[00:06] <jono_> hey all
[00:07] <jono_> I am having problems authenticating with U1 in a free 12.04 install
[00:07] <jono_> are there issues with the auth server right now?
[04:00] <Arun_> hi I am trying to use ubuntu one on my ubuntu 10.04 live USB... but its not working ..can someome plz help me out here
[04:01] <burrito_> i cannot get ubuntuone to sync my documents folder on a 10.04 machine
[04:03] <Arun_> same prob here.. my ubuntu one displays as disconnetec even after I logged into my A/C in firefox browser
[04:05] <burrito_> my client also doesn't have the tab where i can specify to synchronize a file locally like my linux mint 12 has, on this 10.04 i've only got some crabby little gui with a few generic tabs
[10:04] <mandel> morning all!!
[10:04] <mandel> I managed to fix my machine :)
[10:06] <rye> mandel, congrats! What got broken and will it affect me :) ?
[10:07] <mandel> rye, for some crazy reason in my machine lightdm woul nto let me log in, I installed gdm and tested that I could login, that work, but broke unity
[10:07] <mandel> rye, reinstalling lightdm and later unity fixed the problem.. weird
[10:10] <JamesTait> Good morning all, happy new year!
[10:11] <mandel> JamesTait, same to you!"
[10:28] <mandel> coffee for me :)
[10:53]  * mandel back
[11:05] <Chipaca> JamesTait: happy new year! :)
[11:05] <JamesTait> Chipaca: Why thank you, my good man, and the same to you. :)
[11:59] <nessita> hello everyone!
[11:59] <gatox> good morning!
[12:01] <mandel> nessita, gatox morning!
[12:01] <nessita> hello mandel! gatox!
[12:02] <gatox> mandel, i read that you fix your machine
[12:02] <gatox> nessita, hi
[12:02] <mandel> gatox, yeah, took part of the morning but know I'm full track doing some code :)
[12:02] <mandel> gatox, funny thing is that I could just log in the machie via ssh.. but was enough to fix it
[12:05] <nessita> mandel: oh, what happened?
[12:06] <mandel> nessita, lightdm stopped allowing me to log in my machine after an update in P, so I was left out of the box, I played around a little with gdm and got the login working yet it broke unity badly
[12:07] <mandel> nessita, but after reinstalling unity and forcing lightdm to be reconfigured I got it working, I was lucky I though about trying to use sshing the machine
[12:07] <nessita> mandel: oh
[12:17] <gatox> mandel, review please? https://code.launchpad.net/~diegosarmentero/ubuntuone-windows-installer/default-agree/+merge/87282
[12:18] <mandel> gatox, sure, looking
[12:18] <mandel> gatox, why the unpolish call and then a polish call?
[12:19] <gatox> mandel, that's the qt way to reapply the style to some widget
[12:19] <mandel> gatox, ugh, their api there is ugly, because just doing a single polish does not work, right?
[12:21] <gatox> mandel, well..... they have different purposes and can be called without the need of the other one in some cases
[12:21] <mandel> just read http://developer.qt.nokia.com/doc/qt-4.8/qstyle.html#unpolish got what your are doing gatox  :)
[12:21] <gatox> but...... it would be nice to have a: refreshStyle :P
[12:23] <mandel> gatox, yeah, but if they don't have it, lets better use their default pattern :)
[12:23] <gatox> mandel, yep, obviusly
[12:24] <mandel> gatox, on more question, in the test_license_page_agree_button_style you are just asserting that the button is default and enabled, is there a way we could check that the styles was applied?
[12:24] <mandel> gatox, maybe patching the polish and unpolish calls and asserting on them?
[12:25] <gatox> mandel, that is not going to ensure that the STYLE was applied, just that the widget was repainted...... to check for the STYLE, you would need to check the color of the pixels of the widget.....
[12:25] <gatox> mandel, because if i remove the .qss..... the test is going to pass.....
[12:26] <gatox> but the style is not going to be applied
[12:26] <gatox> so...... it doesn't has sense to me....... but maybe i'm wrong
[12:26] <mandel> gatox, so, why do we repaint the widget, it is to set the correct style, right?
[12:27] <gatox> mandel, if you change a property of the widget, and you have a style associated to that property (in this case "default" and "enabled".... you can create your own properties too)..... you need to repolish the widget
[12:28] <gatox> in order to see the new style based on that property in that widget
[12:29] <mandel> gatox, so, if we remove the polish and unpolish calls, the test will pass yet the style will be wrong, right?
[12:30] <gatox> mandel, yep
[12:30] <mandel> gatox, so there is something wrong in the test, right, since we added 2 important lines and we do not test they are performed
[12:30] <mandel> although I can be completely wrong with my assertion :P
[12:31] <gatox> mandel, i understand what you mean...... i'm thinking....
[12:31] <mandel> gatox, maybe checking the style is not what should be done, but assert that it was repainted
[12:31] <mandel> gatox, and assume that the polish and unpolish call do their work correctly :)
[12:32] <gatox> mandel, yep..... we could check that
[12:32] <mandel> gatox, I think that would be a little better for the test, in case an ass like me says, WTF?! polish and unpolish, I'm removing those!
[12:33] <mandel> gatox, lets call it defense programming against smart asses like manuel hehehe
[12:33] <gatox> mandel, ok, i'll add the checks for that
[12:33] <gatox> jejjeje
[12:33] <mandel> gatox, cool! let me know and I'll approve the branch, since the rest looks great :)
[12:34] <mandel> gatox, I hope is not a major PITA
[12:34] <gatox> mandel, nap.... should patch the wizard.....
[12:35] <gatox> in this case i can't just patch the function or the button
[12:35] <mandel> gatox, how come, issues with the C++ code?
[12:35] <gatox> i can patch the button function from the wizard..... that would be better
[12:35] <gatox> mandel, nono
[12:36] <gatox> mandel, i'm going to patch the button..... patching the button function from the wizard..... that will work just fine
[12:36] <gatox> mandel, in the meanwhile: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/fix-links/+merge/86605 :P
[12:36] <gatox> and nessita ^
[12:36] <gatox> please
[12:36] <nessita> gatox: sure!!!
[12:37] <gatox> thanks
[12:37] <mandel> gatox, on it
[12:37] <gatox> thanks
[12:39] <mandel> nessita, can you help me debug a branch, I have a problem with a dirty reactor and I cant see the reason :(
[12:39] <nessita> mandel: sure!
[12:40] <gatox> nessita, mandel just in case.... i've just fix some docstrings in the branch i gave you for review
[12:40] <nessita> gatox: ack
[12:41] <mandel> nessita, the branch is lp:~mandel/ubuntuone-dev-tools/proxy-testcase, the idea is that the MockServer stops the need of the test case to access the internet, but I get a dirty reactor
[12:41] <mandel> nessita, the revno should be 77
[12:41] <mandel> gatox, and adds new lines hehehe
[12:41] <nessita> mandel: did you disconnected all connected clients?
[12:42] <mandel> nessita, hm.. maybe I have to force the disconnection of the WebClient
[12:42] <gatox> mandel, where?
[12:43] <gatox> mandel, ahhhhhhhhh.... those ones was to fix pep8 issues...... i couldn't resist :P
[12:43] <mandel> gatox, 156 in the diff for example, not that it matters :)
[12:43] <mandel> gatox, hehe
[12:43] <mandel> I though so
[12:43] <nessita> mandel: yeah, you need to disconnect everything that you connect
[12:44] <mandel> nessita, let me check down that path
[12:44] <nessita> mandel: sure
[12:47] <nessita> gatox: your fix links branch has removed some flags from some files
[12:47] <gatox> nessita, :S fixing that...
[12:48] <gatox> nessita, ahhhhhh maybe when i copied the files from one machine to another
[12:48] <nessita> gatox: if you switch between windows and linux, yes
[12:49] <gatox> nessita, i switch between one linux to another..... but i think that the problem was the permissions of that disk
[12:50] <nessita> ah... is it a fat32 FS in the disk?
[12:50] <ralsina> gatox: "copied the files" -- cp doesn't preserve permissions by default
[12:51] <gatox> nessita, from ext4 -> ntfs -> ext4 (migrating my data from one laptop to another)
[12:52] <nessita> gatox: right, that will mess up all your file perms
[12:52] <mandel> gatox, and we are in 2012! you would imaging that fs developers should have fixed that already :(
[12:52] <mandel> maybe the mayas were right..
[12:53] <gatox> mandel, jejeje and we are in 2012 and you will expect that windows handles unicode properly :P
[12:55] <mandel> gatox, and that I would not have to write crazy code to disconnect for a twisted server..
[12:55] <mandel> jesus!
[12:56] <ralsina> nessita or mandel: I would love a second review of this: https://code.launchpad.net/~ralsina/ubuntuone-control-panel/even_more_unique_than_the_other_snowflake/+merge/86912
[12:57] <ralsina> if that gets approved, we are down to 1 (one) manually applied branch on windows releases ;-)
[12:57] <mandel> ralsina, did you make the changes regarding glob, which branch was it so I can take a second look?
[12:57] <ralsina> mandel: yep
[12:57] <ralsina> mandel: let me get you a link
[12:57] <mandel> ralsina, thx
[12:58] <ralsina> mandel: https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_809873/+merge/67911
[12:58] <mandel> ralsina, I'll do that one first :)
[12:59] <mandel> gatox, nessita would be nice to have a better review with mine in: https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_809873/+merge/67911
[13:00] <mandel> ralsina, 'Get the VistaLib32.dll and VistaLib64.dll' where do I get those from?
[13:00] <ralsina> mandel: from the share I shared with you yesterday
[13:01] <mandel> ralsina, ok, and If i don't have that share?
[13:01] <ralsina> mandel: then you are not part of the team and should not be building releases? ;-)
[13:01] <mandel> ralsina, we are opensource, we should provide everything, they might not be making canonical releases :)
[13:01] <ralsina> mandel: those files are not necessary to make the software work, they are just needed to build installers. The installer is not open source.
[13:02] <ralsina> mandel: after all, the installer is done using InstallBuilder. You need a license for that, you know.
[13:03] <mandel> ralsina, you can get an opensource license for that.. I mean, our sources should be complete, if not, if we EVER get people to help use the will be like, WTF? where do I get those
[13:03] <mandel> ralsina, I know it will never happen, yet I know the first person that will say WTF about that, and is the new windows guy :)
[13:03] <mandel> but you are the boss, so you decide what to do :)
[13:03] <ralsina> mandel: well, I am ope to suggestions
[13:06] <mandel> ralsina, hm do we have the code for those dlls?
[13:06] <mandel> ralsina, or is it from somewhere else?
[13:06] <ralsina> mandel: we do
[13:07] <ralsina> mandel: but I have never built them myself. We should add a README for them, too, according to the license
[13:07] <ralsina> mandel: I'll do it right away
[13:07] <mandel> ralsina, yeah, sounds reasonable, point to the code, and explain how to build them
[13:07] <mandel> ralsina, we can use the share to simplify our work :)
[13:13] <ralsina> mandel: pushed URL, explanation, and README. The license is a MIT-style "no guarantees, don't remove these paragraphs"
[13:13] <mandel> ralsina, super! I think this is better since if we get ran over by a bus everyone else will know what to do :)
[13:14] <mandel> ralsina, and we would not be the first ones to be ran over by a vehicle :(
[13:14] <ralsina> In an ideal world, we would reimplement those DLLs but the code is gnarly
[13:15] <nessita> rebooting after big update
[13:16] <mandel> ralsina, yeah, I do understand the fact that we do not have the time or resources, documenting sounds good enough for me :)
[13:16] <mandel> lets hope google cache does the rest hehe
[13:23] <mandel> nessita, I'm going to assume that your update went better than mine
[13:25] <nessita> mandel: I updated to P before xmas, it went well :-)
[13:25] <nessita> mandel: I was updating a kernel now
[13:25] <nessita> (and it also went well, luckly)
[13:26] <mandel> nessita, I did the upgrade, that went smooth, but the update after NYE was what broke it
[13:26] <nessita> mandel: ah, hum
[13:26] <nessita> mandel: it worked here :-D
[13:26] <mandel> nessita, I'm glad for you :)
[13:30] <ralsina> nessita: what do you think of https://bugs.launchpad.net/ubuntuone-windows-installer/+bug/910834
[13:30] <ubot4> Launchpad bug 910834 in ubuntuone-windows-installer (and 3 other projects) "The bin/ scripts should have no logic in them (affects: 1) (heat: 6)" [Undecided,New]
[13:31] <nessita> ralsina: looking
[13:32] <nessita> ralsina: what would be "no logic" in that bug report? cuz I see some scripts in bin/ and those are pretty logic-less
[13:33] <ralsina> nessita: script should just be "from whatever import foo; foo.bar(sys.argv)"
[13:33] <ralsina> nessita: we have option processing, for instance
[13:33] <ralsina> nessita: maybe not even passing sys.argv ;-)
[13:34] <ralsina> u1sdtool is far from logic-less, too
[13:34] <nessita> ralsina: agreed on client  bin/  scripts, but does it worth the time but the rest, that has the option parse code there?
[13:34] <ralsina> nessita: sorry, I don't understad
[13:35] <nessita> ralsina: I think is a good idea to do that, specially on the u1client bin/ scripts. But from my POV, I'd say is low priority
[13:35] <nessita> (given the things we have queued up)
[13:36] <ralsina> nessita: I agree it's not a high priority
[13:36] <ralsina> nessita: it does make windows releases considerably more work, though
[13:36] <nessita> ralsina: I read from the bug report that the .exe will not have to be resign, but I'm not sure I follow that
[13:37] <nessita> ralsina: the bin/foo may not change, but all the other pys will
[13:37] <nessita> so why there would be no need of signing?
[13:37] <ralsina> nessita: the exes only contain a loader and bin/whatever's bytecode
[13:37] <ralsina> nessita: the rest is in library.zip
[13:38] <nessita> hum, I, as a end user, would expect that the sign covers everything, included the library.zip, otherwise what guarantee canonical gives me that the library.zip is authenthic?
[13:39] <ralsina> nessita: the second signing is for that
[13:39] <ralsina> the whole installer is signed
[13:40] <nessita> ralsina: sorry, but now you loose me (evidently I'm not familiar with that process). So, we have 2 signs? (I did not know that)
[13:40] <ralsina> but also individual .exes need signing. So right now, it takes about 12 hours and three large uploads to get everything signed. And that can go down to one upload and whatever hours it takes IS to sign
[13:40] <mandel> ok, I'm off to lunch
[13:40] <mandel> bbl
[13:40]  * mandel -> lunch
[13:41] <ralsina> nessita: yes, first all the .exes (and python27.dll) need to be signed. Then the installer is built with the signe files in it, and the installer is signed again.
[13:41] <ralsina> nessita: on windows, you have to sign every .exe, and the installer is an .exe too
[13:41] <nessita> ralsina: I see. So we'd need to sign every exe only when they change?
[13:41] <ralsina> nessita: yes
[13:41] <nessita> and are they changing too much lately?
[13:42] <ralsina> nessita: well, at least windows-installer has.
[13:42] <ralsina> nessita: and u1sdtool
[13:43] <ralsina> and if we ever make u1sdtool --version do something useful, then that will change every version
[13:43] <nessita> ralsina: we can workaround that importing the version from clientdefs, like we do now, no?
[13:44] <nessita> anyways, I see your point. From my POV, I would advice tackling that after feature freeze, but if you consider it more important than the thing we're doing, we can ceratinly schedule it
[13:44] <ralsina> nessita: it can wait until FF, I just wanted to get your OK about doing it
[13:44] <nessita> ralsina: makes sense :-)
[13:53] <nessita> gatox: you let me know when you fix the perms in the branch?
[13:54] <gatox> nessita, ah sorry..... it's already fixed
[14:00] <nessita> ralsina: ping
[14:00] <ralsina> nessita: pong
[14:01] <nessita> ralsina: the code you added to https://code.launchpad.net/~ralsina/ubuntuone-control-panel/fix_906541/+merge/86911 is not a good one, os.path.expanduser('~') does not work well for usernames with non-ascii characters in it
[14:01] <ralsina> nessita: it was used in the same file, IIRC
[14:01] <ralsina> nessita: let me check
[14:01] <nessita> ralsina: gatox implemented a home variable in sso.xdg_base_directory
[14:02] <nessita> ralsina: may be, but we need to fix that (gatox hasn't reached controlpanel yet)
[14:02] <ralsina> nessita: I'll do a new branch and fix all uses on that file, then
[14:02] <nessita> ralsina: well, if it's already used somewhere else, we can wait until diego reach controlpanel to fix it?
[14:02] <ralsina> nessita: I am checking that
[14:03] <gatox> ralsina, nessita, when my last branch of u1-client is approved.... i have another one to propose for controlpanel that should fix the problems in folders
[14:03] <nessita> ralsina: I just wanted to let you know we should not be using that anymore (so is basically a FYI)
[14:03] <nessita> gatox: right, you should also cover in your branch what ralsina added in https://code.launchpad.net/~ralsina/ubuntuone-control-panel/fix_906541/+merge/86911
[14:04] <ralsina> nessita: ack!
[14:04] <gatox> nessita, ah ok!
[14:04] <nessita> ralsina, gatox: thanks!
[14:23] <nessita> dobey: hum, I wanted to ask here, sorry
[14:23] <dobey> oh
[14:35] <nessita> ralsina: were you able to run the linux tests for your is_root branch?
[14:36] <ralsina> nessita: didn't try it since I touched nothing of the linux side
[14:36] <ralsina> nessita: it fails?
[14:36] <nessita> ralsina: can you please try?
[14:36] <nessita> ralsina: no, I'm getting an error re-translations, and I want to see if it's a precise-only thing
[14:36] <ralsina> nessita: sure
[14:37] <ralsina> nessita: I'll try on natty
[14:42] <alecu> uh... hello...
[14:42] <alecu> sorry guys I'm so late today.
[14:43] <alecu> the first two weeks of january there's no kinder, so they are a bit... complicated.
[14:43] <dobey> heh
[14:43] <dobey> young cylons are so troublesome!
[14:45] <alecu> lol
[14:45] <nessita> hello alecu! I missed you
[14:45] <alecu> hola nessita!
[14:48] <dobey> it is too damn cold. :(
[14:50] <ralsina> ayone has the "setting up the windows development environment" wiki page handy?
[14:51] <gatox> ralsina, this one? https://wiki.ubuntu.com/UbuntuOne/Contribute/WindowsTesting
[14:51] <gatox> ralsina, i updated the wiki today adding the comtypes dependency
[14:51] <ralsina> gatox: thanks
[14:52] <ralsina> gatox: if I am lucky, that whole page will re replaced with two commands in a few days :-)
[14:52] <dobey> so few weeks then ;)
[14:52] <gatox> ralsina, that would be awesome!!!..... and a lot easier to attract people to contribute too
[14:53] <ralsina> gatox, dobey: well, the main missing bit is a buildout recipe for PyQt, which is kind of a pain
[14:54]  * mandel back
[14:54] <mandel> alecu, dobey hello :)
[14:54] <dobey> hola mandel
[14:54] <ralsina> OTOH, I do have a "full dev. env. in a zip" kinda working
[14:55] <mandel> dobey, I'm fixing the proxy-testcase code, and while I was implementing the changes you suggested I've decided to change the test so they do not use a real url
[14:55] <dobey> cool
[14:55] <mandel> dobey, there are a few things to change related to cleaning the reactor, as soon as I'm done I'll let you know :)
[15:00] <ralsina> nessita: tests for my branch are running OK on natty
[15:01] <nessita> ralsina: right, seems like something related to translations is broken on precise
[15:01] <dobey> it's because of the new intltool
[15:01] <nessita> dobey: shall we add the clientdefs.py in the .skip file?
[15:01] <gatox> standup?
[15:01] <nessita> ralsina: another question, did you run the tests from your branch in windows?
[15:01] <nessita> gatox: yes!!!
[15:02] <gatox> me
[15:02] <nessita> me
[15:02] <dobey> nessita: i'll get back to you in a minute; i just finished upgrading my laptop's packages, about to test it there
[15:02] <ralsina> nessita: those, yes, I ran:)
[15:02] <nessita> dobey: ack
[15:02] <ralsina> me
[15:02] <ralsina> DONE
[15:02] <ralsina> oops
[15:02] <dobey> me
[15:02] <nessita> ralsina: and is syncdaemon starting for you? I'm getting
[15:02] <nessita> E:\client\review_fix_904300>python bin\ubuntuone-syncdaemon
[15:02] <nessita> Traceback (most recent call last):
[15:02] <nessita>   File "bin\ubuntuone-syncdaemon", line 33, in <module>
[15:02] <nessita>     from ubuntuone.platform import (
[15:02] <nessita> ImportError: No module named platform
[15:02] <nessita> alecu, mandel: standup?
[15:03] <mandel> ups, sorry
[15:03] <mandel> me
[15:03] <ralsina> nessita: I'll check in a few minutes
[15:03] <nessita> DONE: holidays! managed to avoid touching the computer for 10 days :-D
[15:03] <nessita> TODO: catch up, reviews
[15:03] <nessita> BLOCKED: nopes
[15:03] <nessita> NEXT: gatox
[15:03] <gatox> DONE:
[15:03] <gatox> Fixed License page in Installer, fixed problem with dirty reactor in cp, tests improved for some of my branches, couple of reviews.
[15:03] <gatox> TODO:
[15:03] <gatox> Keep fixing bugs.
[15:03] <gatox> BLOCKED:
[15:03] <gatox> No
[15:03] <alecu> me
[15:03] <gatox> mandel, go
[15:03] <gatox> sorry
[15:03] <gatox> dobey, go
[15:03] <dobey> ralsina
[15:04] <ralsina> DONE: reviews, started process or january mini-sprint, canonicaladmin cleanup, bug triaging, started on a buildout config to make windows dev.env. repeatable TODO: calls, reviews, etc. BLOCKED: no
[15:04] <mandel> DONE: Fixed machine, cleaned inbox from mail. Reviews for gatox and ralsina (please let me know when you are done with the changes).
[15:04] <mandel> TODO: Finish fixing proxy-testcase branch to no use a real url and use a localhost one.
[15:04] <mandel> BLOCKED: no
[15:04] <ralsina> dobey?
[15:04] <dobey> λ DONE: holidays
[15:04] <dobey> λ TODO: fix crashers in P, releases, fixes for gireactor from review, initial music store work
[15:04] <dobey> λ BLCK: No.
[15:04] <dobey> then me, then mandel
[15:04] <dobey> now alecu
[15:04] <mandel> dobey, sorry, I show my name and pasted my notes
[15:04] <ralsina> mandel: I have no changes pending
[15:04] <mandel> ralsina, true, sorry, I'll continue the review in a few mins
[15:05] <alecu> DONE: force use proxy in libsoup based webclient
[15:05] <alecu> TODO: use webclient in sso
[15:05] <alecu> BLOCKED: no
[15:05] <nessita> any comments anyone?
[15:05] <alecu> ralsina, january mini-sprint ?
[15:05] <gatox> mandel, updated: https://code.launchpad.net/~diegosarmentero/ubuntuone-windows-installer/default-agree/+merge/87282
[15:06] <ralsina> alecu: syndaemon knowledge transfer and new hire welcome
[15:06] <alecu> oh, and BLOCKED: on mandel's devtools branch!  :-)
[15:06] <ralsina> mini in the sense that it's not the whole team
[15:06] <alecu> ralsina, cool. Any idea on where & when & who?
[15:06] <mandel> alecu, should be ready in a few mins, as soon as it is done, I'll ask you and dobey for a review
[15:07] <alecu> mandel, vamos!!!! :-)
[15:07] <dobey> review will take a while i think :(
[15:07] <ralsina> alecu: nessita, facundobatista, Brian Curtin all week, you and I a couple of days
[15:07] <nessita> ralsina: yes, I'm interested in the when so I can book tickets, if you tell me to
[15:07] <ralsina> nessita: marianna is on swap today, I scheduled it for last week of january but I need to talk t o her
[15:07] <nessita> ack
[15:08] <nessita> ralsina: can brian make it in that week?
[15:08] <ralsina> nessita: yes
[15:08] <nessita> nice
[15:08] <nessita> anyone: I can't start ubuntuone-syncdaemon using u1client\trunk. Can anyone try that, please?
[15:09] <nessita> (that's in windows)
[15:09] <ralsina> nessita: starts here. Maybe you didn't set PYTHONPATH?
[15:09] <nessita> ralsina: I did... http://pastebin.ubuntu.com/791728/
[15:09] <dobey> nessita: ok, yes the 2 clientdefs.py files need to be added to POTFILES.skip
[15:10] <nessita> dobey: ok, I'll propose a branch
[15:10] <nessita> ralsina: hum, I see my mistake
[15:10] <ralsina> yes, ":"
[15:10] <nessita> yes
[15:12] <nessita> alecu: would you have some time to help me with some kinda-integration tests that use dbus?
[15:12] <nessita> alecu: they are driving me crazy, and I (think I) checked everthing
[15:13] <alecu> nessita, for you, always :-)
[15:13] <nessita> alecu: let me know when you have 15-13 minutes (no rush, I have a lot of reviews to do still)
[15:13] <alecu> nessita, can you point me at the code?
[15:13] <nessita> yes, one sec
[15:16] <nessita> alecu: the test suite is test_clients.py. The goal of the suite is to provide multiplatform tests for all the IPC clients that access the sso service, both for the SSOLogin and Credentials sub-services. The code is: http://bazaar.launchpad.net/~nataliabidart/ubuntu-sso-client/unify-signal-broadcaster/view/head:/ubuntu_sso/main/tests/test_clients.py
[15:17] <nessita> alecu: note that SSOLoginProxyTestCase is skipped, since it does not work on linux. You can try running a single test with:
[15:18] <nessita> u1trial -t GenerateCaptchaTestCase.test_success ubuntu_sso/main/tests/test_clients.py
[15:20] <alecu> nessita, but that gets skipped on linux, so I should try that on windows, right?
[15:20] <nessita> alecu: on windows they work perfectly, the problem is linux and dbus
[15:20] <nessita> alecu: UbuntuSSOService, on linux, will start the dbus service, and the get_sso_client will give you an object with clients connected to that dbus service
[15:21] <nessita> alecu: you should remove the .skip and run the test I mentioned, you'll notice it never finishes
[15:21] <nessita> because the dbus call timeouts
[15:21] <alecu> nessita, ok, but the u1trial line you pasted above runs one test that gets skipped.
[15:21] <alecu> ok
[15:21] <alecu> ah
[15:21] <alecu> ok
[15:22] <nessita> alecu: right, you should remove the SSOLoginProxyTestCase.skip = 'foo' line
[15:23] <nessita> alecu: the error I'm getting is ERROR:dbus.proxies:Introspect error on com.ubuntu.sso:/com/ubuntu/sso/accounts: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply. But if I pdb into the setUp, after the UbuntuSSOService() is created, and I connect to the test dbus address session, I see the service running
[15:23] <dobey> nessita: btw, contrib/testing/testcase.py also needs to be added to POTFILES.skip in trunk I think
[15:24] <alecu> nessita, my gut tells me something is not being called async
[15:24] <alecu> looking anyway
[15:25] <alecu> nessita, just so I don't forget: I don't like the "import BaseTestCase as _BaseTestCase"; I think you should rename the BaseTestCase class defined below instead.
[15:25] <nessita> alecu: yeah, absolutely, this branch is on-going
[15:25] <alecu> ack
[15:27] <alecu> nessita,         getattr(self.client, self.method)(*self.params)
[15:27] <alecu> nessita, that bit calls the dbus method, right?
[15:27] <nessita> alecu: yes
[15:27] <alecu> nessita, but is it calling the method in an async way?
[15:28] <nessita> ... no
[15:28] <alecu> right
[15:28]  * alecu googles
[15:28] <nessita> but... it should not block
[15:28] <nessita> the method should return almost immediately -- perhaps the reactor is not giving a "turn"?
[15:28] <alecu> "To make a call asynchronous, pass two callables as keyword arguments reply_handler and error_handler to the proxy method"
[15:29] <alecu> nessita, right now the call is blocking, so the reactor gets no turn to handle the reply from your mock objects
[15:29] <nessita> alecu: I see, so we need the async for the tests... even though we don't need them IRL
[15:30] <alecu> nessita, we *should* have them async too
[15:30] <nessita> alecu: right. Thanks *a lot*, I will continue from here
[15:30] <alecu> nessita, no problem :-)
[15:31]  * alecu adjusts his dbus hat, and rides silver into the sunset.
[15:32] <nessita> lol
[15:32] <alecu> nessita, btw: http://dbus.freedesktop.org/doc/dbus-python/doc/tutorial.html#making-asynchronous-calls
[15:33] <nessita> alecu: I certainly know how to make async calls with dbus, is just that I did not want that in the API...
[15:33] <nessita> in the multiplatform, unified API
[15:33] <nessita> alecu: it complicates the PB code a lot
[15:35] <alecu> nessita, right. Then perhaps your code should add the async handlers... and have the sucess_handler be a No-op, but the error_handler should raise the error signal
[15:35] <alecu> nessita, and the calling code should not worry about adding that.
[15:37] <nessita> alecu: i did not understand the "the sucess_handler be a No-op, but the error_handler should raise the error signal". By 'success_handler' you mean the  cb = lambda *a: d.callback(a)?
[15:40] <alecu> nessita, no, I don't mean the signal handler, I meant the reply_handler that's used to call the dbus proxy method and make it async
[15:40] <alecu> nessita, but I'm thinking of another approach that might be more similar to the pb call
[15:41] <gatox> people.... i'm going to leave now and resume my work tonight.... i'm not feeling very well right now :S
[15:41] <nessita> alecu: right, but then I need the (windows) PB API to also accept the keyword reply_handler and error_handler... and call them somehow (given that the tests are multiplatform)
[15:41] <nessita> gatox: oh, is it the heat?
[15:41] <gatox> nessita, yep..... i couldn't sleep very well last night
[15:42] <alecu> gatox, :-(
[15:42] <nessita> gatox: me too :-/
[15:42] <ralsina> gatox: take care.
[15:42] <nessita> gatox: get better!
[15:42] <gatox> see you later
[15:42] <alecu> gatox, stop spending money on computers, get an air conditioner!
[15:42] <gatox> alecu, jejeje tomorrow i'm going to carrefour to but an air conditioner
[15:43] <gatox> i'm not going to survive the summer
[15:43] <alecu> gatox, don't buy a portable one, they are a piece of crap
[15:43] <ralsina> that heat was surely the alienware's fault
[15:43] <gatox> ralsina, jejejeje
[15:43] <gatox> alecu, roger that
[15:43] <ralsina> gatox: +1 to what alecu said
[15:43] <ralsina> the bad side is, you gonna need an installer
[15:44] <mandel> alecu, one question, I'm using a mock server similar to the one you used in the webclient tests, is there a way to access to the protocol that is used, I want to know when the connection is lost so that I do not get dirty reactor issues
[15:44] <gatox> ralsina, but the installer now is only just one script...... jejejeje deliraba
[15:44] <gatox> ok....... gatox off..... see you later
[15:45] <alecu> nessita, I think I found a better way.... give me a few minutes
[15:46] <nessita> alecu: I give you whatever you need to have that cleverness flowing :-D
[15:46] <alecu> mandel, in my experience, all the dirty reactor issues mean that there was some function that was not called; so it would not be good to wait till the connection is lost
[15:47] <alecu> nessita, ok, here's my idea:
[15:47] <alecu> nessita, we should stop using the dbus "proxies"
[15:47] <dobey> nessita: you are doing releases today as well, right?
[15:47] <alecu> nessita, both dbus proxies and pb proxies
[15:47] <alecu> nessita, and we should use the calls by name instead
[15:47] <alecu> by string name I mean
[15:48] <nessita> dobey: ah, I wanted to talk you about that. I'm not confident on today's trunks to make releases, I haven't been using them. Unless there a specific change to backport, I would release in 2 weeks
[15:48] <alecu> for instance, dbus.call_async("method_name") instead of dbus.method_name(...)
[15:48] <alecu> nessita, ^ and the same for pb
[15:48] <nessita> alecu: we can call by string name in ob?
[15:48] <nessita> pb*
[15:49] <mandel> alecu, I'm getting to the point where my second HttpClientFactory is used to retry the connection after a 407, and then I get a dirty reactor, with the first request only, and getting a 407 or 401 it works..
[15:49] <nessita> dobey: I read about the libsyncdaemon issue, I'm +1 to release that. Are you aware of any other important fix?
[15:49] <alecu> it's the default way to do it in pb; there's a bunch of crap our code does with metaclasses so it looks more like dbus proxies.
[15:49] <alecu> nessita, ^
[15:49] <alecu> mandel, can you point me at the code?
[15:49] <mandel> alecu, sure, let me clean it a little
[15:50] <dobey> nessita: there is a fix in sso that needs to go out; and i think we should make the releases anyway, regardless of what we think is or isn't important to get out. doing releases even if there are no changes, gets us in the habit of releasing on the schedule
[15:51] <dobey> nessita: so for me "releasing on schedule" is an important fix :)
[15:51] <nessita> dobey: I agree to that, but I'd also would like to update stables-3 from trunk, and I don't have the confidence to  do it. DO you?
[15:51] <dobey> sure
[15:52] <dobey> what has changed in the last week with nobody around to work on anything? :)
[15:53] <dobey> maybe sso and cp got a couple changes last week; but for the most part, nothing's changed from the release 2 weeks ago
[15:53] <nessita> dobey: I'll check
[15:54] <alecu> nessita, http://dbus.freedesktop.org/doc/dbus-python/api/dbus.connection.Connection-class.html#call_async and http://twistedmatrix.com/documents/11.0.0/api/twisted.spread.pb.RemoteReference.html#callRemote
[15:54] <nessita> alecu: I ll look into that, thanks!
[15:54] <alecu> nessita, both call_async and callRemote take the method name as a parameter
[15:56] <dobey> in u1client, there were 4 changes since last release, and they are all generally quite small; (plus the fixes i'm about to propose for the crashes)
[15:56] <mandel> alecu, the code is in lp:~mandel/+junk/proxy-testcase if you run the tests you will see where it fails
[15:56] <mandel> alecu, I would really appreciate a hand on this..
[15:57] <dobey> there are 3 changes in cp, though not sure what the scale is
[15:57] <dobey> i think they are also quite small
[15:58] <dobey> and sso is 2 very small changes as well
[15:58] <dobey> so i don't think there's any reason to not be confident :)
[16:00] <nessita> dobey: ack
[16:01] <alecu> mandel, looking right now
[16:01] <mandel> alecu, thx!
[16:01] <nessita> dobey: I will do the releases for sso, controlpanel, installer. Want me to do something else? protocol?
[16:01] <ralsina> I have to go have a long lunch (relatives visiting) will be back in 90 minutes, I hope
[16:02] <alecu> mandel, it's the u1-dev-tools branch, right?
[16:03] <mandel> alecu, yeah, it is in lp:~mandel/+junk/proxy-testcase where I'm looking for fixes :)
[16:03] <dobey> nessita: installer == windows-installer?
[16:03] <nessita> yeap, sorry
[16:03] <dobey> ok
[16:04] <alecu> mandel, btw: the squid(2|3).conf.in files are very similar, right?
[16:04] <mandel> alecu, yeah, there are some minor diffs between them, but that if they are there, the squi3 wont work
[16:04] <alecu> mandel, your branch has some pdbs; but it seems to pass all tests here.
[16:04] <mandel> alecu, really?
[16:05] <mandel> alecu, not in my case.. so there is something wrong for sure, I'll keep looking
[16:05] <alecu> mandel, http://pastebin.ubuntu.com/791786/
[16:06] <mandel> alecu, ok, I'll see wtf is going on in my machine
[16:06] <alecu> mandel, this is oneiric, let me try on the precise vm
[16:06] <alecu> mandel, I'll need to update it first
[16:06] <mandel> alecu, thx
[16:06]  * alecu tried updating his laptop to precise yesterday, but the updater kept failing
[16:08] <nessita> dobey: where in the system should be located thye pylintrc shipped by u1devtools?
[16:08] <dobey> alecu: it worked fine there
[16:08] <dobey> nessita: on linux, or on windows?
[16:08] <nessita> dobey: linux
[16:08] <alecu> mandel, where is "there" ?
[16:08] <alecu> sorry...
[16:08] <nessita> dobey: I'm getting "bad" errors such as
[16:08] <alecu> I mean:
[16:08] <nessita> ubuntu_sso/gtk/tests/test_gui.py:
[16:08] <nessita>     34:  [F0401] Unable to import 'ubuntuone.devtools.handlers'
[16:08] <nessita>     34:  [E0611] No name 'devtools' in module 'ubuntuone'
[16:08] <dobey> nessita: /usr/share/ubuntuone-dev-tools/pylintrc i think is where it is
[16:08] <alecu> dobey, where is "there" ?
[16:08] <nessita> dobey: and I have nightli4es installed and updated
[16:09] <mandel> alecu, he, I was like wtf, spain :P
[16:09] <dobey> nessita: ah, yes; new pylint in precise i think :(
[16:09] <dobey> alecu: sorry "here" not "there" :)
[16:09] <nessita> dobey: and is that file being loaded in our u1lint?
[16:09] <alecu> dobey, cool
[16:09] <dobey> yes
[16:09] <dobey> nessita: the problem is new pylint is being more evil i think
[16:10]  * dobey *really* wants to kick pylint to the curb
[16:10] <nessita> dobey: didn't we have a patch of our own for this?
[16:10] <dobey> nessita: i think for a different problem we had a patch?
[16:11] <nessita> dobey: as far as I recall, the patch was for this. And that may be the issue, since I just confirmed that python-logilab-common is 0.57.1-1ubuntu1 in my computer
[16:11] <dobey> nessita: and if we have to maintain a patch for every new release of pylint, to be able to use it, then even more reason to just not use it.
[16:11] <alecu> mandel, can you paste a run of that branch on your machine?
[16:11] <nessita> dobey: and I should have the package from our ppa, no?
[16:12] <mandel> alecu, sure, give me a sec
[16:12] <dobey> nessita: probably not
[16:12] <nessita> dobey: why not?
[16:12] <dobey> nessita: did you forward port the patch to the new version and get it built in the PPA?
[16:12] <nessita> dobey: no, I'm not sure how to do that, but if you teach me, I'm happy to do it
[16:13] <mandel> alecu, http://pastebin.ubuntu.com/791800/
[16:13] <dobey> nessita: i think we should just move off pylint :)
[16:13] <nessita> dobey: no until we have a *better* alternative
[16:13] <nessita> (and atm, pyflakes is not a better alternative)
[16:13] <dobey> pyflakes is better :)
[16:14] <alecu> mandel, try adding this line to some test file: twisted.internet.base.DelayedCall.debug = True
[16:14] <nessita> dobey: I really don't want to have this talk now, it leads nowhere. Would you please teach me how to do the update of the logilab-common package?
[16:14] <alecu> mandel, globally
[16:14] <mandel> alecu, sure, I was doing that but commented it out :)
[16:14] <mandel> alecu, let me run that again
[16:14] <alecu> mandel, cool
[16:15] <mandel> alecu, here: http://pastebin.ubuntu.com/791806/
[16:16] <mandel> alecu, some of the passing tests are not there due to the buffer size, but the important errors are
[16:16] <alecu> dobey, you ran mandel's test on precise, right?
[16:17] <dobey> nessita: apt-get source the new version, grab the .debian.tar.gz for the older nightlies package, pull the patch file out and put it in the new package, fix the debian/changelog, make sure it applies and builds correctly, test it, and upload it to the PPA
[16:17] <alecu> mandel, the weird lines in the previous pastebin are those above, the ones where squid complains about some process that's missing
[16:17] <dobey> alecu: mandel's test?
[16:17] <nessita> dobey: ack, thanks!
 alecu: it worked fine there
[16:17] <dobey> alecu: updgrading to precise yesterday, worked fine for me, here
[16:18] <alecu> dobey, doh!
[16:18] <dobey> alecu: i upgraded my other laptop yesterday
[16:18] <mandel> alecu, squid is complaining about the ip config, not about processes missing :)
[16:18] <mandel> alecu, for example: 'WARNING: (B) '::/0' is a subnetwork of (A) '::/0''
[16:18] <alecu> mandel, then the only difference is you are using precise and squid3, and I'm on oneiric and squid2
[16:18] <dobey> alecu: i haven't run the squid branch on anything yet; all my reviews so far of it, were purely reading the diff :)
[16:19] <mandel> dobey, can you try to pull from lp:~mandel/+junk/proxy-testcase
[16:19] <mandel> dobey, and see if they pass in P ?
[16:19] <mandel> alecu, yeah, but since it about async stuff maybe I'm just hitting the error more often, ut you will eventually
[16:19] <dobey> not right now; i am doing something else and about to go get lunch :)
[16:20] <alecu> mandel, in the previous paste:
[16:20] <alecu> squid: ERROR: Could not send signal 0 to process 6300: (3) No such process
[16:20] <alecu> Waiting for squid to start..
[16:20] <alecu> (signal 0? wtf!)
[16:21] <mandel> alecu, yeah, is a test to ensure that the process is running, is a command line tool from squid, the message is kinda bad
[16:21] <dobey>        0          0   n/a       exit code indicates if a signal may be sent
[16:21] <dobey> (from man kill)
[16:21] <mandel> alecu, it sends a signal to the process asking if it is ready, if not we wait
[16:24] <alecu> mandel, I think I found the issue
[16:24] <mandel> alecu, if you did, I could love you :)
[16:24] <alecu> mandel, squid 3 knows how to speak http/1.1, but squid 2 does not.
[16:25] <nessita> dobey: could you please trivial-review https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/stable-3-0-update-2.99.1/+merge/87378 and https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/stable-3-0-update-2.99.1/+merge/87379
[16:25] <mandel> alecu, la puta de oros!!!
[16:25] <alecu> mandel, that means that squid 3 will try to keep connections open to web servers, to send more than one request in a given connection.
[16:25] <alecu> mandel, also that means that the mock web server in twisted does not have it's connection terminated.
[16:26] <alecu> mandel, hence the error.
[16:26] <mandel> alecu, yeah, once you mentioned 1.1 it made sense
[16:26]  * alecu is waiting for the link to the picture of "la puta de oros"
[16:27] <mandel> alecu, http://86400.es/2009/02/24/la-puta-de-oros/
[16:27] <alecu> "Squid trunk claims HTTP/1.1 support. Squid v3.1 claims HTTP/1.1 support but only in sent requests (from Squid to servers). Earlier Squid versions do not claim HTTP/1.1 support by default"
[16:27] <alecu> http://wiki.squid-cache.org/Features/HTTP11
[16:28] <alecu> "una mujer blanca, de cabello rubio o entrecano, generalmente joven, de una muy buena posición económica pero descontrolada a la hora de gastar dinero, generosa, apasionada, leal, pero soñadora e imaginativa como pocas."
[16:28] <alecu> Madonna?
[16:29] <dobey> nessita: cp one needs small fix :)
[16:29] <mandel> alecu, hehehe could be, but I always think of the card :P
[16:29] <nessita> dobey: looking
[16:29] <alecu> "¿Quien no ha escuchado esta expresión alguna vez? Bien usada es en España y lo relajado que se queda uno al decirla." -> lol!
[16:30] <nessita> dobey: ah, we should ask ralsina_lunch to change his whoami
[16:30] <dobey> yes that too
[16:30] <nessita> dobey: MP commit message updated
[16:31] <mandel> alecu, so, knowing that, how the hell do I force the connection to be lost so that I get a clean reactor..
[16:31] <mandel> alecu, this is like a dejabu, again the same problem with closing connections
[16:32] <alecu> mandel, the way I'm doing it in my proxy branches is by forcing the disconnection with a header...
[16:32] <alecu> let me find it.
[16:33] <mandel> alecu, please
[16:34] <nessita> dobey: want me to do all the protocol release process up to upload the tarball?
[16:34] <dobey> nessita: if you want to sure; it doesn't have any changes does it?
[16:34] <nessita> dobey: haven't checked, looking
[16:36] <nessita> dobey: right, no change. So no new tarball?
[16:37] <dobey> nessita: sure; version bump, bzr tag, and tarball
[16:37] <nessita> ack
[16:37] <nessita> dobey: I ll do that when doing it for the rest (right now I'm at building stable-3-0 update branches)
[16:39] <dobey> nessita: great, thanks; please prioritize uploading the sso release first. :)
[16:39] <nessita> dobey: of course
[16:40] <nessita> dobey: also, when you have some minutes https://code.launchpad.net/~nataliabidart/ubuntuone-client/ignore-po/+merge/87383 (I targeted it to 2.99.1 as well)
[16:42] <mandel> alecu, found it? need some help?
[16:43] <dobey> nessita: and i have https://code.launchpad.net/~dobey/ubuntuone-client/fix-crashes/+merge/87382 :)
[16:43] <nessita> dobey: looking
[16:43] <dobey> and i am off to get some lunch :)
[16:43] <dobey> bbiab
[16:44] <nessita> me too
[16:44] <alecu> mandel, I can't find it. I'm pretty sure I did something like that... but perhaps I solved it in a cleaner way :P
[16:45] <mandel> alecu, hehehe I love when we have this kind of problems, make me feel stupid :P
[16:45] <mandel> alecu, which branches are you looking at so that I can look at the code
[16:49] <alecu> mandel, this is one of the branches I'm looking at: https://code.launchpad.net/~alecu/ubuntu-sso-client/proxy-integration-tests
[16:49] <alecu> I'm pushing newer stuff to that branch.
[16:50] <alecu> mandel, anyway: IIRC, that branch uses the qt4network http client, that likes to use HTTP/1.1
[16:50] <mandel> alecu, ok, I'll take a look then
[17:08] <nessita> lunchtime!
[17:08] <nessita> and another reboot
[17:33] <mandel> alecu, I sometimes really hate twisted: http://paste.ubuntu.com/791892/
[17:33] <mandel> dobey, ^
[17:34] <dobey> mandel: that's code you wrote, or that's code in twisted?
[17:34] <mandel> dobey, from twisted
[17:35] <mandel> dobey, I'm a bad person, but not that bad
[17:35] <dobey> lovely
[17:35] <dobey> yeah, twisted is nasty
[17:35] <alecu> mandel, that's metaprogramming the documentation, right?
[17:35] <dobey> but it's python, so that generally goes without saying ;)
[17:36] <mandel> alecu, from here: http://twistedmatrix.com/trac/browser/tags/release-2.0.1/twisted/application/internet.py
[17:37] <ralsina> nessita: what's wrong with my whoami?
[17:37] <mandel> alecu, I understand is doing something else
[17:37] <alecu> mandel, "This module (dynamically) defines various Service subclasses that let you represent clients and servers in a Service hierarchy."
[17:37] <mandel> alecu, and I hate them for that :)
[17:38] <alecu> mandel, I think the bit of code you showed is more understandable than 11 classes that are very similar but slightly different.
[17:40] <mandel> alecu, I'm sure something nicer could be done between the 2 extremes..
[17:40] <dobey> mandel: rewrite it in vala? ;)
[17:49] <alecu> dobey, ralsina: do you guys know what would be the best way to read gsettings from a pyqt application?
[17:49] <ralsina> alecu: not the slightest
[17:50] <dobey> alecu: uhm; use the gsettings API?
[17:50] <alecu> I'm looking at libdconf-qt, but I've no idea how to check if there's python bindings
[17:50] <alecu> dobey, would it be safe to use it from a qt mainloop?
[17:51] <dobey> alecu: should be, mostly
[17:51]  * alecu is tempted to use os.system("gsettings get ...")
[17:51] <dobey> if there's a qt api though, use it instead
[17:52] <dobey> alecu: why do you need to even?
[17:52] <mandel> alecu, 'cause you cannotuse dbus, right?
[17:52] <ralsina> alecu: using os.system is discouraged. Other than that, looks simple enough ;-)
[17:52] <alecu> dobey, to put the ubuntu proxy settings into the qt proxy configuration object
[17:53] <alecu> ralsina, yup, not "os.system", but "subprocess.something"
[17:53] <ralsina> alecu: there is a dconf-qt which unity-2d uses
[17:53] <dobey> ralsina: unity-2d isn't written in python
[17:53] <ralsina> alecu: but I am not sure of python bindings for it
[17:53] <alecu> ralsina, right, that's the project I'm looking at, and I see that it defines some bindings for QML
[17:54] <alecu> ralsina, so I was wondering if some of those bindings allowed that classes to be exported to other languages, like GI does
[17:54] <nessita> ralsina: could I please have a review for releasing windows-installer? https://code.launchpad.net/~nataliabidart/ubuntuone-windows-installer/stable-3-0-update-2.99.1/+merge/87385
[17:54] <ralsina> dobey: if dconf-qt is C++, PyQt-style wrappers are easy enough to write, but sounds like way too much work to replace a call to system()
[17:54] <ralsina> nessita: want me to do a IRL?
[17:55] <nessita> ralsina: if you can, that would be great
[17:55] <ralsina> nessita: I can try :-)
[17:55] <dobey> ralsina: well you could just talk to dconf directly over dbus or whatever
[17:55] <ralsina> dobey: yep, makes sense. You can use dbus with PyQt
[17:56] <alecu> dobey, I'm puzzled by this note: "Note: Most applications will not want to interface directly with dconf, but rather with GSettings." http://live.gnome.org/GnomeGoals/GSettingsMigration
[17:57] <mandel> I sometimes feel I have the invisibility super power (I mentioned dbus.. ) :P
[17:57] <alecu> who is saying that!
[17:57] <alecu> I heard a voice!
[17:57] <ralsina> alecu: ghosts!
[17:58] <dobey> alecu: right; GSettings does some extra stuff for you, that you'd have to do yourself if you talk directly to dconf
[17:58] <dobey> alecu: and GSettings uses registry on windows
[17:58] <mandel> the caterville ghost, give me a bucket of paint and you will suffer!
[17:59] <ralsina> nessita: +1 on that branch
[18:02] <mandel> ok, EOD for me, alecu I think I have an idea on how to fix that stupid dirty reactor issue on p with squid3...
[18:02] <mandel> is a PITA but I think I can do it within the code
[18:02] <alecu> mandel, cool
[18:02] <mandel> ralsina, nessita, gatox_away , dobey see you all tom!
[18:03] <nessita> bye mandel!
[18:03] <ralsina> bye mandel!
[18:03] <dobey> cheers mandel
[18:03] <mandel> ralsina, in case I forget, I should feel a admin request, reyes is a holiday in spain
[18:03] <mandel> we are that weird
[18:03] <ralsina> mandel: please file. Or show up ;-)
[18:03] <mandel> ralsina, doing now :)
[18:04] <alecu> mandel, dobey, ralsina: we can't use dbus to read settings from dconf
[18:04] <alecu> only the writer is a dbus service
[18:04] <ralsina> mandel: why?
[18:04] <ralsina> mandel: that makes so little sense :-)
[18:04] <alecu> reading is done in the users process.
[18:04] <ralsina> oops I meant alecu
[18:05] <mandel> ralsina, , hehehe I was trying to think a reason why we believe in three men that get into kids rooms while they are sleep..
[18:05] <mandel> ralsina, I could just think of pedophiles :P
[18:05] <mandel> ok, I need to go and buy presents with extreme urgency, a2
[18:06] <dobey> to give a little kid while he's asleep?
[18:06] <ralsina> mandel: do you rent your camel, or you own?
[18:06] <dobey> lol
[18:15] <alecu> this is why I hate Gnome Introspection: no python exceptions, but fatal errors: http://pastebin.ubuntu.com/791943/
[18:22] <verterok> alecu: nice! oh, a KeyError...let's kill the interpreter! :/
[18:22] <dobey> alecu: that's a bug; it should raise a Gio.Error for that i think
[18:22] <dobey> verterok: not quite; C doesn't have KeyErrors :)
[18:22] <verterok> dobey: I see python code there ;)
[18:23] <verterok> and expect a Key|Value|AnythingError
[18:23] <alecu> dobey, right. But I've seen many of those fatal errors elsewhere while using GI with GTK or Gnome too.
[18:25] <dobey> verterok: import ctypes; foo = ctypes.int("foo");
[18:25] <dobey> although, python might handle that
[18:25] <dobey> alecu: yes, GI is not 100% yet
[18:25] <dobey> and may well never be, sadly :(
[18:26] <verterok> dobey: will be replaced by something else?
[18:27] <dobey> i don't know. not anytime soon
[18:27] <verterok> I asked because of: "may well never be" :)
[18:27] <dobey> well it's software
[18:28] <dobey> it's not like GNOME 3.0 got released, and everyone just stopped working on GNOME :)
[18:29] <dobey> and gsettings in precise is a bit more, uh evil, now; about schemas and keys, not being there
[18:29] <dobey> basically, if there is no schema, it will abort()
[18:37] <dobey> nessita: did you review https://code.launchpad.net/~dobey/ubuntuone-client/fix-crashes/+merge/87382 ? :)
[18:37] <dobey> ralsina: ^^ care to review that?
[18:37] <nessita> dobey: not yet, finishing a former review for diego. On it soon!
[18:38] <ralsina> dobey: looking
[18:38] <ralsina> dobey: isn't indentation wrong there? Just by looking at the diff :-)
[18:38] <dobey> no
[18:39] <dobey> all the other indentation is wrong ;)
[18:40] <ralsina> dobey: hahaha
[18:41] <ralsina> dobey: for C we could adopt using indent as a pre-commit hook someday, so that whatever style we choose we stick to it ;-)
[18:42] <dobey> ralsina: i'm going to adopt not using C in this code :)
[18:42] <ralsina> dobey: that's another way to fix it
[18:43] <dobey> although it would be nice to have a nice "lint" solution for C/Vala/C++/C#/whatever
[18:55] <nessita> dobey: why did you remove the marshaller? no longer used?
[18:56] <dobey> it's not needed or used, yes
[18:58] <nessita> dobey: looks good, +1
[18:59] <dobey> thanks
[19:03] <nessita> dobey: can I build the protocol package and you sponsor it, so I can practice a bit more?
[19:03] <dobey> i guess
[19:18] <nessita> dobey: would you help me solving a lint issue for the ussoc package?
[19:18] <dobey> sure
[19:18] <nessita> W: ubuntu-sso-client: package-contains-readme-for-other-platform-or-distro usr/share/doc/ubuntu-sso-client/README.windows
[19:19] <dobey> ignore it :)
[19:19] <nessita> dobey: you sure?
[19:19] <dobey> yes
[19:20] <nessita> dobey: thanks. How can I know which lint warning are ignorable?
[19:20] <dobey> unless someone complains about that, but nobody should
[19:21] <dobey> nessita: well, ones that start with E: must be fixed :)
[19:21] <nessita> for sure
[19:21] <dobey> nessita: and the W: ones, usually should be fixed.
[19:21] <dobey> that warning makes no sense anyway
[19:22] <nessita> dobey: well, apparently we're installing the README.windows on usr/share/doc, no?
[19:22] <nessita> dobey: what I don't know is how to make a package to "drop" a file
[19:22] <dobey> lintian is making a false assumption that the extension .windows is about the operating system from MS
[19:23] <dobey> nessita: i think debuild or dh_something is installing that file, actually
[19:23] <nessita> but it's! :-)
[19:23] <nessita> (.windows is about installing on windows)
[19:23] <dobey> nessita: it is, but it is not reasonable for software to make that assumption
[19:24] <dobey> setup.py intall itself doesn't actually install the README files anywhere
[19:24] <nessita> right
[19:24] <nessita> but...
[19:24] <nessita> nessita@dali:~$ ls /usr/share/doc/ubuntu-sso-client/
[19:24] <nessita> changelog.Debian.gz  copyright  README.gz  README.windows
[19:24] <dobey> do lintian is complaining about something dpkg itself did
[19:24] <dobey> which is brilliantly stupid
[19:25] <nessita> dobey: is there a way to prevent a file of being "processed"?
[19:25] <dobey> dh_shootselfinfoot
[19:25] <dobey> lint: W: OW! I SHOT MYSELF IN THE FOOT!
[19:25] <dobey> nessita: you can remove it after it's intalled, but i don't see the point in working around dpkg being stupid about itself :)
[19:26] <nessita> ok, let's see what the sponsor say :-)
[19:26] <dobey> besides, we also include the windows .py files
[19:26] <dobey> and if you really wanted to, you could take the resulting .deb, and install it on windows, and use it
[19:26] <nessita> :-/
[19:27] <dobey> a way to tell lintian to shut up for such things, would be nice though
[19:41] <nessita> dobey: https://code.launchpad.net/~nataliabidart/ubuntu/precise/ubuntuone-storage-protocol/ubuntuone-storage-protocol-2.99.1/+merge/87404
[19:45] <dobey> grr launchpad and the stupid broken +download page :(
[19:45] <nessita> dobey: yeah, I know :-(
[19:52] <nessita> dobey: all 4 releases are made now, and the 3 packages proposed
[19:53] <dobey> ok
[19:53] <dobey> i hope we can get sso sponsored quickly
[20:30] <dobey> man, doing releases is so easy when nothing changed
[20:38] <dobey> nessita: uhm; did you do ubuntuone-installer also?
[20:38] <nessita> dobey: I did  windows-installer tarball, no package
[20:38] <dobey> nessita: no, not windows-installer; ubuntuone-installer
[20:39] <nessita> dobey: nopes
[20:39] <dobey> nessita: you uploaded to the wrong place
[20:40] <nessita> dobey: oh, hum, let me look
[20:40] <dobey> nessita: and i didn't realize that the filename conflicted with ubuntuone-installer :-/
[20:40] <dobey> https://launchpad.net/ubuntuone-installer/stable-3-0/2.99.1
[20:40] <nessita> dobey: you're right, I ran this:
[20:40] <nessita> lp-project-upload ubuntuone-installer 2.99.1 ubuntuone-installer-2.99.1.tar.gz
[20:40] <nessita> dobey: will remove the download and re-upload
[20:41] <dobey> nessita: also, please check that the milestones are deactivated for the ones you did release :)
[20:41] <nessita> dobey: yes, got that in my ToDo now
[20:43] <dobey> nessita: also, why is the ubuntuone-windows-installer tarball named as ubuntuone-installer?
[20:43] <nessita> bu! the tarball generated with setup.py is ubuntuone-installer-2.99.1.tar.gz!!!
[20:43] <nessita> dobey: don't know, checking (I just ran ./setup.py clean sdist)
[20:43] <nessita> because of:
[20:43] <nessita>     229 DistUtilsExtra.auto.setup(
[20:43] <nessita>     230     name='ubuntuone-installer',
[20:43] <nessita>     231     version='2.99.1',
[20:44] <dobey> can we change that setup.py to use ubuntuone-windows-installer?
[20:44] <dobey> right. let's change that?
[20:44] <nessita> yeap
[20:44] <dobey> thanks
[20:45] <nessita> dobey: shall I push to trunk so we catch this in next release and just rename the tarball in the stable-3-0 branch?
[20:46] <nessita> or shall I push to stable-3-0 the name change?
[20:46] <dobey> nessita: i think push it to both
[20:46] <nessita> ack
[20:46] <dobey> and stable-2-0 as well perhaps?
[20:46] <nessita> yeah...
[20:48] <dobey> nessita: we can move the release-2_99_1 tag on stable-3-0 also
[20:48] <nessita> dobey: not sure what you mean
[20:49] <dobey> nessita: i presume you already tagged the revision that you modified setup.py to be 2.99.1?
[20:49] <nessita> yes
[20:49] <nessita> how can I move a tag?
[20:49] <dobey> nessita: making a new tarball with the name= change, we should mvoe the tag as well
[20:50] <dobey> i think you'll probably have to delete it and re-tag
[20:50] <nessita> dobey: hum... have you ever deleted a tag?
[20:50] <dobey> yes
[20:50] <dobey> 2 weeks ago even ;)
[20:52] <dobey> nessita: so in the local stable-3-0 copy, you can "bzr tag --force release-2_99_1" to move it to the new revision
[20:52] <nessita> nice
[20:52] <dobey> nessita: then you'll have to "bzr tag --delete -d lp:ubuntuone-windows-installer/stable-3-0 release-2_99_1"
[20:52] <dobey> nessita: then push the branch again with the moved tag
[20:53] <dobey> you have to delete the remote one, because bzr gets a little confused if you moved it locally
[20:53] <dobey> and then try to push
[20:54] <dobey> you can probably do tag --force on the remote branch too; but this way feels more "consistent" to my brain :)
[20:54] <nessita> dobey: ack, done, it worked well
[20:56] <nessita> dobey: tarball uploaded to the proper location now
[20:56] <nessita> and i'm off for the day!
[20:56] <dobey> thanks
[21:02] <nessita> ok, bye crowd!
[22:33] <gatox> back
[22:53] <ralsina> hi gatox!
[22:53] <gatox> ralsina, hi
[22:53] <ralsina> gatox: if you have a clean windows VM, I would love if you could try the stuff I described in my ubunet-discuss mail
[22:53] <gatox> ralsina, i'll install one now
[22:54] <gatox> ralsina, clean clean?
[22:54]  * gatox reading the email
[22:54] <ralsina> gatox: clean is better :-)
[22:55] <gatox> ralsina, i mean.... a fresh install..... or with any dependency?
[22:55] <gatox> i should read the email probablty :P
[22:56] <ralsina> gatox: yes. Basically you need python, pyqt, bazaar and everything else is installed automatically
[22:56] <ralsina> oh, and pywin32
[22:57] <gatox> ralsina, ok..... i'm creating a new vm with win7 right now
[22:57] <ralsina> gatox: cool. Remember to snapshot ;-)
[22:57] <gatox> ralsina, yes! i was thinking exactly thaat! :P
[23:05] <gatox> ralsina, i can't find the email....... which is the subject?
[23:05] <ralsina> gatox: A first attempt at making windows development easier (please try it?)
[23:06] <ralsina> gatox: here it is if you didn't get it: https://pastebin.canonical.com/57688/
[23:06] <gatox> ralsina, i think i didn't get it..... i can't find it
[23:06] <ralsina> gatox: it's ok
[23:07] <ralsina> gatox: BTW, buildout.cfg needs a interpreter = python at the end so you have a python that uses everything that was installed ;-)
[23:07] <gatox> ralsina, ack
[23:09] <gatox> ralsina, this looks really cool!! this truly was needed for newcomers
[23:10] <ralsina> gatox: thanks!
[23:14] <ralsina> gatox: apparently ubunet-discuss takes a while to deliver mail. I just got mine from 2 hours ago
[23:17] <gatox> ralsina, i hope it arrives or i should check if i'm missing other emails
[23:17] <ralsina> gatox: it should, eventually
[23:18]  * gatox taking snapshot.... installing the programs
[23:21] <ralsina> gatox: if it says "internal error" when installing logilab-whatever try again and it may work
[23:22] <gatox> ralsina, the old trick of trying again in windows :P
[23:22] <ralsina> gatox: well, it does look like a bug in pip, but yes ;-)
[23:25] <gatox> ralsina, here is your email :D
[23:28] <ralsina> gatox: after you do all that, bzr branch our code, and run it using build\bin\python and everything should work :-)
[23:29] <gatox> ralsina, ack
[23:45] <ralsina> ok, really really EOD for me. Bye gatox1
[23:45] <gatox> ralsina, bye!
[23:45] <gatox> ralsina, i'll let you know how it was
[23:45] <ralsina> gatox, cool post to ubunet-discuss
[23:45] <gatox> ok!