[09:09] <mandel> morning all!
[09:25]  * mandel back
[09:28] <JamesTait> Morning all! :D
[10:56] <m4n_> guys, can anyone tell me the url for tomboy notes synchronization?
[10:56] <m4n_> https://edge.one.ubuntu.com/notes/ and https://one.ubuntu.com/notes/ gives 404
[11:00] <gatox> good morning!
[11:27] <mandel> m4n_, ask rye, he should/might know
[11:27] <mandel> gatox, morning!
[11:27] <gatox> mandel, hi
[11:27] <mandel> gatox, have I told you I hate proxies ;)
[11:27] <rye> m4n_, checking
[11:27] <gatox> mandel, mmmm i think i heard that once or twice jejee
[11:28] <rye> m4n_, are you using browser to go to https://one.ubuntu.com/notes/ or using directly in tomboy config?
[11:28] <mandel> gatox, great, so, I hate them.. I'm having to write a proxy factory wtf?
[11:29] <gatox> jeje
[11:29] <gatox> mandel, did you read this? http://www.extremetech.com/computing/121015-windows-8-may-drive-me-to-linux
[11:30] <m4n_> rye: my tomboy kept crashing because i had set for auto sync.
[11:30] <m4n_> thats when i went and checked online, but i see that the online UI has been discontinued
[11:30] <m4n_> but the syncing is no longer working for me
[11:30] <mandel> gatox, yep, seen it
[11:31] <mandel> gatox, I'm going to update my developer previews to see what's up with that.. :)
[11:31] <gatox> mandel, :P i'll try to install it later too
[11:31] <m4n_> rye: i also tried running with --debug, but that is also not giving anything relevant
[11:32] <m4n_> am on 11.10
[11:32] <m4n_> tomboy version Version 1.8.0
[11:32] <rye> m4n_, launching vm to test
[11:32] <m4n_> sure
[11:37] <rye> m4n_, no, association is performed correctly, let me give you a link to the script that will check whether notes are ok
[11:37] <m4n_> okay
[11:37] <m4n_> so when you say "notes are ok", you mean any parsing error or stuff like that?
[11:37] <m4n_> i can do a local directory sync - that works fine.
[11:37] <rye> m4n_, yes,   - http://people.canonical.com/~roman.yepishev/us/tomboy-sync-validator.py - download it and run as python tomboy-sync-validator.py
[11:37] <m4n_> if at all that matters
[11:38] <rye> m4n_, no, i mean whether the stored ones on the server are ok
[11:38] <rye> m4n_, the script will check for XML well-formedness and for date format error which are the only two known issues preventing tomboy from syncing
[11:39] <rye> it now has a --fix option which i have just added, so in case it finds something wrong with the dates, it can be used to reset the dates.
[11:45] <m4n_> gimme a sec
[11:45]  * rye is rebooting after lots of updates
[11:45] <rye> brb
[11:46] <m4n_> well, i should have mentioned. i have seen last-sync-date sparse error in manifest.xml
[11:46] <m4n_> i didnt know it was a fatal error
[11:50] <rye> m4n_, ah, yes, it is fatal for tomboy
[11:51] <m4n_> ah no wonder
[11:51] <m4n_> then i tried removing that field altogether in the manifest ;)
[11:51] <m4n_> didnt help :)
[11:51] <rye> m4n_, it was generated by a buggy code on one.ubuntu.com/notes side, and fixed in 3 days or so, but notes that went in during that period became broken
[11:52] <m4n_> btw, the python script is not showing anything nor finishing
[11:52] <m4n_> it is trying to open tomboy notes, i believe from the Gtk warnings that i get on the terminal
[11:53] <m4n_> i searched to see if I can get a valid last-sync-date format from someone.. but all were reports of errors on that
[11:53] <m4n_> with my same time stamp
[11:53] <m4n_> so do you think fixing that timestamp will fix my issue?
[11:53] <rye> m4n_, no, it is not trying to open the notes, it is trying to contact ubuntuone servers
[11:53] <rye> or
[11:53] <rye> it is stuck at credentials getting
[11:53] <m4n_> ah
[11:54] <rye> m4n_, could you please ctrl+c it?
[11:54] <m4n_> sure
[11:54] <mandel> gatox, have you been working with the ./ubuntu_sso/qt/tests/show_gui.py and ./ubuntu_sso/qt/tests/login_u_p.py files?
[11:55] <gatox> mandel, mmmm not recently
[11:55] <gatox> mandel, why? do you need something?
[11:55] <rye> m4n_, did it show where KeyboardInterrupt happen?
[11:56] <mandel> gatox, sorry, I just checked bzr blame, it was not you :)
[11:56] <gatox> mandel, aja!
[11:56] <gatox> jeje
[11:56] <mandel> gatox, there are a number of debug print statements there
[11:56] <gatox> mandel, who was it?
[11:56]  * gatox wants to know
[11:56] <mandel> gatox, such as print "result received", args :P
[11:56] <gatox> mandel, in trunk?
[11:57] <mandel> gatox, do a bzr blame on ubuntu_sso/qt/tests/login_u_p.py
[11:57] <mandel> gatox, AFAIK yes, in trunk
[11:57] <m4n_> rye: yes, in credstorage.loop.run() as you suspected
[11:57] <gatox> mandel, i see :P
[11:58] <mandel> gatox, and in ubuntu_sso/qt/tests/show_gui.py
[11:58] <rye> m4n_, ok, this is not good
[11:58] <m4n_> :)
[11:58] <rye> m4n_, it means that my testing oneiric instance was actually using precise packages, therefore it worked
[11:58] <gatox> mandel, hehe
[11:58] <m4n_> rye: want me to paste the last few lines when i pressed ctrl+c?
[11:58] <rye> which means that ubuntuone-indicator also is broken on oneiric now
[11:59] <m4n_> ah may be i should not do daily updates :|
[11:59] <rye> m4n_, it means that my testing oneiric instance was actually using precise packages, therefore it worked
[11:59] <rye> ah, wrote that already
[11:59] <m4n_> :)
[12:00] <m4n> is this a valid timestamp now? <last-sync-date>2012-03-07T16:51:28.6484780+05:30</last-sync-date>
[12:00] <m4n> that is what i have now
[12:01] <ralsina> good morning!
[12:03] <mandel> ralsina, morning!
[12:03] <ralsina> hello mandel
[12:03] <mandel> ralsina, we might get the ssl dialog support done today!
[12:04] <ralsina> mandel: yay!
[12:04] <gatox> ralsina, hi!
[12:04] <mandel> ralsina, although, the main reason for our problems is that we have not been using a proxy factory and instead we have used a proxy..
[12:04] <ralsina> hola gatox! Feelig better?
[12:04] <rye> m4n, erm, are you using Ubuntu One for file sync?
[12:04] <ralsina> mandel: mmmmmmmkay
[12:05] <m4n> rye: hey it worked this time
[12:05] <m4n> nope, am not using ubuntu one for file sync
[12:05] <gatox> ralsina, yep!!! totally rested
[12:05] <rye> m4n, aha, i should change the priorities of the lookup - tomboy, then gnome-keyring then sso
[12:05] <mandel> ralsina, so the thing is, if a user has set diff proxies for http and https the qtnetwork implemenation was always using the http one no matter which scheme the request had
[12:06] <nessita> hello everyone!
[12:06] <gatox> nessita, hi
[12:06] <mandel> ralsina, I have been swearing about proxies since yesterday when I discovered the issue :)
[12:06] <ralsina> mandel: better yesterday than today!
[12:06] <nessita> how are you feeling today gatox?
[12:06] <m4n> rye: i dont know what exactly fixed this time. but what I believe is, after doing a local sync which fixed the time stamp in the manifest file, the authentication worked
[12:06] <m4n> and then syncing followed
[12:07] <gatox> nessita, splendid! :D
[12:07] <rye> m4n, so syncing with U1/notes worked, right?
[12:07] <m4n> rye: yes. it did
[12:07] <m4n> thanks for your help :)
[12:11] <mandel> nessita, morning!
[12:11] <rye> then i second gatox's reaction on an unrelated event
[12:11] <mandel> nessita, I've seen in the back-log I was mentioned, did I add any new deps? I don't think so..
[12:11] <gatox> rye, second me about what?? what did i miss? :P
[12:11] <rye> gatox, you said "splendid!" :)
[12:11] <nessita> hola mandel!
[12:12] <nessita> mandel: I think you did not, which is great, but I was packaging ussoc and needed to confirm
[12:12] <gatox> rye, ahhhhhh jeje
[12:13] <mandel> nessita, cool, I have been trying not to add a single dep if I cannot, I've been really careful recently ;)
[12:13] <nessita> mandel: that's great, thanks :-)
[12:29] <nessita> which is more correct, "converted to" or "converted into"
[12:32] <mandel> nessita, i think is convert into
[12:33] <mandel> nessita, ref http://www.thefreedictionary.com/convert
[12:38] <nessita> mandel: thanks!
[12:54] <ralsina> nessita, gatox, mandel: I am sleepy and grumpy. Are there any reviews I can do?
[12:55] <mandel> nessita, not from me, sorry
[12:55] <nessita> ralsina: alecu asked several by email, I think
[12:55] <gatox> ralsina, yes sir: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/verification-fail/+merge/95978
[12:55] <gatox> :D
[12:55] <ralsina> nessita: oh, right. Ok, I'll do those.
[12:55] <ralsina> gatox: thanks!
[12:55]  * ralsina gets some mate going
[13:11] <mandel> I'm off to have lunch, catch you in a number of mins!
[13:11]  * mandel lunch
[13:23] <nessita> dobey: hi there, any idea why having the qt man page as we did for the gtk, is not getting installed in debian/tmp/usr/share/man/man1/ubuntuone-control-panel-qt.* ?
[13:24] <nessita> dobey: on setup.py we did nothing special for the gtk man page, as we don't for the qt one
[13:28] <alecu> hello all!
[13:29] <gatox> alecu, hi
[13:29] <ralsina> hello alecu!
[13:45] <mhall119> aquarius: ping
[13:47] <aquarius> mhall119, pong
[13:47] <mhall119> I think I found it myself, but just to double check, is it possible to write a desktop app that can stream music from your U1 cloud?
[13:47] <mhall119> using https://one.ubuntu.com/developer/music/stream_music/cloud
[14:09] <nessita> ralsina, dobey: would you please review? https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/install-man-page/+merge/96359
[14:09] <ralsina> nessita: sure
[14:10] <gatox> nessita, what happend with the bug with uife? should i merge the branch i have with something else?
[14:12] <nessita> gatox: let's ask joshuahoover now that I released ussoc. Hey joshuahoover! how much less uife do we have today? :-) (yesterday I released a lot of them)
[14:14] <dobey> nessita: i think we installed the man page for gtk using the .install file for the package
[14:14] <ralsina> nessita: +1 on man-page
[14:14] <dobey> mhall119: https://launchpad.net/ubuntuone-music
[14:14] <dobey> mhall119: so "yes" :)
[14:15] <nessita> dobey: yes, but the thing is that the manpage from upstream does not get installed in debian/tmp/usr/share/man/...
[14:15] <dobey> nessita: yes, because we *installed* it from the .install file, not the setup.py
[14:16] <nessita> dobey: not sure what you mean (since I still understand what I answered to before :-))
[14:16] <dobey> nessita: one second, i'll show you
[14:16] <nessita> dobey: thanks
[14:17] <dobey> oh, nevermind, apparently we weren't doing what i thought we were doing
[14:17] <nessita> dobey: exactly! (?)
[14:17] <nessita> dobey: I guess distutils-extra was doing some magic for the old man page, that can't apply to the new one (perhaps the "this is a generated file" header is confusing it)
[14:17] <dobey> i don't think it parses man pages
[14:18] <nessita> dobey: so I proposed a branch to install the manpage properly from setup.py, and made a pacth for the package with that, and that is working like a charm
[14:18] <nessita> dobey: the man page for the gtk cp was being installed in {install_prefix}/share/man/man1/...
[14:18] <nessita> and that install was happening by some magic
[14:18] <nessita> we had no code to do it
[14:19] <dobey> yes i know it was; but the magic i thought it was, is apparently not the magic :-/
[14:20] <nessita> dobey: I insist is the disutils-extra. Since we removed the gtk manpage and we have the qt one, we're getting
[14:20] <nessita> WARNING: the following files are not recognized by DistUtilsExtra.auto:
[14:20] <nessita>   docs/ubuntuone-control-panel-qt.1
[14:20] <nessita> and we did not get that for the gtk manpage, so that means that disutils-extra recognized it and did something with it
[14:21] <dobey> hmm
[14:21] <ralsina> nessita: if it's the header, I am removing it on apending branch
[14:21] <dobey> maybe it is parsing it for some reason then and the comment confuses it
[14:21] <dobey> it's easy enough to test it
[14:21] <nessita> ralsina: I kinda like the header, it makes sense (since we're generating the manpage)
[14:21] <nessita> yes
[14:21] <ralsina> nessita: not really, I had to edit it ayway
[14:21] <nessita> done
[14:21] <ralsina> anyway
[14:21] <nessita> is the header
[14:22] <nessita> removing the header made ./setup.py install it
[14:22] <dobey> weird
[14:22] <ralsina> nessita: then do nothing, and when my branch lands it fixes itself
[14:22] <dobey> oh, it probably checks the mime type of the file
[14:22] <nessita> ralsina: hrm. I need to do something for the packaging branch
[14:22] <dobey> and the header breaks the magic for the man
[14:22] <dobey> since the first N bytes are different
[14:22] <nessita> ralsina: what's your fix, so I can apply the same to the packaging branch?
[14:22] <nessita> dobey: right
[14:23] <ralsina> nessita: diff is here https://code.launchpad.net/~ralsina/ubuntuone-control-panel/alerted/+merge/96253
[14:23] <dobey> does man barf with the header?
[14:23] <nessita> dobey: no
[14:24] <ralsina> dobey: it's a limitation on  file, I suppose
[14:24] <dobey> ralsina: well, mime.magic
[14:24] <ralsina> dobey: same thingie
[14:24] <dobey> which says "N bytes are FOO"
[14:24] <dobey> we can also fix the mime
[14:24] <nessita> dobey: how?
[14:25] <dobey> in shared-mime-info i think
[14:25] <ralsina> dobey: basically doesn like *roff files that start with a comment, it seems
[14:25] <dobey> right
[14:25] <dobey> nessita: by adding another magic specification to skip the comment
[14:26] <nessita> dobey: is it woth doing it? removing the comment just works
[14:26] <dobey> nessita: but that's a fix outside of our code, and in shared-mime-info or something
[14:26] <nessita> ah
[14:26] <dobey> nessita: for this specific case? no. for the general case? yes.
[14:26] <nessita> agreed
[14:26] <nessita> ralsina: ok, branch completly changed but same MP
[14:26] <ralsina> nessita: hahaha, keep the +1 ;-)
[14:27]  * ralsina looks anyway
[14:27] <nessita> also updated the commit message
[14:27] <nessita> dobey: you may wanna confirm your +1 https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/install-man-page/+merge/96359
[14:28] <ralsina> nessita: consider it a +2 from me
[14:28] <nessita> nice
[14:28] <nessita> ralsina: you will have to merge this in yours and resolve the conflict, perhaps
[14:28] <ralsina> nessita: no problem
[14:29] <ralsina> nessita: 1-1? Has to be on IRC though, pulseaudio hates me too now.
[14:29] <nessita> sure]
[14:39] <nessita> dobey: also, u1cp needs to depend on installer >= 2.99.90 and u1client 2.99.90, so please let me know when those are uploaded to ubuntu. Thanks!
[14:39] <dobey> it does? why?
[14:42] <gatox> nessita, ralsina i have this branch for review when you have a moment: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/too-verbose/+merge/96364
[14:43] <nessita> dobey: on installer for consistency, on u1client so u1-login prefers the qt UI
[14:43] <nessita> gatox: ack!
[14:43] <dobey> i don't think those warrant a necessary dependency
[14:45] <nessita> dobey: agreed is not 100% necessary, OTOH I think is is cleaner
[14:45] <ralsina> gatox: queued
[14:45] <nessita> anyways, tehre is no problem on holding the u1cp release for a couple of hours, no?
[14:46] <gatox> nessita, do you have any other issue in mind that you would like that i work on??.... i finish with the ones we talk the other day (except one that is in stand by because we need feedback from design)
[14:46] <nessita> gatox: which one? (I fixed the one about login and forgotten page title)
[14:46] <gatox> or I just look in my bug queue?
[14:47] <gatox> nessita, this one: https://bugs.launchpad.net/ubuntuone-control-panel/+bug/942020
[14:47] <nessita> gatox: ah, yes. Can you ping rtgrant and lisette about that?
[14:47] <nessita> gatox: and yes, go thru your bug queue
[14:47] <nessita> gatox: will review it now, just to be sure. Perhaps I will adjust priorites a bit
[14:48] <gatox> nessita, ok
[14:51] <nessita> gatox: I'm re-reading the orange highlight bug... what do you need from design?
[14:52] <nessita> gatox: the bug is about the weird shape of the highlight, in the screenshot you can see if pretty small, it should cover the whole tab, I'd say
[14:55] <mandel> ups, I forgot, I'm back :)
[14:55]  * mandel back
[14:56] <ralsina> gotta go picck up the kid. Will be back in a bit.
[14:59] <mandel> ralsina, can you wait 2 mins, that way we have the stand up :P
[15:00] <gatox> me
[15:00] <mandel> me
[15:00] <urbanape> me
[15:01] <alecu> me
[15:02] <dobey> meh
[15:04] <mandel> nessita, ?
[15:04]  * jalcine wants in
[15:04] <nessita> sorry, on a conversation next door
[15:04] <nessita> me
[15:04] <nessita> but no notes yet
[15:05] <gatox> go for  me?
[15:05] <gatox> DONE:
[15:05] <gatox> Proposed 2 branches for Bug #934500: Qt UI: Logging too verbose (and adds no useful info) and Bug #945065: [UIFe] ResetPassword page has some widgets really close to the right margin. Take half of the day yesterday to rest because i wasn't felling very well.
[15:05] <gatox> TODO:
[15:05] <gatox> Keep fixing the bugs in my Bugs Queue.
[15:05] <gatox> BLOCKED:
[15:05] <gatox> No
[15:05] <gatox> mandel, go
[15:05] <mandel> DONE: Refactored qtnetwork webclient implementation to use a proxy factory. The main reason is that if the user has set up diff proxies for http, https etc.. we have to make sure we use the correct one. I've got tests for the qtnetwork implementation not for libsoup, I think we need to do something similar to the dbus tests so that the gsettings are mocked.
[15:05] <mandel> TODO: Add some extra tests for the ssl dialog. Implement the ssl dialog changes for libsoup (trivial). review alecus branches
[15:05] <mandel> BLOCKED:no
[15:05]  * mandel passes the ball to urbanape
[15:05] <urbanape> DONE: Made a little progress on my branch for ubuntu-sso-client
[15:05] <urbanape> TODO: Figure out the best way to determine network connectivity on darwin, generalize an ignorable test structure for UNIXish environments.
[15:05] <urbanape> BLOCK: None
[15:05] <urbanape> alecu: you're up
[15:05] <alecu> DONE: finished tunnel branches for review, started branch to use tunnel for SD webcalls
[15:05] <alecu> TODO: keep working on webcalls, bug #929212
[15:05] <alecu> BLOCKED: need reviews on the four branches here: https://bugs.launchpad.net/ubuntuone-client/+bug/929207
[15:05] <alecu> NEXT: dobey
[15:05] <dobey> λ DONE: releases and backport to stable branches
[15:05] <dobey> λ TODO: upload new rb, finish releases/uploads, twisted glib2/gireactor debugging
[15:05] <dobey> λ BLCK: none.
[15:06] <dobey> nessita:
[15:06] <nessita> DONE: releases, reviews (lost of those)
[15:06] <nessita> TODO: releases, testing of those, reviews, start with bug #933697
[15:06] <nessita> BLOCKED: nopes
[15:06] <nessita> NEXT: ralsina
[15:06] <nessita> any comments anyone?
[15:08] <alecu> comments:
[15:08] <alecu>  * please, review my branches! :-)
[15:08] <alecu> eoc
[15:09] <dobey> # meh.
[15:10] <alecu> mandel, do you have a branch for the "proxy factory" changes?
[15:10] <alecu> mandel, I'd like to take a look at that, and see if I need to use something like that in the tunnel.
[15:11] <mandel> alecu, let me push it, one min
[15:11] <nessita> alecu: ack, will do
[15:15] <mandel> alecu, here you go: lp:~mandel/qt-ssl-dialog
[15:15]  * dobey needs to send a mail to remind people to target their bugs appropriately it seems
[15:15] <mandel> alecu, sorry, lp:~mandel/ubuntu-sso-client/qt-ssl-dialog :P
[15:16]  * gatox lunch
[15:17] <alecu> mandel, ack
[15:18] <mandel> alecu, the main idea is that setProxy forces the webclient to ALWAYS use the same proxy, which is wrong, if the user did set up http proxy to my_hhtp_proxy and https to my_company_proxy you will always use the http one for the https requests
[15:18] <joshuahoover> ralsina, nessita: here is what our list is showing as of today for FEs: http://is.gd/u1fes
[15:19] <mandel> alecu, which is wrong, what you have to do is create a proxy factory that looks at the scheme of the proxy request and returns the correct proxy according to the settings
[15:19] <mandel> alecu, does it make sense?
[15:19] <alecu> mandel, but remember that *we* always do https requests
[15:19] <nessita> joshuahoover: ok, removing https://bugs.launchpad.net/ubuntuone-control-panel/+bug/940620 from that list since we're not doing that
[15:19] <mandel> alecu, then we have been always the wrong proxy in use because in gsettings we have been reading the http one..
[15:19] <nessita> joshuahoover: and we need to add the one from gatox_lunch from yesterday
[15:20] <alecu> mandel, I agree that querying the proxy for every url is the expected behavior, but there's a gotcha to that, because sometimes it can get too slow... let me find the link.
[15:20] <joshuahoover> nessita, gatox_lunch: once gatox is back, if he gets me the bug # i'll add it
[15:20] <nessita> let me give you that
[15:21] <dobey> background color is sad :(
[15:22] <nessita> bug #940620 marked as won't fix
[15:22] <nessita> dobey: sorry, I agree with you
[15:22] <alecu> mandel, "On Windows platforms, this function may take several seconds to execute depending on the configuration of the user's system", on the bottom of http://doc.qt.nokia.com/4.7-snapshot/qnetworkproxyfactory.html
[15:23] <gatox_lunch> joshuahoover, this one: https://bugs.launchpad.net/ubuntu-sso-client/+bug/945065
[15:23] <joshuahoover> gatox_lunch: thx
[15:23] <dobey> at least this song is a win; it samples from "Altered Beast"
[15:25] <mandel> alecu, yep, but I'm just using that for the linux one, where by implementation of the factory returns the proxies from a dict, you are looking at a static function, I'm implementing http://doc.qt.nokia.com/4.7-snapshot/qnetworkproxyfactory.html#queryProxy
[15:25] <mandel> alecu, you are looking at http://doc.qt.nokia.com/4.7-snapshot/qnetworkproxyfactory.html#systemProxyForQuery
[15:29] <alecu> mandel,  ack.
[15:29] <nessita> mandel: in https://code.launchpad.net/~mandel/ubuntu-sso-client/webclient-use-dialog/+merge/94416, why you need a constant for CREDS_ACQUIRED? isn't the semantics of that return code the same as USER_SUCCESS?
[15:29] <alecu> mandel, a related question:
[15:30] <ralsina> nessita, gatox: +1 on verification-fail. Is ussoc open for merges?
[15:30] <nessita> ralsina: yessir
[15:30] <alecu> mandel, the user/password that's in the gsettings, you are only using it for http, and not for https...
[15:30] <alecu> mandel, should we use it for both instead?
[15:31] <alecu> mandel, I mean I would consider "http.use-authentication" both for the http and https server.
[15:32] <mandel> alecu, and that is a bloody good question.. at the moment I'm just using it for http, the things is, where is that value coming from, the user has no ui in the gnome-control-center for that, right?
[15:32] <mandel> nessita, seems to be the same, yes, has that landed in trunk. I can file a bug and fix it for the next release since that does not require an fe
[15:33] <alecu> mandel, it seems like that UI was gone in O already.
[15:33] <dobey> alecu: i wouldn't think so
[15:33] <alecu> dobey, the thing is that there's no "httpS.use-authentication" in gsettings.
[15:34] <mandel> alecu, we can use it for both, worst case scenario we get a 401 the first time, also, if the user set it as user:password@domain we are ok since it is considered over the gsettings value
[15:34] <dobey> alecu: i understand that. but as they can be different servers, i don't think the auth for one should be used for the other
[15:34] <nessita> mandel: it has not, I'm doing the review for it
[15:34] <alecu> dobey, that makes a lot of sense
[15:34] <mandel> nessita, even better! I can fix that before it lands!
[15:34] <dobey> alecu: i think we should just pop the dialog if we get a 401, and store the results in the keyring
[15:34] <nessita> joshuahoover: which one is the tag for feature freeze expections (no UI)?
[15:34] <dobey> pop the password dialog, that is
[15:35] <alecu> mandel, what dobey says.
[15:35] <mandel> nessita, which branch are you looking at?
[15:35] <dobey> nessita: they are all just u1-fe
[15:35] <alecu> mandel, and on top of that, store different credentials for every proxy configured
[15:35] <joshuahoover> nessita: just use u1-fe
[15:35] <nessita> dobey: there is also a u1-ffe, so I wanted to confirm
[15:35] <nessita> joshuahoover: what's the tag u1-ffe used for?
[15:35] <dobey> typo? or historical fail?
[15:35] <alecu> mandel, so, if the user changes the proxy host or port or type then new credentials are asked always.
[15:35] <mandel> alecu, yes, that is indeed true..
[15:36] <joshuahoover> dobey, nessita: yep, historical fail on my part
[15:36] <nessita> is ok, good to know
[15:36] <joshuahoover> nessita: so just use u1-fe...i'll try to remove the other tags when i see them, feel free to do the same :)
[15:36] <dobey> alecu: i think we should just always ask for credentials if needed, and ignore the use-authentication bits in gsettings
[15:37] <nessita> joshuahoover: thanks!
[15:38] <rye> joshuahoover, https://support.one.ubuntu.com//Ticket/Display.html?id=11867 - Now I am fixing notes w/o the gui
[15:38] <dobey> alecu: the control panel has no ui to set it. and passwords shouldn't be stored in gsettings anyway
[15:38] <rye> web ui
[15:38] <joshuahoover> rye: thx
[15:38] <rye> i mean
[15:39] <dobey> can someone please do quick review of https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/update-from-trunk/+merge/96250 ?
[15:40] <nessita> mandel: let me know when that's changed, the branch is https://code.launchpad.net/~mandel/ubuntu-sso-client/webclient-use-dialog/+merge/94416
[15:40] <mandel> alecu, so thinking about it, passwords should be stored as domain:port or something like that since otherwise we might have issues
[15:41] <mandel> nessita, ok!
[15:41] <alecu> mandel, I think dobey is right and we should not care about authentication settings stored in gsettings; and only use authentication when the proxy requests it, and get it from the keyring.
[15:41] <alecu> mandel, not even user:pass@host
[15:41] <joshuahoover> nessita, gatox_lunch, ralsina: do we have any other ui related changes for sso-client that you know about? if so, i'd like to send one email to the doc/translator email lists
[15:42] <alecu> mandel, and yes, for passwords stored in the keyring we should name them like "type:host:port"
[15:42] <ralsina> none I know of
[15:42] <alecu> mandel, type being http, https or socks
[15:42] <dobey> alecu, mandel: we should store passwords in the keyring, in the same way that eg; epiphany does.
[15:42] <ralsina> joshuahoover: ^ (but wait for gatox's and nessita's ack just in case ;-)
[15:43] <nessita> joshuahoover: thinking, let me review the whole bug queue, will ping you back
[15:43] <mandel> dobey, we are doing that already :)
[15:43] <joshuahoover> nessita: thx
[15:43] <mandel> alecu, is there a diff between https:domain:8888 and http:domain:8888 ? the user could be using to usernames, true.. but how common is that?
[15:45] <alecu> mandel, I'm not sure. But my guess is that if the user changes the proxy type we should ask the password again. What do you think?
[15:45] <dobey> mandel: i am not sure "https" is the right term to use for that proxy setting
[15:46] <dobey> mandel: for instance, in the firefox dialog, it says "SSL Proxy:"
[15:46] <mandel> dobey, I don't think so too.. is very confusing..
[15:46] <dobey> mandel: which means i expect it would also be used for other protocols in use that support ssl
[15:47] <mandel> alecu, is a good question.. if he changes the type it might use the password or it might not.. we could assume that a proxy is uniquely identifies by the combination of the three things you mentioned, is the safest path to follow
[15:48] <mandel> dobey, yeah.. is certainly not very clear overall
[15:49] <alecu> dobey, mandel: I expect it to be https, since we would be using https to connect to the proxy and send the CONNECT request. After that, if the client decides to use SSL or not, it's the client choice.
[15:49] <alecu> mandel, in fact, I'm not exactly sure how the point above works.
[15:50] <alecu> mandel, for the SSL Proxy, what does firefox do?
[15:50] <mandel> alecu, oh, but here is the funny thing about the entire bloody thing that settings is not for that! is to let you set diff proxies according to the scheme, not to use https to connect to the proxy
[15:50] <alecu> mandel, funky
[15:51] <mandel> alecu, I was in libsoup asking why they did not support https proxy and they were like, wtf? that settings doesn't do what you think!
[15:51] <alecu> mandel, you mean "according to the scheme of the url that the user typed on the browser"?
[15:51] <mandel> alecu, yep
[15:51] <alecu> mandel, crap.
[15:51] <alecu> mandel, and what about firefox?
[15:51] <dobey> yes, though socks is a special case
[15:51] <mandel> alecu, let me double check, but I think is the same
[15:52] <dobey> firefox is the same
[15:52] <alecu> damn. somebody is cooking a mighty Pumarola around here.
[15:52] <dobey> and in fact, the default of firefox in ubuntu is to use the system settings
[15:52] <alecu> I can't concentrate on proxies anymore with this smell around :P
[15:52] <dobey> alecu: maybe they are just burning tires at the garage?
[15:53] <alecu> dobey, no, it smells of delicious tomatoes and garlic.
[15:53] <dobey> hehe
[15:53] <ralsina> alecu: hmmmm chorizo a la pomarola...
[15:53] <dobey> mmm chorizo
[15:53] <alecu> lol
[15:53] <ralsina> looks like there is hunger in the air :-)
[15:54] <dobey> quiero parrilla
[15:54] <mandel> alecu, dobey so, I set up a nonauth proxy for http and an auth for https for firefox and it works as we described, if I go to http://google.com I have not password dialog shown, if I go to an https one (u1) I do
[15:54] <alecu> mandel, so: we have connection A "from the browser to the proxy server", and connection B "from the browser to the remote host via the proxy server"
[15:55] <alecu> mandel, but is the connection between the browser and the proxy encrypted in that case?
[15:55] <alecu> mandel, it's connection A we should be worrying about.
[15:55] <mandel> alecu, let me sniff but I don't think so
[15:55] <alecu> mandel, connection B always will use SSL in our case. If the certificate for connection B is wrong, we fail automatically.
[15:56] <alecu> mandel, only for connection A we should allow the user to pin the certificate, right?
[15:57] <mandel> alecu, yes, which means that if that is not the case, ie there is not ssl from browser to server we wasted some time and we can ignore all this crap
[15:57] <ralsina> thisfred: this review is right up your alley https://code.launchpad.net/~ralsina/ubuntuone-control-panel/alerted/+merge/96253
[15:57] <thisfred> on it!
[15:57] <mandel> alecu, if we know 100% sure that we do not use http we just need to change the gsettings parser to parse the https settings and not the http ones
[15:57] <mandel> alecu, which is a trivial fix..
[15:58] <mandel> dobey, line 26 in https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/update-from-trunk/+merge/96250
[15:58] <mandel> dobey, is musc or music
[15:58] <dobey> mandel: someone else might use http though.
[15:58] <dobey> mandel: remember, we're writing a library as well :)
[15:59] <mandel> dobey, la puta de oros!
[15:59] <mandel> dobey, ok, then I'll do it the proper way..
[15:59] <ralsina> mandel: see that corner? Don't cut it ;-)
[15:59] <mandel> ralsina, lol
[15:59] <dobey> mandel: sigh.
[15:59] <ralsina> mandel: specially on something like "let's not read one setting"
[15:59] <alecu> mandel, we should still use the http settings if the https settings are empty.
[16:00] <dobey> alecu: what does firefox do in that case? :)
[16:00] <mandel> alecu, but that means do not use a proxy.. in theory, but I have the impression that no one knows how to set up proxies correctly :P
[16:00] <ralsina> alecu: there should never be an empty https setting. Either the user selected "same server for all protocols" or he wats https to be unproxied. Right?
[16:00] <mandel> dobey, so, is it a typo?
[16:00] <dobey> alecu: i think it just doesn't use a proxy
[16:00] <dobey> mandel: it is a typo, yes
[16:00] <mandel> ralsina, that is what I think
[16:01] <alecu> ralsina, the "same server for all protocols" setting is long gone.
[16:01] <dobey> mandel: and neither nessita nor thisfred caught it :(
[16:01] <ralsina> alecu: it's implicitly true now ;-)
[16:01] <dobey> mandel: which means i now have to fix it in two places
[16:01] <mandel> dobey, oh, well.. sorry for that
[16:02] <thisfred> dobey: oops, sry
[16:02] <mandel> ralsina, we should literally let alecu work upstream to fix this in gnome ;-)
[16:02] <ralsina> mandel: makes sense. But. Timing?
[16:02] <ralsina> I mean, sure, in may :-)
[16:03] <dobey> mandel: doing it right, we might get it done for gnome 4.0
[16:03] <mandel> dobey, lol
[16:03] <mandel> ralsina, well, we can always ask cparrino to let us do it next cycle.. with some help from desktop :P
[16:03] <ralsina> dobey: unless someone decides that proxy configs are out of scope for gnome or something
[16:03] <ralsina> mandel: file a bug on desktop with an offer of assistance in june
[16:04] <dobey> ralsina: i don't think upstream will decide that
[16:04] <mandel> alecu, so getting back to that, we need to ignore the gsettings auth setting, which makes no sense, change the way the passwords are stored to include scheme(type):domain:port
[16:04] <ralsina> dobey: me neither, but hey, I can't decide what happens when I close the freakig lid anymore ;-)
[16:04] <mandel> alecu, use a factory for the qtnetwork and hope libsoup with gnome integration works correctly
[16:04] <alecu> mandel, perfect.
[16:05] <dobey> ralsina: that's ubuntu design's fault :)
[16:05] <alecu> mandel, but we still don't know when is connection A done via SSL
[16:05] <alecu> mandel, or if at all
[16:05] <mandel> alecu, yep, that is something I'm going to be asking in mozilla, the should know :)
[16:06] <mandel> alecu, it might be the case that the ssl dialog is utter useless :(
[16:06] <dobey> alecu: always, and if it fails, then try without
[16:06] <mandel> alecu, that is, A is not done through ssl
[16:06] <dobey> alecu: there is no reason we shouldn't always prefer to use ssl.
[16:06] <mandel> alecu, yes, question about the lib, should we be taking into account ssl errors from the server? even if we know the u1 is correct?
[16:06] <alecu> mandel, I would not be surprised if firefox does sniffing to find out if it connected to a ssl or tcp proxy.
[16:07] <dobey> mandel: https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/fix-typo/+merge/96383
[16:07] <alecu> mandel, and it might be the case that qtnetwork does the same when connecting to a proxy.
[16:07] <alecu> mandel, and in the best of cases we should not do a thing.
[16:07] <mandel> dobey, easy +1
[16:08] <alecu> mandel, so I think we should do some IRL tests to see how both firefox and qtnetwork react to an ssl proxy.
[16:08] <mandel> alecu, yes.. now.. what bloody proxy should we use, ISA?
[16:08] <mandel> alecu, I think that one can do it, but I'm not sure
[16:08] <alecu> mandel, you mean squid compiled with ssl?
[16:08] <mandel> alecu, we should first ask for a raise :P
[16:09] <alecu> mandel, lol
[16:09] <mandel> alecu, true, I have the packages in my machine, I'll do that
[16:09] <alecu> mandel, also we may try some ssl tunnels. Those are processes that start on a port, listening for ssl and fwd the traffic to a tcp socket.
[16:09] <mandel> alecu, have I mentioned that proxies are in the 'things that manuel hate' with the windows file system?
[16:10] <mandel> alecu, similar to the one that rye mentions in the wiki, right?
[16:10]  * alecu is curious about what mandel will hate next cycle. OS X sandboxing?
[16:10] <mandel> alecu, give a +1 to dobey :https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/fix-typo/+merge/96383
[16:11] <mandel> alecu, we are going to end like verterok.. knowing all the boring techs..
[16:11] <dobey> mandel: it's trivial doesn't need a bunch of reviews :P
[16:11] <alecu> mandel, dobey: +1
[16:11] <dobey> it's already merged by now anyway
[16:11] <mandel> alecu, lets girl I chat in a bar I'm going to brake the ice with, do you use a proxy diff proxy for https, that tell a lot about a woman!
[16:12] <mandel> the mother of my kids will use diff proxies for everything!!!
[16:12]  * mandel lost it
[16:13] <dobey> hehe
[16:13]  * dobey needs to take holidays to spain
[16:13] <ralsina> alecu, mandel: asking for raises... did you notice yearly reviews are right around the corner?  ;-)
[16:13] <dobey> heh
[16:13] <mandel> ralsina, nop have we failed? lol
[16:14] <ralsina> mandel: impress me! hahahahaha whip!
[16:14] <dobey> raise NoSoupForYou('You go bread line!')
[16:14]  * alecu scrambles his mail to find his long missed objectives.
[16:14] <mandel> dobey, is there a way to tell apt-get to install a .deb from an specific ppa?
[16:14] <dobey> mandel: apt-add-repository ppa:foo/bar
[16:15] <mandel> dobey, that one I know, but I have squid from a ppa and from main, and I want to reinstall it using the one in the ppa
[16:15] <rye> heh
[16:15] <rye> magicicada is also immune to ctrl+c now
[16:15] <dobey> mandel: they have the same version?
[16:15] <mandel> dobey, yes :(
[16:15] <dobey> rye: or just slow at it
[16:16] <mandel> rye, u1 is immune to ctrl+c buahahahahaha
[16:16] <dobey> mandel: don't do that :)
[16:16] <ralsina> mandel: u1cp is not anymore
[16:16] <rye> dobey, well, it continues running as if it never saw SIGINT coming
[16:16] <dobey> hmm
[16:17] <dobey> rye: does it quit if you do something in the app after the C-c?
[16:17] <dobey> (something other than choosing 'quit' in the app, that is)
[16:18] <rye> dobey, nope, it continues working
[16:18] <dobey> hmm
[16:21] <dobey> mandel: typo should be fixed in https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/update-from-trunk/+merge/96250 now too
[16:21]  * dobey really needs to write that aspell hackery
[16:23] <alecu> mandel, I really like what you did in the branch by adding the WebClientProxyFactory.
[16:24] <mandel> dobey, please do
[16:24] <dobey> ok. i need to get some lunch. bbiab
[16:24] <mandel> alecu, the branch is not done 100% it needs some cleaning regarding the ssl.. but is getting there
[16:24] <mandel> alecu, oh, and thx :)
[16:24] <alecu> mandel, I think that using setProxyFactory to set your class is a good decision.
[16:26] <mandel> dobey, about the choosing a ppa over main: http://jaqque.sbih.org/kplug/apt-pinning.html
[16:27] <mandel> alecu, I think so since we can make it not query the os settings more than once and we can add all the diff proxies in gsettings, shall we consider socks and ftp?
[16:28] <alecu> mandel, socks yes, ftp no.
[16:28] <mandel> alecu, ack
[16:28] <alecu> mandel, anyway, socks is a bugfix that can come next week
[16:29] <mandel> alecu, yeah, I'm going to ensure that the factory works when diff settings are done, then I'll move to that
[16:29] <mandel> alecu, + looking at what happens when you have squid + ssl
[16:29] <alecu> mandel, oh, and to test socks IRL, the easiest way is setting up ssh to be a socks proxy.
[16:29] <alecu> mandel, but before all that... please review my first branch :-)
[16:29] <mandel> alecu, on it right now before I forget :)
[16:31] <mandel> alecu, one question, if you put a # pylint: disable=blah what code block does it affect? does it understand indentation?
[16:34] <rye> dobey, do you happen to know if it is possible to ask nautilus to load ubuntuone plugin from user directory, not system-wide?
[16:35] <alecu> mandel, I'm not sure about pylint. I found little documentation when I was looking for a similar thing.
[16:36] <alecu> mandel, I've seen some places where there's a pylint enable near the disable, and also some places where the disable is indented and not enabled back.
[16:36] <mandel> alecu, ok, I guess we can ignore it, it is because there are several # pylint: disable=C0103 in our code that do not have an enable, so we have # pylint: disable=C0103 # pylint: disable=C0103 which seems odd, unless pylint is very smart
[16:36] <alecu> mandel, but I'm not sure about the behaviour.
[16:37] <alecu> mandel, right
[16:37] <mandel> alecu, is a little noise, nothing to worry about :)
[16:38] <thisfred> pylint != very smart
[16:39] <mandel> thisfred, yeah, I have that feeling too
[16:39] <thisfred> it tries to be too smart, but ultimately it fails
[16:40] <thisfred> that's why I like pyflakes: made to be dumb
[16:40] <thisfred> Zen of Python #443: Dumb is better than Smart
[16:40] <gatox> ralsina, ping
[16:42] <mandel> thisfred, that is why you are just a great python developer ;)
[16:42] <mandel> lol
[16:42]  * mandel loves to make stupid jokes
[16:43] <ralsina> gatox: pong
[16:44] <gatox> ralsina, have you ever been in this situation? when you need to remove the "focus-highlight" that the os does for the widget with the tab focus? like this one: https://launchpadlibrarian.net/94595490/Screenshot%20at%202012-02-27%2009%3A30%3A52.png
[16:44] <gatox> ralsina, i know i had that problem a long time ago..... and i'm not sure if it's possible to remove that
[16:44] <thisfred> mandel: I know, dumb and pretty, that's me
[16:44] <mandel> thisfred, and blonde, right?
[16:45] <thisfred> yep
[16:50] <rye> :/
[16:51] <rye> nautilus plugin never requests public files list for real, i guess the refactoring went a bit wrong in oneiric
[16:51] <mandel> alecu, why do you set get_proxy_settings in RemoteSocketTestCase to be lambda _: {} ??
[16:52] <mandel> alecu, in https://code.launchpad.net/~alecu/ubuntuone-client/proxy-tunnel-server/+merge/95075 470 in lp diff
[16:52] <alecu> mandel, looking
[16:52] <ralsina> gatox: haven't been, let me check
[16:52] <rye> bug #869791
[16:52] <alecu> mandel, it's a different get_proxy_settings, right?
[16:52] <rye> filed by me
[16:53] <alecu> mandel, it's only used by that testcase class and classes that derive from it
[16:53] <mandel> alecu, ah, I see what you did there, really smart :)
[16:54] <alecu> mandel, so it's used by the other testcases....
[16:54] <alecu> well, you got the idea :-)
[16:54] <ralsina> gatox: you mean removing the uderline, or the color change?
[16:54] <gatox> ralsina, orange
[16:54] <gatox> ralsina, orange depending on the os theme really
[16:54] <ralsina> gatox: I don't see any orange here
[16:54] <mandel> alecu, yes, that is why I said smart :)
[16:54] <mandel> with gatos is always orange..
[16:55] <gatox> ralsina, here in your cp..... or the screenshot
[16:55] <gatox> mandel, jejejeej
[16:55] <ralsina> gatox: I am looking at my cp and the current tab ha sno orange, and the hovered tab has no orange
[16:56] <gatox> ralsina, ahhhhh sorry..... you have to navigate pressing tab
[16:56] <gatox> ralsina, when the tab reaches there you will see the orange thing
[16:56] <ralsina> gatox: set the policyfocus for tabs to nofocus
[16:56] <ralsina> gatox: the focuspolicy
[16:56] <gatox> ralsina, groso!
[16:56] <ralsina> gatox: nessita and I had the same problem with buttons last week :-)
[16:57] <ralsina> gatox: details on exactly how the heck to do that, are left to the actual developer
[16:57] <nessita> ralsina: I finally did not do that...
[16:57] <ralsina> nessita: I know
[16:57] <gatox> ralsina, yes..... it's something of the os..... i knew in my previous job..... we fight a lot trying to fix that
[16:57] <ralsina> nessita: but hey, it was a similar problem
[16:57] <nessita> ralsina: if we set the policy to nofocus... can the user tab between widgets and know in which widget he's in?
[16:57] <ralsina> nessita: just for the tabs
[16:58] <ralsina> nessita: and the way to switch tabs is ctrl+tab anyway
[16:58] <nessita> ralsina: but then we have the same visual bug in the rest of the UI... buttons also have a orangish box on top of them when tabbed into
[16:58] <gatox> nessita, ralsina, i think it makes sense to do that for tabs..... not sure for buttons
[16:58] <ralsina> OTOH, ctrl+tab is working intermittently at best
[16:58] <nessita> ralsina: which I think is great and we should not remove, though we need to fix the style for that
[16:58] <nessita> gatox: ^
[16:58] <ralsina> yes, the buttons need another solution
[16:59] <gatox> nessita, so..... do you agree for removing the focus for the tabs and looking for another solution for buttons?
[16:59] <ralsina> nessita: try running in other style, and you will see what it's trying (and failing) to do there
[17:00] <ralsina> nessita: it looks like it's trying to draw a dotted box to highlight the focused widget, and for some reason it's  filling with the wrong color
[17:00] <nessita> gatox: let me try what ralsina proposes...
[17:01] <ralsina> nessita: you may also need to specify a blank stylesheet
[17:01] <gatox> mmmmm i see
[17:02] <ralsina> so, it's really a style bug, I think, but maybe we can find a workaround
[17:02] <nessita> ralsina, gatox: from what I see, I think the problem is that the focused "area" is too small, we're playing with the size of things and we're not adjusting the size of the focused thing
[17:03] <ralsina> nessita: could be widget padding
[17:03] <nessita> ralsina, gatox: for instance, can we adjust the size of a label of a button instead of the button itself, in order to have the same button size as consequence?
[17:03] <ralsina> nessita: also explains the tall tabs
[17:03] <nessita> ralsina: exactly
[17:03] <nessita> ralsina: I'm pretty sure is our fault
[17:03] <nessita> ralsina, gatox: I will share a screenshot of what I mean
[17:04] <gatox> ralsina, the tall tabs is request as design
[17:04] <nessita> gatox: I know, but why not adding the top padding to the tab label?
[17:04] <nessita> instead to the tab itself
[17:04] <nessita> or something analog
[17:04] <ralsina> gatox, nessita: u1cp with gtk style without ubuntuone.qss: http://www.flickr.com/photos/ralsina/6815950482/in/photostream
[17:05] <nessita> ralsina: that's my point, we're not handling button size properly regarding the label in them
[17:05] <ralsina> right
[17:05] <gatox> nessita, ok..... let me keep playing with that..... i was trying to specify some styles to the label that is showing the text inside the tab..... but as i say..... tabbar handles that part in a kind of weird way
[17:05] <nessita> ralsina, gatox: see here http://ubuntuone.com/6Gx6hGIC8E81Xx5ZSalacz
[17:06] <ralsina> without the styling, the orange translucid square is much larger
[17:06] <nessita> ralsina, gatox: leaving tabs aside, I think the dotted square in that screenshot should be as big as the button,
[17:06] <nessita> ralsina, gatox: do you agree?
[17:07] <gatox> nessita, yes
[17:07] <gatox> should be
[17:07] <nessita> ok, I will have lunch now
[17:09] <mandel> alecu, FYI running tests, I don't see any issues in the code, I'll do an extra code review
[17:12] <rye> oopsing launchpad
[17:12] <rye> that's how i file bugs
[17:15] <gatox> i think i found it
[17:16] <alecu> mandel, \o/, thanks!
[17:22] <mandel> alecu, +1 from me
[17:30] <dobey> rye: you have to stop nautilus from self-restarting by the session. then you can LD_PRELOAD the locally built plug-in
[17:31] <dobey> rye: i don't always file bugs, but when i do, i get a TimeoutError from Launchpad.
[17:31] <dobey> mandel: btw, you suck
[17:31] <rye> dobey, yes, that's a proper version
[17:33] <mandel> dobey, thank you! why?
[17:34] <dobey> mandel: i fixed the typo and you haven't re-reviewd my branch
[17:34] <dobey> https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/update-from-trunk/+merge/96250
[17:34] <mandel> dobey, step by step, going
[17:37] <mandel> dobey, done
[17:37] <mandel> dobey, even with a lovely typo :P
[17:37] <mandel> stupid god and good
[17:37] <mandel> who makes two words like that :P
[17:37] <dobey> i feel the same way about that
[17:38] <dobey> oh
[17:47] <alecu> mandel, thanks for the review! :-)
[17:49] <mandel> alecu, reviewing that code makes me feel smart hehe
[17:51] <alecu> mandel, why?
[17:51] <alecu> mandel, also, the following branches are much smaller ;-)
[17:52] <mandel> alecu, because is interesting code to read
[17:52] <mandel> alecu, so much more fun that reviweing ui code :)
[17:52] <alecu> lol
[17:54] <alecu> mandel, oh, and this reminds me
[17:54] <alecu> mandel, I need to discuss "TODO: add authentication here, to replace the empty user/pass" in that branch
[17:54] <alecu> mandel, how should we get the authentication credentials in the tunnel?
[17:55] <mandel> alecu, well I guess that the same way we get them in the rest, right?
[17:55] <alecu> mandel, my guess is that we could get them from the keyring directly if we get the "proxy creds needed" signal
[17:55] <alecu> mandel, hmm
[17:55] <mandel> alecu, we could use the same code that we have in the sso and refactor it out
[17:56] <alecu> mandel, so, we should open the new process just like SSO does?
[17:56] <alecu> mandel, imagines this:
[17:56] <alecu> mandel, imagine this:
[17:57] <alecu> mandel, the user has proxies configured, and starts SD for the first time... so the sso dialog opens
[17:57] <alecu> mandel, (since there were no u1 creds in the keyring)
[17:57]  * mandel follows the train of though..
[17:57] <mandel> alecu, so the user gets the creds, sso needs the proxy so gets the proxy creds and stores them in the keyring..
[17:58] <alecu> mandel, then, sso opens the proxy credentials dialog, because its trying to connect outside, but the proxy requires creds.
[17:58] <alecu> mandel, right.
[17:58] <alecu> mandel, so, this answers my question.
[17:58] <alecu> mandel, the proxy dialog won't be shown twice as I feared.
[17:58] <alecu> mandel, perfect.
[17:59] <mandel> alecu, we need to be careful in the order in which the process are started, that is all
[17:59] <mandel> alecu, as long as one waits for the other on creds, we are fine, when is the tunnel started if I have u1creds? from the control panel?
[17:59] <alecu> mandel, so, when sso returns the creds to SD, SD will try to cross the proxy, and will call the process to get the credentials...
[17:59] <alecu> mandel, I mean, the proxy credentials
[18:00] <alecu> mandel, and what would happen at that point?
[18:00] <alecu> mandel, SD will get the credentials from the keyring by itself, or it will start a new process???
[18:01] <mandel> alecu, the webclient always checks if the creds are in they keyring before it asks for them
[18:01] <mandel> s/they/the
[18:01] <dobey> alecu: all of that should be done by the 'tunnel' process, and not SD itself, no?
[18:01] <alecu> dobey, right.
[18:02] <mandel> alecu, so, if we do have the u1creds and we did not open the control panel there is no guarantee that the proxy creds are there
[18:03] <mandel> alecu, someone could have deleted them or god knows what, so sd, starts, gets u1 creds from keyring and then tries to get the proxy creds, but they are not there!!! we are doomed!
[18:03] <mandel> alecu, in the case I mean
[18:03] <alecu> mandel, most common case, the credentials expired.
[18:04] <mandel> alecu, true.. so we need to show the dialog to get them, the deal is, I don't think executing a diff process gives us anything
[18:04] <alecu> mandel, so, in this case, and when the SSL certificate to the proxy is invalid, I think we should not pop up a dialog from SD
[18:05] <alecu> nor from the tunnel process
[18:05] <alecu> I think we should "wiggle" the u1 icon on the launcher.
[18:05] <mandel> alecu, this is one of those decision we should get design in
[18:06] <alecu> and when that happens, the control panel opens, it tries to connect to get the account details, and then the dialogs are shown.
[18:06] <mandel> alecu, I don't want to write a line of code to later have complains about it.. let them draw a diagram with when to show them
[18:06] <dobey> alecu: that doesn't work on windows :)
[18:06] <dobey> there's no launcher wiggling. not sure what we can do with the tray icon
[18:07] <dobey> or what we'd do on osx
[18:07] <dobey> or if the user isn't running under unity
[18:07] <nessita> ralsina: commented on https://code.launchpad.net/~ralsina/ubuntuone-control-panel/alerted/+merge/96253
[18:07] <mandel> alecu, dobey lets get design in, they get paid for this things :)
[18:07] <alecu> mandel, right, but we need this to land by next friday, so we can make it to beta 2 :-(
[18:07] <ralsina> nessita: ok, looking into it
[18:07] <mandel> alecu,  la puta.. ok we tell them quickly, if needed we do a mumble tom at least to get everyone in the same page
[18:07] <dobey> alecu: then tell your manager to pressure design for it :)
[18:08] <dobey> that's what managers are for
[18:08]  * mandel throws a pokeball
[18:08] <mandel> ralsina appears, attacks with pressure!
[18:08] <alecu> ralsina, ^^^^
[18:09] <gatox> mandel, omg! jeje
[18:09] <dobey> the idaa of mandel playing pokemon, is extreme hilarity in its own right
[18:09] <alecu> ralsina, we need to decide how to show the proxy dialogs from the tunnel that's ran by SD
[18:09] <mandel> :)
[18:09] <alecu> ralsina, usually SD should not pop up dialogs unless the user initiates a connection from the control panel (or from u1sdtool, but it's the same)
[18:10] <alecu> ralsina, but it might be the case that the proxy password has changed or that the proxy certificate has expired.
[18:10] <alecu> ralsina, so we need a way to show those dialogs when SD starts.
[18:11] <mandel> alecu, on windows, we ought not to worry, I think control panel starts sd..
[18:11] <mandel> alecu, due to the little icon launcher thing
[18:11] <alecu> mandel, yup, I'm not worrying about windows just yet.
[18:11] <alecu> mandel, I'm worrying about showing it right on unity
[18:12] <mandel> alecu, yeah.. it might be an issue.. we could tell sd that there is no connection by adding a new evetn in the q, then when control panel appears add the connect one
[18:14] <alecu> mandel, yup, we have a few ways of doing it. But I'm concerned on how we show it.
[18:15] <alecu> mandel, in the worst case we can just pop up the dialog. We still do that for "out of space" situations, iirc.
[18:15] <dobey> alecu: no we don't
[18:16] <alecu> dobey, how are we notifying out of space now?
[18:17] <dobey> alecu: i'm not entirely sure, but i think it involves icon wiggling in unity, and popping up the control panel behind everything. though i'm sure that doesn't work so well in trunk
[18:18] <dobey> if that is in fact what we do
[18:18] <dobey> someone should test it :)
[18:21] <mandel> alecu, EOD for me and I have to go to the hospital, brother went to R and had to stay there the entire day.. what a geek family los de la pena are..
[18:21] <nessita> gatox: commented on https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/too-verbose/+merge/96364
[18:21] <gatox> nessita, ok
[18:21] <mandel> alecu,  can you keep me posted over email?
[18:21] <dobey> yay. switching the series for a bunch of milestones from trunk to a stable series did exactly what i thought it would do
[18:21] <nessita> mandel: what's R?
[18:21] <dobey> now will have to do it for all the other projects too
[18:21] <ralsina> alecu: hmmmm honestly, if people change the passwrd in the proxy and u1 stops working, and we can fix it by telling them "start u1cp" I think I can live with that.
[18:22] <mandel> nessita, alecu maint to say ER, ups
[18:22] <ralsina> alecu: unless the fix is easy, of course.
[18:22] <alecu> nessita, a statistical language?
[18:22] <mandel> nessita, urgencias
[18:22] <nessita> mandel: oh, uh
[18:22] <nessita> good luck there then
[18:22] <mandel> nessita, yeah, we are a weak family ;)
[18:22] <mandel> nessita, also, I updated lp:~mandel/ubuntu-sso-client/webclient-use-dialog/ using the ubuntu_sso vars from __init__
[18:22] <nessita> mandel: I would say inkury-prone :-P
[18:23] <nessita> mandel: thanks
[18:23] <dobey> mandel: did he get a rugby injury too?
[18:23] <alecu> ralsina, but how will the user find out that u1 stopped working because of this?
[18:23] <mandel> nessita, let me know in the comments if there are any other issues
[18:23] <gatox> nessita, ok, i'll do that..... but can this branch wait until i finish with my tests for the highlight focus one?
[18:23] <ralsina> alecu: good point
[18:23] <mandel> dobey, nah, he has some kind of respiratory problem.. and the others just go to the gym :P
[18:23] <ralsina> alecu: so the fix involves what?
[18:23] <mandel> ok, all, see you tom!!
[18:23] <gatox> mandel, bye
[18:23] <dobey> heh
[18:24] <dobey> alecu: so we wiggle the icon if we can (under unity), and just pop the dialog if unity isn't available i guess?
[18:24] <nessita> gatox: for sure
[18:26] <alecu> dobey, ralsina, I see that we are both showing a notification and setting the launcher to urgent when we run out of space.
[18:26] <alecu> that's in alert_user() in u1/status/aggregator.py
[18:26] <ralsina> dobey: yes
[18:27] <ralsina> alecu: yes to that too
[18:27] <alecu> on windows we are doing *nothing* on out of space right now.
[18:27] <dobey> alecu: i think we should only show that notification if we are under unity though. popping a dialog and a notification at the same time without unity is a bit odd
[18:28] <alecu> dobey, afaik we are not popping up a dialog.
[18:28] <dobey> alecu: when unity api isn't available
[18:28] <dobey> the unity case is easy. so we need to figure out the solution for the non-unity cases
[18:28] <ralsina> we are not showing any feedback ever on windows
[18:29] <alecu> dobey, right now we not doing anything for the non-unity case.
[18:29] <dobey> because, we do have people using it on kubuntu, xubuntu, fedora, etc
[18:29] <alecu> dobey, so, we have to issues:
[18:29] <alecu> 1) we need to show alerts when unity is not installed.
[18:29] <dobey> well, > two; but yes :)
[18:29] <alecu> 2) we need to show an alert when the proxy needs credentials.
[18:30] <alecu> for 2 we can use the current code in the aggregator.
[18:30] <dobey> alecu: do we need to solve this *now*, or can it wait until later today/tomorrow?
[18:30] <alecu> dobey, well, yes, a lot more :-)
[18:30] <alecu> dobey, I have enough info for 2
[18:30] <alecu> dobey, and I don't care too much about 1
[18:31] <alecu> :-)
[18:31] <dobey> well, i would like to have /more/ users, and not less. especially if they are going to pay us :)
[18:31] <dobey> which is yet another reason for me to go all hobo on the ubuntuone-installer after precise
[18:32] <dobey> alecu: ok, well if you want to make it work for the unity case now, and we can solve the non-unity cases later, that is fine with me
[18:32] <alecu> great, thanks!
[18:32] <ralsina> nessita: about the test for switch_to: we don't want to test switching to folders_tab first because it succeeds even if if it fails, since it's the default tab, that's why I test each tab explicitly
[18:33] <nessita> ralsina: you can reverse the list, no?
[18:33] <ralsina> nessita: but I am switching to indexes instead of constant integers, though
[18:33] <dobey> but i do want to solve the non-unity case as well; i just have a lot of other work to do at the moment, which is more important. though i think my expreience is useful here :)
[18:33] <nessita> ralsina: or put the list explicitely in the test in the order we want in the test
[18:34] <ralsina> nessita: yes, I can do both. What I do now is test that the default is right, that changing "wrong" is a noop, and then each tab. I can replace the "each tab" with a loop.
[18:34] <nessita> ralsina: perfect
[18:39] <ralsina> nessita: argh, the loop yu suggested is not really doable because the "settings" tab is called "preferences_tab"
[18:39] <ralsina> nessita: so I can only do a loop over two lists of manually specified names and widgets
[18:39] <ralsina> nessita: and then it's more complicated than the manual test
[18:40] <nessita> ralsina: ack then
[18:44] <ralsina> nessita: ready for your re-review whenever you want
[18:46] <nessita> ack
[18:48] <duanedesign> have you all ever heard of a user getting nknown 'uencoding: idna' while logging into Win Client?
[18:49] <alecu> duanedesign, do you have a more detailed error or a screenshot?
[18:49] <dobey> idna? that would be a weird encoding indeed :)
[18:50] <dobey> and definitely unknown
[18:50] <alecu> dobey, idna is the encoding for i18n domain names.
[18:50] <duanedesign> I will gwt one and or noth for you
[18:56] <dobey> weird
[18:58] <dobey> alecu: it's not a character encoding itself, so much as a means to convert between ascii/unicode for DNS, no?
[18:58] <alecu> dobey, right, it's a way to encode unicode in DNS.
[18:59] <alecu> but it's very weird that our client is showing that error.
[18:59] <dobey> it's beyond weird. it's implausible. :)
[19:00] <dobey> we aren't using unicode domain names for anything. does twisted use some idn library that creates some 'idna' character encoding in python?
[19:02] <dobey> nessita: btw, i know how we can fix all the watch files to work correctly again! we can do it for the next release :)
[19:03] <nessita> dobey: teach me!
[19:03] <dobey> i will (later) :)
[19:03] <dobey> it's a tedious bit of work :)
[19:05] <nhaines> Hmm, the ayatana notification for Ubuntu One always tells me "foobar and 199 other files are being downloaded to your computer" (on this new computer I started syncing again) but I have... well, let's say thousands of files.  :)
[19:05] <ralsina> nhaines: known bug
[19:05] <ralsina> nhaines: tricky to fix, we may change the wording to "a truckload of files"
[19:06] <nhaines> ralsina: okay, wasn't sure if it was only queing up 200 files at a time or there was some other weird thing going on.  :)
[19:06] <duanedesign> rye_: have time for a real quick one :)
[19:06] <ralsina> nhaines: everything is queued, we are counting them wrong (sorta, kinda, in a way)
[19:06] <duanedesign> rye_: ticket that is
[19:07] <nhaines> ralsina: I see!
[19:32] <dobey> hrmm; dependencies are hard
[19:33] <ralsina> dobey: let's go shopping?
[19:33]  * ralsina imagines dobey barbie
[19:33] <dobey> i don't think there is any way to do what i want to do, in a debian control file
[19:34] <nhaines> u1sdtool doesn't seem to terminate, either, hrm.
[19:34] <nhaines> 'u1sdtool --current' gives me 9 transfers and then just sits there.  Ctrl-C doesn't work.
[19:34] <dobey> nhaines: use Ctrl-\
[19:34] <dobey> nhaines: also a known issue
[19:35] <nessita> ralsina: would you update the new splash screen to the uife bug for that, please?
[19:35] <ralsina> nessita: sure
[19:35] <nhaines> dobey: lots of those, eh?  Well, at least there's a workaround.  :)
[19:36] <nhaines> nessita: ooh, what's the bug number?  I want to see the new splash screen early.  :)
[19:36] <dobey> nhaines: yeah, welcome to the gobject-introspection world :)
[19:36] <dobey> i don't know why people are saying "splash screen"
[19:36] <dobey> we don't have a splash screen
[19:36] <ralsina> installer
[19:37] <jalcine> like usplash?
[19:37] <dobey> no
[19:37] <dobey> not usplash
[19:39] <jalcine> oy, I read and think too concurrently.
[19:39] <nhaines> UbuntuOne OS!
[19:39] <nhaines> "Now syncing /usr with the cloud."
[19:40] <dobey> i think i need to get a dust cover for my midi keyboard
[19:40] <jalcine> lol, that'd be great.
[19:42] <dobey> i guess i'll just have to add another check to the indicator before we use it.
[19:43] <dobey> and leave as-is-ish for now
[19:57] <nessita> alecu: ping
[20:01] <gatox> people..... EOD! see you tomorrow..... if i'm still alive :P
[20:05] <ralsina> EOD for me too. I think I am catching something. Maybe gatox infected me over IRC.
[20:05] <ralsina> see you all tomorrow, no late night shift from me today...
[20:05] <gatox> ralsina, jeje maybe
[20:17] <nessita> dobey: hi there! any ETA for 2.99.90 installer and client packages (I saw the tarballs already)?
[20:17] <alecu> nessita, I've just seen your ping, but I'm racing to kinde
[20:17] <alecu> r
[20:17] <nessita> alecu: ok
[20:18] <nessita> alecu: ping me when you get back pliz
[20:18] <alecu> nessita, I'll be back later, so mail me please
[20:19] <dobey> nessita: soon. building them in pbuilder isn't the fastest thing in the world.
[20:20] <nessita> dobey: been there, done that. Thanks
[20:20] <dobey> nor is having to triage the bugs :)
[20:20] <nessita> dobey: use a script ;-)
[20:20] <nessita> I have some magic going on... you interested in it?
[20:20] <dobey> i would have to write a script, that was smart enough to triage them correctly
[20:21] <dobey> also, i was looking up how to do a complex conditional dependency; though that is apparently not possible to do
[20:21] <dobey> anyway, just uploaded client
[20:22] <dobey> nessita: btw you can just upload the packages with new dependencies and it will just wait to build
[20:22] <dobey> or won't be installable until the new ones show up
[20:23] <nessita> dobey: yes, but you mentioned that it would be best if I avoid doing that... no?
[20:23] <nessita> dobey: regarding the script, I usually focus the "manual" work on writing a proper changelog in a local file, and then using that file, I have scripts that take the bug numbers from there and do all the needed stuff
[20:24] <dobey> woah. the CD is 12MB bigger than it was yesterday
[20:24] <dobey> oh no. 17MB
[20:25] <dobey> uhm
[20:26] <nessita> dobey: we weren't the cause of that!
[20:26] <dobey> actually, we were!
[20:26] <dobey> but i have no idea how
[20:27] <nessita> dobey: eh?
[20:27] <nessita> dobey: we were? how?
[20:27]  * nessita passes out
[20:27] <nessita> did I screw up an upload?
[20:27] <dobey> nessita: ah. the problem is ubuntuone-control-panel recommends ubuntuone-control-panel-gui
[20:28] <dobey> nessita: can you remove that recommends, or make it a suggests instead?
[20:28] <nessita> dobey: I can, sure, but I wonder
[20:28] <nessita> dobey: what is pulling ubuntuone-control-panel in?
[20:28] <dobey> nessita: because deja-dup needs ubuntuone-control-panel
[20:28] <dobey> for the dbus service
[20:28] <nessita> aahhhahahahhhh
[20:28] <nessita> I see
[20:29] <nessita> ok, will swicth to suggests
[20:29] <dobey> thanks
[20:29] <nessita> prgo
[20:34] <nessita> dobey: just to be sure I got the fix right, is just this, no? http://pastebin.ubuntu.com/873577/
[20:35] <dobey> nessita: yeah, that should do it
[20:38] <dobey> nessita: and installer uploaded now too
[20:38] <nessita> dobey: thanks! will re-review the packaging branch and upload as well
[21:24] <nessita> ok, I'm gone
[21:24] <nessita> bye all, see ya tomorrow
[22:35] <nhaines> Ooh!  "Thriller.mp3 and 198 other files are being downloaded to your computer."