=== zyga-afk is now known as zyga
=== zyga is now known as zyga-gone
[06:54] <mandel> morning all!
=== zyga-gone is now known as zyga
[10:20] <JamesTai1> Good morning, and happy World IPv6 Day, everyone! :D
[11:12] <gatox> good morning
[11:20] <mandel> gatox, morning!
[11:20] <mandel> gatox, I need you :)
[11:20] <mandel> gatox, well, I need you evil vm hehe
[11:20] <gatox> mandel, jeje ok!! shoot
[11:20] <gatox> (to my machine)
[11:21] <mandel> gatox, can you use the branch from https://code.launchpad.net/~mandel/ubuntuone-dev-tools/darwin-unix/+merge/108918 to run the tests?
[11:21] <mandel> gatox, fixes a problem we had in mac os x + tcpactivation
[11:21] <gatox> mandel, yep..... on it
[11:22] <gatox> mandel, just sso or something else?
[11:22] <mandel> gatox, sso and u1-client is enough
[11:22] <gatox> mandel, aack
[11:22] <mandel> gatox, thx a lot!
[11:22] <gatox> mandel, no problem!
[11:50] <gatox> mandel, u1-client is ok, but with sso i'm getting this: http://paste.ubuntu.com/1026675/
[11:50] <gatox> run it 3 times
[11:52] <mandel> gatox, ok.. can you paste that in the mp please, I'm about to finish with some tests and I'll get to it
[11:52] <gatox> mandel, ack
[12:17] <mandel> gatox, a much better python client for the daemon with tests etc... and ofcourse using twisted and async: https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/improve-python-client/+merge/108926
[12:17] <mandel> gatox, I'm moving to the errors you had right now..
[12:18] <ralsina> good morning!
[12:18] <mandel> ralsina, morning!
[12:18] <ralsina> anyone mentioned reviews?
[12:18] <mandel> ralsina, I have a number for the fsevents daemon :)
[12:18] <mandel> ralsina, like, lots
[12:19] <ralsina> mandel: hit me! (with one)
[12:19] <mandel> ralsina, python or objective-c ?
[12:20] <ralsina> mandel: python preferred
[12:20] <gatox> ralsina, hi
[12:20] <gatox> mandel, reviewing....
[12:20] <mandel> ralsina, ok, then here you can find the python side of the daemon https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/improve-python-client/+merge/108926
[12:21] <mandel> ralsina, I want to make the json message smaller (use int enums instead of strings etc..) but that code there does work perfectly
[12:21] <mandel> ralsina, gatox, if you do ./run-tests you will get the object-c tested + the python too (including lint and pep8)
[12:21] <mandel> but I need to configure the lint correctly to not complain with some many stupid things
[12:22] <mandel> tests also create a 'fake daemon' to test the event retrieval and the config interaction (is just a serverfactory running in a unix socket)
[12:37] <mandel> ok, I'm off to have lunch
[12:39] <dobey> mandel: use pyflakes :)
[12:43] <ralsina> Have to run a small errand, will be back in 30' or so
[13:05] <alecu> gatox, ping
[13:05] <gatox> alecu, pong
[13:05] <alecu> gatox, http://wiki.maemo.org/Summer%2712_Device_Program/Qt_5_Mobile_Projects
[13:06] <alecu> gatox, a N950 or N9 for qt developers wishing to test their apps.
[13:06] <gatox> alecu, yes, i tweeted that a couple of times
[13:07] <alecu> gatox, awesome. I've quitted twitter a few weeks ago... it was too taxing on my limited time :-)
[13:07] <alecu> bah, I quitted reading twitter, I still listen to mentions. :-)
[13:27] <dobey> quit :)
[13:43] <mandel> dobey, true :)
[13:44] <mandel> oh, and I'm back
[13:48] <dobey> i hate updating to a newer ubuntu and having default server configs being a wildly different structure :-/
[13:48] <mandel> gatox, so, is there a way I could login in you windows vm over vnc?
[13:48] <gatox> mandel, let me check
[13:49] <gatox> mandel, there was a really simple program for that.... let me check
[13:49] <ralsina> mandel: your branch looks good to me, but I am not at home so have no mac to run the tests. +1 for code review only
[13:50] <mandel> gatox, that way I can look closer at the issue without being a pain in the ass for you
[13:50] <mandel> ralsina, great! I'll be adding some more bugs to that code to increase performance etc.. but I'll probably will integrate with u1-client before I do that
[13:51] <ralsina> hmmmm why would adding bugs improve performance ? Are they very fast bugs? Like dragonflys?
[13:51] <mandel> ralsina, super fast! do not send a string with FSE_EVENT_CREATE_FILE but with 0 for example :)
[13:52] <mandel> ralsina, well make the messages smaller which I think is a plus an very easy to implement, also I want to remove a not used key
[13:52]  * ralsina orders the bug-zapping tennis racket from DX.com
[13:52] <ralsina> http://dx.com/p/electronic-bug-zapper-racket-3441?item=1
[13:53] <mandel> ralsina, I'd love to see someone playing tennis with that :)
[13:53] <ralsina> mandel: you'd need very tough bugs
[13:53] <gatox> mandel, have you ever use remote desktop on windows before?
[13:54] <mandel> gatox, I have used it in the client side, yes
[13:54] <ralsina> gatox: you can just install vnc
[13:54] <mandel> gatox, we can also use http://www.teamviewer.com/en/index.aspx
[13:54] <gatox> mandel, i activate it.... but i don't know what i need to give you, a ip, some key.... what?
[13:54] <ralsina> gatox: easier
[13:54] <mandel> gatox, but yes, I prefer vcn to be hones
[13:54] <ralsina> or logmein.com if it's meant to be used remotely
[13:55] <gatox> checking...
[13:57] <alecu> wow! the security bugs are no longer private! bug #882062 and bug #882055 !!!!
[13:57] <ubot5> Launchpad bug 882062 in ubuntuone-client (Ubuntu Quantal) "ubuntuone-client doesn't validate ssl certificates" [Medium,Confirmed] https://launchpad.net/bugs/882062
[13:57] <ubot5> Launchpad bug 882055 in ubuntu-sso-client (Ubuntu) "ubuntu-sso-client doesn't validate ssl certificates" [Medium,Confirmed] https://launchpad.net/bugs/882055
[13:57] <alecu> that means that the fixes have been released!!!! \o/
[13:58]  * alecu works on a branch for Quantal :P
[13:58]  * alecu installs a Quantal VM, first.
[14:00] <ralsina> alecu: yay!
[14:01]  * alecu should also make branches to propose to every stable release
[14:02] <alecu> uf
[14:03] <ralsina> alecu: for what bug?
[14:04] <alecu> ralsina, for both of the above bugs.
[14:04] <ralsina> alecu: ok, yes :-/
[14:04] <alecu> ralsina, I worked on them as patches, as the security team needed.
[14:04] <ralsina> right
[14:04] <alecu> ralsina, since all branches are public.
[14:05] <alecu> anyway, it's not much work: just tedious :-)
[14:35] <dobey> switching server hardware is always so tedious
[14:53] <ralsina> team, standup in 7'
[14:53] <ralsina> And by team I mean dobey alecu gatox briancurtin mmcc (where are you?) thisfred and mandel
[14:53] <ralsina> and if I forgot one: you too.
[14:53] <mandel> ack
[14:54] <mmcc> ralsina, I'm in texas
[14:54] <briancurtin> writing now
[14:54]  * mmcc forgot to say hi
[14:54] <ralsina> mmcc: weird my autocompletion failed with you for a second :-)
[14:59] <alecu> me
[15:00] <gatox> me
[15:00] <mmcc> me
[15:00] <briancurtin> me
[15:01] <thisfred> me
[15:01] <dobey> meh
[15:02] <mandel> me
[15:03] <mandel> shall we start?
[15:05] <dobey> i would presume so
[15:05] <mandel> @ping
[15:05] <dobey> alecu: go
[15:05] <dobey> ralsina: donde?
[15:05] <ralsina> me
[15:05] <alecu> DONE: patch for bug #1006899, not managed to reproduce #1008662, got security patches released
[15:05] <alecu> TODO: branches for all stables for security issues (bug #882062 and bug #882055)
[15:05] <alecu> BLOCKED: no
[15:05] <alecu> NEXT: gatox
[15:05] <ubot5> Launchpad bug 1006899 in Ubuntu One Windows Installer trunk "[Windows] Ubuntu One syncdaemon tries to connect directly even when proxy is used (3.0.1)" [Undecided,New] https://launchpad.net/bugs/1006899
[15:05] <ralsina> damn phone
[15:05] <gatox> DONE:
[15:05] <gatox> Run some tests in my windows evil machine for mandel, Debugging and debugging on macfsevents, trying to fix the schedule operation.
[15:05] <gatox> TODO:
[15:05] <gatox> Finish with the schedule issue and go back to fix the remaining tests.
[15:05] <gatox> BLOCKED:
[15:05] <ubot5> Launchpad bug 882062 in ubuntuone-client (Ubuntu Quantal) "ubuntuone-client doesn't validate ssl certificates" [Medium,Confirmed] https://launchpad.net/bugs/882062
[15:05] <gatox> No
[15:05] <ubot5> Launchpad bug 882055 in ubuntu-sso-client (Ubuntu) "ubuntu-sso-client doesn't validate ssl certificates" [Medium,Confirmed] https://launchpad.net/bugs/882055
[15:05] <gatox> mmcc go!
[15:05] <mmcc> DONE: reviews, devtools test setup, cleaning up setup-mac.py
[15:05] <mmcc> TODO: confirm packaging proxy-tunnel worked, share setup-mac.py
[15:05] <mmcc> BLCK: NO
[15:05] <mmcc> NEXT: briancurtin
[15:05] <briancurtin> DONE: good progress on building testability (finally!), getting the supplemental parts built as well.
[15:05] <briancurtin> TODO: build an installer with alecu's twisted change, push on testability to get it wrapped up
[15:05] <briancurtin> BLOCKED: no
[15:05] <briancurtin> NEXT: thisfred
[15:05] <thisfred> DONE: https://code.launchpad.net/~thisfred/u1db/home-on-the-range/+merge/108606 ( Bug #999585 ), planning poker TODO: https://code.launchpad.net/~thisfred/u1db/and-the-wild-rocky-mountains-to-roam ( Bug #999585 ), Bug #1009505 BLOCKED: no NEXTH: dobeyh
[15:05] <dobey> λ DONE: releases, uploads, SRU poking, client/cp 3.0.1 in stable PPA, bug #1009117
[15:05] <ubot5> Launchpad bug 999585 in U1DB "support range queries" [High,In progress] https://launchpad.net/bugs/999585
[15:05] <ubot5> Launchpad bug 1009505 in U1DB "get_keys_from_index is useless for multicolumn indexes" [High,Triaged] https://launchpad.net/bugs/1009505
[15:05] <ubot5> Launchpad bug 1009117 in ubuntuone-dev-tools (Ubuntu Quantal) "Cannot run tests on Mac OS X" [Undecided,Triaged] https://launchpad.net/bugs/1009117
[15:05] <dobey> λ TODO: finish releases/uploads, SRU poking, tarmac tweakery
[15:05] <dobey> mandel
[15:05] <mandel> DONE: Added all new python client for fsevents daemon. Worked on fixing stucked tests on sso for mac (work on u1devtools)
[15:05] <mandel> TODO: propose the fix for the u1devtools. Start looking at writing some integration tests for the daemon. Move to u1-cleint.
[15:05] <mandel> BLOCKED: no, but I hate that bloody dirty reactor error.
[15:05] <dobey> λ BLCK: None.
[15:05] <mandel> ralsina, please
[15:06] <ralsina> DONE: 1-1 with chipaca, briancurtin, started work on wizard bug, learning about bunch of things, did my site passes
[15:06] <ralsina> some bug triaging
[15:06] <ralsina> TODO: fix that bug
[15:06] <ralsina> BLOCKED: no
[15:06] <gatox> i forgot... todo: 1-1 with ralsina
[15:07] <gatox> comments? eom?
[15:08] <mandel> mmcc, so I fixed the blockage from the tests, yet I get a dirty reactor error now :(
[15:08] <ralsina> eom it is
[15:08] <mandel> mmcc, you can find the current code at lp:~mandel/ubuntuone-dev-tools/darwin-unix
[15:10] <mmcc> mandel: ok - did you see the default reactor changes dobey made yesterday? is this related?
[15:10] <mmcc> looking
[15:11] <mandel> mmcc, no, that was a diff change, by the way, dobey I reviewed and gave a +1 but did not approve the mp
[15:13] <elopio> alecu, mandel, do you have any links or mails about how to set up proxies for testing with u1?
[15:13] <elopio> I have seen some instructions, but until now I'm starting to work on proxy. Sorry about that.
[15:14] <mandel> elopio, hm.. I think I wrote some notes.. I'll have to take a look
[15:15] <dobey> mandel: i know
[15:15] <dobey> mandel: i set it to approved and it's merged now
[15:16] <mandel> dobey, ok :)
[15:16] <elopio> mandel: thanks. Today I'll be trying to document the proxy set up so we can have the proxy testing in MozTrap. Please send me anything you think that can be useful.
[15:18] <mandel> elopio, I can send you ok instructions for squid on linux and windows to have auth and nonauth running in diff ports with the same instance, does that sound useful?
[15:18] <mmcc> mandel, your branch fixes my sso test hang.
[15:18] <elopio> mandel: sounds great.
[15:19] <mmcc> mandel, can you CC me on that email to elopio, just for future reference?
[15:19] <mandel> mmcc, sure :)
[15:19] <mandel> mmcc, so, it fixes the hang but breaks leaving a dirty reactor on slow windows machines..
[15:19] <mandel> I'm trying to fix that atm
[15:20] <mmcc> ok
[15:23]  * gatox lunch
=== gatox is now known as gatox_lunch
[15:30] <mandel> I need to reboot, I shall be back!
[15:30] <dobey> lunch. bbiab
[15:32]  * briancurtin brb, i need to run to the copy shop to fax some stuff. maybe they'll turn my gas back on after this and i can have hot water!
[15:56] <joshuahoover> ralsina: any idea what would cause this error on windows? http://paste.ubuntu.com/1027073/
[15:57] <ralsina> joshuahoover: looking
[15:57] <ralsina> joshuahoover: something that should be visible in the syncdaemon exceptions log
=== gatox_lunch is now known as gatox
=== salgado is now known as salgado-lunch
[16:26] <ralsina> alecu, dobey: it seems our update is failing in lucid, bug #1009573 bug #1009582
[16:26] <ubot5> Launchpad bug 1009573 in ubuntuone-client (Ubuntu) "package python-ubuntuone-client 1.2.2-0ubuntu2.2 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/1009573
[16:26] <ubot5> Launchpad bug 1009582 in ubuntuone-client (Ubuntu) "package python-ubuntuone-client 1.2.2-0ubuntu2.2 failed to install/upgrade: Unterprozess installiertes post-installation-Skript gab den Fehlerwert 1 zurück" [Undecided,New] https://launchpad.net/bugs/1009582
[16:27] <ralsina> that's just the SSL fix, isn't it?
[16:27] <alecu> looking
[16:28] <alecu> ralsina, "except pycurl.error as e:
[16:28] <alecu>                          ^
[16:28] <alecu> SyntaxError: invalid syntax"
[16:28] <alecu> :-(
[16:29] <alecu> hmmmmf
[16:29] <ralsina> Old python in lucid?
[16:29] <ralsina> I thought it supported that :-(
[16:29] <alecu> right.
[16:29] <alecu> but I thought we had 2.6 in lucid.
[16:30] <alecu> checking now.
[16:34] <alecu> ralsina, lucid came with 2.6 and 3.1
[16:34] <ralsina> alecu: then I don't get it
[16:34] <alecu> ralsina, I've read comments that some people were installing 2.5 on Lucid to develop for the google app engine, because that was the version supported at the time.
[16:35] <ralsina> alecu: that may trigger something with /use/bin/env python perhaps
[16:36] <alecu> ralsina, but our code was installed in a python2.5 path... so that's weird...
[16:36] <alecu> Compiling /usr/lib/python2.5/site-packages/ubuntuone/utils/curllib.py ...
[16:36] <alecu>   File "/usr/lib/python2.5/site-packages/ubuntuone/utils/curllib.py", line 134
[16:36] <alecu>     except pycurl.error as e:
[16:36] <alecu>                          ^
[16:36] <alecu> SyntaxError: invalid syntax
[16:36] <ralsina> ugh
[16:36] <alecu> so... "?"
[16:36] <ralsina> then it may be something local, and not a packaging/code problem
[16:37] <briancurtin> if it's installed on a 2.5 path wouldn't that be a packaging problem?
[16:38] <briancurtin> or do you mean someone messed with the system beforehand, which caused u1 to install to 2.5?
[16:39] <ralsina> briancurtin: the second one
[16:39] <ralsina> as in they have a really screwed system
[16:41] <alecu> ralsina, but now I see that the users had 2.6 installed...  https://launchpadlibrarian.net/107012304/Dependencies.txt
[16:41] <ralsina> alecu: confusing and confusinger
[16:41] <alecu> so I think briancurtin may be right
[16:42] <alecu> in both bugs the system had 2.6 installed
=== salgado-lunch is now known as salgado
[16:47] <dobey> hrmm
[16:47] <dobey> i think it's a packaging problem
[16:49] <dobey> XS-Python-Version: >= 2.5
[16:50] <dobey> because previously, we were still supporting 2.5 in that very old version of u1
[16:52] <ralsina> alecu: also bug #1009598 seems to be a bug in our proposed package
[16:52] <ubot5> Launchpad bug 1009598 in u1sync "U1 cannot sync anymore" [Undecided,New] https://launchpad.net/bugs/1009598
[16:52] <ralsina> and/or dobey ^
[16:54] <alecu> ralsina, that looks like ubuntuone-storage-protocol was updated, but ubuntuone-client was not.
[16:54] <ralsina> alecu: ok then, adding a comment
[16:54] <alecu> since get_ssl_context takes two args in the new version of storage-protocol
[16:54] <ralsina> alecu: oh but have we uploaded that storage-protcol to proposed?
[16:55] <alecu> ralsina, the user has the new storage protocol (that takes two args), but has the old u1-client
[16:55] <alecu> ralsina, so it seems we have uploaded it, yes.
[16:55] <ralsina> alecu: haha, ok
[17:03] <dobey> and ugh. api breaks are pain
[17:04] <dobey> particularly in python
[17:04]  * dobey hugs compiled languages
[17:04] <mmcc> dobey, be careful you don't cut yourself on dynamic libraries
[17:05]  * mmcc waves to LD_PRELOAD
[17:05] <dobey> mmcc: but dependency-based rebuilds mean the things that depend on it will fail to build :)
[17:06] <dobey> python doesn't care at all, until that specific api actually gets called by something
[17:07] <mmcc> true
[17:09] <ralsina> lunchtime for me
[17:09] <gatox> ralsina, let me know when you are back, and we can have our 1-1
[17:14] <mandel> EOD here.. I bloody hate twisted sometimes..
[17:14] <mandel> I'll catch with you all tom!
[17:14] <alecu> mandel, lols
[17:14] <alecu> mandel, bye!
[17:14] <gatox> mandel, i understand your pain...... jeje bye!
[17:18] <mmcc> lunch
=== Guest50087 is now known as zumbi
[17:21] <dobey> brain overload
[17:23] <ralsina> gatox: let's say at 3
[17:23] <gatox> ralsina, ack
[17:50] <dobey> alecu: btw. there isn't also a patch for ubuntuone-control-panel for the ssl thing is there?
=== yofel_ is now known as yofel
[17:59] <gatox> ralsina, 1-1 irc or mumble'
[17:59] <gatox> ?
[18:00] <ralsina> gatox: irc
[18:15] <briancurtin> alecu: here's an installer with your dont-open-console-window branch and the twisted branch - http://ubuntuone.com/1y4nTEL7pEcRJ8TgOg3MqW
[18:29] <alecu> briancurtin, awesome, I'll test it right away.
[18:36]  * briancurtin lunch
[19:14] <alecu> kinder run!
[19:21] <ralsina> we seem to still have a bunch of lucid users
[19:27] <dobey> yes we do :-/
[19:45] <dobey> ralsina: are all the bugs coming in for lucid you're referencing with that comment, the python 2.5 issue?
[19:46] <dobey> ralsina: the fix for it was just uploaded to lucid-security, so those should go away soon
[19:52] <ralsina> dobey: they are all from the post-install script and they all look the same
[19:52] <ralsina> dobey: also, should we create distros for windows/mac?
[19:52] <ralsina> dobey: for bugs like 984964 where we have it fixed on ubuntu, but not on windows/mac
[19:54] <dobey> ralsina: i have been wanting to do that, and also for our PPAs, but haven't got round to it
[19:54] <ralsina> dobey: it's ok, not a priority
[19:55] <dobey> i want it. but i think the way lp works wrt distros also makes it a bit weird to do
[20:00] <alecu> briancurtin, the installer is working perfectly. I'm testing it on a vm than can only reach the internet via proxies, and it connected properly, and did not open the dreaded console window.
[20:01] <briancurtin> alecu: awesome. is that twisted branch something that will make it into a release, or is that something we'll need to start including in all of our windows releases?
[20:01] <briancurtin> s/into a release/into a twisted release/
[20:01] <alecu> briancurtin, I'm proposing it for twisted, but I'm not sure if they are going to like the patch as is.
[20:01] <gatox> alecu, the problem with the missing event..... was in the algorithm of the lib that try to detect the file changes using stat...... i'm fixing that...... just to let you know
[20:02] <alecu> gatox, in what file/function was it?
[20:02] <gatox> alecu, fsevents.py > FileEventCallback > __call__
[20:04] <alecu> gatox, I've opened that file. So... what was the bug?
[20:04] <ralsina> dobey: we can use some of the weird ones as codenames, like, say, make Elbuntu mean windows ;-)
[20:04] <dobey> let's not do that
[20:05] <ralsina> dobey: yes, was a joke, will not do
[20:05] <ralsina> unless you also meant let's not do elbuntu, which I also agree with
[20:06] <gatox> alecu, i'm still debugging to see how it fix it properly..... but the event is reaching that method correctly, the problem is that there is some conflict in the snapshot of the folder, and it gets confuse adding the new file to the snapshot, and then using it to compare a modify event, which it shouldn't because it shouldn't be in the snapshot, and because the stat is the saame, the modifications is rejected, and the file is discarded so the
[20:06] <gatox>  create event is not created in the next for-loop
[20:07] <gatox> alecu, i think i almost have it
[20:14] <dobey> ok, i need to go now. have an appointment. may do a little more later tonight. later everyone
[20:15] <gatox> alecu, ahhh the problem seems to be in the "snapshot" function...
[20:15] <gatox> in the first exxeccution, that  modifies the snapshots var with some data that i shouldn't
[20:17] <alecu> gatox, I'm not following, but don't worry, go ahead, and then you can tell me later :-)
[20:18] <gatox> alecu, ok.... i'm looking at 2 possible solutions.... i'm going to try it and then i can explain what happend
[20:20] <gatox> ok..... eod for me..... and i need a rest from this before start testing the same things again :P
[20:20] <gatox> see you tomorrow people!
[20:21] <ralsina> bye gatox!
[20:21] <gatox> ralsina, bye o/
=== AJenbo_ is now known as AJenbo
[20:51] <ralsina> EOD for me
[20:51] <ralsina> bye people!
[20:51] <briancurtin> bye ralsina, also a late bye to gatox
[21:13] <mmcc> so it looks like 100 stack frames isn't enough to print to see what executable is printing this Qt warning
[21:14]  * mmcc goes back to editing my private copy of  dtruss
[21:16] <mmcc> 122 stack frames is the correct answer
[21:27] <alecu> briancurtin, when you have a few minutes, let's think how we can make jenkins use the patched twisted.
[21:28] <alecu> briancurtin, if you have any idea, please add it to the comments in https://code.launchpad.net/~alecu/ubuntuone-windows-installer/tunnel-use-console/+merge/108279
[21:28] <alecu> and this is an EOD for me!
[21:28] <briancurtin> alecu: what i did on my machine was just manually apply the change to the twisted we have within buildout. i'll think about a better way
[21:29] <alecu> briancurtin, great, thanks.
[21:29] <alecu> briancurtin, I think we should not approve that branch till we have jenkins building with the patched twisted
[21:30] <alecu> ok, EOD for me.
[21:30] <alecu> bye all!
[22:46] <mmcc> my god, I think I have to build a custom Qt
[22:54] <mmcc> which I'll do tonight and let it run over night. For reference, I'm hitting Qt bug #24541: https://bugreports.qt-project.org/browse/QTBUG-24541
[22:54] <ubot5> Launchpad bug 12893 in Baltix "duplicate for #24541 Shouldn't put .Trash-$USER on removable devices" [Undecided,New] https://launchpad.net/bugs/12893
[23:01] <mmcc> heh.
[23:03] <mmcc> the problem is that the ubuntu-sso-login executable doesn't read its qt.conf even when we put it in the right place, because QTCore QLibraryInfo won't look for it there on OSX unless the app instance has been created.
[23:04] <mmcc> the bug suggests a patch to always look for qt.conf in the app wrapper, which looks like it'll work.
[23:05] <mmcc> I think we might be able to work around it if we create the QCoreApplication instance early enough, but I don't yet know enough about Qt to know if that'd cause problems, or even how to do it