[01:02] <ceed^> Is it normal that the sync daemon uses around 500 mb ram when idle?
[08:08] <mandel> rye_, ping
[12:04] <gatox> good morning!
[12:05] <alecu> and... hello week!
[12:05] <gatox> alecu, hi
[12:05] <alecu> hey gatox, how was your weekend at ubuconla?
[12:09] <gatox> alecu, awesome!!! the ubuconla was really nice..... and during the weekend i ate a lot of sushi, visit the japanese garden and went to a museum with japanese and chinese things :D
[12:19] <alecu> mandel, so, you say that if we run the tunnel within a .bat, it would not show the console window?
[12:19] <alecu> mandel, my guess is that it will show it anyway, but I'll give that a try.
[12:20] <alecu> mandel, I think the right solution would be patching twisted, so it uses the "CREATE_NO_WINDOW" flag when calling CreateProcess: http://docs.activestate.com/activepython/2.4/pywin32/win32process__CreateProcess_meth.html
[12:28] <ralsina> good morning!
[12:28] <gatox> ralsina, hi
[12:31] <mandel`> ralsina, gatox, morning!
[12:31] <gatox> mandel`, hi
[12:32] <ralsina> alecu: since we use a "private" twisted on windows, we can patch as much as you want
[12:33] <mandel`> alecu, I private twisted on windows?
[12:34] <alecu> mandel`, you private twisted!
[12:35] <mandel`> ups, a :P
[12:37] <mandel`> ralsina, gatox, alecu, although is my review day I cannot review the fsevents branches because I did them, can I have some help with them, this is looking like a Windows initial branches deja vu where they will be there for ages :(
[12:38] <ralsina> mandel`: so, let's make it not be that way!
[12:38] <alecu> mandel`, I promise to review em today.
[12:38] <ralsina> I am not at home today so no mac, but gatox has his
[12:38] <mandel`> thx!
[12:38] <alecu> mandel`, can you point me at the urls again?
[12:40] <gatox> mandel`, ack.... links?
[12:41] <mandel`> alecu, gatox: http://paste.ubuntu.com/1023011/
[12:41] <mandel`> those are the diff mps
[12:41] <gatox> ack
[12:41] <mandel`> I'm off to have lunch
[12:41]  * mandel` lunch
[12:51] <facundobatista> alecu, never saw this before (but probably is not new) Failure: dbus.exceptions.DBusException: org.gnome.zeitgeist.EngineError.DatabaseError: sql.vala:263: Can't start transaction: 1, cannot start a transaction within a transaction
[12:51] <facundobatista> alecu, in any case, it didn't break SD at all
[12:51] <facundobatista> alecu, hola, btw
[12:51] <alecu> facundobatista, right: no error calling the zg libraries should break SD.
[12:51] <alecu> facundobatista, how often did you see that message?
[12:52] <facundobatista> alecu, just created 10 files, I saw that once per each file
[12:52] <facundobatista> alecu, I can send you whole logs
[12:53] <alecu> facundobatista, yes please.
[12:55] <facundobatista> alecu, http://ubuntuone.com/2KigvQ9ZEKL96mg5uLsbHl
[12:56] <facundobatista> gatox, this one is for you: http://pastebin.lugmen.org.ar/7522
[12:56] <gatox> facundobatista, did you try to put the path between "...."
[12:56] <gatox> facundobatista, u1sdtool --publish-file="~/Ubuntu One/temp/log-dir.tar.bz2"
[12:57] <facundobatista> gatox, why would I? anyway, just tried, doesn't work either
[12:58] <gatox> facundobatista, weird...... could you file the bug?
[12:58] <gatox> please
[13:02] <gatox> mandel`, is it ok if i review your branches after lunch?? i want to get up to speed with fsevents here... it's hard to stop right now
[13:03] <alecu> facundobatista, what version of zeitgeist does "dpkg -l zeitgeist" says you've got installed?
[13:03] <alecu> facundobatista, the bug in SD seems to be bug #1001484
[13:03] <alecu> facundobatista, but I've tracked it to zeitgeist upstream: bug #937991
[13:04] <facundobatista> alecu, 0.9.0-1ubuntu1
[13:04] <dobey> gatox: all of the fsevents branches are blocking on the ocmock outcome at the moment, i think
[13:05] <gatox> dobey, ahh ok
[13:05] <dobey> alecu: interesting. i was *just* thinking "can we just get rid of this zeitgeist stuff we aren't using anyway"
[13:06] <facundobatista> dobey, I'd buy you a beer for that
[13:08] <ralsina> I am tempted to just agree with the ocmock author that we only use ocmock when developing
[13:08] <ralsina> after all, he wrte the damn thing and agrees we don't need to put it n our ads or anything
[13:09] <facundobatista> gatox, https://bugs.launchpad.net/ubuntuone-client/+bug/1008499  (didn't set importance)
[13:09] <ralsina> I can even make mmcc and mandel say "all hail ocmock" when they sit down to code
[13:09] <gatox> facundobatista, thx!
[13:09] <ralsina> facundobatista: but ~/whatever should be expanded by the shell!
[13:09] <dobey> ralsina: it doesn't matter if that's true, given the license
[13:10] <dobey> ralsina: ocmock has no use outside development. :)
[13:10] <ralsina> dobey: noone said the license makes sense (not even the author ;-)
[13:11] <ralsina> facundobatista: using ~/whatever in long options is not supposed to work  :-)
[13:11] <ralsina> facundobatista: because the shell doesn't expand it
[13:12] <alecu> ralsina, only if you don't use =
[13:12] <ralsina> alecu: exactly
[13:12] <ralsina> alecu: well, using = it doesn't expand it
[13:13] <alecu> facundobatista, so, the ~ issue is with bash, not with u1sdtool
[13:14] <alecu> facundobatista, if you want to use ~ you need to use --publish-file ~/Ubuntu\ One\abc.txt
[13:14] <alecu> facundobatista, if you want to use ~ you need to use --publish-file ~/Ubuntu\ One/abc.txt
[13:14] <ralsina> alecu, gatox, facundobatista: -1 to making u1sdtool expand paths, so bug is invalid.
[13:14] <gatox> ralsina, ack
[13:15] <alecu> facundobatista, dobey: I don't see anybody using our zeitgeist logs... but that means that I don't know what our users think of them.
[13:16] <alecu> facundobatista, dobey: what I do for a fact  know is that the zeitgeist devs don't like the amount of stuff we log.
[13:16] <alecu> facundobatista, dobey: so I support the idea of axing all of that.
[13:18] <dobey> alecu: the whole reason we're using zeitgeist isn't for users to use our info within gnome-activity-journal or anything. it's so we could use it to index activity, and write some app to present the data in a meaningful way
[13:18] <dobey> alecu: right. we've had this conversation before :)
[13:19] <dobey> basically, for us to present a meaningful UI to users about what happened when, in U1, we have to do a custom thing anyway, because we need multi-platform, and the server.
[13:23] <alecu> dobey, right. But we don't have a complete plan to achive this in a multi-platform way, and the log in zeitgeist makes sense even when we use something else in other platforms.
[13:23] <ralsina> the only way to make that cross-platform is to implement it once on every platform
[13:23] <alecu> dobey, so, from that angle I'm still not sure we should get rid of this till we have a multiplatform plan.
[13:23] <ralsina> according to the roadmap in my head, that's like 5 mountains ahead from here
[13:24] <alecu> ralsina, or just log everything in sqlite in all platforms.
[13:24] <ralsina> alecu: yikes
[13:24] <ralsina> alecu: sqlite tends to frown at you once you go over a certain logsize
[13:24] <alecu> ralsina, to us, zeitgeist is just a sqlite wrapper.
[13:25] <dobey> alecu: well, i think u1db is the way to acheive that
[13:26] <ralsina> alecu: which is maybe one of the reasons why zh devs hate us a little ;-)
[13:26] <ralsina> zg
[13:42]  * mandel` back
[13:43] <mandel`> gatox, review the branches whenever you want :)
[13:43] <mandel`> gatox, if you do not review them today, I'll ask you again tom, and if not tom+1 etc.. ;)
[13:44] <gatox> mandel`, thx... after lunch will be
[13:57] <joshuahoover> ralsina: when do we plan on doing another win release? i'm not pressing for one, but would like to know when we plan on doing the next one :)
[13:58] <ralsina> joshuahoover: next monday looks reasonable, but I need to talk with alecu and brian
[13:59] <joshuahoover> ralsina: k, that sounds good...let me know when you confirm a day and i'll put in the rt request...in the mean time, qa will test specifically for the proxy issue and potential regressions in the releases coming out of jenkins
[13:59] <ralsina> joshuahoover: cool, I should have a date late today
[14:04] <gatox> need to restart! brb
[14:18] <ralsina> oh launchpad, why are you showing duplicates of fix-committed bugs in the default bug listing now :-/
[14:23] <ralsina> WTF, it's showing bugs marked invalid! Ohhhhh the page is not marked as not cacheable. Nice.
[14:29] <ralsina> dobey: I just tried u1cp with turkish locale and I get no errors, is that enough to mark bug #467397 as fix-released?
[14:30] <ralsina> Also, we have a decimal import in sd with a comment saying something like "this is so we don't get the RoundCEiLiNG error"
[14:30] <dobey> right
[14:31] <dobey> it looks like part of it is still failing in natty though, so perhaps we need to SRU a workaround to fix that
[14:31] <urbanape> mmcc: ping
[14:31] <mmcc> hi folks
[14:31] <dobey> ralsina: not sure what it has to do with control-panel though
[14:31] <mmcc> what's up urbanape
[14:32] <urbanape> heya. Just wondering if you guys have a punch list started of things you'd like asked about in the labs
[14:32] <urbanape> This week will be a lot of auditing of the Music and Files apps for that kind of stuff.
[14:33] <mmcc> urbanape: ok - I've been thinking about it occasionally but
[14:33] <mmcc>     nothing big has come up yet. I'll go back through my notes
[14:34] <urbanape> k, no hurry, but maybe something to consider this week.
[14:34] <urbanape> I know you guys have been talking about sandboxing, &c.
[14:34] <mmcc> if you can get Finder devs drunk and find out what dropbox is
[14:34] <mmcc>     patching for their file badges :)
[14:34] <mmcc> what is going on with my IRC
[14:34] <urbanape> heh
[14:35] <dobey> ralsina: it seems i have a linked branch which probably is applicable to natty, as well
[14:35] <mmcc> so far I don't think sandbox is a problem for us. I might ask if
[14:35] <mmcc>     the app-store is picky about the layout of app wrappers, but I'm
[14:35] <mmcc>     not sure you'd get a real answer
[14:36]  * mmcc arg
[14:36] <dobey> what irc client are you using?
[14:37] <mmcc> dobey, emacs' ERC. it just started doing this today. this is the same buffer that worked before, even. I think I have word-wrap set accidentally.
[14:37] <dobey> looks like it's wrapping and tabbing, yeah
[14:38] <mmcc> yeah, auto-fill was set on. the connection dropped over the weekend, and I think that messed up the modes somehow
[14:39] <dobey> ah
[14:39] <dobey> mmcc: any news about the ocmock license?
[14:40] <mmcc> dobey: sifting through emails now. I need to get some filtering set up
[14:41] <mandel`> mmcc, the handler that is created by laynchd, is it already binded for the communication?
[14:41] <mmcc> dobey: no new news. I'll ping him.
[14:42] <mmcc> mandel`: looking...
[14:46] <dobey> ok
[14:50] <ralsina> dobey: AFAICS, if there ever was a manifestation of this bug in u1cp, it's gone
[14:52] <ralsina> dobey: marking as fix-released in u1cp, and maybe we can merge your natty branch eventually one of these days
[14:53] <ralsina> dobey: should I target it to each release and mark it fixed for, say, O P Q?
[14:53] <ralsina> dobey: for ubuntuone-client
[14:53]  * mmcc afk for a minute, something's wrong with one of my dogs
[15:01] <dobey> well, the u1client bug is against the project, not the package
[15:01] <dobey> not sure what to do at the moment
[15:01] <mmcc> mandel`: yes, launchd calls bind() and listen(), you just need to call accept(). I suppose you also need to duplicate the fd.
[15:02] <mmcc> mandel`: in my earlier project, I just passed the launchd-provided fd to python's socket.fromfd and that did the trick
[15:02] <ralsina> dobey: I say we close it with the classical "if it still happens let us know"
[15:02] <mmcc> mandel`: for reference: https://bitbucket.org/mikemccracken/py-launchd/wiki/Home
[15:03] <mandel`> mmcc, cool, I'll take a look
[15:04] <dobey> ralsina: also works
[15:04] <gatox> standup?
[15:04] <mandel`> me
[15:04] <gatox> me
[15:04] <mmcc> me
[15:04] <briancurtin> me
[15:05] <ralsina> me
[15:05] <dobey> meh
[15:06] <mandel`> go?
[15:06] <ralsina> go
[15:06] <mandel`> DONE: reviews, reviews, reviews. Added more tests and proposed a new branched with them (fsevents daemon).
[15:06] <mandel`> TODO: Add tests for python code that gets events from daemon. Make a nicer api.
[15:06] <mandel`> BLOCKED: no
[15:06] <mandel`> gatox, please
[15:06] <gatox> DONE:
[15:06] <gatox> Conference day on Friday, assist to UbuConLA on friday and saturday! fun! Working on fsevents move-(from/to) event.
[15:06] <gatox> TODO:
[15:06] <gatox> Finish with the fsevents last remaining issues for test_filesystem_notifications. Review mandel's branches.
[15:06] <gatox> BLOCKED:
[15:06] <gatox> No
[15:06] <gatox> mmcc, go
[15:07] <mmcc> DONE: control-panel packaged, SSO works from its spot within
[15:07] <mmcc> TODO: need to fix QT plugin support in py2app
[15:07] <mmcc> BLCK: NO
[15:07] <mmcc> NEXT: briancurtin
[15:07] <briancurtin> DONE: created some one-off installers for a few branches, can't remember what else I did :/
[15:07] <briancurtin> TODO: pick up where i left off with testability, building from scratch
[15:07] <briancurtin> BLOCKED: no
[15:07] <briancurtin> NEXT: ralsina
[15:07] <ralsina> DONE: bug triaging, askubuntu, canonicaladmin TODO: lots more bug triaging, perhaps fix one, looks like tech leads is not going to happen. BLOCKED: no NEXT dobey
[15:07] <ralsina> briancurtin: I recommend taking notes :-)
[15:07] <dobey> λ DONE: reviews, triage, SRU poking, fix
[15:07] <dobey> λ TODO: releases, SRU poking, tarmac tweakery
[15:07] <dobey> λ BLCK: None.
[15:08] <dobey> thisfred, alecu: fight
[15:08] <briancurtin> ralsina: will do
[15:08] <thisfred> me
[15:08] <ralsina> NEXT:thisfred :-)
[15:08] <alecu> me
[15:08] <thisfred> DONE: bug #1006954
[15:08] <ralsina> dobey: when is our next scheduled release? (are they in the calendar?)
[15:08] <thisfred> TODO: planning poker, bug #999585
[15:09] <thisfred> BLOCKED: NO
[15:09] <thisfred> NEXT: alecu
[15:10] <dobey> ralsina: today
[15:10] <alecu> DONE: py2exe tests for bug #1006899
[15:10] <alecu> TODO: working on a patch for twisted for said bug
[15:10] <alecu> BLOCKED: no
[15:10] <ralsina> dobey: grmbl
[15:10] <ralsina> dobey: we need alecu's proxy fix for windows
[15:10] <ralsina> dobey: can we delay it a couple of days?
[15:11] <dobey> ralsina: this is the stable-4-0 releaes, not the 3.0
[15:11] <dobey> ralsina: we haven't scheduled future 3.0.x releases yet :)
[15:11] <briancurtin> alecu: if you need help on that py2exe/twisted thing, let me know
[15:11] <ralsina> dobey: then it doesn't matter, really. Let's schedule them :-)
[15:11] <dobey> sure
[15:11] <ralsina> dobey: one for early next week as 1st step
[15:11] <mandel`> eom?
[15:11] <ralsina> eom it seems
[15:12] <mandel`> ok, I'm back with mac
[15:12] <alecu> briancurtin, I think I've got it. I ended up with some patches for twisted that I'm reporting upstream.
[15:12] <briancurtin> alecu: awesome
[15:12] <dobey> ralsina: ok. hopefully the remaining SRUs can get accepted by then :)
[15:12] <alecu> briancurtin, after I finish this, I'll point you to the patch so you can do a build with a modified twisted like that.
[15:13] <briancurtin> alecu: sounds good
[15:13]  * briancurtin brb, someone knocking on the door
[15:14] <mmcc> briancurtin: when you're done talking to the land shark: how are you handling the testability plugin in py2exe? I ran into some plugin issues with qtnetwork that it looks like I can't ignore, like I did with the image plugins
[15:27] <dobey> need to get lunch. bbiab
[15:31] <ralsina> mmcc: looks like it was NOT a candygram
[15:33] <mmcc> ralsina: :) in Chicago, it'd be a brat-gram?
[15:33] <briancurtin> mmcc: i'm not currently handling them. i'll let you know what i figure out
[15:34] <briancurtin> ha, i wish they'd start leaving me candy or brat grams rather than notes and shutting off the gas
[15:34] <popey> i have some files in my ~/Pictures on my computer, and some files in "My Storage / ~/Pictures online in one.ubuntu.com/files. I have synced in the past but right now they are not in sync. if I tick the box to sync locally, what will happen? Will they merge?
[15:34] <popey> or will I lose files?
[15:34] <mmcc> briancurtin: ok, cool. btw I noticed there's some code in sso-client that seems to be working around not having Qt image plugins on windows, using PIL instead of Qt images
[15:35] <briancurtin> mmcc: yeah i did see that as well the other day
[15:35] <mmcc> what I've noticed is that QtNetwork also loads plugins, but just doesn't crash...
[15:36] <mmcc> (yet)
[15:37] <ralsina> popey: they will merge
[15:37] <popey> yay!
[15:37] <ralsina> popey: however, if there are files with matching names, that differ, you may get conflicts
[15:38] <ralsina> popey: or, if you have old metadata in the device you will start to sync, they may get overwritten
[15:38] <popey> i probably have old metadata
[15:38] <popey> i have synced this machine before
[15:38] <ralsina> popey: backup recommended
[15:39] <popey> so backup ~/Pictures, enable sync, wait for it to finish then blat my backup over the top again?
[15:39] <popey> to ensure I have everything
[15:39] <popey> (I don't want to spend ages analysing whether it has synced, I just want it to be in sync)
[15:39] <ralsina> popey: depends on whether ~/Pictures or the server has the newer version of the matching files
[15:40] <ralsina> popey: that's the tricky part, since we don't know which ones you prefer to keep
[15:40] <mmcc> my new Qt plugin issue seems to be coming from proxy stuff in ubuntu_sso/utils/webclient/qtnetwork.py:121: -- QNetworkProxyQuery looks like it causes loading plugins libqcorewlanbearer.dylib and libqgenericbearer.dylib
[15:41] <mmcc> does that look familiar to anyone?
[15:41] <mmcc> on windows, s/dylib/dll/
[15:41] <ralsina> mmcc:  nope
[15:41] <mandel> mmcc,no, we (alecu and I) added that code but we did not take a look at what was loaded, sorry
[15:42] <ralsina> mmcc: I suspect those are not exactly required http://qt-project.org/doc/qt-4.8/bearer-management.html
[15:43] <mmcc> mandel: no big deal. probably not simple to know to trace if you don't expect this problem
[15:43] <mandel> mmcc, no, we certainly ignored such a problem
[15:43] <mmcc> I only notice it because I get this message: "On Mac OS X, you might be loading two sets of Qt binaries into the same process. Check that all plugins are compiled against the right Qt binaries. Export DYLD_PRINT_LIBRARIES=1 and check that only one set of binaries are being loaded."
[15:43] <alecu> ralsina, the "bearer" surely looks useful.
[15:43] <mmcc> mandel: this does cause the moveToThread warning for me, too
[15:43] <ralsina> alecu: yes, but we are not using it, AFAIK
[15:43] <alecu> ralsina, mmcc: I would include that in our packages.
[15:44] <mmcc> alecu: agreed.
[15:44] <alecu> ralsina, probably nam uses it. (qnetworkaccessmanager)
[15:44] <alecu> ralsina, which we use.
[15:44] <mmcc> has windows proxying been working OK without these plugins included?
[15:45] <gatox> mandel, +1 to the two branches i needed to review
[15:45] <ralsina> alecu: seems like it doesn't
[15:46] <alecu> mmcc, are those .dlls on windows as well?
[15:47] <alecu> ralsina, awesome then, let's not include them on mac unless they are used.
[15:47] <mmcc> alecu: I don't know.
[15:47] <ralsina> alecu: but that was a quick look
[15:47] <ralsina> mmcc: it's broken but for other reasons :-)
[15:47] <mmcc> alecu: ralsina: something's loading these plugins. might be easier to include them than to figure out how to avoid loading them
[15:47] <mandel> gatox, superb! thx a lot!
[15:47] <ralsina> mmcc: yes
[15:48] <ralsina> mmcc: they are tiny anyway. So if you figure out how to include them that works, I say go ahead
[15:48] <mandel> dobey, is this still blocked until we hear about the OCMock license changes?
[15:48] <alecu> ralsina, how did you find out it was not used?
[15:48] <alecu> ralsina, I see this:
[15:48] <alecu> ./src/network/access/qnetworkaccessmanager.cpp:#include "QtNetwork/qnetworksession.h"
[15:49] <mmcc> ralsina: ok.
[15:49] <alecu> ralsina, so I guess the qnetworkaccessmanager *does* use the bearer.
[15:49] <ralsina> alecu: that doesn't mean it uses the plugin. It apparently only uses it if you call capabilities()
[15:49] <ralsina> alecu: if it used it all the time it would not be a plugin :-)
[15:50] <alecu> ralsina, right :-)
[15:51] <alecu> mmcc, how big are those dynlibs?
[15:52] <alecu> mmcc, if they are small enough we should include them, because they sound useful, and we might be missing some other usage of it.
[15:52] <mmcc> 206K for both
[15:52] <mmcc> so I vote include
[15:53] <elopio> briancurtin: hola. Any news about the testability plugin in windows?
[15:53] <briancurtin> elopio: not yet :/ i'm working out building everything from source this time to get it working
[15:53] <alecu> ralsina, I think we should include them. And perhaps open a bug to test what happens if we don't include them on windows and mac.
[15:53] <ralsina> alecu: we are not including any plugins on windows
[15:54] <ralsina> alecu: but sure, let's include them. And if we can make the image plugins work and get rid of PIL, even better
[15:54] <alecu> ralsina, but are those included as plugins too, or are those statically compiled?
[15:54] <ralsina> alecu: we ship no plugin whatsoever
[15:54] <elopio> briancurtin: ok, thanks for your help :)
[15:54] <ralsina> alecu: and we need the jpg plugin to make the recaptcha work
[15:54] <mmcc> alecu: the Qt image format support is plugins
[15:54] <alecu> right.
[15:55] <elopio> briancurtin: I'll report a bug so we can follow the status of it. Is it a problem on control panel or the installer?
[15:55] <mmcc> fwiw, all the Qt plugins are 4.2 MB total
[15:55] <briancurtin> elopio: i would say installer
[15:56] <ralsina> mmcc: yes, but that includes thing like postgreql support, we can trim it after we make it work
[15:56] <mmcc> ralsina: yeah, for sure we don't need it all
[16:02] <elopio> briancurtin: bug #1008583.
[16:02] <briancurtin> elopio: cool, thanks
[16:09]  * gatox lunch
[16:11]  * ralsina wonders how many bugs are there filed against the packages that are not filed against the projects, thus he was not ware of them at all
[16:19]  * alecu lunches
[16:37] <dobey> mandel: yes
[16:37] <mandel> dobey, ack
[16:38] <dobey> ralsina: bugs where?
[16:38] <ralsina> dobey: launchpad
[16:38] <ralsina> dobey: sil had reported a bug on u1cp-ubuntu andnot on u1cp and I never saw it
[16:38] <ralsina> dobey: I only found it because I was looking for it specifically
[16:39] <dobey> ralsina: never saw it on lp, or never saw it in e-mails?
[16:39] <ralsina> dobey: probably saw the email a while ago
[16:39] <ralsina> dobey: but I never saw it when I do triage
[16:39] <ralsina> dobey: because I am triaging on the project. lesson learned
[16:40] <dobey> ralsina: at the top of the list of bugs there's a link to the other set of bugs, btw
[16:40] <mandel> ok, I need to do a some quick errand, will me back in 10'
[16:43] <dobey> mandel: you can probably reject and submit new proposals for the license-headers and kernel-h branches, and not make them depend on the ocmock-dependent branch, so they can land
[16:52]  * mmcc lunch
[16:58] <mandel> dobey, is not a terrible thing, I can wait
[16:59] <mandel> dobey, if indeed we cannot use it I'll do that
[16:59] <mandel> by the way, I'm back :)
[17:10] <mandel> alecu, one question about twisted protocol, if there is an error in the protocol, that is, the message is malformed, what is the correct way to raise an error?
[17:10] <dobey> yay disrespectful people demanding things
[17:10] <mandel> alecu, is it just raising an exception?
[17:10] <dobey> they are my favorite kinds of people
[17:10] <mandel> dobey, I WANT CHOCOLATE ICECREAM!!!
[17:10] <mandel> :P
[17:10] <dobey> mandel: fix your banks, then maybe you'll be able to afford it, again :)
[17:11] <mandel> lol
[17:11] <mandel> dobey, I might be able to buy a bank with the price of a cone hehe
[17:11] <dobey> hehe
[17:11] <ralsina> mandel, someonelse: I would like a review for https://code.launchpad.net/~ralsina/ubuntuone-control-panel/sort-shares/+merge/108594
[17:11] <mandel> ralsina, sure
[17:12] <ralsina> mandel: it's trivial, except for the test fixes which are easy anyway ;-)
[17:12] <mandel> ralsina, which test fixes? I only see a 2 line check (lp diff)
[17:13] <mandel> ralsina, which revno did you push?
[17:13] <ralsina> mandel: oops, no push
[17:13]  * ralsina is a svn guy at heart
[17:13] <ralsina> mandel: there, revno 331
[17:13] <mandel> ralsina, ok
[17:14] <dobey> ralsina: it just makes it alphabetical?
[17:14] <ralsina> dobey: yes
[17:14] <ralsina> dobey: which broke like 5 tests ;-)
[17:14] <dobey> i also don't think it solves the problem :)
[17:15] <ralsina> dobey: but it does
[17:15] <ralsina> dobey: aquarius has 58 shares or wmething like that. Befre, they were random. Now they are in alphabetical order :-)
[17:15] <ralsina> dobey: I actually discussed the solution with him, the bug reporter
[17:16] <dobey> i know, i was there. and you also talked about having it select/scroll to the specific share automatically, for when clicking on the messaging menu
[17:16] <dobey> but there are cases where the fix is still a problem
[17:17] <ralsina> dobey: that's something else
[17:17] <ralsina> dobey: that involves having the messaging menu pass u1cp the share name and adding another cli option
[17:18] <dobey> amongst other complexity, yes
[17:18] <ralsina> dobey: precisely. So this is about one part of the problem, the part that can be fixed ;-)
[17:18] <dobey> but if i have 50 shares, and there are 5 new ones, how do i find all 5 of them easily?
[17:19] <ralsina> dobey: I am not saying this makes it easy to find new shares. It makes it easy to find shares in general.
[17:19] <mandel> dobey, ralsina, we have some bad ux bugs in the control panel to be honest.. alphabetical order does seem a step forward in one of them
[17:19] <ralsina> dobey: so the answer is "faster than before!" ;-)
[17:19] <dobey> also, does the qt list view not have type-to-search?
[17:19] <ralsina> dobey: no, it doesn't
[17:19] <dobey> s/makes it easy/makes it easier/
[17:19] <dobey> and not having type to search sucks
[17:20] <ralsina> dobey: not all that hard to add, really
[17:20] <dobey> mandel: not arguing that :)
[17:20] <ralsina> dobey: correction, it does have it :-)
[17:20] <ralsina> dobey: never knew it!
[17:20] <dobey> ralsina: already enabled in u1cp?
[17:21] <ralsina> dobey: yep
[17:21] <ralsina> dobey: click on the folder list, type a letter
[17:21] <ralsina> or several letters really quick
[17:21] <mandel> how quick? nessita quick? :P
[17:21] <dobey> ralsina: can we make the folder list have focus by default when that tab is selected?
[17:21] <ralsina> dobey: we might
[17:22] <ralsina> elopio: AFAIK we are using the trashcan (for bug #375833)
[17:23] <dobey> ralsina: also, can we get people to properly triage bugs better?
[17:23] <ralsina> dobey: specifics?
[17:23] <elopio> ralsina: that works? Awesome. I'll try it and add it to the smoke in moztrap.
[17:24] <dobey> ralsina: targeting to project and distro series when triaging
[17:24] <ralsina> dobey: sure
[17:25] <ralsina> dobey: AFAIK from now on we are targetting to trunk unless we are ging to do it n 3-0 too
[17:25] <dobey> ralsina: well, there's also stable-4-0 now too
[17:25] <ralsina> dobey: ok, so targeting everything to trunk + 4-0
[17:25] <dobey> ralsina: and things that need to be SRUed to older versions need to have their series targeted
[17:26] <ralsina> dobey: yes, but that's case by case, and we are the ones who should do that
[17:26] <dobey> ralsina: plus, things need to target the ubuntu package as well, and the series we're going to fix/release things in
[17:26] <ralsina> we have to always target the package? really?
[17:26] <dobey> we should
[17:27] <ralsina> half the time when I do that, launchpad breaks
[17:27] <dobey> eh?
[17:27] <ralsina> and the bug ends tracking wrong
[17:27] <ralsina> dobey: I have had a dozen bugs that were fix-released in the branches and in the package, and in trunk and the bug itself never updated
[17:27] <ralsina> dobey: and then there is no way to fix that unless you go, untarget everything, then target it again
[17:27] <mandel> ok EOD for me, I'll catch you all up tom!
[17:28] <dobey> ralsina: fix released in the project doesn't mean it is in the package
[17:28] <mandel> laters!
[17:28] <ralsina> bye mandel!
[17:28] <ralsina> dobey: I never set fix-released in the project, only on the branches. But yes, will target the packages when we target stable branches
[17:29] <dobey> ralsina: well, *you* should probably generally never set fix-released on anything
[17:29] <ralsina> he
[17:29] <ralsina> and in fact, I usually don't, except for ancient bugs
[17:29] <elopio> ralsina, dobey: joshuahoover told us we should mark the bugs as affecting the package on the distro it's happening. We weren't doing that, but we are starting now. Any additional think you would like us to do while marking bugs as confirmed?
[17:30] <dobey> whoever makes the tarball release should set it for the project series it's being released on. and whoever makes the package updates should mention the bug # in the changelog, and lp will automatically fix-release it in the package bug for the series, when the upload is accepted into the appropriate archives
[17:30] <joshuahoover> ralsina: i do it sometimes just to see if dobey catches me setting a bug to fix released ;)
[17:32] <ralsina> dobey: so, in short, target the package and the series where the bug happens.
[17:32] <dobey> elopio: we really need to verify what all supported versions are affected, and target the ones that are
[17:32] <ralsina> gotta run an errand/lunch will be back in about one hour
[17:35] <elopio> dobey: ok. We'll try. If you notice we are missing something or doing something wrong, please ping and complain :)
[17:36] <dobey> elopio: of course. i am good at complaining. just ask joshuahoover ;)
[17:36] <joshuahoover> :)
[17:37] <joshuahoover> elopio: don't be fooled...dobey isn't good at complaining, he's great at it! ;)
[17:38] <dobey> heh
[17:49] <elopio> briancurtin: do you know how can I create a file with random contents on windows?
[17:50] <elopio> what I do on linux with dd if=/dev/urandom
[17:50] <briancurtin> elopio: you'd just have to open a file and write some random stuff to it. there's no easy way like on linux
[17:52] <elopio> briancurtin: sad.
[17:52] <briancurtin> like open(blah, 'w').write(os.urandom(500)) or something
[17:53] <rmcbride> elopio: what I do is install cygwin so I can use posix tools on windows. Perhaps not a great universal solution, but it's something
[17:53] <elopio> hey ralsina, can you please check that the delete test checks for every feature we have around it? http://moztrap.elopio.net/manage/case/70/
[17:54] <elopio> rmcbride: yes. Maybe there's a web file generator, I'll look for it.
[17:59] <dobey> elopio: do it on a linux and use putty or something to scp it over to the windows?
[18:20] <popey>  3286 alan      20   0 7158m 4.2g 2416 R   96 54.8  80:21.90 ubuntuone-syncd
[18:20] <popey> I would really rather u1 didn't eat all my RAM...
[18:20] <popey> what is it possibly doing?
[18:20] <popey> (note: it's also killing my cpu)
[18:21] <dobey> wow, that RSS is insane
[18:21] <dobey> what is it doing?
[18:21] <dobey> ~/.cache/ubuntuone/log/syncdaemon.log
[18:21] <popey> http://paste.ubuntu.com/1023651/
[18:21] <popey> a _lot_ of that
[18:23] <popey> u1sdtool -q tells me it's still running
[18:23] <popey> and i can see that by my cpu still being eaten
[18:23] <dobey> popey: hmm. i'd say you should kill it
[18:23] <dobey> popey: do you also have a syncdaemon-exceoptions.log ?
[18:24] <popey> yes
[18:24] <popey> will kill and provide logs
[18:24] <popey> aaaand.. breathe!
[18:24] <dobey> ralsina: care to review https://code.launchpad.net/~dobey/libubuntuone/update-stable-4-0/+merge/108605 ?
[18:25] <popey> http://paste.ubuntu.com/1023629/ dobey
[18:25] <popey> there was a lot of that going on
[18:25] <popey> dobey: want any particular logs?
[18:26]  * briancurtin lunch/fresh air
[18:26] <dobey> looks like zeitgeist went crazy
[18:26] <elopio> dobey: yes. Actualy it would be easier to publish it with U1, that way you just need a browser on windows.
[18:26] <popey> just a bit dobey
[18:26] <dobey> alecu: ^^
[18:27] <dobey> popey: alecu could probably use the logs
[18:27] <dobey> popey: is this in quantal or precise?
[18:27] <popey> precise
[18:27] <popey> tried to restart the control panel..
[18:27] <popey> "Sorry, an error has occurred and Ubuntu One needs to close."
[18:27] <popey> "IPCError"
[18:28] <popey> starting again.. okay
[18:29] <dobey> yeah, dbus breaking will do that
[18:29] <dobey> popey: did you just drop like 5 billion files into u1?
[18:31] <alecu> popey, I'd really like to see the full syncdaemon.log and syncdaemon-exceptions.log... I wonder what's making it grow so much...
[18:31] <ralsina> dobey: sure, looking
[18:32] <ralsina> elopio: looking...
[18:34] <ralsina> dobey: +1
[18:34] <popey> alecu: not on this machine
[18:36] <alecu> popey, so, the "lots of files added to u1" were on a different machine?
[18:38] <ralsina> elopio: random files http://www.random.org/files/
[18:38] <ralsina> elopio: the internet is a wonderful place :-)
[18:39] <ralsina> elopio: also, the test looks complete to me
[18:39] <popey> alecu: well, technically I synced down on another machine
[18:39] <popey> alecu: and also then deleted some files but not a huge number, all on another machine
[18:39] <elopio> ralsina: gr8. Thanks.
[18:40] <popey> alecu: the machine that had the issue is now saying I'm all in sync, i guess because the machine doing syncing hasn't finished doing all its stuff
[18:42] <alecu> popey, right: "all in sync" means from the perspective of the machine doing the upload vs. the server. It does not take into account the state of "other" machines.
[18:43] <alecu> popey, and what about the machine doing all those uploads? is syncdaemon taking a lot of ram there too?
[18:44] <alecu> popey, also, would you mind creating a private bug for ubuntuone-client and uploading there the compressed logs so I can take a look?
[18:44] <popey> alecu: i dont think I'm doing any uploads
[18:44] <popey> oh, yes i am, new photos, my bad
[18:45] <popey> alecu: on the machine doing uploads, ubuntuone-syncd is using 312M RES
[18:45] <alecu> popey, is that an up-to-date 12.04 too?
[18:46] <popey> no, its 12.10
[18:47] <popey> will make the bug now
[18:47] <alecu> popey, thanks!
[18:51] <popey> alecu: bug 1008662
[19:06] <ralsina> dobey: can I get you as 2nd review for https://code.launchpad.net/~ralsina/ubuntuone-control-panel/sort-shares/+merge/108594
[19:07] <alecu> popey, I'm taking a look at the logs, but it seems that the messages went by too fast, and the logs files were rotated in a couple of seconds, so the interesting info on how this problem started was chopped.
[19:07] <ralsina> dobey: I can even target stable-4-0 for the bug ;-)
[19:07] <alecu> popey, I can only find a reference to a failing zeitgeist module, which sounds similar to a problem facundobatista found today.
[19:08] <popey> bummer
[19:09] <alecu> popey, I'll try to reproduce it later today, by syncing a lot of new photos from a different computer.
[19:09] <alecu> but first I need to run to kindergarten :-)
[19:10] <dobey> ralsina: can you triage the bug to include the branches we need to ensure it gets merged to, and the versions of ubuntu we need to SRU/relaese it on
[19:11] <ralsina> dobey: done, I think
[19:11] <ralsina> dobey: it's not letting me target releases for the package, or I am not finding the right place to click for it
[19:11] <dobey> ralsina: it is. you can only nominate them though (which is fine)
[19:12] <ralsina> dobey: ok then, it's done
[19:12] <dobey> ralsina: but do we not want to SRU it to precise and put it on stable-3-0 as well?
[19:12] <ralsina> I only nominated for Q, this is not SRU-worthy
[19:12] <ralsina> dobey: I am not sure it
[19:12] <ralsina> 's worth the trouble
[19:12] <ralsina> unless we get the MRE
[19:12] <dobey> well, it's a simple fix
[19:12] <dobey> we'll get the MRE
[19:12] <ralsina> k, will put it up for 3-0 then
[19:12] <dobey> and even without the MRE, it's an easy fix
[19:13] <ralsina> I also need to target trunk?
[19:13] <dobey> it doesn't break any freezes
[19:13] <dobey> yes
[19:13] <ralsina> there then, I think I didn't miss any
[19:14] <ralsina> Ok, I should add Precise then
[19:14] <ralsina> but I can't nominate for more releases, it seems
[19:14] <dobey> ?
[19:14] <dobey> sure you can
[19:14] <dobey> though the UI is confusing
[19:15] <ralsina> dobey: I used to have a "nominate for release" link and now I don't?
[19:15] <dobey> ralsina: you have to be looking at the bug from the ubuntu point of view (the /ubuntu/+source/ has to be in the URL)
[19:15] <ralsina> oh tricky
[19:15]  * briancurtin i think i need to restart...
[19:15] <dobey> yeah, the UI is not great for that
[19:17] <ralsina> ok, it's easiest if I do that right after I affect the package, when I am in the right facet
[19:18] <ralsina> because I see no way to get from one to the other except by searching for the bug in the package's bug page. Man this is annoying :-/
[19:18] <ralsina> nominated it is
[19:19] <dobey> hrmm. having dates in evolution be only partially translated in evolution, is annoying in this language
[19:19] <dobey> ralsina: you can hit Ctrl-L and just tweak the URL by hand. it's faster :)
[19:19] <ralsina> dobey: yeah, I just keep forgetting where to stick the +source ;-)
[19:20] <ralsina> I think there is a stone carve somewhere that says "if thy user edits the URL manually, ye are doing it wrong"
[19:20] <dobey> ralsina: /ubuntu/+source/ goes at the beginning of the path. :)
[19:20] <ralsina> dobey: I know *now*, I will forget it when I need it ;-)
[19:20] <dobey> if you can't tell. i've done this a lot
[19:20] <ralsina> hahaha
[19:21] <ralsina> no wonder you want QA to do it ;-)
[19:22] <dobey> well i want whoever triages a bug to do it. whether it's qa, or someone grabbing a bug to fix it. :)
[19:25] <ralsina> dobey: doing a few more just for you, dude
[19:32] <mmcc> does anyone have tips on debugging Qt's qt.conf? It looks like it's lying about qt.conf overriding the default plugin path...
[19:33] <ralsina> dobey: something that adds a translatable string is a nono for 3-0 right? Even if it's something that surely is already translated, like "KiB" ?
[19:33] <mmcc> and this business about an empty qt.conf stopping it from loading image format plugins - that's nowhere in the Qt docs
[19:33] <ralsina> mmcc: it's probably an unintended side effect
[19:33] <dobey> ralsina: are you asking about your KiB fix?
[19:33] <ralsina> dobey: yes
[19:34] <ralsina> dobey: bug #985333
[19:34] <mmcc> ralsina: I LOVE those
[19:34] <ralsina> mmcc: anyway, we decided we want to load them right?
[19:34] <mmcc> ralsina: that's what I'm trying to do :)
[19:34] <ralsina> mmcc: ok then :-)
[19:34] <dobey> ralsina: the translation probably isn't the issue there. the issue is that the UI is changed. we can discuss with release team about getting in an SRU if we want to ship it on precise though
[19:35] <ralsina> dobey: I don't really care all that much for this one
[19:35] <ralsina> dobey: I am more trying to learn the limits here
[19:35] <mmcc> ralsina: I have py2app recognizing them and editing the linker commands, but I don't have qt loading them from the right place, and I can't confirm that my edits to qt.conf actually work
[19:35] <mmcc> s/work/do anything at all/
[19:35] <ralsina> mmcc: can you pastebin your current qt.conf?
[19:35] <dobey> ralsina: string and ui changes are bad, yes :)
[19:35] <mmcc> ralsina: echo /dev/null | pastebinit :)
[19:36] <ralsina> mmcc: well, that's supposed not to load them so if we want to load them it's bad? ;-)
[19:36] <dobey> ralsina: oh, actually, that one is probably not that bad. where are those strings already translated at?
[19:36] <ralsina> dobey: I am betting there is *something* else that says "KiB" somewhere :-)
[19:36] <mmcc> ralsina: I kid, I'm writing a quick summary in a pastebin
[19:36] <dobey> ralsina: i thought you were asking about the kilobits bug before. sorry :)
[19:37] <ralsina> dobey: I had two very similar ones ;-)
[19:37] <dobey> ralsina: but is that something anything control panel depends on?
[19:37] <briancurtin> i need to get away from the computer. be back in a bit.
[19:37] <dobey> iow, if it's not qt, then we don't just get to use the other translations
[19:37] <ralsina> dobey: ok, dropping this one
[19:37] <dobey> basically if you add the _(), and run it without any other changes, it should work. if it doesn't, then the strings aren't already translated :)
[19:37] <ralsina> dobey: since it would have no visible effect
[19:37] <joshuahoover> ralsina: is it possible bug #1006899 impacts ubuntu in addition to windows? i have a precise user behind a proxy with no authentication and says it's not working
[19:38]  * alecu is back
[19:38] <ralsina> joshuahoover: no, that one is very windows specific
[19:38] <dobey> also. are those translatable?
[19:38] <ralsina> joshuahoover: so it must be a new bug
[19:38] <ralsina> dobey: yes
[19:38] <joshuahoover> ralsina: k, i'll get logs :)
[19:38] <dobey> i don't see any point to translating them
[19:38] <dobey> KiB in any other language is KiB?
[19:38] <ralsina> dobey: nope. Horribly, it's not.
[19:39] <dobey> example?
[19:39] <ralsina> dobey: I thought it was used diferently in french but can't find a reference
[19:40] <ralsina> Kibinytes in french are Kibioctets but have not found the abbr for it
[19:40] <ralsina> Kibibytes*
[19:41] <ralsina> KiO in french: http://en.wikipedia.org/wiki/Octet_(computing)
[19:42] <ralsina> Ok, looks like Kibioctet in english is one thing that is not the same as Kibioctet in french, which is a Kibibyte
[19:42]  * ralsina is dizzy now. Dobey's fault.
[19:42]  * dobey blames IEC
[19:43] <sysdoc> Hey all Ubuntu one tells me in 11.10 that there is a Thunderbird contacts sync, however in 12.04 I I don't see a contacts sync at all, and on the site there is just instructions for Evolution. Where is the Thunderbird sync in 12.04?
[19:43] <sysdoc> oops,lol
[19:44] <ralsina> Also, octet means both "byte of 8 bits" and "byte of 3 bits", in english.
[19:44] <dobey> ralsina: no it doesn't
[19:45] <ralsina> dobey: yes it does
[19:45] <dobey> proof?
[19:45] <ralsina> it's the smallest byte that can represent octal numbers
[19:46] <dobey> that is true of a 3 bit byte, but it is not an octet
[19:47] <ralsina> dobey: http://en.wikipedia.org/wiki/Talk%3AOctet_(computing) if you want to argue :-)
[19:47] <dobey> also, we are using IEC standard as specified by Ubuntu Units Policy, which means KiB/etc are not translated
[19:48] <ralsina> then that branch should be reverted
[19:48] <dobey> probably
[19:49] <ralsina> proposing a revert then
[19:49] <dobey> though
[19:49] <dobey> "bytes" probably does need to be translated
[19:49] <ralsina> argh
[19:49] <dobey> let's see
[19:49] <ralsina> yes because it's not a unit
[19:50] <ralsina> OTOH, the user will never ever see bytes there
[19:50] <dobey> also
[19:50] <dobey> i wish people would stop using "base-2" to mean "base 8"
[19:50] <ralsina> hehe
[19:50] <ralsina> I think I will just let it be, it's harmless
[19:51] <ralsina> and not merge it anywhere
[19:51] <dobey> ?
[19:51] <ralsina> not stable-3-0, at least
[19:51] <dobey> well, everyone can translate KiB to KiB for their favorite languages I guess :)
[19:51] <ralsina> dobey: you branched stable-4-0 after this one merged?
[19:52] <dobey> i'll have to see if it's in stable-4-0 already or not
[19:52]  * dobey goes back to releasing things
[19:52] <ralsina> dobey: right, because I need to not target 4-0 for bugs merged to trunk before it branched. Life is hard-ish.
[19:53] <dobey> ralsina: well, we do sort of
[19:53] <ralsina> dobey: I would target *and* mark as fix-committed at least
[19:53] <dobey> so we can target to the milestone/etc
[19:54] <dobey> sure
[19:55] <ralsina> yep, revno 327, 8 revnos after this one. Will target correctly each one.
[19:55] <mmcc> ralsina and anyone who wants to look at qt.conf fun: http://paste.ubuntu.com/1023834/
[19:56] <dobey> cool
[19:56] <dobey> thanks
[19:56] <ralsina> mmcc: looks like qt.conf has to be somewhere else
[19:56] <ralsina> mmcc: is there anything like strace to figure out what paths the app is trying?
[19:56] <mmcc> dtrace. good idea
[19:57] <ralsina> mmcc: for example, on windows qt.conf has to be in the python.exe folder
[19:57] <mmcc> ralsina: however, recall that an empty qt.conf *does* stop it from loading the image plugins
[19:57] <mmcc> that is, an empty myapp.app/Contents/Resources/qt.conf
[19:57] <mmcc> so...
[19:57] <mmcc> and that's where the docs say it should be
[19:58] <mmcc> but I wouldn't mind looking at some IO traces
[19:58] <ralsina> mmcc: hmmm then I don't understand line 14 of your paste
[19:58] <ralsina> oh, it fixes it for image plugins and not for network plugins?
[19:58] <mmcc> what am I saying, I'll mind. I'll just do it
[19:58] <mmcc> ralsina:  right.
[20:00] <mmcc> ralsina: line 14 means that in that case, when we call QtNetwork proxy stuff, it loads /usr/local/Cellar/qt/4.8.1/plugins/bearer/libqcorewlanbearer.dylib, which depends on /usr/local/cellar/qt/4.8.1/lib/QtNetwork.framework, instead of the QtNetwork.framework we have in the app wrapper
[20:00] <mmcc> , which is already loaded, so we get this warning about loading the same library twice
[20:00] <ralsina> mmcc: ok, so for that plugin, the empty .conf is not doing much
[20:00] <mmcc> although I should note that it doesn't seem to cause any problems, yet. It did crash with the image plugins though
[20:02] <ralsina> briancurtin: sorry to bother, but did this revision have a bug attached? http://bazaar.launchpad.net/~ubuntuone-control-tower/ubuntuone-control-panel/stable-4-0/revision/327
[20:03] <mmcc> ralsina: right. it doesn't seem to affect the network loading. I'll go look at tracing IO
[20:03] <ralsina> mmcc: cool, good luck!
[20:12] <gatox> eod here!! time to rest! see you tomorrow!
[20:12] <ralsina> bye gatox!
[20:15] <dobey> hmm
[20:16] <dobey> crikey it is that late already
[20:18] <dobey> ralsina: 2 failures on your branch, btw
[20:18] <ralsina> dobey: really?
[20:18] <ralsina> dobey: I did run tests
[20:18] <dobey> yep
[20:18] <dobey> something comparing 'bar' to u'My Ubuntu'
[20:19] <dobey> in 2 tests
[20:20]  * briancurtin back
[20:20] <briancurtin> ralsina: no, that one didn't as far as i know but i'll double check
[20:20] <ralsina> dobey: ok, got them. Only happens with LANG=C for some reason
[20:20] <ralsina> briancurtin: don't bother, no need
[20:21] <ralsina> Oh, fun, LANG=C I get two failures and a segfault
[20:23] <dobey> nice
[20:23] <ralsina> sigh
[20:23] <ralsina> everything is difficult today.
[20:23] <dobey> actually, i think tarmac is running with LANG unset
[20:23] <ralsina> dobey: probably same thing
[20:24] <dobey> actually, it's weirder
[20:24] <ralsina> I am sorting ascii strings why would it be locale-dependent :-(
[20:24] <dobey> ralsina: well, notice that one string is a str() and one is a unicode()
[20:24] <ralsina> ohhhh
[20:25] <ralsina> but anyway, should still work
[20:25]  * ralsina debugs
[20:30] <alecu> guys, gals, I must take Amelia to a doctor's appointment. I'll see you in an hour or a bit more.
[20:31] <alecu> ttyl!
[20:32] <sysdoc> My ubuntu one does not have a services tab, what is up with this in 12.04?
[20:32] <ralsina> sysdoc: that's not necessary anymore
[20:32] <sysdoc> So how to install the Thunderbird contacts sync?
[20:35] <ralsina> sysdoc: just install the package for it from software center
[20:35] <ralsina> dobey: found it, Qt sorts differently on different locales
[20:36] <ralsina> dobey: on "C" "b" > "C" but on "en" "b" < "C"
[20:36]  * ralsina is about to kick something
[20:36] <dobey> yep
[20:37] <ralsina> even though it's technically correct and everything
[20:37] <ralsina> python disagrees though
[20:37] <ralsina> Oh, because on python we are sorting unicode
[20:37] <dobey> ls agrees
[20:38] <dobey> ls does the same thing
[20:38] <ralsina> if we were sorting ascii, it's the same. Of course then it would fail on "en"
[20:38] <ralsina> possibly evil fix is unsetting LANG on run-tests ;-)
[20:38] <dobey> why do the tests care about sorting?
[20:39] <dobey> i mean, either way, it will still also fail on other languages
[20:39] <ralsina> dobey: because it has to check if the items in the folder list have the correct text in them
[20:39] <ralsina> dobey: no, it will work in any language except on "C"
[20:39] <dobey> no, it won't
[20:39] <ralsina> dobey: because we are sorting two things alphabetically and comparing them
[20:39] <sysdoc> ralsina, any idea what the package might be called?
[20:40] <dobey> well, most any language
[20:40] <dobey> sysdoc: thunderbird-couchdb
[20:40] <ralsina> dobey: works in tr_TR :-)
[20:40] <ralsina> dobey: only fails when python and Qt disagree on what "alphabetical" means
[20:40] <sysdoc> ralsina, thank you!
[20:40] <dobey> it will succed in any language which doesn't change the semantics of sorting related to the latin alphabet :)
[20:41] <ralsina> dobey: no, we are sorting the same strings twice, once using Qt, once using python
[20:42] <dobey> anyway, tests shouldn't depend on sorting
[20:42] <dobey> or language
[20:42] <ralsina> dobey: we are (in part) testing that we sort them correctly. I am working now on making it not depend on locale, too.
[20:47] <mmcc> for future reference: QT_DEBUG_PLUGINS exists, but it's not really any more informative than DYLD_PRINT_LIBRARIES=1...
[21:02] <dobey> ralsina: https://code.launchpad.net/~dobey/ubuntuone-client-gnome/update-stable-4-0/+merge/108638
[21:03] <ralsina> dobey: +1 trivial
[21:20] <mmcc> interesting: according to this Qt bug, qt will never pay attention to a qt.conf in the app package in OS X: https://bugreports.qt-project.org/browse/QTBUG-24541?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
[21:59] <salgado> :q
[22:03] <dobey> that's a weird smiley. or this isn't vim.
[22:14] <briancurtin> Qt sucks, i'm out. wasted day.
[22:16] <dobey> heh
[22:16] <dobey> oh crikey
[22:26] <mmcc> I think brian and I had the same day...
[22:38] <mmcc> ok something is forking deep in twisted, and *that* is loading plugins, and it doesn't have the right qt.conf
[22:40] <mmcc> might just be a theory, but the dtruss (aka strace) traces show that the qt.conf is definitely getting read by the process that loads the image plugins, but later on, something else still looks in the wrong place.
[22:42] <dobey> sigh. fml.
[22:44] <mmcc> dobey: I thought it'd be funny to run emacs' 'doctor' mode on your comment, but it just asked me if I was afraid of sex
[22:45] <mmcc> :(
[22:45] <mmcc> is anyone still around who's familiar with QNetworkAccessManager?
[22:45] <dobey> heh
[22:51] <dobey> well i'm glad i didn't go try to cut grass
[22:51] <dobey> no forecast of rain at all
[22:51] <dobey> what's it doing
[22:52] <dobey> sigh
[22:52] <dobey> ok, i am gone.
[22:53] <mmcc> bye dobey, send some rain this way
[22:56] <mmcc> for the benefit of future readers: Here's the status - somewhere in the QtNetwork code, it appears that something is forking a new Qt process which doesn't know about the qt.conf I put in the .app resources dir. Therefore, it starts loading bearer plugins and gets them from the wrong place. The stat() call and stack trace where we can see what it's doing is here: http://paste.ubuntu.com/1024099/
[22:56] <mmcc> I have to leave now, but I will probably be back later tonight
[22:57] <mmcc> from looking at the Qt code, I don't think this exists, but if I've missed an env var to set the location of qt.conf, I'd love to hear that. That'd be a nice way of solving this. Right now I'm not even sure what executable is getting forked
[22:58]  * mmcc exits, stage left