[07:56] <mandel> morning all!
[08:23] <JamesTait> Good morning all! :-D
[11:13] <ralsina> good morning!
[11:14] <mandel> ralsina, morning!
[11:14] <mandel> ralsina, and.. I hate unity trunk, they changed again something in the preview container that broke the payment preview.. I already fixed it but it is a PITA
[11:17] <ralsina> mandel: :-(
[11:17] <ralsina> mandel: well, our trunk is broken too
[11:17] <mandel> ralsina, what?! how come?
[11:18] <ralsina> something in the sync menu breaks when there is no sync menu installed
[11:18] <mandel> ralsina, unity trunk I don't care,, but ours? how did that happen?
[11:18] <mandel> ralsina, mierda :(
[11:19] <ralsina> mandel: surely an easy fix, I am about to do it
[11:19] <ralsina> mandel: plus nightlies are broken in Lucid, which affects the server guys, so I am looking into that as well
[11:19] <ralsina> mandel: and apparently windows is very broken
[11:19] <ralsina> mandel: so, FUN DAY!
[11:20] <mandel> ralsina, I'm not that busy with the preview, send some bugs my way :)
[11:29] <gatox> good morning!
[11:30] <ralsina> mandel: sure will
[11:30] <ralsina> good morning gatox
[11:30] <gatox> ralsina, hi
[11:35] <ralsina> gatox: do you have a lucid box?
[11:35] <ralsina> gatox: it seems like the sync menu breaks when there is no sync menu: http://pastebin.ubuntu.com/1273005/
[11:37] <gatox> ralsina, no...... no lucid..... i can install one if you want
[11:38] <ralsina> gatox: if you can just start a branch fixing that one, and pass it onto facundobatista to test, that's good enough
[11:38] <gatox> ralsina, ok.... i'll do that
[11:39] <gatox> the problem seems to be with dbusmenu, not the sync menu
[11:40] <ralsina> gatox: yes, but dbusmenu is only used by the sync menu, isn't it? (I don't know)
[11:41] <gatox> ralsina, yes
[11:41] <ralsina> gatox: the import is guarded, but then the usage is not
[11:41] <ralsina> gatox: so the import fails, sets a flag, and the flag is ignored :-)
[11:42] <gatox> ralsina, we are using the flag for the other part of the menu, i'll extend it to take care of that part too
[11:43] <ralsina> gatox: awesome
[11:47] <ralsina> gatox: also, trunk on P: http://pastebin.ubuntu.com/1273024/
[11:47] <ralsina> gatox: I think you can do both in one branch
[11:47] <gatox> ralsina, yap :S
[12:01] <mandel> ralsina, any bus numbers for me
[12:02] <ralsina> mandel: let me find a special one :-)
[12:02] <mandel> ralsina, I don't want bug #1 thx ;-)
[12:03] <ralsina> mandel: ok bug #3 ;-)
[12:03]  * ralsina loves vague mission statement bugs
[12:04] <mandel> ralsina, I write by bugs as todos :)
[12:04] <gatox> ralsina, can you review it? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/menulucid/+merge/129169
[12:04] <ralsina> gatox: sure
[12:05] <gatox> let me know if we need two reviews for that
[12:07] <ralsina> gatox: I would like facundobatista to take a look
[12:07] <gatox> ralsina, ah right
[12:07] <gatox> facundobatista, can you please review this one? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/menulucid/+merge/129169
[12:08] <mandel> ralsina, gatox, do you guys have any idea of why the following is not merging: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/u1-cp-publishapi/+merge/128316
[12:08] <gatox> mandel, i was about to check that..... because those tests are working here
[12:08] <facundobatista> gatox, from the diff, if you do
[12:09] <facundobatista> + Menuitem = object
[12:09] <facundobatista> and then
[12:09] <mandel> gatox, yes, same here, which is weird
[12:09] <facundobatista> + self.go_to_web = Menuitem()
[12:09] <gatox> facundobatista, i know.... fixing that
[12:09] <gatox> flakes fail
[12:09] <facundobatista> gatox, you know what?
[12:10] <gatox> facundobatista, i thought you were talking about something else
[12:10] <gatox> facundobatista, what happend?
[12:10] <facundobatista> gatox, from the diff, if you do
[12:10] <facundobatista> + Menuitem = object
[12:10] <facundobatista> and then
[12:10] <facundobatista> + self.go_to_web = Menuitem()
[12:10] <facundobatista> gatox, how do you then do
[12:10] <facundobatista> 41	self.go_to_web.property_set(Dbusmenu.MENUITEM_PROP_LABEL, GO_TO_WEB)
[12:10] <facundobatista> ?
[12:10] <gatox> facundobatista, if you look at the bottom of the file, we have: UbuntuOneSyncMenu = UbuntuOneSyncMenuLinux if use_syncmenu else DummySyncMenu
[12:11] <gatox> so, if we reach the point to do:
[12:11] <facundobatista> gatox, I mean, object() has no property_set
[12:11] <gatox> Menuitem = object
[12:11] <gatox> that part of the code is never used
[12:11] <ralsina> gatox: shouldn't that be TransfersMenu()?
[12:12] <ralsina> gatox: so, you replace the base class of TransfersMenu but you use the actual class everywhere?
[12:13] <gatox> ralsina, the only place where TransfersMenu is being used, is inside UbuntuOneSyncMenuLinux..... and if the import fail and we have use_syncmenu = False the real use_syncmenu = False is not used..... so we are not using TransfersMenu either
[12:13] <ralsina> gatox: yes, but that code feels bad :-)
[12:13] <gatox> the real UbuntuOneSyncMenuLinux i mean
[12:13] <mandel> gatox, ralsina, what was the exact error, getting an import error from gi.repository?
[12:13] <facundobatista> gatox, if you have code that is broken, and never used, *delete it*
[12:13] <ralsina> gatox: if it's not used, then why set it?
[12:14] <gatox> facundobatista, it is used...... not in lucid
[12:15] <ralsina> gatox: is there any chance of Menuitem being instantiated from object?
[12:15] <gatox> ralsina, the other thing was to declare TransfersMenu and DummyTransfersMenu inside an if using use_syncmenu..... but it seems ugly...... maybe it just my opinion
[12:15] <gatox> ralsina, sorry.... didn't understand that question
[12:16] <ralsina> gatox: you are setting sometimes Menuitem = object and in other cases Menuitem is something else
[12:17] <ralsina> gatox: so you have a class that sometimes has an API and sometimes doesn't
[12:17] <ralsina> gatox: which is an accident waiting to happen :-)
[12:17] <ralsina> gatox: I would be happier if it were sometimes the real thing and sometimes the dummy but always with a consistent api
[12:18] <gatox> ralsina, yes..... if we assume that TransfersMenu is going to be used for someone else instead of UbuntuOneSyncMenuLinux...... do you prefer the option with the class declared inside an if?
[12:18] <ralsina> gatox: yes
[12:18] <gatox> ralsina, ok.... i'll change that
[12:19] <ralsina> gatox: or declare it always and use an if to assign it to the "real" name?
[12:19] <ralsina> gatox: either way is good
[12:19] <facundobatista> gatox, wasn't simpler to do this? class TransfersMenu(Dbusmenu.Menuitem if use_syncmenu else object)
[12:20] <gatox> facundobatista, but that has the same problem that ralsina was saying
[12:20]  * facundobatista reads backlog
[12:21] <facundobatista> gatox, +0 to the both classes definition inside a big if
[12:22] <gatox> facundobatista, i don't like that option either..... but......
[12:22] <ralsina> facundobatista: I don't love it either but having an instantiable object feels just bad :-(
[12:22] <mandel> facundobatista, ralsina gatox, why not leave it like it is, do the import from within  the UBuntuOneSyncMenuLinux with a try except and then if you got an except set it to none and check when needed
[12:23] <facundobatista> gatox, ralsina, the cleanest way to do it is to import the class you'll use from two separated files, but this is good enough
[12:23] <gatox> mandel, +1
[12:23] <gatox> i like mandel suggestion
[12:23] <ralsina> facundobatista, gatox: I trust your good taste guys :-)
[12:27] <mandel> ralsina, I'm done with the preview changes, any bugs for me.. please...
[12:30] <ralsina> mandel: in 1'
[12:31] <ralsina> mandel: bug #1065272
[12:31] <ralsina> mandel: should be easy-ish
[12:32] <gatox> ralsina, facundobatista here is the branch using mandel 's suggestion (which i think is the cleanest one): https://code.launchpad.net/~diegosarmentero/ubuntuone-client/menulucid/+merge/129169
[12:32] <ralsina> mandel: then bug #1062729 which is semi-trivial too
[12:33] <ralsina> gatox: much nicer
[12:33]  * mandel ego +1
[12:33] <mandel> hehehe
[12:33] <gatox> thx mandel
[12:33] <gatox> jeje
[12:34] <ralsina> gatox: still feels bad to declare a whole class we will never use :-)
[12:34] <gatox> ralsina, but we use it!
[12:34] <ralsina> gatox: not if it inherits object we don't
[12:34] <gatox> ralsina, in that case goes back to None
[12:34] <gatox> at the end
[12:34] <ralsina> gatox: exactly, so, we define a class that inherits object and then we set it to None :-)
[12:35] <ralsina> gatox: that code is like pedaling on a fixed bike! (but yes, alternatives are verbose and worse)
[12:37] <gatox> yay...... tests just ends..... couple of fails.... fixing
[12:43] <gatox> ralsina, ouch..... i'm having problems with the new OrderedDict :S...... fixing....
[12:46] <ralsina> gatox: really?
[12:46] <gatox> ralsina, yap..... i'm trying to figure it out why
[12:47] <gatox> maybe is something related to other stuff...... but failing with that (and just a coincidence)
[12:51] <mandel> ok, time for my lunch
[12:56] <gatox> ralsina, i found the problem.... it seems that the OrderedDict from twisted doesn't have the proper method implemented to be used with for
[13:04] <ralsina> gatox: how could that not trigger any test failures?
[13:04] <ralsina> gatox: in any case, maybe t can be used with for k in d.keys()
[13:04] <gatox> ralsina, that's what i did
[13:04] <gatox> facundobatista, ralsina now is ready: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/menulucid/+merge/129169 and with the orderedict from twisted fixed too
[13:05] <ralsina> gatox: cool
[13:09] <facundobatista> gatox, you keep changing the branch?
[13:09] <gatox> facundobatista, i found another problem with the OrderedDict, that's why
[13:10] <gatox> facundobatista, but now is final
[13:10] <facundobatista> gatox, sure?
[13:10] <gatox> facundobatista, 100%
[13:10] <facundobatista> gatox, ok
[13:10] <ralsina> gatox: sorry about the ordereddict breakage!
[13:11] <gatox> ralsina, no problem.... it wasn't a big thing
[13:11] <gatox> pretty easy to find and fix
[13:14] <ralsina> gatox: it's what happens when I try to fix bugs I can't see :-)
[13:20] <ralsina> and ... http://t.co/VdYiSl8Z
[13:20] <ralsina> congrats gatox, mandel, and the rest of the team who's late ;-)
[13:20] <gatox> \o/
[13:20] <chaselivingston> congrats everyone!
[13:32] <gatox> mandel|lunch, ralsina when you have a moment please: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/clickshare/+merge/129188
[13:37] <ralsina> gatox: sure
[13:37] <ralsina> gatox: you proposed it against the wrong package ;-)
[13:37] <ralsina> gatox: therefore 77390 lines (+76238/-0) 209 files modified (has conflicts)
[13:37] <gatox> ralsina, ohhhhhhh
[13:38] <gatox> ralsina, brain fart
[13:38] <gatox> ralsina, is a good thing that the weekend is close
[13:39] <chaselivingston> ralsina: remind me the command to run u1sdtool on mac?
[13:40] <ralsina> chaselivingston: well, it's u1sdtool I just have no idea where it lies ;-)
[13:40] <chaselivingston> ralsina: i've found it, but when i try to run it i just get -bash command not found
[13:40] <ralsina> chaselivingston: if you are in that folder, do ./u1sdtool
[13:41] <chaselivingston> ralsina: ah, I think that's right
[13:41] <gatox> ralsina, mandel|lunch now in the proper project: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/clickshare/+merge/129199
[13:41] <karni> ralsina: hiya! my friend can't sign in to Mac beta. did check his username/pass by signing in to o.u.c. on mac, getting "The authentication failed"
[13:41] <alecu> hey, I forgot to say "hi everyone!"
[13:41] <karni> ralsina: any hints?
[13:41] <ralsina> alecu: hello!
[13:42] <ralsina> karni: not really
[13:42] <karni> ralsina: any logs he can provide/turn on?
[13:42] <karni> ralsina: ok.. he changed his password, and it worked this time o_O
[13:42] <ralsina> karni: MAGIC
[13:42] <karni> ralsina: I don't know where the problem was, but yeah.. problem solved ;]
[13:43] <karni> he did sign in to o.u.c with previous password no prob.
[13:43] <karni> Kewl
[13:44] <karni> FWIW he changed his password for the same one, and it worked afterwards.
[13:44] <karni> also, congratulations on Mac beta launch, team!
[13:51] <gatox> dobey, ping
[13:51] <ralsina> alecu: no need for 1-1, right? We talked already?
[13:52] <gatox> dobey, do you know why this branch is bouncing?? tests are ok for mandel and me but failing on tarmac: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/u1-cp-publishapi
[13:53] <alecu> ralsina: right
[13:55] <ralsina> gatox: dobey is on vacation
[13:55] <gatox> ralsina, oops.......
[13:56] <gatox> ralsina, do you know if tarmac is running with an older version of qt or something?
[13:56] <alecu> gatox: we should try asking sidnei regarding tarmac
[14:00] <dobey> i am on holiday, but the version of qt in quantal is 4.8.3
[14:01] <dobey> gatox: those are the same 67 failures in QThread I asked about in another cp branch of yours a week or two ago
[14:01] <gatox> dobey, thx.... keep enjoying your vacation.... i'll check what is going on.....
[14:03] <ralsina> dobey: go away and vacate
[14:04] <ralsina> dobey: or however the thing one does on a vacation is called
[14:04] <dobey> i am
[14:05] <ralsina> gatox: +1 but get a good second review, mine was soft :-/
[14:06] <gatox> ralsina, yes.... i would ask mandel|lunch when he finish his lunch
[14:06] <gatox> now
[14:06] <gatox> jeje
[14:06] <mandel> gatox, you want what?
[14:06] <mandel> gatox, I finished in u1-internal first hehe
[14:07] <gatox> mandel, a small review please: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/u1-cp-publishapi/+merge/128316
[14:07] <mandel> gatox, I already gave a +1 to that, tarmac is complaining about qthread
[14:08] <ralsina> mandel: here's a real bug that's been pissing me for a while bug #1065513
[14:08] <mandel> ralsina, ok, I'll tackle the other one first, then will more to that one :)
[14:08] <gatox> mandel, sorry, this one: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/clickshare/+merge/129199
[14:09] <ralsina> mandel: ok
[14:13] <mandel> gatox, would be great if connect form pyqt could take extraparameters that would be appended to the ones emitted, like self.button.clicked.connect(self.set_name, '') I hate seeing lambdas everywhere ;)
[14:14] <gatox> mandel, i'm starting to notice that.....
[14:14] <gatox> jejee
[14:14] <gatox> mandel, but you can't do that in pyqt
[14:15] <gatox> mandel, or i'm not understanding what you say
[14:15] <mandel> gatox, I know.. that is why I used the second conditional, would be great if...
[14:15] <gatox> ahhhhhhh
[14:15] <mandel> gatox, it would make the api nicer :)
[14:15] <gatox> mandel, i thought you were requesting a change
[14:15] <mandel> no no no
[14:15]  * gatox start doing a pull request into pyqt because mandel ask
[14:15] <gatox> jejjee
[14:15] <mandel> lol
[14:16] <ralsina> mandel: can you give me an example? Because there is quite a bit of flexibility in connect :-)
[14:17] <mandel> ralsina, for example: http://paste.ubuntu.com/1273294/
[14:18] <mandel> ralsina, just to avoid using lambdas, but I agree with gatox, there is nothing we can do, right?
[14:19] <mandel> alecu, can you merge the following: lp:~mandel/avani/unity-lens-music.ubuntuone-purchase-take2 with your branch?
[14:20] <mandel> alecu, some of the string are wrong (the ui ones) and I fixed them according to the given design
[14:23] <alecu> mandel: ack
[14:24] <mandel> gatox, branch lp:~diegosarmentero/ubuntuone-control-panel/clickshare have you ran the tests on !
[14:24] <alecu> mandel: did you see the new checkbox in th design? it's called "forget payment details" (or something like that)
[14:24] <mandel> gatox, on Q I meant, sorry for the !
[14:24] <alecu> gatox: I was not planning on including that.
[14:24] <alecu> sorry
[14:25] <gatox> mandel, yes..... i'm on Q
[14:25] <alecu> mandel: I was not planning on including that.
[14:25] <mandel> alecu, wtf? when was there? and no, I have a design without it
[14:25] <gatox> mandel, what are you seeing?
[14:25] <mandel> gatox, let be > to a file
[14:26] <mandel> alecu, and I agree, we are not doing that, we don't even have that in the rest api, do we
[14:26]  * mandel does not want to have to do yet another ui element in nux
[14:26] <alecu> mandel: right!
[14:26] <alecu> mandel: it's on "the one and only 12.10 screen"
[14:26] <mandel> alecu, ok, you can ignore that
[14:27] <alecu> mandel: it's called "remember my credit card details"
[14:27] <mandel> alecu,  let me fwd the real design
[14:27] <mandel> alecu, I already said no to that in the sprint :)
[14:27] <alecu> mandel: oh, awesome.
[14:27] <mandel> alecu, just sent you the real thing
[14:29] <mandel> gatox, http://paste.ubuntu.com/1273319/
[14:29] <mandel> gatox, but maybe is my system that got bananas
[14:30] <gatox> mandel, i'm not being able to run lint for u1-cp in Q neither
[14:30] <gatox> mandel, but the tests pass
[14:30] <ralsina> mandel, alecu: scared me there for a bit :-)
[14:30] <mandel> gatox, yes, that I know..
[14:31] <mandel> ralsina, nah, I already made sure we were not doing that hehehe
[14:31] <mandel> I was very very clear
[14:31]  * mandel said no in nearly all languages => NO!
[14:35] <mandel> gatox, +1 hoping that if there are real lint issues tarmac will tell us we are stupid
[14:36] <gatox> mandel, na, they are not real if you check
[14:36] <mandel> gatox, I got lost in the verbal diarrhea that lint has.
[14:46] <mandel> gatox, that sync menu has lots of blackguard compatibility issues =>  bug #1065272
[14:47] <mandel> easy to fix though
[14:58] <briancurtin> mumble in 2'?
[15:00] <joshuahoover> briancurtin: just the question i was going to ask
[15:00] <gatox> brb
[15:00] <ralsina> oh,mumble, right
[15:00] <ralsina> sorry, been a little distracted
[15:02] <mandel> sorry, joining
[15:02] <ralsina> gatox, alecu, mmcc mumble
[15:22] <gatox> mmcc, ralsina when you have a moment, another review: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/deflated-size/+merge/129215 (now i'm moving to windows)
[15:23] <briancurtin> brb, need to grab some coffee
[15:23] <gatox> briancurtin, i'm going to have lunch now..... and we can talk after that what do you need..... or tell me what you have been trying so i start debugging that here too.... do you agree?
[15:30] <ralsina> briancurtin, alecu: I have assigned to you a few bugs that are the ones support is getting questions about
[15:30] <ralsina> briancurtin: yours look easy
[15:30] <ralsina> alecu: yours looks hard
[15:31] <alecu> ralsina: ack
[15:31] <mmcc> ralsina: I don't see a welcome page bug in my assigned bugs, did you say you assigned it to me?
[15:31] <ralsina> briancurtin: one interesting bit is that we are still getting valicert questions. I am not sure if we ever released with the cert bundled on windows. If we did, then it's not working thugh
[15:31] <ralsina> mmcc: let me track it
[15:31] <mmcc> also, gatox I am looking at your deflated-size branch now
[15:32] <gatox> mmcc, great, thankx
[15:32] <ralsina> mmcc: https://bugs.launchpad.net/ubuntuone-control-panel/+bug/1065514
[15:32] <mmcc> btw, I think I'd like to look at client test running soon. it's slow everywhere, and totally broken on osx
[15:33] <ralsina> mmcc: fixing it on osx and then getting a darwin jenkins souds like something we really should be doing
[15:33] <mmcc> ralsina: yes.
[15:34] <ralsina> and remember, team, the bug situation can always be worse http://feeds.gawker.com/~r/gawker/full/~3/1gnDhJEMjp4/heres-footage-of-the-live-roach+eating-contest-that-killed-the-winner
[15:34]  * mmcc is NOT clicking that link
[15:34] <mmcc> thank god for readable URLs
[15:34] <briancurtin> ralsina: once we get this other win stuff figured out, i'll look at how the installer is built in regards to certs
[15:35] <ralsina> briancurtin: awesome
[15:35] <ralsina> briancurtin: since we are not going to do another release before the 4-beta ...
[15:36] <mmcc> ralsina: that welcome screen bug -> I think he just means the control panel main window. it has a rounded white rect inside a grey background. Although I agree that it's not like the default mac look, it is on-brand, and there isn't a real guideline for 'mac look' anymore anyway.
[15:37] <briancurtin> gatox: i've been poking around the code and following log files and trying to see how this CredentialsError is coming up. i'm going to go back and look at the dirspec changes to binary finding, then try to debug SSO in isolation (inside credentials.py) again
[15:39] <gatox> briancurtin, ok..... lunch and i'll start trying to execute u1 in windows and track the bug
[15:39] <briancurtin> gatox_lunch: sounds good
[15:41] <mmcc> gatox +1 on that deflated-size branch, thanks
[15:42] <mmcc> briancurtin: how are you running things when you hit problems? I'm about to kill my credentials and try running just control-panel
[15:42] <briancurtin> mmcc: i'll put my details on a paste in 1', too much to type here
[15:43] <mmcc> briancurtin: cool ,thanks
[15:50] <mmcc> where do the logs go in windows? I see the special folder names in dirspec but I don't know where to find those in the filesystem :\
[15:51] <briancurtin> mmcc: C:\Users\brian\AppData\Local\xdg\cache\ubuntuone\log
[15:51] <mmcc> oh, that's hidden in the explorer?
[15:52] <briancurtin> mmcc: yeah i think AppData is hidden
[15:52] <ralsina> mmcc: ok!
[15:52] <mmcc> oh it's even hidden from 'dir', tricky!
[15:52] <ralsina> mmcc: put that same explanation and a wontfix status
[15:53] <mmcc> ralsina: ok
[15:53] <ralsina> mmcc you can use %LOCALAPPDATA% in windows explorer (or something similar)
[15:54] <briancurtin> mmcc: still working on your doc...now u1cp is being weird as i try to document what im doing
[15:54] <mmcc> ralsina: oh on second thought he probably means this welcome screen: https://lh3.googleusercontent.com/-wpAFFGeAPzKkY7KFyiNmj01uz7KKrnkzVqoy25LuoJuIdXU1tRKfOnjDASSKo6ycwLmAiCyL7RmjFufhU1KsCQmrdBkhikgC8fxRTY2CTH9R_W0hKI
[15:55] <mmcc> the one that says welcome :)
[15:55] <ralsina> hahaha
[15:55] <mmcc> same response though
[16:00] <mandel> ralsina, gatox_lunch, please can you review the following: https://code.launchpad.net/~mandel/ubuntuone-client/spawn-cp/+merge/129230
[16:01] <gatox_lunch> mandel, will do after lunch
[16:02] <briancurtin> mmcc: here's the setup: https://pastebin.canonical.com/76354/
[16:02] <mmcc> gym, I have no logs in %LOCALAPPDATA%\ubuntuone
[16:02] <mmcc> gym was supposed to be hrm.
[16:04] <mmcc> so briancurtin, are environment variables inherited by child processes on windows?
[16:04] <briancurtin> mmcc: this is what i was just thinking about 1' ago...no, you need to pass it on, otherwise apps get a clean slate
[16:04] <mmcc> the import errors you see when just running cp by itself should be avoided because it uses 'python ' + sso whatever
[16:04] <mmcc> oh snap
[16:04] <mmcc> well there you go
[16:05] <mmcc> the whole deal with pre-pending 'python' to the binary paths was so that we get the buildout python
[16:05] <briancurtin> yeah it kind of clicked in my head when i saw it complain about twisted
[16:05] <mmcc> but if the child process doesn't have the path set right, then boom
[16:05] <mmcc> and that's why sso-qt isn't working either
[16:05] <mmcc> because sso uses the same trick to use the buildout python to run sso-qt
[16:06] <mmcc> well that's what you get when a guy who has avoided windows for 30 years tries to write windows code :\
[16:07] <briancurtin> mmcc: not your fault. i should have been on the lookout (also should have been able to diagnose this a while ago...)
[16:07] <briancurtin> mmcc: your changes were in dirspec, or was that just finding the binaries and not running them?
[16:07] <mmcc> briancurtin: dirspec finds 'em, other code runs it
[16:08] <mmcc> briancurtin: I'll dig up a list of places now. I think there is one in sso and one in cp or client
[16:10] <mmcc> sso-client - utils/__init__.py, line 126 , and ubuntuone-client syncdaemon/utils.py
[16:11] <briancurtin> mmcc: yep i'm looking at the sso one now
[16:11] <mmcc> just those two places. control panel imports from those places
[16:11] <mmcc> should be the exact same fix in both places. they're almost identical
[16:11] <mmcc> ok, brb with more coffee
[16:44] <mandel> ok, EOD, catch you all tom!
[16:45] <ralsina> bye mandel!
[16:45] <gatox> mandel, bye
[16:47] <mmcc> why do we only have upload info for the sync menu and not downloads also?
[16:48] <gatox> mmcc, you should ask the designers i think
[16:49] <mmcc> gatox: not just the ubuntu sync menu but also the control-panel sys tray one…
[16:49] <ralsina> mmcc the data is the same
[16:50] <ralsina> mmcc: syncdaemon collects the data, describes a menu and either u1cp or sync-menu draws it
[16:50] <mmcc> ralsina: yeah, I was thinking maybe the decisions had been made differently for each case
[16:50] <gatox> mmcc, yes, what i mean is that the design team creates the spec like that..... let me give you the design doc in pm
[16:51] <mmcc> since the sys tray one was older…
[16:51] <mmcc> thanks gatox
[16:52] <mmcc> that osx screenshot is just photoshop, right?
[16:53] <gatox> mmcc, yes
[16:53] <mmcc> hmm, that slide set just says 'files that are currently in transfer', it doesn't say if they're uploads or downloads
[16:55] <mmcc> ok, so as far as I can tell, using QWidgetAction to show a progress bar doesn't work on osx
[16:55] <gatox> mmcc, maybe lisettte can clarify that
[16:55] <mmcc> even if I just show the progress bar (no layout, no label, just the bar), it shows nothing
[16:56] <lisettte> gatox: what do you want me to clarify?
[16:56] <gatox> mmcc, can you send me an screenshot..... maybe we are having the same problem as in ubuntu where we needed to use the number instead the progress bar
[16:56] <gatox> lisettte, mmcc has some question about the upload/download info in the menu
[16:57] <mmcc> lisettte: the flyout menu that shows current transfers, in the code now it only shows uploads, no downloads. was that intended, or should it show both?
[17:00] <mmcc> gatox: I will get a screenshot for you in a minute, but it's easy to imagine - as it is, it will show the label and text percentage but no progress bar, and no space for one. if I only put in the progress bar, it just shows nothing (there's vertical space for a menu item but nothing is drawn there)
[17:01] <gatox> mmcc, yes..... the same problem..... so we need to use the saame code in mac as in ubuntu
[17:02] <chaselivingston> mmcc: the mac client is for intel macs only right? no power pc support?
[17:03] <mmcc> chaselivingston: yes, no ppc
[17:03] <chaselivingston> mmcc: cool, thanks
[17:04] <mmcc> gatox: which code is that?
[17:05] <ralsina> mmcc, gatox: on mac we have the same problem as in ubuntu, the tray menu is actually out of process
[17:06] <mmcc> ralsina: that seems unrelated, I'm confused
[17:07] <gatox> mmcc, is actually the same code.... if the progress bar doesn't work it should show just the text..... i'm debugging u1 on windows now..... if you can send me the screenshot it would be helpful
[17:08] <mmcc> gatox: it does show just the text, that's the problem. I want it to show the progress bar
[17:09] <mmcc> gatox: what are you debugging on windows now?
[17:10] <mmcc> I have a branch mostly done that fixes the sync transfers menu on osx, and need to test it on windows. I was just hoping to also get the progress bars showing, but it sounds like that's a known problem with qt?
[17:11] <gatox> mmcc, ahhhh..... well.... i think ralsina is right about that...... and maybe is not possible to do it when integrating with the systray
[17:11] <gatox> mmcc, the credentials problem.... as was discussed in the weekly call
[17:12] <mmcc> gatox: check with briancurtin, I think we figured that out while you were at lunch
[17:12] <gatox> mmcc, briancurtin really? please tell me
[17:13] <mmcc> gatox, I still don't understand what ralsina said! :) The progress bar seems to be an issue with qt menus on macs, but whether or not it's in a separate process shouldn't matter
[17:13] <mmcc> gatox: see the scroll back, it's because windows apps don't inherit environment variables, so prepending 'python' to the path does not give us the buildout python as intended
[17:14] <briancurtin> mmcc: i'm an idiot and mixed myself up...they *do* inherit
[17:14] <mmcc> gatox, so if you start sso separately like briancurtin did, control panel will find it via the TCPactivation, but it'll eventually fail when sso tries to launch sso-login-qt
[17:14] <mmcc> briancurtin: oh no :( and there I was all convinced it was solved
[17:14] <gatox> mmcc, the thing that is a different process it's because it seems that mac handle the systray menus in a different way (as ubuntu does)...... so you probably are not having a qt menu there...... but something that makes the qt menu compatible with mac, so you can't use all the things from qt
[17:15] <gatox> briancurtin, so i should keep debugging this?
[17:15] <briancurtin> mmcc: yeah, im sure its something around that, though. it really cant be anything else
[17:15] <mmcc> briancurtin: well, I still think that's the source of the issues. I'd poke at the get_bin_cmd stuff and see if maybe prepending 'python -v ' gives clues or whatever
[17:15] <mmcc> let me know if I can be of more use. for now I'll go back to sys tray stuff
[17:16] <briancurtin> gatox: i think for now you can probably go back to whatever else. if *this* doesn't turn out to be the real issue, then maybe i'll try to bring you back to the windows dark side
[17:16] <gatox> briancurtin, ok, no problem... i'll keep fixing syncmenu and shares issues...... let me know as soon as you need me :D
[17:43] <gatox> alecu, ping
[17:44] <gatox> ralsina, ping
[17:47]  * briancurtin lunch
[18:04] <mmcc> ok, so I can get the cp systray menu to work on osx the same way it should on linux - except that nothing will update while you hold the menu open like it does on linux, because of how Qt interacts with the osx menu system
[18:04] <gatox> mmcc, :S
[18:05] <mmcc> gatox: yup
[18:12] <mmcc> also, the menu items can't be set to disabled, because of a qt bug that will be fixed in 4.8.4: https://bugreports.qt-project.org/browse/QTBUG-25544
[18:13]  * gatox is waiting for Qt5...... so excited! \o/
[18:16] <mmcc> hrm, I'm not convinced that their fix actually fixes this
[18:17] <gatox> mmcc, well.... at least digia has declared to improve the integration with mac
[18:17] <mmcc> gatox, that's good news I guess. Doesn't help me fix this stuff now, though :\
[18:18] <gatox> mmcc, jeje no..... for that you have to blame the people who didn't invent the time machines yet
[18:19] <mmcc> gatox or I could blame the people who broke menu enabling in qt/mac, because apparently it used to work :)
[18:19] <gatox> or that
[18:25] <mmcc> interesting, qt's systemtrayicon can be compiled with support for growl
[18:26] <mmcc> can't tell if it was in the bottle from brew though
[18:29] <alecu> gatox: pong
[18:30] <mmcc> oh yeah, I see what's happening here. Qt is re-creating the menu items on every update, even when I don't re-create the python menu items :(
[18:31] <gatox> alecu, hi....... i'm looking at this bug: https://bugs.launchpad.net/ubuntuone-client/+bug/1065272
[18:32] <gatox> alecu, it seems that spawn_command_line_async is not supported in older versions..... but that was there from before..... although that report is not using the last version because we are not spawning ubuntuone-installer anymore
[18:32] <gatox> alecu, do you know what should be used for backward compatibility instead of spawn_command_line_async?
[18:33] <alecu> gatox: no. But I see that mandel has proposed a branch for that bug.... what do you think of it?
[18:35] <gatox> alecu, ahhhhh..... i had that on hold to start working on it.... i didn't know that mandel propose a branch for that because he show me the issue and it didn't has anything at that moment
[18:35]  * gatox refresh the page
[18:36] <alecu> gatox: :-)
[18:36] <gatox> alecu, +1 from me
[18:37]  * gatox goes back to his cp branches
[18:39] <alecu> gatox: please, when reviewing this kind of branches make sure to run the tests before approving, and since this affects old versions, try to run them in a VM with the old version too.
[18:40] <gatox> alecu, ack...... i'll install a lucid vm now
[18:41] <gatox> alecu, at least for Q they work
[18:43] <mmcc> gatox, btw please let me know if you're working on anything in cp that affects the sys tray stuff. thanks!
[18:43] <alecu> gatox: for the backporting work I had to make a lot of VMs, and now I have from L to Q. It's really handy to test this kind of stuff.
[18:43] <mmcc> and shared folders! use shared folders
[18:44] <gatox> alecu, ack..... i was on precise with a Q vm..... but now i'm in Q too......i'll install some old vms
[18:44] <gatox> mmcc, no.... i'm working in stuff related to the shares tab
[18:44] <mmcc> I was using shared folders, then upgraded and forgot to reinstall guest additions, and it broke, I just set it back up and I'm so happy
[18:45] <mmcc> gatox: cool, thanks
[18:54] <alecu> mmcc: yes, shared folders is great. I have the source directory mapped to every vm, so the vms get to run the latest version of the code every time.
[18:55] <alecu> gatox: ^^^
[18:55] <mmcc> alecu: it also saves me space on my tiny SSD. :)
[18:55] <alecu> mmcc: how tiny?
[18:55] <mmcc> unfortunately it doesn't work super cleanly on windows yet, since buildout generates a different set of executables
[18:55] <gatox> alecu, nicceeeee
[18:56] <gatox> i do that too
[18:56] <mmcc> alecu, 120GB of usable space, I think 128 in the chips
[18:56] <gatox> mmcc, it works fine on windows..... i've always use that on windows
[18:56] <gatox> mmcc, ahhhhh talking about other thing
[18:56] <alecu> mmcc: I've got the same on my mac, and I kind of regret it :-(
[18:56] <mmcc> gatox, yeah, running from source on windows because of the buildout 'python.exe' and such
[18:57] <mmcc> alecu, I don't regret it yet, but I do need to set up some networked storage or something
[18:57] <alecu> mmcc: now I'm looking at this: http://eshop.macsales.com/shop/SSD/OWC/Aura_Pro_Air_2011
[18:58] <mmcc> alecu: egads!
[18:58] <mmcc> not going near this thing with a screwdriver, at least until the AppleCare warranty expires
[18:59] <alecu> mmcc: lol
[18:59] <alecu> mmcc: I think I may shell out for the 240
[18:59] <alecu> my warranty expires on may next year though...
[18:59] <mmcc> I had a job in high school that was mostly unscrewing PC cases and re-seating boards, and reinstalling windows. I'm happy with my sealed boxes now :)
[19:00] <mmcc> Although I do wish they were upgradeable. Life is complex.
[19:00] <mmcc> So, I'm beginning to think I should just do this sync menu in pyobjc. anyone want to talk me back from the edge?
[19:01] <mmcc> I already did a test that works with showing the sync status via the twisted IPC, so twisted and pyobjc do work
[19:01] <alecu> mmcc: do you have that in a branch I can peek?
[19:02] <alecu> mmcc: a priori it sounds like a reasonable idea to me.
[19:02] <mmcc> alecu, :) it's a hacky mess on top of an unrelated pyobjc example - do you just want to look at the relevant source or do you want to try it out?
[19:02] <alecu> mmcc: but perhaps you can send a mail to u1-discuss with a further description of the problem and why not using Qt is a better idea.
[19:03] <mmcc> alecu, but that will force me to think it through clearly!
[19:03] <mmcc> which is an excellent idea, so I'll go do that
[19:03] <alecu> mmcc: I just want to have an idea at how much code we'll have to maintain going forward :-)
[19:03] <alecu> mmcc: so, a broken branch sounds ok too
[19:04] <mmcc> alecu: an entirely reasonable question. in particular, the reactor is of some concern. it doesn't use the "CFReactor" which is listed as unmaintained, it uses the 'threaded select' reactor
[19:05] <alecu> mmcc: oh, and this would be a separate process? I was guessing it could be a objc library that was used from the SD process, but yes, reactors.
[19:06] <mmcc> alecu, yes it's a separate process, which is necessary on osx anyway, because we want to not show the dock icon all the time
[19:06] <alecu> or ran inside the control panel process.
[19:06] <alecu> mmcc: right, we want the dock icon when the control panel is shown only.
[19:07] <mmcc> alecu: yep, so the menu will be a separate process that is background and the control panel will be foreground
[19:07] <alecu> mmcc: my opinion is that qt is great for most ui, but it gets hairy when doing platform specific bits like this. (well, not only qt, but every cross platform toolkit I've worked with)
[19:09] <alecu> mmcc: on the other hand, we may take a look at qt / pyqt innards, and if it's easier, try to come up with a patch that we can submit upstream.
[19:10] <mmcc> well, I'm looking at the innards now… live updating is probably possible but a pretty big change
[19:10] <alecu> mmcc: perhaps ralsina or gatox or any other qt fan can shed some insight on that.
[19:10]  * alecu just realizes he's wearing his Qt shirt today :P
[19:10] <alecu> literaly.
[19:10] <mmcc> ha
[19:10] <ralsina> alecu: you have  a QT shirt? I have to complain to the secret norwegian maling list
[19:11] <gatox> ralsina, i have one that too
[19:11] <ralsina> DAMN YOU ALL
[19:11] <gatox> mmcc, the patch should be submitted to Qt....
[19:11] <ralsina> I hav one of a giant octopus destroying a lighthouse
[19:11] <ralsina> anyway
[19:11] <ralsina> reading backlog...
[19:11] <mmcc> gatox: if it were up to me, I will avoid changing qt code. I don't want to have to recompile it even once
[19:12] <mmcc> but I'll try to make a rational decision :)
[19:12] <ralsina> mmcc: why in pyobjc? really, tell me :-)
[19:13]  * gatox has been debugging and touching qt code in the past..... and he find out weird things
[19:14] <mmcc> ralsina - here are the menu item things that don't work in Qt/mac: live updating while you hold the menu open, setting a menu action to be disabled (really!), QWidgetAction,
[19:14] <ralsina> gatox: Qt is over 15 years old. It's full of evil stuff. Evil C++ stuff.
[19:14] <mmcc> I think I'm forgetting one more
[19:14] <ralsina> mmcc: that's because the sync menu is out of process. It's done by the system's UI
[19:14] <ralsina> mmcc: it has more or less the same limitations sync-indicator has
[19:15] <mmcc> no that's it, a couple of other things are broken in the menu but not qt's fault
[19:15] <ralsina> mmcc: athough disabling items is surprising
[19:15] <mmcc> ralsina: disabling is a straight up qt bug,  https://bugreports.qt-project.org/browse/QTBUG-25544
[19:15] <ralsina> mmcc: ok, the others sound more or less reasonable as limitations
[19:16] <ralsina> mmcc: ie: while popping the menu, you are probably in a local event loop and can't update stuff
[19:17] <mmcc> hrm. on osx, qsystemtrayicon just uses the NSStatusBar API, and I don't think it's in a separate process
[19:17] <mmcc> you can't update stuff because that code just doesn't handle updates - when the menu is shown, it creates a one-off copy of your QMenu hierarchy as NSMenu's and shows them
[19:17] <ralsina> mmcc: ok, in that case, it does make sense to do it as pyobjc
[19:18] <ralsina> mmcc: you can just reimplement most of systray
[19:18] <ralsina> systray.py that is
[19:18] <mmcc> at the beginning of menuNeedsUpdate:, it clears everything and regenerates all the NSMenuItems… and doesn't give us hooks to change them :(
[19:18] <mmcc> yeah, it should be very similar
[19:19] <mmcc> btw, for what it's worth, dropbox now has a very different non-menu thingy for their notification icon, at least on osx
[19:19] <mmcc> it's in their version 1.5
[19:21] <mmcc> oh, yeah, looking at qsystemtrayicon_mac.mm, no wonder QWidgetAction doesn't work, if you set the widget on an item it'll just get ignored… 'widget' doesn't even appear in that file
[19:21] <mmcc> wish I'd looked at their code before trying it.
[19:21] <mmcc> ok, I'm still going to write up that email, just to have all the pros/cons in the same place
[19:22] <gatox> mmcc, ralsina when you have a moment please: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/not-shared-with-me/+merge/129284
[19:23] <ralsina> gatox: next monday is good? ;-)
[19:23] <ralsina> gatox: j/k on it
[19:24] <gatox> ralsina, is good..... because nowwwwwww.... i'm going to stop proposing branches.... and try to figure it out why tarmac the tests in tarmac are failing in a way that only happend there..... i have 3 u1-cp branches on hold at this momento for that
[19:24] <ralsina> gatox: qt version looks like a reason
[19:25] <gatox> ralsina, but it's the same one i have here
[19:25] <ralsina> gatox: since it seems to be a change in QThread
[19:25] <ralsina> oh, well
[19:25] <gatox> ralsina, and is running on Q.... as i do
[19:37] <gatox> if anyone has a minute, please review this branch (REALLY SMALL)... mike already approve it: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/deflated-size/+merge/129215
[19:40] <mmcc> holy cow it's raining
[19:44]  * gatox tries to find sense to the universe again
[19:47] <ralsina> gatox: in not-shared-with-me ... maybe we could use "ROOT" and whatever UDFs say, instead of just "ROOT"? Would make the tests a bit more robust
[19:47] <gatox> ralsina, ack..... udf says UDF
[19:47] <gatox> ralsina, i'll change one of them to udf
[19:48] <ralsina> gatox: cool, since you usually have 2, one root, one udf is good
[19:48] <gatox> in the name of everything that is good and pure, tell me why! please tell me why!
[19:48]  * gatox lost his temper
[19:48] <gatox> jejejje
[19:48] <gatox> that is related to the bouncing branches
[19:48] <ralsina> gatox: guessed as much
[19:48] <ralsina> gatox: you leave in 38 minutes. Get back to it in the morning.
[19:48] <mmcc> gatox: ralsina I was looking at those strings too. I couldn't find what generates 'SHARE' anywhere -- I see this in volume_manager.py -- self.__dict__['type'] = 'Share'
[19:49] <ralsina> mmcc: ... interesting...
[19:49] <mmcc> I figured my ack skills were failing me. But when I IRL test it, it does work!
[19:49] <mmcc> getting uppercased somewhere I guess?
[19:50] <ralsina> mmcc: yeah, probably
[19:51] <ralsina> 'Root' 'Share' 'Shared' and 'UDF'
[19:51] <gatox> ralsina, the thing with the bouncing branches...... is that i can't reproduce it
[19:51] <mmcc> ok, time for my lunch here
[19:51] <gatox> ralsina, that's my biggest problem
[19:52] <beuno> gatox, a branch is bouncing in tarmac
[19:52] <beuno> but not locally?
[19:52] <ralsina> beuno: yes, tests are failing only on tarmac
[19:52] <beuno> ralsina, gatox, poke sidnei to give you ssh access to tarmac
[19:52] <gatox> beuno, yep..... tests pass here in 2 machines and in mandel's too.... but not in tarmac..... and tarmac is in q (as me) and using the same qt versio
[19:52] <beuno> run manually, debug
[19:52] <gatox> beuno, will do that
[19:53] <ralsina> gatox: best idea would be to just ssh into tarmac ,but I don't have creds
[19:53] <gatox> ralsina, yes...... i'll poke sidnei as beuno said
[19:54] <gatox> ralsina, the branch has been updated to include UDF
[19:58] <ralsina> gatox: awesome, lookging in 1'
[20:04] <briancurtin> YESSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS
[20:05] <gatox> briancurtin, WHAT HAPPEND??!!
[20:05] <gatox> TELL!!
[20:05] <gatox> briancurtin, you figure it out?
[20:05] <briancurtin> gatox: yep
[20:05] <gatox> briancurtin, awesome!!! what was it?
[20:05] <briancurtin> gatox: at least to login with creds. of course this is a one-line change
[20:06] <briancurtin> gatox: has to do with the qt spawn_process inserting the python.exe to the front of the args rather than replacing it
[20:07] <briancurtin> i think it may be better to change this in another spot, but at least its actually working
[20:21] <ralsina> I have to leave for a doctor's appointment, so slightly early EOD for me.
[20:22] <ralsina> See you guys tomorrow!
[20:23] <gatox> ralsina, bye!
[20:34] <gatox> ok..... eod here too....... see yyou tomorrow!
[21:07] <mmcc> briancurtin: hooray. and yeah, that makes sense. it blindly inserts python.exe, huh
[21:08] <briancurtin> mmcc: yep. i think this is going to be the fix but could use a look on mac: http://bazaar.launchpad.net/~brian.curtin/ubuntu-sso-client/correct-subprocess-args/revision/1006
[21:09] <mmcc> briancurtin: well, that certainly looks like it won't change the darwin behavior at all :)
[21:10] <briancurtin> i am shaking my head right now...
[21:11] <briancurtin> i think i need one or more beers
[21:13] <mmcc> yeah, those cases where it takes a week to find a one-line fix - I'm not a real doctor but in this case I can prescribe dark beer.
[21:14] <briancurtin> all i drink are stouts and porters so i think this can be cured
[21:48] <mmcc> hrm, method names like "addItemWithTitle_action_keyEquivalent_" were not made for pep8-compliance
[23:31] <mmcc> ok, time to go