20:06] <gatox> ok..... eod for me!! bye people! :D
[20:08] <ralsina> bye gatox!
[20:08] <gatox> ralsina, see you
[20:09] <dobey> ralsina: which branch?
[20:09] <alecu> ralsina, dobey: is GTK_MODULES set when running on tarmac?
[20:10] <alecu> if not, then the tests will probably break.
[20:10] <dobey> alecu: no; there's no user session.
[20:14] <ralsina> alecu: hmmm let me check
[20:14] <dobey> alecu: if the tests depend on the user's environment being a certain way, then the tests are broken anyway :)
[20:15] <ralsina> dobey: it's the restoring that may break
[20:17] <alecu> ralsina: I think it's "old_modules = os.environ['GTK_MODULES']" that will break if there's no GTK_MODULES set
[20:17] <ralsina> alecu: yes, I am replacing that with a get('GTK_MODULES', '')
[20:17] <dobey> that should use os.environ.get()
[20:17] <ralsina> alecu: that may change the variable from None to '' though
[20:17] <dobey> but good catch :)
[20:18] <ralsina> which should make no difference
[20:21] <ralsina> alecu: there, done using get()
[20:22] <dobey> ralsina: so, which branch should go in quantal?
[20:22] <dobey> ralsina: this gtk_modules hack?
[20:22] <ralsina> dobey: lp:~ralsina/ubuntuone-control-panel/no-scroll-overlay
[20:22] <ralsina> dobey: yes
[20:22] <ralsina> dobey: without it u1cp i completely broken
[20:22] <dobey> i'll do another upload with it as a patch after it lands
[20:22] <ralsina> dobey: awesome, thanks
[20:32] <thisfred> ralsina, been googling but I can't find how to do this: I have a treewidget, with editable items (yay!) but any time I start editing an item, its text is cleared. That seems like a strange default.
[20:33] <thisfred> I found something about openPersistentEditor which may be related, but I can't find how to hook that up either
[20:33] <ralsina> thisfred: haven't actually done that before :-)
[20:33] <thisfred> ah
[20:33] <ralsina> thisfred: so, if you push, I'll pull ;-)
[20:33] <thisfred> will do
[20:33] <dobey> thisfred: it clears the text before you start typing, or when you type?
[20:34] <thisfred> dobey, when you double click the item, before it "turns editable"
[20:34] <dobey> that sounds horribly broken, yeah
[20:35] <thisfred> ralsina: lp:~thisfred/u1db/u1todo-design-1
[20:35] <ralsina> thisfred: getting
[20:35] <thisfred> then cd into that, and do PYTHONPATH=. python cosas/ui.py
[20:35] <ralsina> thisfred: ack. Will have an answer in ... [shakes magic 8-ball] 12 minutes
[20:36] <thisfred> outlook not so good
[20:36] <thisfred> that's what mine says. And yesterday's Microsoft announcement hasn't changed its mind.
[20:37] <alecu> thisfred: your ball might have bad taste!
[20:37] <ralsina> alecu: stop licking your magic ball!
[20:37] <ralsina> sorry couldn't help myself
[20:38] <thisfred> hey, you're the manager, if that's the standard you want to set... :P
[20:38] <ralsina> thisfred: got a fix forya
[20:38] <thisfred> wow
[20:39] <ralsina> thisfred: https://pastebin.canonical.com/71337/
[20:39] <ralsina> thisfred: this is the stuff I am **good** at, remember ;-)
[20:39] <alecu> ralsina: +1
[20:39] <thisfred> ah so I was screwing it up because of the order of tests
[20:40] <ralsina> thisfred: it looks like it was returning None for editRole
[20:40] <ralsina> not that I actually *tried* to see what it was returning
[20:40] <ralsina> alecu: awesome, thanks!
[20:40] <thisfred> ralsina, I should have just returned from super there, maybe
[20:40] <ralsina> thisfred: could be
[20:41] <ralsina> thisfred: this fix is explicit and trivial though
[20:42] <thisfred> ralsina, yeah, and super doesn't work
[20:42] <thisfred> ralsina, thanks!
[20:42] <ralsina> thisfred: you're welcome!
[20:51] <ralsina> EOD for me
[20:52] <ralsina> I will, as usual, pop up at night, feel free to mail me if you need anything
[21:00] <thisfred> ralsina, good night, and when you're bored (or tomorrow) I'd appreciate a review on https://code.launchpad.net/~thisfred/u1db/u1todo-design-1/+merge/117796
[21:02] <dobey> alright, gotta run. later
[21:04] <alecu> it's EOD for me too. Bye all!
[22:35] <briancurtin> alrighty, well enough unsuccessful debugging for the afternoon. EOD.
[22:58] <mmcc> hrm, I guess we all have 64-bit macs. fsevents daemon doesn't run on old 32-bit objc runtime...
[22:58] <mmcc> hooray for keeping "old" (2010) hardware around