[00:03] <mmcc> ok, that sync menu / indicator bug appears nice and fixed now on os x, but for some reason I can't reproduce the bug on ubuntu with recent trunk.
[00:04] <mmcc> have to go now, will come back later tonight to try again and then propose the merge
[00:05] <mmcc> what should happen: start CP with --with-icon, and then do something outside the CP to change the status (add a file, etc). with the current code, the menu should update but the CP main window indicator won't, because the menu's handler overwrote the main window's handler
[00:06] <mmcc> but on linux just now the main window's indicator worked fine when called --with-icon, so maybe the menu registered its handler first? or something else is going on…
[00:06] <mmcc> to be continued…
[07:12] <mmcc> ok, it looks like that issue is because of dbus - in the dbus implementation of the syncdaemon proxy we ADD callbacks for the status-changed signal, while in the perspective_broker implementation, we just SET the callback, so There Can Be Only One (callback)
[07:13] <mmcc> so my question for the night is - is the sys tray status menu broken on windows like it is on OSX now, and no one noticed? or is there something else I need to figure out?
[07:13] <mmcc> but it's time for sleep now, so I look forward to seeing your answers in the morning :)
[08:32] <JamesTait> Good morning all! :D
[12:03] <alecu> hello!
[12:06] <facundobatista> Hola alecu
[12:13] <mandel> alecu, buenas!
[12:14] <mandel> alecu, so I'm nearly done with clicking download and showing a empty generic preview with a cancel and buy now actions
[12:14] <mandel> alecu, debugging at the moment why the second preview does not get shown
[12:14] <mandel> is something like this: http://paste.ubuntu.com/1187152/
[12:15] <alecu> mandel: is that all on the unity side? I've been working on some bits on the lens side.
[12:15] <alecu> mandel: oh, no: that's on the lens side too.
[12:16] <mandel> alecu, is in the unity-lens-music side in the musicstore scope to get it done, the next step is to do the actual preview in the cpp side (unity)
[12:16] <mandel> alecu, yes, atm the unity preview I have is a summy one
[12:16] <mandel> s/summy/dummy
[12:17] <alecu> mandel: yes, this is similar to what I've been playing with.
[12:18] <mandel> alecu, ups.. well, I got to the point where I got an answer of the type: 'this was never tested'
[12:19] <mandel> alecu, I just wanted to be able to see the u1paymentpreview so I had to do this glue anyway
[12:24] <alecu> mandel: right! I didn't get that far, either: I only showed the same preview, but with a new button added.
[12:24] <alecu> mandel: but the dash would not show the new button.
[12:25] <alecu> mandel: also, I got a lot of trouble getting the file.from_url to work.
[12:25] <mandel> alecu, where you using the same instance of the preview?
[12:25] <alecu> mandel: yes
[12:26] <mandel> alecu, that might be the reason why it shows a new button.. from here: http://people.canonical.com/~mhr3/python/Unity-5.0.html I understand that with_preview takes a new instance of a preview and does the right thing
[12:26] <alecu> mandel: it seems that gvfs is broken on Q, so getting a url is broken. I had to replace every webcall with a reading from a local file.
[12:26] <mandel> alecu, but it seems that it does not..
[12:26] <alecu> mandel: right
[12:26] <mandel> alecu, fuuuuu to gvfs
[12:27] <alecu> mandel: even nautilus has problems opening the trash:// url!
[12:27] <mandel> alecu, dednick seems to be they guy to ping about the preview problem
[12:27] <alecu> anyway, I'm scared of updating, because I updated yesterday morning and then spent half of the day chasing this issue :-(
[12:28] <alecu> mandel: right, let's ask him
[12:28] <mandel> alecu, I'm in theory up to date, let me try to get to the trash
[12:29] <mandel> alecu, in my system trash:// on  nautilus (or files how they call it know) works
[12:29] <alecu> mandel: if the webcalls work for you then it's not broken.
[12:29] <alecu> mandel: when did you update your vm?
[12:29] <mandel> alecu, is my actual machine
[12:29] <mandel> like a few mins ago
[12:30] <alecu> mandel: awesome, updating too.
[12:42] <ralsina> good morning!
[12:43] <ralsina> mandel: unless you have gnome from git, you still have nautilus :-)
[12:43] <ralsina> mandel: ubuntu didn't update it
[12:44] <mandel> ralsina, yes, we are in 3.4
[12:44] <alecu> ralsina: good morning to you, boss
[12:45] <dobey> hmm
[12:49] <mandel> lunch time for me, I need energy to deal with unity ;-)
[12:49] <alecu> morning to you too, dobey!
[12:49] <alecu> mandel: buen provecho
[12:49] <dobey> buenas dias alecu
[13:59]  * mandel back
[14:00] <mandel> alecu, I'll be doing the preview for the payment, or at least the closest thing possible
[14:01] <alecu> mandel: let's settle on what we will send in that extended struct, so we can work in parallel.
[14:01] <mandel> alecu, sure, mumble?
[14:01] <mandel> alecu, talking seems to be faster
[14:01] <alecu> mandel: let me find my mumbling laptop
[14:02] <mandel> sure
[14:20] <mmcc> hey folks
[14:21] <mmcc> does anyone use the windows sys tray icon/menu?
[14:21] <mmcc> I think it should break the main window's status indicator the same as on darwin…
[14:21] <briancurtin> mmcc: yes
[14:22] <briancurtin> it doesnt really do a lot, but you have to use it to quit so that's about all i use it for
[14:22] <mmcc> briancurtin: have you ever noticed if the control panel main window status indicator is correct or not?
[14:23] <briancurtin> mmcc: you mean the green checkbox thing, or something else? (i'm not familiar with 'main window status indicator')
[14:24] <mmcc> briancurtin: sorry, yeah the green checkmark in the top right of the control panel window.
[14:25] <mmcc> briancurtin: I think that on windows, with the sync menu running, the checkmark in the main window should be stale - never updated after the initial status
[14:25] <mmcc> I'm going to go test it in my windows VM, but I was just wondering if that was a known bug
[14:26] <briancurtin> mmcc: one of my tests after we make a new installer is to make the installer, then sync the installer to my u1 with that installer, and i watch that status change from green to synching to green, then check it online, as a smoke test
[14:26] <briancurtin> mmcc: is this stale status a recent change/issue?
[14:26] <mmcc> briancurtin: I don't think it's recent. hmmm.
[14:26] <briancurtin> mmcc: because i haven't been watching status change since i did the last installer a few weeks ago...otherwise i just have u1 off in the tray and i ignore it
[14:27] <mmcc> briancurtin: the sys tray icon isn't running in its own process or anything weird, right?
[14:28] <briancurtin> nope
[14:29] <mmcc> the issue is that the IPC version of the syncdaemon proxy replaces the status update handler when you tell it to set the handler, while the dbus version *adds* the handler - so on linux, both get called but on darwin, only the menu's handler gets called, because it was set last
[14:29] <mmcc> It *should* be the same on windows, but I guess there's more to it than that
[14:31] <briancurtin> mmcc: hmm. let me resync everything with trunk and try it out
[14:32] <mmcc> briancurtin: ok, thanks. I'm testing it too, so if you're busy you can skip it…
[14:35] <briancurtin> nvidia driver just crashed three times, so that's nice. "when in doubt, reboot"
[14:50] <mmcc> brb
[14:55] <mmcc> b
[14:59] <ralsina> mmcc, briancurtin, mandel, thisfred, alecu, dobey: standup in 1'
[15:00] <alecu> thanks!
[15:00] <mandel> me
[15:01] <ralsina> me
[15:01] <thisfred> me
[15:01] <briancurtin> me
[15:01] <alecu> me
[15:01] <dobey> me
[15:01] <ralsina> mmcc is last, go mandel!
[15:01] <mandel> DONE: Tried to get a preview after a preview in the dash, it looks like calling Unity.ActivationResponse.with_preview does not work as expected, we will hardcode previews to always show the payment confirmation until we get that to work. Understood how we can pass the data between the processes so that, 1 we get data from the lens scope, 2 we send the password back by reading from the text field.
[15:01] <mandel> TODO: Create a mock of the payment preview with hardcode data but correct layout + sending the password back (will add the proper dict once alecu has written the doc so that we are in the same page).
[15:01] <mandel> BLOCKED: no
[15:01] <mandel> ralsina, please
[15:01] <ralsina> DONE: calls, worked on fixing a bug, talked a lot with design, 1-1s, etc. TODO: propose branch, fix other stuff BLOCKED: headache NEXT: thisfred
[15:02] <alecu> DONE: music-lens: research, understanding and mumbling
[15:02] <alecu> TODO: more of the same
[15:02] <alecu> BLOCKED: no
[15:02] <alecu> oops
[15:02] <briancurtin> DONE: replaced getPage and the HTTPServer's twisted implementations with standard library "equivalents" in https://code.launchpad.net/~brian.curtin/ubuntuone-dev-tools/py3-getPage-replacement
[15:02] <briancurtin> TODO: finish the HTTPS version
[15:02] <briancurtin> NEXT: dobey
[15:02] <dobey> DONE: icon magic, music store poking
[15:02] <dobey> TODO: release/package icon magic?, more music store poking
[15:02] <dobey> BLCK: None.
[15:02] <briancurtin> back to thisfred
[15:02] <dobey> mmcc or thisfred: fight and go
[15:02] <thisfred> DONE: u1db basic auth | fixed get_docs bug TODO: u1db playlists BLOCKED: no NEXT: briancurtin
[15:02] <mmcc> DONE: tweaked daemon branches, review, investigate icons, sync
[15:02] <mmcc> indicator bug
[15:02] <mmcc> TODO: sync indicator, start new daemon
[15:02] <mmcc> BLCK: no
[15:03] <mmcc> I am confused by our new ordering scheme
[15:03] <mandel> thisfred, is that an or | a pipe?
[15:03] <ralsina> Not bad, no twitter paddlings today!
[15:06] <alecu> briancurtin: I like how you solved the HTTPWebServer issue in that branch. Have you tried running the sso tests with it?
[15:07]  * alecu is really curious on how this will interact with our varied WebClient implementations.
[15:11] <briancurtin> alecu: i haven't tried yet. i have one small branch to fix that i can do right now, then i can check SSO
[15:15] <mmcc> better status note: so I have a fix for the sync indicator inconsistencies, and I understand why the fix wasn't necessary on linux, just testing now on windows to be sure I understand what's going on there
[15:20] <dobey> ugh
[15:21] <dobey> png2icns doesn't support the @2x stuff
[15:29] <dobey> ok, need to get lunch. bbiab
[15:34] <ralsina> dobey: when you come back, review of https://code.launchpad.net/~ralsina/ubuntu-sso-client/change-your-mind/+merge/122904 please (sadly, they are new strings)
[15:50] <mmcc> so, what am I forgetting on windows to get the debug output? SET U1_DEBUG=1 doesn't seem to be having any effect
[15:50] <briancurtin> set that, and then run the app with --debug as well
[15:51] <mmcc> briancurtin: that seems to have no effect either…
[15:52] <briancurtin> weird. what exactly are you running?
[15:52] <mmcc> control-panel. I'm trying to reproduce the sync indicator inconsistencies I mentioned earlier
[15:53] <mmcc> and I'm running from the source tree
[15:53] <briancurtin> ohh, i'm not sure if u1cp has a debug output. SSO and syncdaemon do for sure
[15:54] <mmcc> ah, ok. so we need to do something special to enable sending the log to the console on windows?
[15:55] <mmcc> because that'd be handy, as running from source isn't finding the sso gui :( (It's looking for ubuntu-sso-login-qt.exe)
[15:57] <mmcc> but it shouldn't be… urgh
[15:59] <mmcc> OK - this was unexpected - I noticed there was still an icon in the tray, so I quit from there, then tried starting again, and control-panel complained about my command line args
[16:00] <briancurtin> how are you starting it?
[16:00] <mmcc> they were wrong the whole time - but it was showing the control panel despite the args being wrong, until I quit the running instance - does it trigger some kind of "show running program" hook ?
[16:01] <mmcc> I was running it with 'python bin\ubuntuone-control-panel-qt --with_icon' (note underscore instead of dash)
[16:02] <mmcc> and now that I quit the old control panel, I'm getting debut prints from new runs
[16:03] <briancurtin> mmcc: yeah it'll just show the running one, since we only allow one instance at a time
[16:03] <briancurtin> mmcc: nice, i didn't think we had debug output in cp for some reason, but i guess i never tried that one. i know for sure its in SSO and SD
[16:04] <mmcc> huh, I guess I was expecting it to check the args before checking the unique instance, but I see we do it the other way around
[16:04] <mmcc> ok, not so mysterious anymore :)
[16:18] <ralsina> mmcc, briancurtin: either one could do a review of https://code.launchpad.net/~ralsina/ubuntu-sso-client/change-your-mind/+merge/122904 please?
[16:18] <ralsina> to test it run "PYTHONPATH=. bin/ubuntu-sso-login-qt [--login_only] --app_name testing"
[16:19] <mmcc> ralsina: I'll do it on darwin
[16:19] <briancurtin> i'll check it
[16:20] <mandel> mmcc, jut got the email about the latests branch of the daemon (sorry crappy beta email client in Q)
[16:20] <mmcc> mandel: which email was that?
[16:21] <mandel> mmcc, the latests is jenkins-daemon-main but some tests are missing
[16:21] <mmcc> oh the one about proposing the jenkins-daemon-main branch, right
[16:21] <mandel> mmcc, the one with the subject: 	fsevents daemon most recent branch
[16:21] <mandel> mmcc, yes, so, tests are missing but is nearly done
[16:22] <mmcc> ok. well, we can wait on that until you have time, I was just wondering if it was ready yet
[16:22] <mandel> mmcc, no, not yet
[16:23] <mmcc> ok, thanks
[16:28] <briancurtin> ralsina: hm, for some reason i end up getting an ImportError for qt4reactor (which is there, and i can import it). looking into it
[16:28] <briancurtin> it happens after the window comes up, then it gets "Getting information, please wait..."
[16:34] <mmcc> briancurtin: are you running from source? I'm seeing import errors for twisted.internet, same problem. I wonder if the fixes from a while back for running from source didn't quite do it
[16:35] <briancurtin> source
[16:35] <mmcc> specifically, prepending 'python' to the command line for IPC
[16:39] <briancurtin> i dont know what the "for IPC" part means, but i'm running "python bin\ubuntu-sso-login-qt --app_name testing"
[16:41] <dobey> ralsina: needs fixing :)
[16:41] <mmcc> briancurtin: I mean that when the code generates exe names to start subprocesses, it is supposed to be prepending 'python' if sys.frozen isn't defined
[16:41] <mmcc> in my case, it seems like it's not doing that correctly for syncdaemon, sometimes :\
[16:41] <briancurtin> yeah it does. i see get_bin_cmd returning [u'python', u'C:\\u1\\working\\change-your-mind\\bin\\ubuntu-sso-login']
[16:42] <briancurtin> weird that it's not doing that for you
[16:42] <dobey> maybe an issue on osx only?
[16:42] <mmcc> no, it's working for sso-login
[16:42] <dobey> or are you both actually talking about running on windows?
[16:42] <mmcc> dobey, right now I'm in windows
[16:43] <dobey> hmm
[16:43] <mmcc> I still haven't tried ralsina's branch
[16:45] <briancurtin> i'm on windows and using ralsina's branch
[16:46] <briancurtin> i'm not sure that this is specific to his branch, though. trying on trunk now
[16:48] <briancurtin> ehh, doesn't work from source there either
[16:49] <briancurtin> i'll try building this out to executables and look at it that way for the time being
[17:05]  * briancurtin lunch+doctor run, back in a bit
[17:09] <dobey> and there was much rejoicing!
[17:17] <mandel> ok, EOD for me
[17:19] <mmcc> well, running from source is just broken on windows for me
[17:20] <mmcc> I suspect it has something to do with this bug about inconsistencies between windows and unix in os.exec: http://bugs.python.org/issue8557
[17:20] <mmcc> if shell=False in subprocess.Popen, windows uses a function that won't look in $PATH
[17:20] <mmcc> sounds suspicious but I'm not sure that's really what is killing this for me
[17:21] <mmcc> bleh
[17:21] <ralsina> dobey: right, the app name. On the other bit, apparently user testing has caused us to avoid using "sign up" everywhere
[17:26] <dobey> ralsina: i'm not picky; but it can't say ubuntu one there :)
[17:30] <ralsina> dobey: agreed, removed the "Ubuntu One"
[17:30] <ralsina> dobey: BTW, next cycle it's going to say Ubuntu One everywhere because SSO is being rebranded
[17:30]  * ralsina is ahead of the curve
[17:33] <dobey> anyway, couple more comments on the proposal
[17:38] <ralsina> dobey: the different strings in the .ui files are meaningless, they are just placeholders
[17:45] <dobey> ralsina: assuming we don't use them on windows/mac to do translations through the Qt stuff in the future, I guess?
[17:47] <ralsina> dobey: if we do that, we are in for a world of pain, so I propose we don't do that
[17:47] <dobey> ok
[17:47] <ralsina> dobey: we can just always do the translation generation on linux and just use the files on windows
[17:50] <dobey> right
[17:52] <dobey> ralsina: can you make the strings be the same anyway? So it's clear when someone else edits the ui in the future?
[17:52] <ralsina> dobey: sure
[17:52] <dobey> thanks
[17:55] <ralsina> dobey: pushed
[17:59] <mmcc> hey ralsina, do you still have a branch pending working on the share_links panel?
[17:59] <mmcc> s/panel/tab
[17:59] <ralsina> mmcc: no
[17:59] <mmcc> oh, it's merged?
[17:59] <ralsina> mmcc: I have been doing some exploration there but nothing is working well
[17:59] <ralsina> mmcc: yes
[18:00] <mmcc> ah, ok. I noticed test_share_links_search has some paths that are wrong on windows. I'll just file a bug
[18:00] <mmcc> was going to bug you if you were messing with it already anyway
[18:00] <ralsina> mmcc: ok!
[18:03] <dobey> ralsina: looks ok to me now.
[18:05] <ralsina> dobey: I know it's light on tests but I would have to introduce whole new tests cases because this is not tested at all :-(
[18:05] <ralsina> dobey: so thanks
[18:09] <mmcc> ralsina, it's bug 1046466. I also noticed that the code it's testing returns paths with ~ in them - does that work on windows?
[18:09] <ralsina> mmcc: yes, if you apply expanduser to it
[18:10] <ralsina> mmcc: or rather "on any OS that only works if you apply expanduser to it"
[18:10] <mmcc> ralsina: ok then, it should be a trivial test fix
[18:21] <dobey> ralsina: and we have the FFe approval now; so as soon as it's approved/landed in trunk, i'll get the patch into the Ubuntu package as well :)
[18:22] <ralsina> dobey: awesome
[18:27] <dobey> ralsina: btw, your math is wrong. :)
[18:27] <dobey> ralsina: if you don't pay your $100 the first month, the second month you would owe $202, not $102, no? ;)
[18:28] <ralsina> dobey: not if you pay the second month's bill :-)
[18:29] <ralsina> dobey: a possible strategy is to pay every other month, for example
[18:29] <ralsina> very sun tzu, confusing the enemy, etc.
[18:30] <dobey> "only pay on months when the moon is full on the eve of the bill due date"
[18:30] <ralsina> dobey: sure
[18:30] <dobey> sounds less Sun Tzu, and more OCD. :)
[18:31] <ralsina> you have only two moves, pay or not pay
[18:31] <ralsina> the city as sue or not sue
[18:31] <ralsina> on each owner, of course
[18:42] <mmcc> aha, so the windows tray icon didn't have file sync status until just recently
[18:43] <mmcc> so I'm not crazy and I can propose this branch
[18:43] <mmcc> as soon as I test it on linux
[18:45] <mmcc> have we had any serious windows testing since that status menu stuff went in? (ralsina ?)
[18:52]  * briancurtin back
[18:52] <briancurtin> mmcc: i dont know anything about "status menu stuff" so i would think that answer is probably no
[18:53] <ralsina> mmcc: nope
[18:54] <mmcc> briancurtin: the "status menu stuff " landed between august 14 and 20, it adds extra stuff to the sys tray menu and breaks the main window sync status indicator on mac and windows
[18:54] <ralsina> mmcc, briancurtin: we should make windows nightlies work again and ask QA to run some
[18:54] <ralsina> I will put it on my plate for tomorrow
[18:54] <mmcc> ralsina: any plans to eventually have tarmac run tests on mac and windows?
[18:55] <mmcc> I guess I know a bit about the plans for mac
[18:55] <dobey> well, no, not really
[18:55] <mmcc> not that the tests caught this
[18:55] <dobey> tarmac running tests on everything would be crazy complicated :-/
[18:56] <mmcc> fair enough. be nice to have automated tests on windows at least, though. seems like no one (including me) enjoys firing up windows when reviewing branches
[18:56] <ralsina> mmcc: running tests pre-commit on tarmac in more than on eOS is very very difficult
[18:56] <ralsina> mmcc: yes, we relaxed on that a bit and we shouldn't have
[18:57] <briancurtin> i didn't know windows nightlies weren't working. i figured since QA hadn't said anything that they were fine
[18:57] <mmcc> I'll start with the man in the mirror… gotta solve my VM shared folder problems
[18:58] <mmcc> that'll make it less annoying to run windows tests
[18:59] <dobey> mmcc: we do have automated tests on windows in jenkins; but i think they are failing for some reason
[19:00] <mmcc> dobey: ok. where do I look to see that reason? not promising I'll fix it, just curious if it's the same thing I just fought
[19:01] <ralsina> mmcc: there has been a broken test for a few weeks we never got around to fixing :-/
[19:02] <mmcc> ok, so not the same thing.
[19:04] <ralsina> mmcc: https://jenkins.errormessaging.com/view/Windows/ 20 days or so down
[19:05] <dobey> ah
[19:05] <dobey> search box tests are failing on cp
[19:05] <dobey> a couple of u1sdtool tests are failing on client
[19:06] <dobey> hrmm
[19:08] <mmcc> ok those search box failures are bug 1046466
[19:09] <dobey> ok
[19:09] <mmcc> and the other one is something that's been really annoying me on darwin too
[19:09] <dobey> i have a branch to fix the windows-installer issue
[19:09] <mmcc> I meant the u1-client one . weird that that's where it shows up
[19:10] <dobey> oh, the on_download_finished issue?
[19:10] <mmcc> yeah
[19:10] <dobey> https://code.launchpad.net/~dobey/ubuntuone-windows-installer/fix-jenkins/+merge/122928
[19:11] <mmcc> annoying when testing because it spams the log output
[19:11] <dobey> ralsina, mmcc, briancurtin: ^^ that should fix the installer build anyway
[19:11] <ralsina> dobey: checking
[19:15] <ralsina> dobey: looks good to me
[19:16] <mmcc> dobey, looks ok. just running setup-mac now to assuage my paranoia
[19:17] <briancurtin> looks fine, building now
[19:17] <briancurtin> cross platform jinx
[19:18] <dobey> heh
[19:18] <dobey> mmcc is buying coke for everyone?
[19:19] <mmcc> huhwha
[19:20] <mmcc> briancurtin: I'll let you sign off, since the windows change is more significant
[19:26] <dobey> hrmm, i wish there was an easier way to pull all the translations out of ubuntu for something, off launchpad. the whole getting a link in an e-mail thing is a bit tedious
[19:30] <mmcc> ps briancurtin if you are building trunk to test that jenkins fix, can you also just poke at the sync indicator? start cp, add a file to your synced folder, see if it changes the status shown in the window, and see if it's in sync with the menu that the sys tray icon shows…
[19:30] <briancurtin> mmcc: will do
[19:30] <mmcc> briancurtin: thanks!
[19:36] <dobey> gah; the cloudy shroud of darkness is moving in
[19:42] <briancurtin> dobey: installer branch is good
[19:42] <dobey> thanks
[19:50] <dobey> rain rain, GTFO
[20:10] <briancurtin> mmcc: installer built from trunk doesn't work. "getting information" forever, status shows "loading" after it started up disconnected. weird. i cant really look into it right now though since i'm behind on py3 stuff
[20:12] <mmcc> briancurtin: ok. thanks for looking. :(
[20:14] <mmcc> I also get the 'getting information' forever on windows
[20:16] <briancurtin> mmcc: maybe tied to the issue i was having earlier with ralsina's branch where sso-login-qt was starting sso-login and when it did that, it could no longer import qt4reactor
[20:17] <briancurtin> er, not an issue with his branch, just an issue seen while trying to test it
[20:17] <mmcc> briancurtin: maybe. I was hoping that was confined to running from source and not having the right buildout
[20:17] <ralsina> mmcc, briancurtin: good thing we are not planning any windows releases
[20:17] <mmcc> er, not having the right name for python in the new buildout
[20:17] <ralsina> mmcc, briancurtin: and I know that's not "good" but hey, we are stretched
[20:17] <mmcc> right right
[20:46] <mmcc> lunch
[20:50] <dobey> hrmm
[20:50] <dobey> we use python-distutils-extra on windows/mac also?
[20:51] <briancurtin> dobey: i havent looked into what it's used for, but it's there
[20:51] <briancurtin> and i seem to remember not having it once and it causing issues, but that was long ago
[20:52] <dobey> oh, i see
[20:52] <dobey> the build_i18n is replaced in sso
[20:53] <ralsina> dobey: it's there but it's completely broken, at least on windows
[20:54] <dobey> broken how?
[20:55] <ralsina> dobey: well, it does nothing useful
[20:55] <dobey> build_i18n?
[20:55] <ralsina> dobey: and when I was doing the windows packaging, nothing -extra is supposed to do worked
[20:55] <ralsina> dobey: build_i18n doesn't work because we lack tools on windows
[20:55] <dobey> ralsina: then how is ubuntu-sso-client installed?
[20:56] <dobey> ralsina: right, which is why there is a dummy override in ubuntu-sso-client's setup.py
[20:56] <ralsina> dobey: it's probably just copied into a folder
[20:56] <ralsina> dobey: since "installing" it on windows is useless anyway
[20:56] <dobey> because we're not installing intltool and everything it needs as well
[20:56] <ralsina> dobey: because that's not how we ship it
[20:57] <dobey> ralsina: well, it has options for py2exe stuff defined in it
[20:57] <ralsina> dobey: yes, they never worked, either
[20:57] <dobey> if all that stuff doesn't work, we should probably remove it all
[20:57] <ralsina> dobey: py2exe needs so much hand holding, the generic stuff -extra does never worked at all
[20:57] <ralsina> yes
[20:57] <ralsina> we should not even import it on windows
[21:00] <dobey> the problem is i think we have to :(
[21:00] <dobey> hrmm, or maybe not
[21:00] <dobey> but extra ugly stuff if not
[21:01] <ralsina> dobey: that setup.py needs lots of love from someone who actually understands that kind of thing
[21:01] <ralsina> which AFAIK is noone in the world
[21:02]  * ralsina has hunted years for that mythical developer who understands crossplatform python setup scripts
[21:04] <briancurtin> everyone who has tried to tackle distutils eventually wishes to be disassociated with the project
[21:05]  * dobey sort of understands that stuff; though not py2exe
[21:05] <briancurtin> a few years ago i might have said to get tarek ziade...i think he retired from distutils forever
[21:05] <dobey> to be fair, i have hacked on automake before.
[21:06] <ralsina> dobey: automake should have twisted your brain enough for this to make sense
[21:06] <ralsina> briancurtin: tarek ran away after a while yes
[21:06] <ralsina> like everyone else
[21:06] <ralsina> dobey: py2exe adds a lot of funny bits
[21:07] <ralsina> dobey: and AFAIK py2app adds a barrel of *other* funny bits
[21:07] <ralsina> dobey: in fact, any py2exe thing we have in the individual project's setup.py can be killed because we are not using it
[21:07] <ralsina> the only py2exe/py2app stuff in use in real life is in ubuntuone-windows-installer AKA the-project-that-should-be-renamed
[21:08] <ralsina> I may add cxfreeze support some friday to that to make it more interesting
[21:08] <dobey> ok
[21:11] <dobey> well i have a cross platform setup.py working for ubuntuone-client-data now; but it doesn't install anything for !linux yet
[21:11] <dobey> but of course, installing just an .ico or .icns file isn't very useful
[21:12] <ralsina> dobey: indeed
[21:12] <ralsina> dobey: on windows we don't even install the icon, it's embedded in the binary
[21:12] <dobey> well, it is on linux too
[21:13] <dobey> but we have some extra magical magic
[21:20] <alecu> om$d, I really hate C++, and the time it takes to compile.
[21:20] <alecu> on the other hand, I have a new found love in Vala.
[21:21] <alecu> ever since realizing that vala has yield and async methods that are very similar to inlineCallbacks.
[21:24] <dobey> heh
[21:31] <ralsina> alecu: there is a chance to send mandel to a nux sprint next week
[21:31] <alecu> ralsina: that would be awesome
[21:31] <ralsina> alecu: still pending approval, but hope that helps you guys
[21:31] <ralsina> alecu: I am reading a go book, it's like twisted, but sane
[21:32] <alecu> ralsina: we have split our work so he gets to do the ui, and would benefit a lot from going to that sprint.
[21:32] <ralsina> alecu: that's what I heard, so fingers crossed we get it
[21:33] <alecu> ralsina: I'm working on the bits that talk to the servers... so, as much I would love to go to London on summer, I won't benefit so much from it :-(
[21:33] <ralsina> alecu: haha, was not going to send you. Mandel is much closer.
[21:34] <dobey> alecu: can you review https://code.launchpad.net/~ralsina/ubuntu-sso-client/change-your-mind/+merge/122904 ? it's your review day after all :)
[21:35] <ralsina> alecu: and I *know* it lacks tests, but those areas have none and I can't do a hugebranch with completely new testcases :-(
[21:35] <alecu> dobey: oh, it's my review day, right! I spent all day reading c++ and vala :P
[21:36] <ralsina> alecu: you lazy dev, having fun all day ;-)
[21:36] <alecu> ralsina, dobey: are you guys in a hurry with this branch? I have to go now; I can do the review tonight or first thing tomorrow.
[21:37] <ralsina> alecu: early tomorrow is ok, we have 3 days
[21:37]  * alecu takes note.
[21:37] <ralsina> I would rather finish that thing tomorrow though :-)
[21:38] <ralsina> EOD for me
[21:38] <ralsina> bye people see you all tomorrow!
[21:38] <alecu> ralsina: my first review would be "please add a bug for the tests"!
[21:39] <ralsina> alecu: feel free to even file the bug and assign to me, but that's not something we are going to get a FFe for :-)
[21:39] <alecu> oh, right :-)
[21:41] <dobey> yes, the sooner that branch lands, the better. but tomorrow am review is ok. just want to get it in ubuntu and get it done :)
[21:41] <dobey> have other stuff i need to get FFe for as well :-/
[21:49] <alecu> dobey: ack.
[21:49] <alecu> ok, this is EOD for me! see you all tomorrow!
[21:50] <dobey> cheers alecu
[21:53] <dobey> same here; later all!
[22:03] <mmcc> back...
[23:11] <mmcc> OK, fixed some dbus tests and wrapped that sync indicator fix up. For review whenever: https://code.launchpad.net/~mikemc/ubuntuone-control-panel/fix-sync-status/+merge/122974