[09:15] <mandel> morning all!"
[09:15] <paranoidphreak> hi everybody, how do i install an "ubuntu one" client?
[09:15] <paranoidphreak> in debian
[09:29] <JamesTait> Happy Monday, all!
[09:34] <mandel> paranoidphreak, on debian.. well there are lots of dependencies that you have to double check that we have not added patches etc..
[09:34] <mandel> paranoidphreak, I have never tried to run it on debian to be honest.. it is going to be a 'little' adventure
[09:38] <paranoidphreak> mandel: ok...........so, the short answer is it needs alot of experimenting?
[09:38] <mandel> paranoidphreak, yes, and I know that rye (who is not here know) is the kind of guy that likes to do that
[09:39] <mandel> paranoidphreak, it should be possible since I know we have people using it on Fedora :)
[09:39] <paranoidphreak> mandel: thanks.........
[09:39] <mandel> paranoidphreak, sorry I cannot be more helpful, but got be coding at full speed to get to the bext ubuntu beta 2 with proxy support, sorry
[09:40] <paranoidphreak> mandel: i'll probably do some googling
[09:41] <paranoidphreak> mandel: goodluck on your coding
[09:42] <mandel> paranoidphreak, thx!
[09:52] <ralsina_> good morning!
[09:53] <alecu> mandel, morning!
[09:53] <alecu> ralsina_, morning to you too, sir!
[09:53] <mandel> alecu, ralsina_ estais los dos locos!
[09:53] <ralsina_> good morning alecu, mandel
[09:54] <alecu> mandel, it's only 7 am here ;-)
[09:54] <mandel> ralsina_, you've got mail with the answers that you needed about windows + a review request for a simple branch
[09:54] <alecu> mandel, and we only got 8 more hours to land stuff today
[09:54] <ralsina_> mandel: yes, trying it now
[09:55] <mandel> alecu, I'm trying to get the update of the proxy of network access manager working unless you have something else for me
[09:55] <mandel> alecu, I'm reading the c++ implementation to see if we can do something..
[09:56] <alecu> mandel, I've got one big thing we should do today: disabling ssl
[09:56] <alecu> mandel, and a few more smaller ones.
[09:57] <mandel> alecu, that one is very easy.. specially since the support for libsoupd was not landed, shall I disable it, remove it or add a param to the webclient to state that you will no allow ssl errors?
[09:58] <alecu> mandel, "disable it" sounds right. We might need to use that support much later.
[09:58] <mandel> alecu, ok, simple work, but then.. what do we do with the ssl support branches that have not landed? there are a few
[09:59] <ralsina_> mandel: QCoreApplication.instance() is not None if you have a QApplication, so no need to test both (line 21 of your branch's diff)
[09:59] <mandel> ralsina_, cool, I did not know that, will fix a in a min
[10:00] <ralsina_> mandel: also, typo in line 38 "cres"=> "creds"
[10:00] <mandel> ralsina_, can you put both in a comment?
[10:00] <ralsina_> mandel: sure!
[10:00] <ralsina_> Oh, great, launchpad is down
[10:02] <ralsina_> and bzr server is down. Come on!
[10:04] <alecu> mandel, "not yet landed": we don't worry about landing them today, we land them during the week, or next week even.
[10:04] <mandel> alecu, ok
[10:05] <alecu> mandel, I'm talking about ssl support branches, right? :-)
[10:05] <mandel> alecu, yes :)
[10:06] <alecu> ralsina_, is it april 1st already?
[10:06] <ralsina_> alecu: not yet!
[10:06] <alecu> ralsina_, I mean, because of LP being down... it sounded like a really annoying joke
[10:07] <ralsina_> alecu: it's back up, so it was probably an update
[10:08] <alecu> great
[10:09] <karni> Good morning #ubuntuone!
[10:09] <alecu> ralsina_, it seems more than a few things are broken: "apt-get update" is unable to reach http://archive.ubuntu.com/
[10:10] <alecu> hey there, karni!
[10:10] <ralsina_> alecu: fun!
[10:10] <karni> hola alecu !
[10:10] <ralsina_> mandel: when IRLing your brach https://pastebin.canonical.com/62534/
[10:11] <ralsina_> mandel: but then again, that build has sommay things layered it is a club sandwich
[10:11] <mandel> ralsina_, hm.. let me do IRL on windows, one sec I'm trying to get my head around the proxy cache..
[10:11] <ralsina_> mandel: ok, thanks, but check my email, doing a IRL on windows means merging like 6 branches
[10:18] <ralsina_> Ok, I am off to dress kid, send him to be educated, and have breakfast. Should be back in ... 72 minutes, give or take
[10:21] <alecu> crap! archive.ubuntu.com works fine. It's my P vm that's working awfully.
[10:23] <alecu> mandel, because of what ralsina_ says (merging 6 branches to do IRL), I would say that we should not worry about windows today, and instead focus on fixing the issue on linux.
[10:24] <mandel> alecu, that is what I'm doing atm
[10:25] <alecu> mandel, do you want me to take a look at that too? sounds like a big issue
[10:25] <mandel> alecu, let me try on more hack and I will, I want to try a number of different ideas I have been gathering over the weekend
[10:26] <alecu> mandel, do you have a .txt list of those? or let's mumble after your hack.
[10:26] <mandel> alecu, mumble after the hack if you don't mind
[10:31] <mandel> alecu, mumble?
[10:33] <alecu> mandel, coming
[10:39] <mandel> alecu, you sound like a rapper..
[10:39] <mandel> alecu,  a bad one, muy entrecortador
[10:51] <alecu> mandel, https://code.launchpad.net/~alecu/ubuntuone-client/proxy-tunnel-auth/+merge/97763
[11:12] <gatox> good morning!
[11:13] <mandel> alecu, PYTHONPATH=. ./bin/ubuntu-sso-login-gtk  --tc_url http://www.google.com --app_name test
[11:13] <mandel> gatox, morning!
[11:14] <gatox> mandel, hi! o/
[11:14] <mandel> alecu, use a real tc_url if you want :)
[11:23]  * mandel coffee 
[11:38]  * mandel back
[12:04]  * mandel rebooting due to updates
[12:06] <mandel> alecu, did they t&c test work?
[12:25] <ralsina_> good morning gatox!
[12:25] <gatox> ralsina_, hi!
[12:25] <gatox> ralsina_, how are you?
[12:25] <ralsina_> gatox: full of energy and high spirits
[12:25] <ralsina_> gatox: it feels like speed + vodka
[12:25] <gatox> ralsina_, jeje nice way to start the week!
[12:26] <ralsina_> gatox: did you see my first message about windows? It seems I get disabled login buttons on windows
[12:26] <gatox> ralsina_, yes..... i read it..... do you want me to take a look at that?? i'm kind of block with a tests right now anyway
[12:27] <ralsina_> gatox: if you are blocked, yes please
[12:27] <gatox> ralsina_, yap!.... on it then!
[12:27] <ralsina_> gatox: remember that right now the only way to get that far is to usemy doing-windows branch and build exes
[12:27] <gatox> ralsina_, woww....... no way to test it from sourceS?
[12:27] <ralsina_> gatox: or maybe you can just run the scripts for the sso ui and the same thing happens, I have not tested that
[12:28]  * ralsina_ checks
[12:29] <ralsina_> gatox: yes, same thing happens from source
[12:29] <gatox> ralsina_, ok great..... i'll test it from sources
[12:40] <nessita> hello everyone!
[12:40] <gatox> nessita, hi
[12:40] <nessita> I'm sorry I'm late, I set the alarm clock to the wrong time :-/
[12:44] <ralsina_> hello nessita!
[12:44] <nessita> hola ralsina_!
[12:47] <mandel> lunch time for me
[12:49] <ralsina> alecu, nessita, gatox who needs reviews?
[12:49] <gatox> ralsina, not me
[12:49] <nessita> ralsina: me not yet... I guess we need to define what to do with cloud-to-computer
[12:49] <ralsina> nessita: I think just use the real Folders panel
[12:49] <nessita> ralsina: your last comment is confusing me a bit, perhaps you meant "if the user has 'non subcribed' folders?
[12:49] <ralsina> nessita: not the remote-only version
[12:49] <ralsina> nessita: no, "no suscribed folders"
[12:50] <ralsina> nessita: the idea is, this only needs to happen for new users
[12:50] <ralsina> nessita: so, a user that already suscribed a UDF is not new
[12:50] <ralsina> nessita: but leaving that aside, just switching to the full folders list is enough for this release, I think
[12:50] <nessita> ralsina: well, my understanding is that we need to show the wizard every time the user adds the device
[12:51] <ralsina> nessita: ok, then we switch to the full folders list, and ask for UX input on that. Sounds good?
[12:51] <ralsina> nessita: it's just that having the folder you just clicked on "disappear" is scary
[12:52] <nessita> ralsina: I'm still not sure we're talking about the same... I would like first to understand what you meant and then go to design with something  unified :-)
[12:52] <ralsina> nessita: sure, quick mumble?
[12:53] <nessita> ralsina: oh, I'm too sleepy, I just jumped from  the bed (PICNIC while setting the alarm clok - the "pm" is too small :-/)
[12:53] <ralsina> nessita: ok, when you had a few mates :-)
[12:53] <nessita> ralsina: so, may I ask you a few questions about what confuses me?
[12:53] <ralsina> nessita: sure
[12:54] <nessita> ralsina: so, why do youy say the wizard is supposed to be shown only for "new" users? on windows is shown every time there is no credentials (which can be after an existing user removed his device). I matched that behavior in linux
[12:55] <ralsina> nessita: agreed. Actually, it was supposed to check for things like whether you had a ~/Ubuntu One but we never got around to that
[12:55] <ralsina> nessita: so, you are right, there
[12:56] <nessita> ralsina: ok, no thaqt sleepy then :-P. So, given that, I'm +1 to show the current folders list (*without* the explore button)
[12:56] <ralsina> nessita: right
[12:56] <nessita> great, thanks
[12:56] <ralsina> cool then, with that, you have a +1 fromme
[12:56] <nessita> will make that chnage asap so while that lands, I review for you
[12:56] <ralsina> that also makes a bug from thisfred on friday invalid
[12:57] <thisfred> tss
[12:57] <ralsina> the one about having "header" items with nothing under them
[12:57] <nessita> gatox: holas! question, are you working on the translations for u1cp?
[12:58] <thisfred> ralsina: so we will always show all the folders? I think that's much
[12:58] <thisfred> better
[12:58] <ralsina> thisfred: yep
[12:58] <thisfred> cool
[12:58] <thisfred> my bug's still valid,  this will just fix it :P
[12:59] <gatox> nessita, no, that issue was actually the same as sso.... the first screen is just sso stuff.... i'm working in the windows problem that ralsina mention in his branch
[12:59] <ralsina> thisfred: although it will still "flicker" a bit when you check/uncheck one, which I need to fix someday
[12:59] <thisfred> yeah, but that's minor
[12:59] <ralsina> thisfred: yeah
[12:59] <nessita> ralsina: what I know is missing but will do in another branch for the release after this, as bug fix, is to skip the page if there are no volumes to show, since actually I want to use this to wait for syncdaemon's volumes list: https://code.launchpad.net/~brian.curtin/ubuntuone-client/851810-notify-on-volumes
[12:59] <ralsina> thisfred: we have much bigger fish to fry :-)
[12:59] <nessita> gatox: so you 100% sure there are no missing translations on controlpanel? :-)
[12:59] <ralsina> nessita: +1 on that, that will help the UX a lot
[13:00] <thisfred> Whale fry!
[13:00] <nessita> ralsina: yes, but need brian to actually fix the things I asked there ;-)
[13:00] <ralsina> thisfred: with veggies (I am on a diet :-)
[13:00] <ralsina> nessita: after this release ;-)
[13:00] <ralsina> nessita: I need brian al for me today ;-)
[13:00] <nessita> ack
[13:01] <thisfred> ralsina: steamed whale for you then
[13:01] <ralsina> thisfred: you know, whale is technically a fish for lent-related purposes
[13:01] <gatox> nessita, well.... if you put it like this.... i think that is a trap jeeje..... but at least for this screenshot: https://launchpadlibrarian.net/96191627/ubuntuone-controlpanel-qt.png you can see that all the non-translated things are from sso.... i didn't check the other files.... i could do that if you want
[13:01] <thisfred> ralsina: clever catholics
[13:02] <ralsina> thisfred: panda, the other vegetable
[13:02] <thisfred> ha!
[13:02] <gatox> nessita, also..... i'm really stuck with the tests for main for the unicode branch..... it keeps giving me a seg fault (the asserts are ok.... but it through a seg fault)
[13:03] <ralsina> gatox: what branch? I can give you a hand there
[13:03] <ralsina> let's all try to help each other instead of tupac-amaru-izing nessita today :-)
[13:03] <gatox> ralsina, let me push the changes
[13:05] <nessita> gatox: pinging ralsina for that specific thing is the best decision ever :-D
[13:05] <nessita> gatox: and yes please to check the other files
[13:06] <gatox> nessita, ok..... as soon as i finish with the windows thing, i'll check those files
[13:06] <ralsina> unicode is my brêäd and bùttër
[13:06] <nessita> gatox: ack
[13:07] <gatox> ralsina, this branch..... after the reviews, i talk to nessita and we agree to decode all the args: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/unicode-args/+merge/97936
[13:07] <ralsina> gatox: ack
[13:07] <gatox> ralsina, so i decode everything in the main before sending that data to the ui.... but i having problems to test that
[13:08] <ralsina> gatox: so you have a segfaulting test?
[13:08] <gatox> yep
[13:08] <gatox> test_main
[13:08] <ralsina> gatox: ack, will get back to you as soon as I figure something out
[13:08] <gatox> ralsina, PYTHONPATH=. u1trial --gui --reactor=qt4 ubuntu_sso/qt/tests/test_main.py
[13:09] <ralsina> gatox: I am again roberto, the one that has windows abierto. But I will try it! :-)
[13:09] <gatox> ralsina, ahhhhhhhh jeje
[13:10] <gatox> ralsina, if you want, paste the changes and i can try it here
[13:10] <ralsina> gatox: ack
[13:10] <ralsina> gatox: no segfault here!
[13:10] <ralsina> gatox: so, YOU FIXED IT! ;-)
[13:10] <gatox> ralsina, cuack..... the test execute and say pass?
[13:10] <ralsina> gatox: yep
[13:11] <ralsina> then u1trial gets stuck
[13:11] <ralsina> gatox: https://pastebin.canonical.com/62540/
[13:12] <gatox> ralsina, i see this: http://paste.ubuntu.com/890580/
[13:12] <ralsina> gatox: one f my branches fixes this test!
[13:12] <ralsina> gatox: let me find it
[13:13] <ralsina> gatox: https://code.launchpad.net/~ralsina/ubuntu-sso-client/ubuntu-font-in-windows/+merge/98010
[13:14] <ralsina> gatox: merge manually the test fixes
[13:14] <gatox> ralsina, are you going to add more stuff to that branch..... because i could review it so it can be merged in trunk.....
[13:15] <gatox> already has 1 approval
[13:15] <ralsina> gatox: no, that branch is good to go
[13:15] <gatox> ralsina, ok..... reviewing...
[13:21] <alecu> ralsina, nessita: any of you knows when are nightlies rebuilt?
[13:21] <nessita> alecu: after each tarmac build
[13:21] <ralsina> alecu: not me
[13:21] <nessita> alecu: after each tarmac land I meant :-)
[13:21] <nessita> alecu: so, checking https://code.launchpad.net/~ubuntuone-hackers/+recipe/client-dailies I can tell builds are failing for client
[13:22] <nessita> alecu: click on any of the build failed and check the error
[13:22] <alecu> nessita, what does it means "Successful build" but stop signs?
[13:22] <nessita> alecu: error on tests.proxy.test_tunnel_server.MainFunctionTestCase.test_on_proxies_enabled_prints_port_and_cookie
[13:23] <nessita> alecu: where do you see that?
[13:23] <nessita> alecu: ah, I understand your question
[13:23] <alecu> nessita, in the page you just pasted...
[13:23] <nessita> let me explain
[13:23] <alecu> ok
[13:24] <nessita> alecu: we have a single "source" package, and from that we built several binary packages. The outer success means the *source* apckage built ok, but the "stop sign" means that both binary packages failed
[13:24] <alecu> ok, great.
[13:24] <alecu> nessita, so: from dbus.mainloop.qt import DBusQtMainLoop
[13:24] <alecu> exceptions.ImportError: No module named qt
[13:25] <nessita> alecu: seems like python-qt4-dbus is a missing dep?
[13:25] <alecu> nessita, right
[13:25] <nessita> alecu: is that code not in the default install of SD (need to check that)
[13:25] <alecu> nessita, I added that to use the keyring from the tunnel
[13:25] <gatox> ralsina, +1 to this one: https://code.launchpad.net/~ralsina/ubuntu-sso-client/ubuntu-font-in-windows/+merge/98010
[13:25] <nessita> alecu: where was that code added?
[13:25] <ralsina> gatox: awesome!
[13:26] <alecu> nessita, rev 1211
[13:26] <nessita> alecu: which file, I mean :-)
[13:26] <alecu> nessita, and that's why 1210 is the last package available on nightlies!
[13:26] <alecu> nessita, ubuntuone/proxy/tunnel_server.py
[13:27] <alecu> nessita, so, it would be a dependency for the proxy tunnel binary
[13:27] <nessita> alecu: did someone already built the separated binary proxy  package?
[13:28] <alecu> nessita, very good question.
[13:28] <alecu> nessita, I can't answer it!
[13:28] <alecu> nessita, let's ask ralsina.
[13:29] <nessita> alecu: if you give me the complete list of dependencies for that python package, I can make the binary package build
[13:29] <ralsina> ?
[13:30]  * ralsina has clue -4 about packaging
[13:30] <alecu> ralsina, that's +1 than what I have.
[13:31] <ralsina> alecu: having played munchkin with you, you have a "takes all packages from the enemy" card.
[13:31] <nessita> alecu: just check on the nightlies packaging branch and there is no binary for proxy
[13:31] <nessita> alecu: so, I can add it but need the full dependency list... could you give me that? (alphabetically ordered please)
[13:32] <ralsina> nessita: should that be something other than "ubuntuone-client" and "python-dbus-qt4" ?
[13:33] <nessita> ralsina: yes, every module that the python package depend on. From the top of my head, python-twisted-core is needed
[13:33] <alecu> nessita, I think the dependencies would be: python-qt4, python-qt4-dbus, ubuntuone-client
[13:33] <nessita> alecu: what about all the twisted deps?
[13:33] <nessita> do not assume you get twisted from u1client
[13:33] <ralsina> nessita: we don't need to support installing it withut u1-client
[13:33] <alecu> nessita, they would be pulled by ubuntuone-client
[13:34] <alecu> nessita, why not twisted from u1client?
[13:34] <nessita> ralsina: u1-client definitely will be a dep
[13:34] <nessita> alecu: because that's how packages deps are listed
[13:34] <alecu> nessita, how do I find the list of packages that u1client depends on?
[13:34] <nessita> alecu: preferred and recommended and appreciated way is not to assume deps from another package you depend on
[13:35] <alecu> makes sense
[13:35] <nessita> alecu: I would prefer that you grep over imports in proxy package and build the list from there, since u1client list may be incomplete
[13:35] <nessita> packaging is not easy :-)
[13:35] <dobey> huh
[13:36] <nessita> hola dobey
[13:36] <dobey> hi
[13:36] <urbanape> morning, folks
[13:46] <ralsina> morning urbanape!
[13:46] <urbanape> heya
[13:46] <alecu> nessita, http://pastebin.ubuntu.com/890617/
[13:47] <nessita> alecu: thanks! no python-twisted-web? (did not check the code)
[13:47] <alecu> nessita, only to run the tests
[13:47] <nessita> alecu: perfect, thanks, will let you know
[13:47] <alecu> nessita, but not in the modules that should go in the package.
[13:47] <nessita> ack
[13:47] <alecu> nessita, great. Let me know if I can help with any other thing.
[13:47] <nessita> alecu: sure
[13:48] <alecu> nessita, one thing: we are importing some modules that may have some other dependency.
[13:48] <nessita> alecu: such as?
[13:48] <alecu> nessita, say: from ubuntu_sso.keyring import Keyring
[13:48] <alecu> nessita, and that module in turn *may* use python-dbus.
[13:49] <nessita> alecu: ah, that's ok: the dep in *our* proxy code is python-ubuntu-sso-client
[13:49] <alecu> perfect.
[13:49] <nessita> is "hidden" for us the deps that python-ubuntu-sso-client requires
[13:49] <alecu> I guessed so, but wanted to check anyway :-)
[13:49] <nessita> thanks :-)
[13:49] <ralsina> gatox: my sso branch is now merged, so your segfault should have gone away (please check!)
[13:50] <gatox> ralsina, ok, thanks..... i'm looking at the problem that you have in windows.... is the same that was happening in some linux... and this is really weird... the good part...... we can now reproduce the bug
[13:50] <gatox> so i'm looking at that
[13:51] <nessita> dobey: any reason not to have python-twisted-core as a u1client build-depends?
[13:51] <nessita> (is not listed ATM)
[13:53] <dobey> nessita: i guess it's not there because we're pulling it in via ubuntuone-dev-tools
[13:53] <nessita> dobey: ok, adding it then to be explicit
[13:53] <urbanape> briancurtin: ping
[13:53] <briancurtin> urbanape: what's up
[13:54] <urbanape> hey a, can I steal some of your time this afternoon?
[13:54] <ralsina> hi briancurtin!
[13:54] <briancurtin> urbanape: depends on how far i get with this windows release. currently working through some of ralsina's weekened branches
[13:54] <urbanape> kk
[13:54] <briancurtin> hey ralsina
[13:55] <ralsina> briancurtin: please check them out, they are full of sunday goodness
[13:55] <ralsina> as in, I did two while eating a diet sundae and everything
[13:55] <briancurtin> haha
[13:56] <nessita> dobey: so, once we have the python-ubuntuone-client-proxy bin package, shall python-ubuntuone-client suggests it?
[13:57] <dobey> why is it python-ubuntuone-client-proxy?
[13:57] <nessita> dobey: how would you like to name it?
[13:57] <dobey> it should be ubuntuone-client-proxy i think
[13:57] <dobey> as it includes the tunnel script
[13:57] <nessita> dobey: why? (interested in the convention and rationale)
[13:58] <dobey> and it's not a "python library" really
[13:58] <nessita> ok, makes sense, changing
[13:58] <nessita> dobey: so, ubuntuone-client should suggests it then?
[13:59] <dobey> yes, only Suggests: at most. it can't Recommends or Depends it
[13:59] <nessita> right
[14:01] <ralsina> nessita and everyone else: sorry abut pep and lint issues in my branches, but it's hard to get that done right on windows :-(
[14:01] <nessita> ralsina: yes, I know, is ok
[14:02] <nessita> alecu: so, what should the -proxy package ship? the binary ubuntuone-proxy-tunnel, and the python package under ubuntuone/proxy/... anything else?
[14:03] <nessita> alecu: is there any other needed file, man page, dbus service file, icon, desktop file, etc? (just doing a checklist)
[14:03] <alecu> nessita, I think we don't need any other file.
[14:03] <nessita> alecu: great, thanks
[14:04] <alecu> nessita, we don't have a manpage, and anyway it's an internal binary, not supposed to be called by users.
[14:04] <nessita> yes
[14:04] <alecu> nessita, and no dbus service file, icon nor desktop file.
[14:04] <nessita> perfect
[14:10] <briancurtin> launchpad's working nicely today. only on my 7th try to "setup.py fetch" and get sources
[14:10] <ralsina> briancurtin: ugh
[14:11] <ralsina> briancurtin: I had a bit more luck earlier. But it's mostly bzr for windows that sucks :-/
[14:11] <ralsina> briancurtin: you getting permission denied errors?
[14:12] <briancurtin> ralsina: bzr error 3, whatever that means
[14:12] <briancurtin> even opening branches to view in the browser is hit-or-miss this morning
[14:12] <ralsina> briancurtin: ok, never saw that one. It was actually down this morning, too
[14:12] <briancurtin> ah
[14:13] <nessita> ralsina: https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/cloud-to-computer-page/+merge/97990 updated and irl tested
[14:14] <ralsina> nessita: yuju! Will review right away!
[14:15] <nessita> dobey: can you please review this diff? http://pastebin.ubuntu.com/890669/
[14:15] <nessita> (for u1client packaging branch)
[14:15] <ralsina> nessita: just noticed there is no way to test that on windows, though, so just code review from me
[14:15] <nessita> ralsina: you could run it IRL, I think...
[14:16] <ralsina> nessita: no, because on windows you can't login right now
[14:16] <nessita> dobey: still pending to write the changelog though :-)
[14:16] <nessita> ralsina: is that on every windows or just yours? (curious)
[14:16] <alecu> mandel, ping
[14:17] <ralsina> nessita: all windows, reported it on sunday, diego is working on it
[14:17] <nessita> ralsina: interesting :-/.
[14:17] <ralsina> nessita: the sso ui ever enables the login button
[14:17] <ralsina> s/ever/never/
[14:17] <nessita> ralsina: yeah, read the email... and also saw some reports on linux as well
[14:17] <ralsina> nessita: so, if the issue is the same, hey, windows helps us fix it :-)
[14:18] <nessita> yes
[14:18] <ralsina> nessita: gatox has reproduced it on windows
[14:18] <nessita> nice
[14:18] <nessita> will make some mate to see if I can finish waking me up :-d
[14:18] <dobey> ugh
[14:18]  * nessita slaps herself
[14:18] <gatox> ralsina, nessita yes.... there is a problem with the credentials and the backend is returning none.... i'm looking at the logs right now
[14:18] <nessita> cd ..
[14:18] <nessita> oops
[14:18] <dobey> i really don't like having to list each python package individually
[14:19] <gatox> nessita, ralsina i'll let you know as soon as i have something
[14:19] <dobey> nessita: also, the paths are wrong
[14:19] <nessita> dobey: any suggestion how to list everything but "proxy"?
[14:19] <dobey> nessita: it would be ubuntuone-client/ubuntuone/foo
[14:19] <dobey> nessita: no. i don't think dpkg has an "everything but the things that are in other binary packages" thing
[14:19] <nessita> dobey: you're right, will change. Any solution to not to list everything?
[14:19] <nessita> ok
[14:20] <nessita> dobey: the breaks-replaces is fine, no?
[14:22] <dobey> nessita: should be. i'd probably do it as << ${binary:Version} or source:Version (whichever one works there)
[14:23] <nessita> dobey: any reason for that? last time we fixed something like this you suggested using the exact version, to avoid breaking/replacing in "the future"
[14:24] <mandel> alecu, pong
[14:24] <mandel> alecu, sorry took me longer to have lunch, had to deal with the family, is father day here..
[14:24]  * mandel hates family things..
[14:25] <dobey> nessita: because it's nightlies and sticking the +r43432 in there feels weird
[14:25] <alecu> mandel, roberto reported two small issues in this branch: https://code.launchpad.net/~mandel/ubuntu-sso-client/use-qt4-webclient-backend/+merge/97934
[14:25] <nessita> dobey: ok then, as you wish. Will change
[14:25] <mandel> alecu, yes, I know fixing in a few minx
[14:25] <mandel> alecu, mins*
[14:25] <alecu> ok
[14:25] <dobey> nessita: for proper stable packages, i'd use specific versions, but for nightlies, i'd generally prefer a minimum of that, in favor of automation
[14:26] <nessita> dobey: duly noted
[14:26] <nessita> changed to (<< ${binary:Version})
[14:26] <dobey> cool
[14:27] <dobey> nessita: we use zope.interface in u1client?
[14:27] <nessita> dobey: yes, on trunk/ubuntuone/syncdaemon/interfaces.py
[14:28] <nessita> dobey: and in the tests now
[14:33] <ralsina> alecu: I answered your needsinfo in https://code.launchpad.net/~ralsina/ubuntuone-client/find_exes/+merge/98107 as "you are right" ;-)
[14:34] <alecu> ralsina, ok, I'll re-review.
[14:34] <mandel> alecu, I think that setting QNetworkRequest::AuthenticationReuseAttribute to QNetworkRequest::Manual might do the trick
[14:35] <alecu> mandel, awesome
[14:36] <ralsina> nessita: ay reason to create are_you_sure.ui instead of using QMessageBox.question()?
[14:38] <alecu> mandel, that means that each webcall would be done once, the proxy will return a 407, and then it will be retried with the credentials.
[14:38] <mandel> alecu, in theory.. I'm writing some tests to make sure it does what the can says..
[14:39] <alecu> mandel, but it sounds like a reasonable compromise until the credentials cached by NAM are truly erased on failure.
[14:39] <nessita> ralsina: you tell me, you added that in the installer :-) (just copied it)
[14:39] <ralsina> nessita: really? Whoa, was I in a dumb day! :-)
[14:39] <ralsina> nessita: ok, I'll file a bug for it and give this one a +1
[14:39] <nessita> ralsina: I think you needed to customize the button texts...
[14:39] <mandel> alecu, it is just good enough, specially when we can catch them and  then tell the requests after the first successful one to use it in automatic mode
[14:40] <nessita> ralsina: and I m not sure you can do that with a yes no dialog?
[14:40] <ralsina> nessita: could be. Whatever, it works ;-)
[14:40] <mandel> alecu, since this in an attribute at the request level and not the network access manager
[14:40] <alecu> mandel, I think we should not care about re-setting it to automatic.
[14:42] <gatox> nessita, ralsina yes.... it was because of the buttons-text
[14:43] <ralsina> gatox: thanks!
[14:43] <mandel> alecu, that could in an optimzation done later for sure
[14:45] <ralsina> briancurtin: fllowed your suggestion on find-exes, please rereview
[14:46] <urbanape> ralsina: if you're looking for them, I believe they all live in Texas.
[14:46] <urbanape> At least, that's how the song goes.
[14:46] <ralsina> urbanape: ?
[14:46] <urbanape> “All my exes live in Texas”
[14:47] <ralsina> urbanape: groovesharking it :-)
[14:48] <ralsina> urbanape: omfg is that a crappy song (sorry country fans :-)
[14:48] <urbanape> hahah
[14:48] <ralsina> I wonder how his san antonio ex is called. Antonio?
[14:48] <alecu> that's why I lay may hat in Tennessee
[14:49]  * alecu heard that song a lot while playing San Andreas.
[14:51] <alecu> mandel, I'm not finding an easy way to enable socks
[14:51] <mandel> alecu, in webclient?
[14:51] <alecu> mandel, I'll start with the SSL-disable branch if you think it's fine.
[14:51] <alecu> mandel, in webclient, yes.
[14:51] <mandel> alecu, if you have nothing on your plate, sure
[14:52] <alecu> mandel, also, it seems that this test in webclient would not work right: test_https_request
[14:52] <alecu> mandel, since self.assertFailure needs to be yielded
[14:53] <alecu> or failUnlessFailre, as it's called in the test.
[14:54] <alecu> mandel, also, the indenting of that method suggests that a yield was there at some point
[14:54] <mandel> alecu, really.. I cannot recall, but I'm sure it was not
[14:54] <mandel> alecu, as in, I don't remember it like that, but ypu are right, there is a yield missing
[14:55] <alecu> mandel, if I add a yield the indenting of the parentesis matches the parameters on the line below :-)
[14:56] <alecu> mandel, anyway, I think I've found the smallest patch that can disable ssl
[14:56] <mandel> alecu, not connect to the signal?
[14:56] <alecu> mandel, yup
[14:56] <mandel> alecu, that is the one I had in my head
[14:57] <alecu> great
[14:57] <mandel> alecu, and if we wanted to re-enable it would be very easy..
[14:57] <mandel> alecu, but the patch also has to skip the ssl tests
[14:57] <alecu> right
[14:58] <alecu> mandel, I'll handle the ssl branch.
[14:58] <mandel> alecu, thx!
[14:58] <alecu> mandel, but please, *please* fix the qt-enable branch so it finally lands.
[14:58] <mandel> alecu, ok, I'll move to that and will get back to the cache bug
[14:58]  * mandel switches..
[14:59] <nessita> standup in 1 minutes crowd
[15:01] <ralsina> me
[15:01] <briancurtin> me
[15:01] <urbanape> me
[15:02] <alecu> mandel, btw: it looks like the handler for ssl errors cannot be an inlineCallbacks either. similar issue.
[15:02] <mandel> me
[15:02] <alecu> me
[15:02] <mandel> alecu, hm.. stupid qt signals..
[15:03] <nessita> me
[15:03] <ralsina> dobey: standup?
[15:03] <dobey> oh
[15:03] <dobey> me
[15:03] <ralsina> DONE: proposed some branches to fix windows peculiarities, helped gatox with his segfault (I hope), did reviews, some 1:1s, lots'o'stuff TODO: help around, keep on pushing the windows side of things, tech leads call BLOCKED: no NEXT: briancurtin
[15:03] <briancurtin> DONE: lots of debugging, got most things running in some form for the windows release
[15:03] <briancurtin> TODO: if bzr ever starts working properly i can get this installer environment up to date with ralsina's branches then continue on
[15:03] <briancurtin> BLOCKED: bzr/launchpad is awful this morning
[15:03] <briancurtin> NEXT: urbanape
[15:03] <ralsina> thisfred: standup you too
[15:03] <gatox> me
[15:04] <urbanape> DONE: More tests for ubuntu-sso-client
[15:04] <urbanape> TODO: Finish it up, propose
[15:04] <urbanape> BLOCK: None
[15:04] <urbanape> briancurtin: You're up
[15:04] <nessita> urbanape, briancurtin: you're in a endless loop?
[15:04] <nessita> :-)
[15:04] <alecu> DONE: mad rush to fix and land pending sso branches
[15:04] <alecu> TODO: reach the deadline alive
[15:04] <alecu> BLOCKED: no
[15:04] <alecu> NEXT: nessita\
[15:04] <briancurtin> i think it's mandel
[15:04] <nessita> NEXT: alecu
[15:04] <nessita> DONE: proposed cloud-to-computer u1cp branch, reviews
[15:04] <nessita> TODO: reviews, land the aforementioned branch, propose branch for computer-to-clound u1cp wizard page, help alecu building a new binary package for u1client-proxy
[15:04] <nessita> BLOCKED: nopes
[15:04] <nessita> NEXT: dobey
[15:05] <dobey> λ DONE: bug #934206, reviews, askubuntu
[15:05] <dobey> λ TODO: bug #932103, bug #953119
[15:05] <dobey> λ BLCK: none.
[15:05] <nessita> alecu: oops, I think you skip mandel
[15:05] <urbanape> nessita: endless
[15:05] <alecu> nessita, you skipped him too!
[15:05] <dobey> gatox: go
[15:05] <gatox> DONE:
[15:05] <gatox> Landed a branch that fix translation problems worked on a branch for unicode issues (blocked with the tests... still segfault with ralsina's branch), working on the issue about the none backend in some linux and windows.
[15:05] <gatox> TODO:
[15:05] <gatox> Finish with the backend thing, and check translations in u1cp.
[15:05] <gatox> BLOCKED:
[15:05] <gatox> Just the test thingy in the u1-sso-unicode branch
[15:05] <nessita> alecu: I was just copying you :-P
[15:05] <mandel> I feel left out
[15:05] <nessita> mandel: go!
[15:05] <mandel> DONE: Reviews, look at cache creds issues in the QNetworkAccessManager.
[15:05] <alecu> mandel, come!
[15:05] <mandel> TODO: fix the qt webclientimplementation to deal with bad creds.
[15:05] <mandel> BLOCKED: is a hard problem to solve because is a bug from qt itself... :(
[15:05] <mandel> next, alecu
[15:05] <mandel> :P
[15:06] <mandel> alecu, now is your turn hehe
[15:06] <alecu> thisfred, go
[15:06] <nessita> this has been a messy standup :-/ (I guess this shows the state of our minds!)
[15:06] <thisfred> DONE: reviews, hunting segfault
[15:07] <thisfred> TODO: hunt and kill u1db segfault
[15:07] <thisfred> ALSO: will
[15:07] <ralsina> nessita: +1 on cloud-to-computer
[15:07] <thisfred> be on holiday wed, thu, fri
[15:07] <nessita> ralsina: nice!
[15:08] <ralsina> everyone TODO: not kill yourselves, ok? Let's not fall in the same thing as with windows.
[15:08] <ralsina> work hard, but not TOO hard
[15:08] <urbanape> I promise not to kill myself this week.
[15:08] <ralsina> urbanape: thanks!
[15:09] <ralsina> urbanape: but guys with laser vision always tend to have tortured souls. I learned that from DC comics.
[15:09] <urbanape> I also promise not to level a city this week.
[15:10] <mandel> alecu, I just had an EVIL idea!! regarding the signals etc..
[15:10] <mandel> alecu, also, updated the use_qt branch
[15:11] <dobey> ralsina: dc or marvel?
[15:11] <alecu> mandel, and you've pushed it too, right?
[15:11] <briancurtin> ralsina: find_exes looks good...im just waiting on bzr/launchpad to stop breaking so i can actually run with it
[15:11] <mandel> alecu, runnings tests, I'll push when done
[15:12] <alecu> dobey, both supes and cyclops.
[15:12] <ralsina> briancurtin: he
[15:12] <ralsina> briancurtin: I sometime end doig all the fetching manually :-/
[15:12] <ralsina> dobey: superman is a basket case
[15:13] <Chipaca> I saw somebody in here talking about hasattr the other day
[15:13] <ralsina> Chipaca: it's gone!
[15:13] <Chipaca> hasattr is *less* evil now, but "getattr(thing, attr, flag) is not flag" is still better
[15:14] <Chipaca> hasattr now is "only" try: getattr except StandardError: return false else: return true
[15:14] <Chipaca> before it was a raw except
[15:18] <mandel> alecu, ralsina FYI updated: https://code.launchpad.net/~mandel/ubuntu-sso-client/use-qt4-webclient-backend/+merge/97934
[15:21] <ralsina> mandel: ack, checking!
[15:23] <ralsina> nessita, dobey: looks like we need to update sso in tarmac?
[15:23] <dobey> ?
[15:23] <nessita> ralsina: why?
[15:23] <ralsina> nessita: your branch failed to merge because of a sso import
[15:23] <nessita> ralsina: it gets automatically updated
[15:23] <nessita> ralsina: let me look
[15:24] <dobey> what branch?
[15:24] <nessita> ralsina: I guess that's a missing build dep
[15:24] <nessita> dobey: https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/cloud-to-computer-page/+merge/97990
[15:25] <dobey> it's automatic i think, but not guaranteed to be up-to-date when you necessarily need it
[15:25] <dobey> also, it does depend on nightlies building successfully and timely :)
[15:26] <dobey> hrmm
[15:26] <nessita> dobey: we're missing a build dep
[15:26] <nessita> ubuntu-sso-client-qt
[15:26] <dobey> right
[15:26] <nessita> will add now
[15:30] <nessita> ralsina: after nightlies re-build will re-approve
[15:30] <ralsina> nessita: cool, thanks
[15:33] <ralsina> mandel: +1 on use-qt4-webclient-backend
[15:33] <mandel> ralsina, superb!
[15:36]  * gatox quick lunch and keep fighting with the backend thing....
[15:38] <dobey> hrmm. if we care about the hasattr thing at all, shouldn't we make tests fail if it exists?
[15:39] <ralsina> mandel: is there a reason why this is not approved? https://code.launchpad.net/~mandel/ubuntu-sso-client/merge-code/+merge/97627
[15:39] <mandel> ralsina, I missed it
[15:39] <mandel> ralsina, approved now :)
[15:39] <ralsina> mandel: cool :-)
[15:42] <ralsina> dobey: +1 on highlander
[15:48] <Chipaca> dobey: I think it's more a styling thing; pylint should probably flag it, if we used pylint :)
[15:49] <Chipaca> dobey: it's try/except; my problem with it is that that is what it is. It is documented as such, and often it's fine
[15:49] <nessita> alecu: -proxy binary built! https://code.launchpad.net/~ubuntuone-hackers/+recipe/client-dailies
[15:49] <dobey> right
[15:50] <dobey> Chipaca: but if our "style" is to not use it; then we should probably fail when it gets used
[15:50] <ralsina> Chipaca: we do use pylint, and it doesn't, and maybe we should make it do it.
[15:50] <dobey> also we should get rid of pylint
[15:50] <dobey> ralsina: we don't use pylint everywhere
[15:50] <ralsina> dobey: or that ;-)
[15:51] <Chipaca> dobey: that sounds more like *work*
[15:51] <Chipaca> dobey: I
[15:51] <ralsina> dobey: yes, but this one was in a place where we do
[15:51] <Chipaca> dobey: I'm generally opposed to the idea of work, as you know
[15:51] <briancurtin> ralsina: your find_exes branch isn't working for me. 1) start u-sso-login.exe, 2) start u1-cp-qt.exe, 3) doesn't find SD
[15:51] <dobey> well, pyflakes is much easier to extend than pylint
[15:51] <ralsina> briancurtin: no, that will not work
[15:51] <dobey> Chipaca: as am i
[15:51] <ralsina> briancurtin: start syncdaemon manually, too
[15:51] <dobey> Chipaca: i much prefer hedonism
[15:51] <ralsina> briancurtin: instructions in one of the emails
[15:51] <briancurtin> ralsina: ok, then it works :)
[15:51] <Chipaca> dobey: indeed
[15:51] <briancurtin> i think i missed the manual part
[15:51] <ralsina> briancurtin: yay :)
[15:51] <Chipaca> dobey: and cake
[15:52] <briancurtin> (as an official step)
[15:52] <ralsina> briancurtin: yes, sd will only start from the installed location. Need to fix that someday
[15:52] <dobey> well, hedonism without cake is just vice.
[15:52] <dobey> or vicing
[15:52] <ralsina> briancurtin: is that the find-exes in ussoc or in u1-client?
[15:53] <briancurtin> i'm running with all of your branches merged
[15:53] <ralsina> briancurtin: ack
[15:53] <ralsina> mandel: is there a chance for a re-review of https://code.launchpad.net/~ralsina/ubuntu-sso-client/find_exes/+merge/98108?
[15:53] <mandel> ralsina, yes, let me finihs the IRL I'm doing
[15:54] <ralsina> mandel: sure, thanks!
[15:56] <ralsina> mandel: https://code.launchpad.net/~mandel/ubuntu-sso-client/merge-code/+merge/97627 was too long there, got a bit of rot (needs the prereq removed)
[15:59] <dobey> alright i should get lunch
[15:59] <dobey> bbiab
[16:13] <mandel> ralsina, ok, we will deal with that branch later, is not a big issues atm
[16:13] <ralsina> mandel: ack
[16:17] <mandel> alecu, ralsina got a min?
[16:17] <alecu> mandel, got
[16:17] <ralsina> mandel: I am on tech leads call, will be fre in about 45'
[16:17] <alecu> mandel, momble?
[16:17] <mandel> alecu, no mumble needed
[16:18] <mandel> alecu, I'm doing my IRL of my evil hack and I get blocked in getting information, any idea on how to debug that?
[16:19] <alecu> mandel, what steps are you following?
[16:20] <mandel> alecu, I got to my branch and do U1_DEBUG=1 PYTHONPATH=. bin/ubuntu-sso-login and on control panel: U1_DEBUG=1 PYTHONPATH=~/Projects/Canonical/ubuntu-sso-client/use-qt/:. bin/ubuntuone-control-panel-qt
[16:21] <alecu> mandel, I assume u1cp is trunk; can you send me again the link to your branch?
[16:21] <alecu> I mean, the sso one.
[16:22] <mandel> alecu, sure, give me a sec and I'll push all the changed and merge with trunk to reduce the diff
[16:22] <alecu> mandel, also: your branch depends on the enable-qt one, right?
[16:22] <mandel> alecu, yes
[16:22] <alecu> well, that has already landed, so yes.
[16:27]  * briancurtin needs coffee by the gallon today, brb/lunch
[16:28] <mandel> alecu, lp:~mandel/ubuntu-sso-client/nam-is-buggy
[16:29] <alecu> mandel, thanks. Please review this one while I take a look: https://code.launchpad.net/~alecu/ubuntu-sso-client/disable-ignore-ssl-option/+merge/98226
[16:29] <alecu> mandel, it's tiny.
[16:30] <mandel> alecu, ok, I changed the order of the QObject, is wrong when calling super you will get an exception.. ups!
[16:30] <alecu> mandel, was that all the problem?
[16:31] <mandel> alecu, no no, it has nothing to do with that, I changed it because I'm stupid
[16:31] <mandel> alecu, merge issues: https://code.launchpad.net/~alecu/ubuntu-sso-client/disable-ignore-ssl-option/+merge/98226
[16:31] <alecu> crap
[16:38] <alecu> mandel, now it's merged with trunk: https://code.launchpad.net/~alecu/ubuntu-sso-client/disable-ignore-ssl-option/+merge/98226
[16:39] <mandel> alecu, is wrong, we should not be listening to proxyAuthenticationRequired, right?
[16:39] <mandel> alecu, AFAIK it is removed in trunk
[16:40] <alecu> mandel, le'crap.
[16:40] <mandel> lol
[16:40] <alecu> mandel, my mind is broken. I should stop waking up at this time :-)
[16:40] <alecu> mandel, I've pushed that branch three times, and every time I made a mistake.
[16:40] <mandel> alecu, take a nap, I'll be here 'til late.. I don't know if that is possible regarding the deadline..
[16:41] <mandel> alecu, but, power nap for the win!
[16:41] <alecu> mandel, I think not: we are 1:20hs from it.
[16:41] <mandel> alecu, la puta..
[16:41] <mandel> ack
[16:43] <nessita> mandel: would you please re-review https://code.launchpad.net/~ralsina/ubuntu-sso-client/find_exes/+merge/98108 ?
[16:43] <mandel> nessita, ralsina on it
[16:44] <mandel> nessita, no tests at the end?
[16:44]  * briancurtin back, mm coffee
[16:46] <nessita> mandel: did you read ralsina's comment?
[16:46] <nessita> mandel: the code is a refactor, covered by the existent tests
[16:46] <mandel> nessita, oh, I skipped that one, then +1
[16:46] <ralsina> mandel: thanks!
[16:49] <nessita> dobey: when you come back, would you help me understand why, after installing latest nightlies for u1client, I'm getting:
[16:49] <nessita> >>> import ubuntuone.syncdaemon
[16:49] <nessita> Traceback (most recent call last):
[16:49] <nessita>   File "<stdin>", line 1, in <module>
[16:49] <nessita> ImportError: No module named syncdaemon
[16:49] <nessita> even though /usr/lib/python2.7/dist-packages/ubuntuone-client/ubuntuone/syncdaemon/ does exist. I'm guessing a .pth issue?
[16:50] <ralsina> nessita: looks like the ubuntu default is all buttons on the right
[16:50] <ralsina> nessita: so let's do that for all platforms
[16:50] <nessita> ack
[16:56] <mandel> nessita, this line has nothing wrong: 2012-03-19 17:56:21,712:712.683916092 - ubuntu_sso.main - DEBUG - shutdown!, ref_count is 1.
[16:56] <mandel> right?
[16:57] <nessita> mandel: right, perfectly fine
[16:58] <mandel> nessita, ok
[16:58] <nessita> mandel: that means the sso service will not shutdown because it has an ongoing request
[16:58] <mandel> ack
[16:58] <alecu> mandel, can you re-review? https://code.launchpad.net/~alecu/ubuntu-sso-client/disable-ignore-ssl-option/+merge/98226
[17:05] <mandel> alecu, +1
[17:12] <nessita> dobey: hum, why there is no ubuntuone.client.pth in u1client source tree?
[17:19] <briancurtin> ralsina: to what degree should SSO work? i see in your mail that it's not complete, but this is what i'm seeing after install http://i.imgur.com/jdfFs.png (can't click Sign In)
[17:20] <briancurtin> (my installer was built with all of the branches you listed in yesterday's email)
[17:20] <ralsina> briancurtin: yes, that's what I ran into, gatox is looking at it
[17:20] <briancurtin> ok cool
[17:20] <gatox> briancurtin, yes, i'm in crazy debug mode :P
[17:21] <briancurtin> gatox: i can't offer much on the GUI stuff, but if you have branches you want me to test out or anything, let me know
[17:21] <gatox> briancurtin, yes thanks..... actually the problem seems to came when we want to get the port for an instance already running..... i'm debugging that
[17:21] <dobey> nessita: oh, you need to include the *.pth in python-ubuntuone-client
[17:22] <dobey> nessita: sorry i didn't catch that one. it was pulled in by the ".../*-packages/*" line before
[17:23] <dobey> nessita: it doesn't need to be in the source tree; the Makefile creates it on install though
[17:30] <nessita> dobey: ah, ok, so:
[17:31] <nessita> will add debian/tmp/usr/lib/python2.*/*-packages/*.pth to python-ubuntuone-client.install\
[17:31] <nessita> dobey: ack?
[17:31] <dobey> yeah
[17:32] <dobey> this is why i hate the splitting of stuff in python packages :)
[17:32] <dobey> also, why i hate having to use the standard python lib dirs for things that don't belong in there
[17:34] <nessita> ok, u1client packaging branch updated, and a new build was requested. Will have lunch now
[17:34] <nessita> dobey: ^
[17:34]  * nessita -> lunch
[17:40] <alecu> thisfred, gatox, ralsina: can I ask any of you guys for a really small review?
[17:40] <alecu> https://code.launchpad.net/~alecu/ubuntu-sso-client/disable-ignore-ssl-option/+merge/98226
[17:40] <ralsina> alecu: sure
[17:40] <gatox> alecu, shoot
[17:40] <ralsina> alecu: I got it
[17:40] <thisfred> alecu: yep
[17:41] <gatox> alecu,  the three of us?
[17:41] <alecu> only one is needed, and ralsina has already volunteered.
[17:41] <gatox> alecu, roger that
[17:41] <dobey> sigh
[17:42] <dobey> my eyes are not well
[17:43] <ralsina> dobey: your fonts are too big.
[17:43] <dobey> my eyelids are too big right now
[17:43] <dobey> thanks to the plant kingdom
[17:43] <ralsina> dobey: unless you are actually feeling bad, in which case previous comment is to be considered rude and insensitive and replaced with "wanna go see a doctor?"
[17:45] <dobey> i think i just need some water for my eyes
[17:47] <ralsina> mandel: I am stumped with ubuntu-sso-proxy-creds-qt I just can't make deferToThread ever return
[17:47] <ralsina> mandel: tried adding qt4reactor, makes no difference
[17:48] <mandel> ralsina, hm.. can I have your branch?
[17:48] <ralsina> mandel: you can start from lp:~ralsina/ubuntu-sso-client/fix-958884 I have done no progress :-(
[17:49] <ralsina> mandel: I am this close to just using blocking calls for the keyring on windows
[17:49] <mandel> ralsina, let me try
[17:49] <ralsina> mandel: thanks, you don't have other more urgent stuff?
[17:50] <mandel> ralsina, I need to rest a little with the auth issues I have.. I nearly got it solved but needs some maceration
[17:50] <briancurtin> ralsina: is there anywhere else we need help with on the windows release?
[17:50] <ralsina> briancurtin: hmmm.... can you test a non-authed proxy?
[17:51] <briancurtin> i don't know anything about the proxy stuff that has been going on so i'd need some test steps, but yes i can do it
[17:51] <ralsina> briancurtin: just get access to a proxy (there are a bunch of simple ones for windows) and see if it works :-)
[17:51] <briancurtin> ralsina: i'm on it
[17:52] <ralsina> alecu, mandel: I have a vague memory that we used to use deferToThread before and it had a hack o windows because twisted's did not work
[17:53] <ralsina> and then we got rid of it everywhere except on keyring/windows.py
[17:54] <mandel> ralsina, give me a sec
[17:54] <alecu> ralsina, did not recall about that really.
[17:55] <alecu> ralsina, in fact, deferToThread still seems to be used in action_queue.py, iirc
[17:55] <alecu> ralsina, are we talking about SD or U1CP or SSO?
[17:55] <ralsina> here's the change I remembered http://bazaar.launchpad.net/~ubuntuone-control-tower/ubuntu-sso-client/trunk/revision/850
[17:55] <nessita> aaaagggggrrrrhhhhhhhhh my branch keeps being rejected! /me cries (I know what the problem is, no worries, but needed to scream a bit)
[17:56] <ralsina> so no, we always used defertothread on windows, and we have always been at war with eurasia
[17:57] <dobey> nessita: with the same qt import issue?
[17:57] <alecu> ralsina, we got rid of the threads there, when we finally got rid of lazr.restfulclient
[17:58] <ralsina> alecu: right
[17:58] <nessita> dobey: no, now is the busted u1client... so I will wait for the new u1client to build and then ask for package updates in tarmac
[17:58] <alecu> ralsina, but I've just checked and we are still using twisted threads in SD
[17:58] <nessita> dobey: since every branch using u1client will fail with the same
[17:58] <ralsina> alecu: yes, but that has no qt :-)
[17:58] <dobey> nessita: ah right, because of the missing pth
[17:58] <alecu> (not deferToThread, but the very related callInThread and callFromThread)
[17:58] <nessita> dobey: yeap
[17:58] <ralsina> alecu: this is in sso in the proxy-creds-qt script
[17:58] <alecu> ralsina, riiight.
[17:58] <alecu> ralsina, ack
[17:59] <mandel> ralsina, can you IRL the following: lp:~mandel/ubuntu-sso-client/fix-958884
[17:59] <ralsina> mandel: sure!
[17:59] <mandel> ralsina, I have not tested, did a quick fix to what I think should be done
[17:59] <ralsina> mandel: ack
[18:00] <mandel> ralsina, let me know if it works (you might need to fix some stuff)
[18:00] <nessita> gatox: is your ussoc unicode branch ready for review?
[18:00] <nessita> gatox: I'm lost with the status of that
[18:00] <ralsina> mandel: no, doesn't work
[18:01] <mandel> ralsina, what does it get?
[18:01] <ralsina> mandel: clicking "save and connect does nothing"
[18:01] <mandel> ralsina, may I see the debug messages?
[18:01] <gatox> nessita, not yet.... the test was still segfaulting.... and i'm with the sso not working in windows/linux issue right now.... some crazy twisted thing it seems
[18:02] <ralsina> mandel: says  Save credentials as for domain 1.2.3.4. and never closes (and creds are not saved) so exact same behaviour as my branch
[18:03] <mandel> ralsina, hm.. deferToThread works, and I know that because we have sued that code from the very very beginning
[18:03] <briancurtin> simple test of the proxy - open up google, see what country it thinks im in. germany!
[18:03] <mandel> ralsina, try to callInThread instead
[18:03] <ralsina> mandel: what can I say?
[18:03] <mandel> ralsina, bubaloo? let me launch my vm in a few mins, I'm nearly done with the proxy issue
[18:05] <ralsina> mandel:  thanks
[18:05] <ralsina> briancurtin: cool, now let's see if you can sync stuff :)
[18:06] <dobey> anyone want to give a second review on https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/highlander/+merge/98214?
[18:06] <briancurtin> ralsina: it'll be a few minutes, i need to run the old installer first so i can get U1 to run and get me SSO creds first
[18:06] <ralsina> briancurtin: ack
[18:11] <ralsina> Fellows, I have to go to a doctot's appointment, will be back in about 80 minutes
[18:11] <ralsina> so, ay IRL testing on windows, briancurtin can do it
[18:16] <alecu> ralsina, mandel: ping
[18:16] <mandel> alecu, catapong!
[18:17] <ralsina> alecu: pong
[18:17] <alecu> ralsina, mandel: deferToThread should not work any longer on sso nor u1cp
[18:17] <mandel> alecu, why?
[18:17] <alecu> since we are not using twisted for those two on linux.
[18:17] <ralsina> alecu: this is on windows
[18:17] <ralsina> alecu: but on a program without a reactor
[18:17] <alecu> there you go!
[18:18] <ralsina> alecu: tried adding qtreactor, still didn't work, but you know my twisted is awful
[18:18] <alecu> ralsina, shall I take a look?
[18:18] <alecu> ralsina, hit me with the branch!
[18:18] <ralsina> alecu: talk with mandel, I am leaving as soon as the taxi arrives :-)
[18:18] <alecu> ack
[18:19] <ralsina> alecu: lp:~ralsina/ubuntu-sso-client/fix-958884 the problem is creds-qt script
[18:21] <alecu> mandel, ralsina: that seems to be the only file in that branch.
[18:21] <alecu> mandel, ralsina: creds-qt, I mean.
[18:27] <alecu> mandel, what's the status of the credentials branch?
[18:29] <mandel> alecu, stuck :( I though I had a hack around it using the QEventLoop but aint working, I can get the signal to block and execute all the deferred stuff in the child eventloop and quit it, but the creds dialog appears too many times..
[18:43] <gatox> nessita, i'm going to finish with the unicode branch right now, and then keep working on the windows-backend thing.... because i'm super block with that one
[18:44] <nessita> gatox: yes please, given that the unicode one is absolutely needed for the release
[18:44] <gatox> nessita, ok
[18:46] <mandel> alecu, I'm running out of ideas.. I'm going to walk the dog, get some fresh air and try a diff approach..
[18:46] <alecu> mandel, ack
[18:52] <dobey>   758 dobey     20   0 1215m 1.2g 1.0g S    0 14.7   0:58.67 gdb
[18:52] <dobey> :(
[18:57] <gatox> nessita, ralsina with trunk i still have: http://paste.ubuntu.com/891082/ are you able to run successfully that test?
[19:01] <briancurtin> ralsina, proxy people: while running with a proxy i just get "file sync in progress..." forever from u1-cp-qt.exe, and then in its console window i get endless amounts of https://pastebin.canonical.com/62570/
[19:03] <roadmr> Hi folks! Is there anywhere I can get the U1 qt-based frontend's source? I branched lp:ubuntuone-client but that doesn't seem to be it
[19:04] <dobey> roadmr: it's ubuntuone-control-panel
[19:05] <roadmr> dobey: ah! great, thanks
[19:06] <roadmr> dobey: one more question, is this all programmed using Python? no C++ for the Qt frontend part?
[19:06] <alecu> roadmr, ubuntuone is 100% C++ free.
[19:06] <dobey> yes it's python
[19:06] <roadmr> awesome, thanks!
[19:10] <alecu> briancurtin, and you do not get the error from the paste at all if running with no proxies?
[19:10] <alecu> briancurtin, I can't think of a reason why on_download_finished would be called differently if going thru a proxy.
[19:10] <briancurtin> alecu: nope. the paste is coming from an proxied XP VM. my Win7 non-proxy box is working nicely
[19:11] <briancurtin> alecu: i can't either, i've been poking around and don't see where the divergence would be
[19:11] <briancurtin> maybe it's another time for super debugger (python tools for visual studio)
[19:11] <gatox> alecu, can you do me a favour please?
[19:11] <alecu> gatox, sure. But depends.
[19:11] <alecu> :-)
[19:12] <gatox> alecu, :P can you run in sso-trunk (updated) this: PYTHONPATH=. u1trial --gui --reactor=qt4 ubuntu_sso/qt/tests/test_main.py
[19:12] <gatox> and let me know if that works for you
[19:12] <gatox> i'm having a segfaul
[19:13] <alecu> gatox, I'm getting that as well :-(
[19:14] <gatox> ralsina, ping! the segfault thing seems to be still broken
[19:14] <gatox> alecu, thanks
[19:14] <alecu> gatox, run-tests seems to work perfectly, though.
[19:14] <gatox> alecu, cuack...... ok....... trying that
[19:15] <alecu> gatox, you can try "./run-tests path/to/test_file.py"
[19:16] <gatox> alecu, thanks!..... that seems to be broken too
[19:16] <alecu> gatox, run-tests passed perfectly here.
[19:16] <alecu> gatox, I'll try updating again.
[19:16] <alecu> gatox, I'm on a P vm dist-upgraded just today.
[19:17] <gatox> alecu, updating now....
[19:22] <gatox> alecu, nop.... i've just updated, and i'm getting this: http://paste.ubuntu.com/891133/
[19:28] <alecu> gatox, not to worry.
[19:28] <gatox> alecu, i'm going to write the test..... and hope it works :P
[19:28] <alecu> gatox, qt tests are not supposed to be run while running the GTK test suit
[19:29] <alecu> gatox, the thing is that when running the GTK test suite the Qt tests are usually excluded
[19:29] <alecu> gatox, but you are forcing it
[19:29] <alecu> gatox, hence the explosion.
[19:29] <alecu> gatox, so...
[19:29] <alecu> gatox, two options:
[19:29] <alecu> gatox, "run-tests" with no parameters
[19:30] <alecu> gatox, or use the right command line to start only your test... let me find it.
[19:31] <alecu> gatox, the fastest way if you only want to run tests for that bit is to comment the first instance of u1trial in run-test
[19:31] <alecu> gatox, but remember to run the whole test suite when you are done.
[19:31] <gatox> alecu, ok, thanks! i'm going to do that
[19:40] <nessita> gatox: u1trial -r qt4 --gui ubuntu_sso/qt/tests/test_main.py
[19:41] <nessita> alecu: ^
[19:41] <gatox> nessita, that is not working for me
[19:41] <briancurtin> does one more person have time for a small review (windows "do not run as root" thing)? https://code.launchpad.net/~brian.curtin/ubuntuone-client/930398-windows-is_root/+merge/96224
[19:41] <nessita> gatox: why not?
[19:41] <gatox> nessita, http://paste.ubuntu.com/891164/
[19:42] <nessita> gatox: yes, is segfaulting for me too
[19:42] <nessita> gatox: that's why the test was skipped
[19:42] <nessita> gatox: but is not related to how you're running the test
[19:43] <nessita> the test itself is segfaulting
[19:43] <gatox> nessita, yes.... what should i do? it makes sense to write the test that is not going to run?
[19:43] <nessita> gatox: no, but didn't we mention we will gonna patch something to make it work?
[19:43] <gatox> the change of the branch is just adding the decode('utf-8')
[19:44] <gatox> nessita, yes, but it doesn't work with that either
[19:44] <nessita> gatox: yes, but semantically is a huge change
[19:44] <nessita> gatox: where is segnfaulting? have you used strade? gdb?
[19:44] <gatox> nop
[19:44] <nessita> gatox: what are you patching in the test?
[19:45] <alecu> nessita, gatox: if I leave the --gui out, it works.
[19:45] <nessita> alecu: right, the --gui creates a QApplication
[19:45] <nessita> gatox: you 100% you're not creating a QApplication?
[19:45] <gatox> nessita, yes
[19:45] <nessita> gatox: can I see the code?
[19:45] <gatox> alecu, yes! without the --gui works
[19:46] <gatox> nessita, it's in trunk
[19:46] <gatox> trunk is not working
[19:46] <nessita> gatox: exactly, but trunk is creating a QApplication
[19:46] <gatox> but without the --gui as alecu says, it works
[19:47] <nessita> gatox: you branch was supposed to patch QApplication so is not created
[19:47] <nessita> isn't that what we discussed?
[19:48] <nessita> gatox: you understood my point? in trunk, ubuntu_sso/qt/tests/test_main.py calls         main.main(**kwargs) *which creates a QApplication*, and this that segfauls
[19:48] <gatox> nessita, actually, it segfault when creating the ui
[19:48] <nessita> gatox: that does not answer my question :-D
[19:49] <gatox> nessita, at least in the tests i made on friday
[19:49] <nessita> gatox: in C, segfaults may happen a little after the violation happened
[19:49] <gatox> nessita, let me try again.... i tested several things on friday but none of the patches works.... i'll try again
[19:50] <nessita> gatox: just patch Qapplication, and nothing else, and let me know, we'll debug further
[20:11] <gatox> nessita, review pleaseeeeeeeeeeeeeee
[20:11] <gatox> https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/unicode-args
[20:12] <nessita> gatox: 66+#    test_main.skip = 'Failing with QObject::startTimer: QTimer can only be ' \67+#                     'used with threads started with QThread'
[20:13] <gatox> nessita, mmmmmm that is not there anymore
[20:13] <nessita> ah, was an old LP scan then
[20:14] <nessita> gatox: oh hum... remember what I mentioned in one review about doing this? :-/
[20:14] <nessita> 54+    def desktop(self):
[20:14] <nessita> 55+        """Fake Desktop."""
[20:14] <nessita> 56+        return self
[20:14] <gatox> nessita, lp is updated now
[20:14] <nessita> gatox: we should have a FakedGeometry class and return that
[20:15] <gatox> ok
[20:15] <alecu> nessita, I used that test pattern a few branches ago too. Can you elaborate?
[20:17] <nessita> alecu: when doing fakes, we should try to mimic the faked object API as much as possible. So, if the API for Foo.dektop() returns an instance of an object X with a specific API, we should not return on desktop an instance of Foo, but an instance of an X (ideally) or a faked X
[20:17] <nessita> alecu: if we don't, out production code could added methods on objects that will fail IRL but will pass all green in the test suite
[20:18] <nessita> an example would be having this in production:
[20:18] <nessita> QApplication.instance().desktop().setStyleSheet()
[20:18] <nessita> that will pass in our tests (as how the branch from gatox is right now) ^
[20:19] <nessita> alecu: IRL, an instance of a QDesktop (or whatever the class is), do not have a method called setStyleSheet
[20:19] <nessita> alecu: makes sense?
[20:19] <gatox> nessita, branch updated
[20:20] <alecu> nessita, right, it makes sense now. And I would argue that it's a bit unlikely, but I agree that it might happen, and we better be careful. Thanks!
[20:21] <nessita> alecu: I try to aboid having fakes as much as possible, weighing having a fake vs how comples is to use the "real" class
[20:21] <gatox> nessita, ok, i've updated the branch and run the whole suit, everything is ok
[20:21] <nessita> avoid* (sorry!!!)
[20:21] <nessita> gatox: ack!
[20:22] <gatox> well people......  EOD for me...... time to see the walking dead season finale! :P
[20:22] <alecu> nessita, yup: generally it's better to use patch on the real class than to use a fake, since patch fails if the method being patched is no longer there.
[20:23] <nessita> alecu: right
[20:32] <mandel> alecu, did you get the ssl dialog running on windows?
[20:32] <alecu> mandel, what?
[20:33] <mandel> alecu,  the issue that ralsina had with the deferToThread on windows
[20:33] <alecu> mandel, no. I've been looking at the issue with proxy credentials.
[20:34] <mandel> alecu, any clue? I've started running out of ideas..
[20:34] <alecu> mandel, I'm thinking of replacing WebClientProxyFactory
[20:35] <alecu> mandel, and doing the request twice
[20:35] <alecu> mandel, just like I did on the SD branch
[20:35] <mandel> alecu, I have considered that to, but updating the problem is that the CachableObject implementation of the QNetworkAccessManager keeps track if there are replies that are using the proxy...
[20:36] <mandel> alecu, so, the issue would be that in the second request you will try to update the creds and the stupid cache will complain..
[20:36] <alecu> mandel, oh, right.
[20:37] <alecu> mandel, I remember now.
[20:37] <alecu> mandel, and that credential cache is global, so creating a different nam would not help.
[20:37] <mandel> alecu, bingo!
[20:37] <alecu> mandel, anyway, I'm continuing with some irl for this.
[20:37] <mandel> alecu, I have been trying a gazillion diff things and I always get to the same bloody problem..
[20:38] <mandel> alecu, I'm going to have a dinner and will be back for a 3rd or 4th round against the nam
[20:38]  * alecu somehow has just read "gazillon" as "a few"
[20:38] <mandel> alecu, well, 4/5 hehe
[20:38] <alecu> mandel, from now on, you'll be able to say "I was in 'Nam!"
[20:39] <mandel> alecu, the annoying thing.. I keep typing name instead of nam
[20:43] <mandel> alecu, I think that if we do not pass the parent of the QNetworkAccessManager to be the instance of the application we could remove and use a new one
[20:43] <mandel> alecu, why did you pass the app instance as the parent?
[20:44] <alecu> mandel, because every instance derived from qobject needs some reference to another qobject.
[20:44] <briancurtin> alecu: forget what i said earlier...that "on_download_finished" exception is happening even without a proxy setup, but only on XP. Win7 everything is fine proxy or no-proxy
[20:44] <briancurtin> which is even more weird
[20:45] <mandel> alecu, I don't think so.. parent can be 0
[20:45] <alecu> mandel, and here we don't have a reference to the UI objects
[20:45] <alecu> mandel, ok. But when are those refcounted?
[20:47] <alecu> briancurtin, it sounds like the version of u1-client does not match the version of u1cp. Or that u1cp is not pointing PYTHONPATH to the same version of u1-client.
[20:47] <mandel> alecu, is part of the pyqt magic.. we should ask ralsina, specially because google returns this: http://lateral.netmanagers.com.ar/weblog/posts/BB990.html
[20:47] <alecu> briancurtin, otherwise, cool that you managed to make proxies work on win7! congrats :-)
[20:48] <dobey> thisfred: want to review https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/highlander/+merge/98214 ?
[20:48] <thisfred> dobey: sure thing
[20:48] <briancurtin> this is all out of a py2exe'ed bundled installer, so the versions should be matching since it's all pulled from the latest of each project. arghhh
[20:48] <mandel> alecu, I'll have lunch and will do some irl with the idea of no parent and removing the nam when we have an error.. is worth a shot
[20:48] <alecu> mandel, there you go: references to QObjects. I had some of those issues, and that's how I solved it :-)
[20:49] <alecu> mandel, why no-parent?
[20:49] <alecu> mandel, I don't think that's a proper solution.
[20:50] <alecu> mandel, a better one is to forcefully delete the nam when it's not used anymore.
[20:50] <alecu> mandel, that article by ralsina gives some hints.
[20:51] <thisfred> dobey: +1
[20:52] <mandel> alecu, but that is talking about to use QObject to look a signals
[20:52] <mandel> alecu, and use sender to know what finished, you can perfectly have a QLabel('test') QLabel.show()
[20:53] <mandel> alecu, the label has no parent, and is a QObject
[20:55] <mandel> alecu, I'll have lunch an will look down that path, I have a feeling that we might get something out of there..
[20:56] <alecu> mandel, QtNetwork.QNetworkAccessManager.cache().clear() ?
[20:56] <mandel> alecu, wrong cache..
[20:56] <dobey> i should set up a squid cache on my server
[20:56] <mandel> alecu, has nothing to do with the creds one :(
[20:56] <dobey> just so i don't have to download bloody libwebkitgtk-dbg N times
[20:56] <mandel> alecu, https://bugreports.qt-project.org/browse/QTBUG-17312?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#issue-tabs
[20:57] <dobey> it's almost as big as the cd image, just by itself
[20:57] <mandel> alecu, it was not a gazillion things, but I have considered a few ;)
[20:57] <alecu> :-)
[20:57] <mandel> alecu, the evil bastards is called: QNetworkAccessCache
[20:57]  * mandel hates that guy
[20:58]  * mandel dinner -> back after chinese food
[20:59] <dobey> heh
[21:07] <alecu> mandel, I got an inkling for a solution.
[21:09] <nessita> thisfred: would you like to do a review similar to the one you did last Friday?
[21:10] <nessita> thisfred: it has tons of instructions and details in the MP description :-)
[21:10] <thisfred> nessita: sure, gimme the link!
[21:10] <nessita> thisfred: https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/computer-to-cloud-page/+merge/98286
[21:10] <thisfred> on it
[21:10] <nessita> thisfred: depends on the one from Friday (tarmac is down since canonistack was down), but is already merged in the the MP
[21:11] <thisfred> ok
[21:12] <nessita> will do the dishes and come back :-)
[21:12] <dobey> oh my it is 17:12 already
[21:13] <dobey> wish laptop dist-upgrade would go faster
[21:35] <nessita> ok, I'm crowd
[21:35] <nessita> thisfred: will check email later today, so any change/issue, just add it to the MP
[21:35] <nessita> thanks!!!
[21:35]  * nessita waves
[21:37] <thisfred> nessita: everything works, now going through the code
[21:37] <thisfred> gone throu
[21:37] <thisfred> gh I mean
[21:37] <thisfred> approved with minor critique
[21:39] <thisfred> now gonna walk the dog
[22:05] <dobey> later all
[22:18] <briancurtin> ralsina: well, after tons of errors from "on_download_finished" TypeErrors only seen on XP, u1 does end up syncing files
[22:18] <briancurtin> and with that, i'm out of here
[22:21] <mandel> alecu, I'm back!
[22:21] <mandel> alecu, too my time to watch the walking dead.. bloody tv show..
[22:23] <alecu> mandel, hey there.
[22:24] <mandel> alecu, have you tried removing the nam when there is a proxy auth error?
[22:26] <alecu> mandel, take a look at this branch: https://code.launchpad.net/~alecu/ubuntu-sso-client/fix-mandels-mess
[22:26] <alecu> mandel, it's incomplete, and tests fail
[22:26] <alecu> mandel, but it works.
[22:26] <alecu> mandel, I've been trying it irl for the past 10 minutes
[22:26] <alecu> mandel, and it keeps asking for the proxy creds, but it lets you authenticate.
[22:27] <mandel> alecu, I got to that point too in my mess ;)
[22:27] <mandel> alecu, the problem is to have everything :(
[22:27] <alecu> mandel, this is no longer exhibiting the NAM issue.
[22:29] <mandel> alecu, why did you remove the if not retry statement?
[22:30] <alecu> mandel, perhaps I should leave that :-)
[22:30] <mandel> alecu, you have remove exactly the code that loads the creds from the keyring and probably that is the reason why you allways get the dialog
[22:30] <mandel> alecu, maybe ;)
[22:31] <mandel> alecu, I also don't uderstand why you added the str in the creds = yield keyring.get_credentials(str(domain)) I think common should always get ptyhon strings, if a qt string was getting there the problems comes from somewhere else
[22:32] <mandel> alecu, as in, you should be passing domain as str(QString) to the base class
[22:34] <mandel> alecu, and you are cheating, this is the initial solution I had, you are always using the same proxy for all the diff requests, no matter if they are http or https, that is way the hack of setting the application proxy works and the issues does not longer happen
[22:35] <alecu> mandel, and that's good, since we are using https for even the ping.
[22:35] <alecu> mandel, and I guess for the captcha too.
[22:35] <mandel> alecu, as long as it is document I'm fine, the only thing missing then is getting the keyring code back
[22:35] <alecu> no
[22:35] <alecu> I've just restored the keyring code
[22:35] <mandel> no?
[22:35] <mandel> alecu, he :P
[22:36] <alecu> and I'm not getting asked for the proxy credentials anymore
[22:36] <alecu> mandel, that is: if any proxy creds are in the keyring, those are retried, and they never are asked back.
[22:37] <mandel> alecu, as in, even you get an error?
[22:37] <alecu> mandel, so, there's a bit of logic missing somewhere.
[22:37] <alecu> mandel, as in: the connection is retried with the old proxy credentials, and they are never requested again.
[22:37] <alecu> mandel, go ahead, IRL it.
[22:39] <mandel> alecu, ok, let me see if I can fix that
[22:41] <alecu> mandel, just a sec, I've just merged with trunk, and I'm pushing.
[22:42] <mandel> alecu, ok, let me know, I think I know how to fix it
[22:43] <alecu> pushed
[22:43] <alecu> mandel, ^
[22:48] <mandel> alecu, doing IRL tests with the fix, gime me 5 mins
[22:56] <mandel> alecu, first thing, this comment: https://code.launchpad.net/~mandel/ubuntu-sso-client/use-qt4-webclient-backend/+merge/97934/comments/211886 is wrong, you need to check for QCoreApplication and QApplication since they are diff
[22:58] <alecu> mandel, why? It seems to be returning the QCoreApplication just fine.
[22:58] <mandel> alecu, from IRL I'm getting libsoup and not qt
[22:58] <alecu> mandel, we should ask ralsina why he suggested that.
[22:58] <alecu> mandel, how are you testing?
[22:58] <mandel> alecu, running ubuntu-sso and then control panel
[22:59] <alecu> mandel, which ubuntu-sso command?
[22:59] <mandel> alecu, and I get this: http://paste.ubuntu.com/891465/ which is very easy to solve
[22:59] <alecu> mandel, if you just run ussoc-login-qt then the ussoc-login that will be started would be the one from the system.
[23:00] <mandel> alecu, the command is: U1_DEBUG=1 PYTHONPATH=. ./bin/ubuntu-sso-login
[23:03] <alecu> damn ^W
[23:12] <mandel> alecu, hehe
[23:12] <mandel> alecu, I'm pushing a branhc for you to do IRL that has everything working, or so I think :)
[23:12] <alecu> cool
[23:12] <alecu> mandel, even the tests you've skipped? ;-)
[23:13] <mandel> alecu, IRL, I need to look at the tests but I want you eyes on it too :)
[23:13] <alecu> :-)
[23:13] <alecu> url!
[23:13] <alecu> url!
[23:13] <alecu> url!
[23:13] <alecu> url!
[23:16] <mandel> alecu, you already know the url ;)
[23:16] <mandel> alecu, lp:~mandel/ubuntu-sso-client/fix-alecus-mess :P
[23:20] <mandel> alecu, by the way, this 'authenticator.user() != self.proxy_username' is not the same as 'self.proxy_username != authenticator.user()' I was getting a warning due to the expected type.. puto qt :P
[23:22] <alecu> lols
[23:23] <mandel> alecu, and all the tests pass except one :)
[23:23] <mandel> alecu, and I did not even touch them, and the issue is with the tests that choose which proxy regarding the request (http or https)
[23:23] <mandel> alecu, so, we can skip that one and without touching the tests, everything works!
[23:24]  * mandel dances before he becomes a pumpkin 
[23:24] <alecu> mandel, wait
[23:24] <mandel> alecu, no dancing :(
[23:24] <mandel> ??
[23:24] <alecu> mandel, if the password is wrong on the keyring, then I'm not asked for it again
[23:25] <alecu> mandel, and the wrong one is used.
[23:25] <mandel> alecu, I have tested that and worked
[23:25] <mandel> alecu, what have you done?
[23:25] <alecu> mandel, my bad!
[23:26] <alecu> mandel, I think I was checking the wrong branch :P
[23:26] <mandel> lol
[23:27]  * alecu makes a note not to work 13 hours in a day.
[23:27] <mandel> alecu, I just updated the branch with a tiny small fix on force_use_proxy
[23:27] <mandel> alecu, yeah... you should have taken a break.. buy a dog to have an excuse to take a walk or something
[23:28] <alecu> mandel, it seems to be working fine IRL
[23:29] <alecu> mandel, I'll do a little more testing by disabling auth in the proxy.
[23:29] <mandel> alecu, run the tests, nearly all of them should be working
[23:39] <alecu> mandel, it seems to be working lovely!
[23:39] <mandel> alecu, I just ran the tests and all passed!
[23:39] <mandel> alecu, let me push this to a branch with a nicer name ;)
[23:40] <alecu> mandel, nahhh.... that's good enough :-)
[23:41] <alecu> mandel, twit it so ralsina can review it too.
[23:41] <mandel> alecu, dough! some stupid pep8 issues.. lines too long, fixing :)
[23:42] <alecu> mandel, I'm getting this:
[23:42] <alecu>   CorrectProxyTestCase
[23:42] <alecu>     test_http_request ... Traceback (most recent call last):
[23:42] <alecu>   File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 1039, in _inlineCallbacks
[23:42] <alecu>     result = g.send(result)
[23:42] <alecu>   File "/media/sf_canonical/ubuntu-sso-client/fix-alecus-mess/ubuntu_sso/utils/webclient/tests/test_webclient.py", line 996, in test_http_request
[23:42] <alecu>     self.assertIn('request_proxy_auth_credentials', self.called)
[23:42] <alecu> twisted.trial.unittest.FailTest: 'request_proxy_auth_credentials' not in []
[23:42] <alecu> [FAIL]
[23:42] <alecu>  let me paste it we know nothing was lost:
[23:43] <mandel> alecu, yes, that is the test I was talking about, sets the http to use auth and the https not to, and expects to get a request for the creds
[23:43] <alecu> great then.
[23:43] <mandel> alecu, since we no longer use a factory, we don't have that feature :(
[23:47] <mandel> alecu, oh.. there is one use case we did not test.. but can be left as a bug
[23:47] <mandel> alecu, when the user cancels the creds dialog..
[23:47] <mandel> alecu, shall I try to fix it know?
[23:48] <mandel> s/know/now
[23:50] <alecu> mandel, what happens if he closes the dialog?
[23:50] <mandel> alecu, apears again and then we get an error in control panel stating that WebClientError has not attr content
[23:51] <mandel> alecu, I think is a bug in control panel.. in a way..
[23:51] <alecu> mandel, and there's no way to close it!
[23:51] <alecu> it keeps respawning :-)
[23:51] <mandel> alecu, yep
[23:52] <alecu> mandel, I think we can consider this a bug
[23:52] <mandel> alecu, so, I propose get your review and ralsinas and we go for this?
[23:52] <alecu> mandel, since whoever has configured proxies on gnome settings will surely try to enter valid proxy settings, and most people using this won't cancel.
[23:53] <alecu> mandel, sure, go ahead!
[23:53] <alecu> pro-pose
[23:53] <alecu> pro-pose
[23:53] <alecu> pro-pose
[23:56] <mandel> alecu, https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-alecus-mess/+merge/98308
[23:56] <mandel> alecu, even with a bug report linked :P