=== jalcine is now known as webjadmin_ [08:54] hola [10:18] * mandel restarting due to updates [10:51] * mandel back from update [10:58] appport-bug compiz [10:58] Dammit. :( [11:01] good morning! [11:10] And https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/952887 is the bug I was filing. [11:10] Ubuntu bug 952887 in compiz (Ubuntu) "Transparent terminal windows and black backgrounds in other windows after kernel update" [Undecided,New] [12:00] hello all! [12:03] alecu, hi! [12:16] wow... moving a folder in evolution has managed to screw up my system.. wtf! [12:17] alecu, gatox morning! :) [12:17] mandel, hi [12:17] ralsina, mandel, et al: Gotta run to a follow-up eye exam after dropping Lex off at school, so I might miss our standup. Here's my update: [12:17] DONE: Not much, I'm afraid. Floundered a bit, so started spelling out an actual ubuntuone-darwin-installer project to make some forward progress. [12:17] TODO: Dunno if briancurtin is back from PyCon this week or if he's participating in sprints, but maybe pair with him, otherwise keep going. [12:17] BLOCK: In the process of upgrading my machine to Lion, I seem unable to get the tests that we had running going again, so I'm a bit blocked on getting new tests for the darwin-specific code written. [12:18] urbanape, ok, good luck with the eyes tests, I hope your sight is as bad as it was :) [12:41] I need to go to the back to solve some little problem, Ill be back in a few mins [12:49] hello everyone! [12:52] nessita, hi! [12:55] good morning! [12:55] ralsina, hi o/ [12:56] urbanape: if you are still around, because of flight combinations, brian will only be back tomorrow [12:56] urbanape: and good luck with the eye-doctor! [12:56] hello gatox! [13:04] Noone has commented on my app indicator. I am so sad I am going to start doing schedules and phonecalls like a manager should instead of coding. [13:05] ralsina: do i really need to comment? [13:05] dobey: haha [13:05] dobey: what would you say if I told you it has no twisted? That's a +1 from yu, right? ;-) [13:05] but it's qt [13:06] dobey: yes, that it is [13:06] dobey: it's what I know. I am not doing vala on sundays. [13:06] doesn't need to be vala. if we're going to provide an app indicator, it should be in syncdaemon, as the current messaging indicator integration is, i think [13:07] but meh [13:08] dobey: but the indicator should work even if syncdaemon is stopped [13:08] dobey: IMVHO [13:08] New -> Opinion [13:08] :) [13:08] dobey: haha [13:08] it's time to change our name to "GNU/Linux One" [13:09] dobey: well, we are packaged for Fedora now! [13:09] not exactly [13:09] i wouldn't call running alien a .deb and turning it into an rpm "packaged" [13:10] if you squint and make loud noises, it feels like we are ;-) [13:10] * mandel back [13:18] ralsina, dobey wow... people have very interesting opinions. Also, I wonder if they know how many of us are in the desktop+ team, is not that we have the time to package this for everything [13:18] ralsina: i'd definitely rewrite my indicator in vala, having 50Mb of python process just showing a cloud is an overkill [13:19] rye_: the indicator I did is totally redoablein C++ [13:19] ralsina: but, I guess you need to write it in Qt [13:19] rye_: except for the OAuth signing, which I have no idea how to do outside python [13:19] ralsina: ubuntuone-indicator is also totally redoable in anything that has DBus and HTTP library and sha1 somethings [13:19] ralsina: we should never write C++. [13:19] ralsina: unless it's to patch Qt or something upstream. [13:19] dobey: ok, C# ;-) [13:20] dobey: it's my own toy, don't be a killjoy ;-) [13:20] 36 Mb now, in IDLE [13:20] ralsina: write it in go! [13:20] dobey: which brings us to the question - "HOW?" :) [13:20] dobey: I was thinking cobra [13:20] dobey, go + dbus.. is not ready yet, right? [13:20] go + libindicator? [13:20] rye_: it has dbus client bindings [13:20] go + oauth? [13:21] ralsina: oauth is kind of simple [13:21] once you have the tokens [13:21] mandel: you can't use it as a server, but you can as a client, at least [13:21] rye_: I need signed URLs, how's that done? [13:21] dobey, oh, I did not know that [13:22] ralsina: exec oauth-sign which signs a url and prints the results on stdout. it's the unix way. [13:22] dobey: the unix way is bumpy? [13:23] ralsina: sign them? http://oauth.net/core/1.0/ [13:24] but still it is kind of weird that we are using qt binding for python to tak to indicator libarary which uses gtk [13:24] rye_: and implementing oauth is exactly the kind of things that is not fun to do and for which a library written by security-aware people should already exist ;-) [13:24] rye_: it was easy. [13:25] rye_: and I am on uity-2d so qt all the way ;-) [13:25] ralsina: if only oauth was written by security-aware people… ;) [13:26] dobey: well, I at least want I library full of /* */ and /* WTF!? */ [13:27] ralsina: unity-2d is not using qt for the panel as far as I remember [13:27] dobey: ^ ? [13:27] rye_: no idea, really [13:28] ralsina: only launcher and dash [13:28] hold on. disk i/o [13:28] Now onto real work and my freaky sunday project sleeps until sunday. Any pending reviews? [13:28] rye_: ok, so Qt is on memory anyway [13:29] rye_: uh. well, it links to both [13:29] rye_: but i think it's rendering with qt. not sure where the gtk link comes from [13:31] ralsina: btw, were you checking bug #946626 with trunk, or plain ubuntu 12.04? :) [13:31] Launchpad bug 946626 in ubuntuone-control-panel (Ubuntu) "ubuntuone-control-panel does not adapt to length of translated messages" [Undecided,New] https://launchpad.net/bugs/946626 [13:31] dobey: trunk [13:32] ralsina: maybe it's fixed in trunk already, but not in the release? [13:32] dobey: you tried it on last release and it fails? [13:32] dobey: I can't go back to release because I am on nightlies [13:32] dobey: and have no other P-box [13:32] ralsina: no, i'm just saying because i think the guy isn't running nightlies [13:33] ralsina: bzr branch lp:ubuntuone-control-panel -r release-2_99_90 ? [13:33] dobey: booooooring [13:33] dobey: testing it in 1' [13:34] * dobey shsould change the system language on his computers [13:34] dobey: bzr branch lp:ubuntuone-control-panel -r release-2_99_90 do you remember the revision name? [13:35] oops, I cut it off, it says that revision doesn't exist? [13:35] err? [13:36] ralsina: oh. right. lp:ubuntuone-control-panel/stable-3-0 :) [13:37] so you can just grab tip because i'm sure we haven't merged anythng else [13:37] dobey: hehe, stable branches dude! ;-) [13:37] it's early, dude [13:37] dobey: yeah. Sorry, I am a little manic until I find the right sugar level [13:38] * dobey puts on that bit of hendrix [13:39] gatox: this branch https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/ugly-reset/+merge/96130 looks great, but any chance you can apply this diff to the .ui file? I was curious why the width was getting increased, and why you were adding right margin to the page, so I played a little and found that that the fixed size for the ballong was too wide [13:39] gatox: the diff is http://pastebin.ubuntu.com/880337/ [13:39] sigh [13:40] * gatox looking [13:40] dobey: looks good on 3-0 [13:42] nessita, just asking: did you try that IRL and the widgets don't move when things appear? [13:42] gatox: yes [13:43] nessita, ok then [13:43] ralsina: weird [13:43] dobey: may be related to having a really really long name, though. Can I change mine somewhere? [13:43] ralsina: looks like he filed it against 2.99.5 [13:44] ralsina: and he doesn't have a long name, though it looks like in his screenshot, the window is maximized [13:44] so, I suppose we fixed it [13:45] u1cp looks surprisingly nice maximized [13:45] gatox: hum, wait, I think I tested a not compiled version, let me re try [13:45] nessita, ok ok [13:46] ralsina: you can't see the rest of the world to compare it to? :) [13:46] dobey: haha the problem is yours is too tiny [13:47] anyway, gatox can I get a review for https://code.launchpad.net/~ralsina/ubuntuone-control-panel/translate-all-the-dialogs/+merge/96889 please? [13:47] ralsina: the fonts in qt control panel are hard coded :( [13:47] ralsina, yep, on it! [13:48] dobey: really? file a bug please [13:48] well they're all bold, and gigantic [13:48] so i presume they are [13:48] dobey: the "bold" thing is a known bug (and not ours, really) [13:49] dobey: the size looks good here, but I have everything set as default, so I suppose I wouldn't notice [13:53] ralsina, +1 [13:53] gatox: thanks! [13:54] gatox: confirmed this is the diff http://pastebin.ubuntu.com/880367/ (no need to change the ubuntu_sso/qt/reset_password_page.py file) [13:54] nessita, ok [13:54] * Chipaca returns from lunch [13:54] gatox: hola! what did the machinalis guys do that was so awesome? [13:54] * Chipaca curious :) [13:55] gatox: I noticed the label for the password assistance was bigger than the image, which is 198x98 [13:55] Chipaca, about what? when? :P [13:55] gatox: so if we want the text to fit the ballon image, we need the label to be smaller than that, and to wrap [13:55] gatox: oh, wait, old news [13:55] gatox: machinalis responding to a tweet of you from ... february? [13:56] Chipaca, ahhhh they assign some hours from their developers to code ninja-ide [13:56] assigned [13:56] :) [13:56] Chipaca, it was really helpful :D [13:57] I can imagine :) [13:57] glad they were carrying that spirit forwards [13:58] Chipaca, yes, it's really nice [13:59] gatox: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/forgotten-link/+merge/96814 globally approved [14:00] nessita, great, thanks [14:04] ok, lunch time for manuel [14:04] * mandel lunch [14:06] ralsina: regarding the translate-all-the-dialogs, isn't there a more "generic" call that can be made? the code you added looks like it requires a lot of "manual" work, speacially the building of the translator.load("qt_" + locale, ...) [14:06] nessita: nope,that's straight from the Qt docs [14:06] ralsina: bummer :-/ [14:07] nessita: yeah, not the prettiest possible code. I tried to test it as well as I could. [14:07] ralsina: looks good, approving [14:07] nessita: thanks! [14:12] nessita, the ugly-reset has been updated [14:12] gatox: awesome, will look after too-verbose [14:12] nessita, ok [14:13] hey, all. [14:13] back [14:13] Looks like I'm still good to go for LASIK on Thursday. [14:13] Should be back and fully functional after the post-op follow-up on Friday morning. [14:13] ralsina: in tab-tab-tab you have some leftovers (#self.is_processing = True). Also, question, is it really needed to move this code: [14:13] child.icon_obj = icon # hack! [14:13] child.setIcon(FOLDER_NAME_COL, icon) [14:13] item.addChild(child) [14:13] before the if like the branch is doing? [14:17] nessita: looking [14:18] nessita: yes, you can't add widgets to a QTreeWidgetItem that is not added to a QTreeWidget [14:18] ralsina: ah, ack [14:18] nessita: if you do, they all end invisible stacked in the top-left corner of the QTreeWidget [14:19] nessita: that took me 2 hours to figure out, BTW ;- [14:19] ralsina: would have taken more to me ;-) [14:20] gatox: too-verbose approved [14:20] nessita, cool [14:20] :D [14:20] nessita: removed the comment, will check the diff in a bit to find other randomness [14:20] ralsina: ack [14:22] gatox: not sure if you got it from before, but with the new changes to the ui file there is no need to add changes to ubuntu_sso/qt/reset_password_page.py [14:22] nessita, i'll check that right now [14:23] gatox: thanks! === JanC_ is now known as JanC [14:24] ralsina, can you review this: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/too-verbose/+merge/96364 when you have a moment? [14:24] gatox: of course! [14:24] ralsina, thanks [14:28] uhm [14:28] i won't say anything :) [14:28] dobey: failed at that! ;-) [14:32] nessita, :S did you really test it IRL? ugly-reset? [14:32] gatox: yes, I did, but le me try again? [14:32] gatox: wanna push the latest you have? [14:32] nessita, the balloon is not aligned with the password..... a spacer was removed [14:32] in your diff [14:32] gatox: how are youy running this? [14:33] I removed a spacer that was doing nothing (I think :-)) [14:33] nessita, jejee was aligning the balloon with the line edits :P [14:34] ralsina: https://bugs.launchpad.net/ubuntuone-control-panel/+bug/953062 [14:34] gatox: ASCO TOTAL :-D [14:34] Ubuntu bug 953062 in Ubuntu One Control Panel "Font and sizes are hard-coded" [Undecided,New] [14:34] dobey: yes! that's true :-) [14:34] nessita, what?? [14:35] gatox: using spacers to align stuff that may changed because line wraps or font changes is ulgy, IMHO [14:35] nessita: I know we hardcoded them for windows, but on Ubuntu we should follow the desktop. The default is ubuntu font anyway. [14:35] ralsina: the .qss is compiled into the control panel regardless of what platform we're on [14:35] ralsina: we need to have the same looks in every OS... I understand that that means do not change the looks if the theme changes, no? [14:35] dobey: I know [14:35] nessita, but it looks uglier now [14:35] dobey: yes, because of that ^ [14:36] ralsina: and it is where the fonts are defined [14:36] nessita: but suppose for example, a locale where ubuntu font doesn't have the glyphs... [14:36] nessita: i don't think it means "don't follow the theme" [14:36] gatox: then we need a grid there instead of what we have, and have the ballon in the "second" row instead of using a spacer to align it [14:36] nessita: since the default is ubuntu, I think we can be a bit lax in Ubuntu [14:36] nessita: to me, it means "follow the same general layout and workflow" [14:37] dobey: well, we can ask, but I understood is "also follow the same colours, fonts, sizes, and shapes" [14:37] dobey: we may add "keep branding color scheme" but I am with you on the fonts. Let me ping upstairs. [14:37] theme != branding [14:37] i understand how marketroid types can get that confused, though [14:38] nessita, ok..... i'll change the layout of that page then [14:39] gatox: ack, try to keep it as simple as possible [14:39] dobey: let's not have that argument now, when I am on your side on the fonts issue mmmmkay? ;-) [14:40] gatox: +1 on verbose [14:40] ralsina, thanks [14:40] it's not an argument. it's a fact. simply because i'm not on the design team, doesn't make me not a designer :) [14:41] * ralsina goes back to coding mumbling grmblcocacolaredwhiteyaddambrgl [14:42] we can just bring back the old applet and preferences :P [14:43] dobey: wanna rotate to design in the R cycle? [14:43] dobey: you'd have to move to london and work in an office though. With people. [14:44] no [14:48] dobey: cool, we need you (group hug!) [14:52] nessita: tab-tab-tab looks clean to me now (revno 291) but maybe I am just tired of looking at it [14:52] ralsina: will look after a couple of on going reviews [14:52] nessita: thanks [14:52] * nessita will review all morning and will hack undergrounf after lunch [14:53] https://code.launchpad.net/~dobey/ubuntuone-client/transient-notify/+merge/96815 can has reviews? [14:53] dobey: on it! [14:53] nessita: remember tech leads, but if you want to skip it, do me a briefing [14:54] ralsina: may attend and do noth, lowering attention to teach leads [14:54] both* [14:54] will not type correctly today, though [14:54] nessita: cool, will ping you on IRC if I need you [14:55] ack [14:55] * mandel back [15:00] me! [15:00] me [15:01] me [15:01] what? [15:01] oh [15:01] damned dst. [15:03] meh [15:03] nessita: standup! [15:03] alecu, nessita? [15:03] oh [15:03] me [15:03] urbanape is at the doctor's [15:03] so gatox go! [15:03] DONE: [15:03] Propose some branches to improve the logging operations in u1-sso (landed), propose a branch that fixes a unicode bug in current user sign in page in u1-sso (landed). [15:03] TODO: [15:03] Refactoring the ui layout for Reset Password page, started working on Bug #940392. [15:03] BLOCKED: [15:03] (alecu and I are debugging a weird issue, sorry) [15:03] No [15:03] Launchpad bug 940392 in Ubuntu Single Sign On Client "Qt UI: policy_link gets cut off at the end" [High,In progress] https://launchpad.net/bugs/940392 [15:03] mandel, go [15:03] alecu is last [15:03] DONE: Implemented ssl error handeling in webclient libsoup. Got to deal with errors correctly I need to find out how to load the cert details from the pem file (probably I just need to look at openssl) [15:03] TODO: Read cert details from pem. Propose branch move to pinned certs. [15:03] BLOCKED: no [15:03] COMMENTS: nessita I added bug 952880 at the moment appname is a kwarg so that we do not brake control panel but we should make it an arg. [15:03] Launchpad bug 952880 in Ubuntu One Control Panel "The app name is not passed to the webclient" [Medium,Triaged] https://launchpad.net/bugs/952880 [15:03] ralsina, please! [15:03] DONE: a few branches, wrote a prototype indicator, reviews TODO: finish pushing tab-tab-tab tech leads call, pick a fewmore bugs, more reviews, administrivia, BLOCKED: no NEXT: dobey [15:04] λ DONE: initial installer changes, bug #887369 [15:04] λ TODO: finish installer, bug #951425, bug #934206 [15:04] λ BLCK: none. [15:04] Launchpad bug 887369 in Ubuntu One Client trunk "Please use transient hint so notifications in Gnome Shell stay out of the way" [Low,In progress] https://launchpad.net/bugs/887369 [15:04] Launchpad bug 951425 in Ubuntu One for Rhythmbox "Ubuntu One plugin has problems when being enable/disabled" [Undecided,Confirmed] https://launchpad.net/bugs/951425 [15:04] Launchpad bug 934206 in rhythmbox-ubuntuone (Ubuntu Precise) "track duplication from RB-U1" [High,Confirmed] https://launchpad.net/bugs/934206 [15:04] nessita [15:04] no notes, writting [15:04] nessita: DONE: got bitten by mistery animal [15:04] jejeje [15:04] or more than one, perhaps [15:04] :-) [15:04] me [15:04] DONE: started adding wizard pages to controlpanel, tons of reviews [15:05] alecu: you are after nessita [15:05] DONE: fixes for reviews: refactored tunnel_runner so tunnel_client can be in a different package [15:05] TODO: finish bug #929212 [15:05] BLOCKED: no [15:05] Launchpad bug 929212 in Ubuntu One Client "[FFE] Tunnel webservice calls if proxy is enabled" [High,Confirmed] https://launchpad.net/bugs/929212 [15:05] TODO: finish all the things I started in the DONE :-) [15:05] BLOCKED: no [15:05] NEXT: alecu but he already went :-D [15:05] alecu: precoz! [15:05] :-) [15:05] sorry for being distracted, but we have a weird issue with the new tunneling process in syncdaemon [15:05] ok, eom? [15:05] nessita: it's ok [15:05] eom! [15:06] NOTE: tomorrow I'm not coming to work! [15:10] nessita: before you hibernate if you want to throw a few bugs my way, I will appreciate it! [15:10] ralsina: wanna fix the tests using the es lang? :-) [15:11] nessita: happy to [15:11] ralsina: most of those need to use unicode :-) [15:11] nessita: it's just adding unicode() around things [15:11] ralsina: hem... not sure if is that... [15:11] nessita: at least the ones I looked are comparing unicode objects and QStrings [15:12] ralsina: ok, I'll leave it to you [15:12] nessita: cool, thanks [15:19] alright, need to get lunch, bbiab [15:28] * gatox lunch! [15:50] nessita, gatox, alecu,mandel: whenever you use strings you get from the UI on tests, convert to unicode first: https://bugs.launchpad.net/ubuntuone-control-panel/+bug/951716/comments/1 [15:50] Ubuntu bug 951716 in Ubuntu One Control Panel "Runing tests with LANG set makes some tests fail" [Low,In progress] [15:50] I will do a check and fix all the ones I find, but from now on, keep it in mind. [15:51] ralsina, yes [15:51] ralsina, perhaps we can remind about that on the team meeting. [15:51] ralsina, because writing a method in the test case in ubuntu one devtools for that is not an option? [15:51] ralsina, I mean, we can ensure that the self.assertEquals does the right thing [15:51] mandel: what would the right thing be? [15:52] mandel: not typecheck and convert all QStrings beforehand, right? [15:52] ralsina, I don't know what the right thing is, but certainly do not go hunting after string everywhere when you can do it in a single place, right? [15:53] ralsina, and looking at that message = we need to manually convert the QString to unicode prior to comparison. [15:53] ralsina, type checking in an assert is not THAT ugly [15:53] mandel: but to typecheck that, we need to import QtCore in devtools [15:53] and dobey will have a stroke ;-) [15:54] ralsina, make a QtTestCase and just import it when you need it, like the qtreactor thing [15:54] mandel: naaaaaaah [15:54] :-) [15:54] we already have so many base test cases :-) [15:55] mandel: and then I have to see all the things that use this and change their ancestors? That's even trickier :-) [15:55] ralsina, nessita ok, as you please, I just don't like to type, the more I type the more typos! [15:55] it's not even that important since tests always succeed in english ;-) [15:56] I assume setting the locale to chinese will trigger all possible failures, too [16:00] alecu, I'm getting a dirty reactor: with libsoup but not with qtnetwork.. have you seen that already? [16:02] nessita: tech leads! [16:09] what? [16:25] mandel, have you made sure that libsoup has disconnected? [16:25] mandel, usually it's the library (qt or libsoup) still connected for the next request (http/1.1) [16:26] alecu, I found the issue, we are just storing a single protocol instance, by appending them and then closing the connection from there is sorted out [16:26] mandel, ralsina: why would we need to add a utility function to devtools? [16:26] alecu, I'm done with the libsoup + ssl branch, will propose in a few mins [16:27] dobey: worry not, we will not [16:27] dobey, I was proposing to add an assert that will deal with the issue that ralsina pointed out about unicode and QStrings in ui tests.. [16:27] dobey, so, ignore it :) [16:29] we could add an assertUnicodeEqual() without the need for adding a QtCore import [16:31] hi, i published a file but when i go to th elink i just get could not locate [16:31] http://ubuntuone.com/2Buoib4zz4Abolyrp7SXla [16:31] lamalex: it opens it here. shinke.zip [16:31] wth [16:31] ok [16:32] guess ijust didnt wait long enough [16:32] perhaps [16:32] could use a better error message [16:34] lamalex: is it just a plain text error, or stylized page? [16:34] plain text [16:34] its not happenign now, now i get the zip [16:35] lamalex: file a bug that the error message could be better. we should probably have the stylized error page pop up, and say "Please try again in a few minutes." or something [16:35] indeed [16:42] nessita, alecu proposed the ssl for libsoup implementation here: https://code.launchpad.net/~mandel/ubuntu-sso-client/libsoup-ssl-dialog/+merge/97057 I added both of you for the reviews, I'm moving to the pinned certs (I don't know if there is a bug for that) [16:42] mandel: ack [16:42] mandel: perhaps you wanna takcle the passing the app+name to the webclient? [16:42] nessita, sure is a very easy one, is there a constant for the control panel app name? [16:48] gatox, do you know if there is an appname var in control panel? [16:48] mandel, nop, i don't know that about control panel [16:48] mandel, but, i don't think so.... [16:48] at least for me, it doesn't have much sense, because it's only an app.... [16:49] unless you want to reuse "Ubuntu One" string in several places maybe [16:49] mandel, in that case..... check at the __init__ with the translations [16:49] gatox: why would there be? [16:49] err [16:49] gatox, is just to show the Ubuntu One thing on the ssl errors.. I'll use the ubuntu one string [16:49] mandel: why would there be? cp doesn't work for other services :) [16:50] dobey, that's what i said [16:50] dobey, I'm asking just in case [16:50] mandel: the app name string that gets passed to SSO is in ubuntuone.credentials which is currently in ubuntuone-client [16:50] dobey, ok, thx! [16:50] mandel, look hoow the cp call the login from sso [16:50] mandel, that's the place where you will see if there is any, or how is being done right now [16:53] alecu: ping [16:53] bah, perhaps mandel? [16:53] sso dailies are failing with: [16:53] File "/build/buildd/ubuntu-sso-client-3.1+r907/ubuntu_sso/utils/webclient/tests/__init__.py", line 20, in from twisted.internet import ssl File "/usr/lib/python2.7/dist-packages/twisted/internet/ssl.py", line 23, in from OpenSSL import SSL ImportError: No module named OpenSSL [16:53] oops, that does not look good [16:54] mandel, alecu: did we add a new dep on ussoc? http://pastebin.ubuntu.com/880624/ [16:54] nessita, AFAIK no because openssl is already a dependency of twisted [16:55] mandel: perhaps is a specific twisted package we're not depending on? [16:55] nessita, yes, I'm checking right now [16:55] hum, no :-/ [16:55] nessita@dali:~$ dpkg -S /usr/lib/python2.7/dist-packages/twisted/internet/ssl.py [16:55] python-twisted-core: /usr/lib/python2.7/dist-packages/twisted/internet/ssl.py [16:57] nessita, the ssl namespace of twisted needs to have pyopenssl installed [16:57] nessita, I'm surprise that the twisted package does not get that as a dependency [16:57] seems like python-openssl is needed, looking at the packaging branch [16:57] mandel: same reason squid doesn't ship ssl support in ubuntu [16:58] ralsina: tab tab tab approved [16:58] i think [16:58] nessita: cool, thanks! [16:58] * nessita fixes [16:58] dobey, but.. I mean, it should be marked as a dependency in the .deb, right? [16:58] mandel: not necessarily [16:59] we could act gracefully if that is not available? [16:59] mandel: twisted doesn't Depends: on gtk3 gtk2 gtk1 and glib2 and glib1 and all the various python bindings, for example [16:59] gatox: could I get a second review for https://code.launchpad.net/~ralsina/ubuntuone-control-panel/tab-tab-tab/+merge/96687 ? [16:59] mandel: sometimes it makes, if there is an issue with dependencies [17:00] ralsina, of course..... [17:00] dobey, AFAIK twisted.internet.ssl is in python-twisted [17:00] python-twisted-core [17:00] as is gtk2reactor.py [17:01] dobey, yes, in python-twisted-core we have /usr/lib/python2.7/dist-packages/twisted/internet/ssl.py witch uses pyopenssl.. [17:01] mandel: yes. and many other things, which use many other things, though the twisted package itself doesn't depends on them [17:01] but I have no idea on packaging.. maybe there is an issue with that, but I find it odd that we let people install that without the required openssl lib [17:01] mandel: because ssl isn't a requirement for the twisted library itself to work [17:02] mandel: if you want to use gireactor in your app, you need to depends on python-gi in your app package. and if you want to use ssl, you'd need to depend on what the twisted ssl module depends on, to work [17:02] dobey, as I said, I have no idea, but if it is like that I would have imaging there was a python-twisted-ssl package with that dependency [17:02] dobey: so, there is not licensing issue with python-openssl? can I add it to the depends, right? [17:03] nessita: add it to the depends of what? [17:03] dobey: python-ubuntu-sso-client [17:04] nessita: ubuntu-sso-client is GPL right? [17:04] let me confirm [17:04] dobey: yeah gpl v3 [17:04] nessita: i think that's a problem then [17:05] dobey: so the window release we're making also have this issue? [17:05] if it's using openssl, yes [17:06] is today Licensing Problems Monday or something? [17:06] dobey: maybe, why? :-) [17:06] dobey, nessita yes, the windows release is also using openssl because it was stated as a dependency of twisted on windows [17:06] ralsina: ^^ we have another problem :( [17:07] dobey: add to that u1client, which also uses [17:07] dobey, so having an apache license library is a problem? [17:07] from twisted.internet import ssl [17:07] ralsina, +1 [17:07] nessita, where is that in control panel? [17:07] mandel: where did you read coontrol panel? :-) [17:07] I said "u1client" :-) [17:07] nessita: *sigh* :( [17:07] *crap* [17:08] nessita, oh, miss read! [17:08] ralsina, alecu, dobey, mandel (if you can, I know may be late for you): let's have a talk in 1 hours, please, regarding this licensing issue? I need to have lunch === JanC_ is now known as JanC [17:08] nessita: ack [17:08] nessita, I can make in an hour, if is not terribly long :) [17:09] mandel: in the mean time, can you analyze if we can make the ssl dependency optional, both in sso and in u1client? [17:09] nessita, certainly [17:09] well actually [17:09] it's an easy problem to solve for us [17:10] since we own all the copyright [17:10] so not a *huge* problem [17:17] yes, we may have to add an exception [17:18] s/may// [17:18] but since all the base are belong to us, it's easy [17:18] dobey: he [17:18] dobey: indeed. There is even a standard "gpl but let me use openssl" license somewhere [17:26] nessita: I can't make it at 3PM ART I have to pick up the kid [17:27] nessita: what's the meeting about? Getting an exception for it, or something technical? If it's just the exception I cna just contact the lawyers to get approval. [17:28] http://people.gnome.org/~markmc/openssl-and-the-gpl.html [17:28] heh [17:32] dobey: got the ok about soft-coding the font family && size [17:32] dobey: should not need a UIFe since the default is what we have now, IIUC [17:33] ralsina: i don't think the font sizes are all the same as default. the font face is, but sizes not so much [17:34] dobey: ok, so at least we can unhardcode that [17:34] hrmm, i wonder if some of the font sizes are set outside of the qss [17:35] dobey: wouldn't surprise me [17:35] dobey: I will do some experimenting. If I can get a screenshot without visible changes, I'll just do it. Then I need to think how to re-enable it for windows only. [17:35] '
%s!' [17:36] grr [17:36] dobey: grr indeed [17:36] dobey: but hey, if everything was perfect, we would be on foodstamps. [17:36] and there are no fodstamps here! [17:37] i have a branch that removes the font-face and sizes from the qss, already [17:37] dobey: care todo a before/after shot? [17:43] uploading now [17:43] but i am not using the default fonts [17:46] http://ubuntuone.com/0wGgKhfzlY3YeZBD7OGCFJ [17:46] lp:~dobey/ubuntuone-control-panel/system-font is the branch [17:46] ralsina, dobey: aren't we using openssl+twisted in u1-client since forever? [17:46] alecu: shhhhhh [17:47] dobey: ok, so two hardcoded fontsizes at least [17:47] dobey: care to try switching them to something like 200% instead of 16pt? (or whatever) [17:47] ralsina: i left the PERCENT and NAME bits alone [17:47] sure i'll try [17:49] ralsina: it doesn't like % [17:49] dobey: frak [17:50] dobey: right, only pt and px [17:50] dobey: which is a pain in the ass of apocalyptic proportions [17:51] we could precalculate them but then it will react badly on theme changes [17:51] ralsina: doesn't seem to like em either :( [17:51] dobey: can you push thatbranch somewhere so I ca play with it? [17:51] 13:46 < dobey> lp:~dobey/ubuntuone-control-panel/system-font is the branch [17:52] dobey: thanks [17:52] bonus is this makes the "giant translations" bug eaiser to deal with, too :) [17:52] WHOA school run now! [17:52] haha [17:53] nessita, when you return, I've added a small fix for the "second connection attempt" issue that you've found. [17:53] nessita, so, please re-review when you can. === beuno is now known as beuno-lunch === beuno-lunch is now known as beuno [17:57] alecu: ack [17:57] * nessita is back from lunch [17:57] dobey: so, how can we workaround/fix the ssl licensing issue? [17:58] dobey: I read backlog and you perhaps mentioned there is a chance/way? [17:59] nessita, gatox can I have a very easy review: https://code.launchpad.net/~mandel/ubuntuone-control-panel/add-app-name/+merge/97071 [18:00] mandel: sure [18:00] mandel, yes sir [18:01] nessita, can we have the quick mumble now? [18:01] nessita, the one regarding openssl [18:01] mandel: ralsina is gone in school run, so I was waiting for him. Feel free to eod if you have to, really :-) [18:02] mandel: besides I m waiting on what dobey can say about a possible workaround? [18:02] nessita, I can wait a little, I'll give him 10 min, then I'll have to go [18:03] nessita: there isn't much to say. we have to add the exception to our license [18:04] dobey: is that doable? can you please teach me so I update nightlies and then the ubuntu packaging? [18:04] mandel, HEY!! WHERE IS THE TEST?! jeje [18:05] gatox, kahdlashd [18:05] nessita: i think ralsina is wanting to mail legal [18:05] mandel, jejeje just kiding....... or not ¬¬ jeje [18:05] gatox: NF!!! [18:05] that definitely needs tests [18:06] mandel, what happens if i call that without the appname? [18:06] ok..... nf [18:06] nessita, gatox atm it will work, in my next sso push it will brake and all tests will brake too! [18:06] muejejej [18:07] nessita: oh, i guess alecu's branch landing also broke the client builds [18:08] mandel, i think you can test what is being done with that argument that the class is receiving or something..... maybe i'm wrong.... but that's my idea [18:08] nessita: are you also fixing that, or should i? [18:09] gatox, yeah, I can assert that the factory does indeed get the parameter, but it the webclient brake form the sso side it does not make sense.. [18:09] dobey: I can fix both, if you teach me about me the "exception" we need to add in our end [18:09] gatox, maybe asserting that the correct app_name is passed is a better test.. [18:10] mandel: what thing will break? (sorry, not following) [18:11] nessita, atm the app_name in the webclient is a kwarg so that I did not brake the work done already in control panel when it landed, now once the control panel passes it appname will be an arg [18:11] nessita, so, if you don't pass it you won't get a webclient [18:11] gah, i think my current usb keyboard is usb 1.0 [18:11] :( [18:11] nessita, is to ensure that the correct appname is in the sso dialogs [18:12] mandel: ok, but why the fact that the app_name is not mandatory does not allow you to add tests for the controlpanel? (or at least that's what I understood) [18:13] indeed it is [18:13] nessita, no, what I meant to say is that I did not add them because if you do not pass the app_name in the next version of sso everything will brake and I did not considered the tests to be needed [18:13] dobey, you type so fast you need it to be 3.0? [18:14] mandel: any change must have it's test/s :-) [18:14] mandel: it has a built-in usb hub, so the 2 GB of data transferring of this usb stick was going extremely slow [18:14] nessita, ok, I'll add tests then.. [18:14] dobey, oh, makes sense [18:14] mandel: ~16M/s is much faster than ~1M/s :) [18:14] nessita, ok EOD for me, I'll add the tests tom [18:14] it's a usb 2.0 device though, so a usb 3.0 port probably wouldn't help :P [18:15] dobey: so, any advices/links on what should I add besides the new dep on python-openssl? [18:15] nessita: wait for ralsina [18:15] nessita: he said he was going to mail legal [18:15] nessita: so we should wait for that. but should go ahead and make nightlies build i think [18:16] dobey, he, you know things are going to go fast when you mail legal ;) [18:16] heh [18:16] ok, time to go and do all the errands I call life, see you tom! [18:16] mandel: it only took ~4 years to get Novell legal to approve relicising tango-icon-theme :P [18:16] dobey, lol [18:17] dobey: ack then, will push to both nightlies [18:21] dobey, nessita I am back [18:21] dobey, this is a slightly less controversial legal matter. Not to mention that we *do* want to ship it ;-) [18:22] well yes, we basically have no choice in the matter [18:37] dobey: what do you think, needs a UIFe? http://ubuntuone.com/0dHmwZ1BiRlM909ZVtBnSM [18:38] ralsina: heh, the fonts we hard-coded it to are smaller than the defaults in ubuntu? nice. [18:39] dobey: but *bolder* ;-) [18:39] right [18:39] ralsina: the screenshot in the back does not have ubuntu font, no? [18:40] nessita: I am doubting ow if I tweaked fonts here [18:40] I will try in a clean account [18:40] so, given that cp-qt isn't in the default install, and i don't think there are any screenshots of it in the docs, i'd say it doesn't need one [18:41] Right, I have andale mono 11 as font for some reason! [18:41] but that's not a mono font.Weirder and weirder. [18:42] andale mono? [18:42] weird [18:42] i don't think so [18:43] clean account time, I suppose [18:43] i don't think the default font is ubuntu, on ubuntu. [18:44] also, the bold text everywhere in the control panel is an issue with the ubuntu font, it seems [18:44] yes, and mostly with only the ubuntu font [18:44] dobey, ralsina: so, I think that we need to ensure that the u1cp gets the ubuntu font... from my POV is part of the branding [18:45] nessita: on windows, yes. [18:45] dobey, ralsina: regarding the bold, I think is an issue with the qt theme and gtk [18:45] dobey: from my POV, not only on windows. On every OS. [18:45] dobey: but we can ask design to confirm that [18:45] nessita: yes, but it only happens for the ubuntu font because it uses a "advanced" way to specify the weight of the font variants [18:46] 13:32 < ralsina> dobey: got the ok about soft-coding the font family && size [18:46] dobey: I'm pretty sure the approved was on the base that ubuntu is the edfault font on ubuntu :-) [18:46] ralsina: o no? ^ [18:47] nessita: there's a big difference between "default font" and "forcing the font family" [18:48] dobey: maybe, will not argue about that. My point is that I think design expects the font in the u1cp to be UBuntu font, on every OS (no matter if it's hardcoded, default, or foo). But again, I haven't ask so explicitly and perhaps we shou;d [18:48] nessita: I have asked cristian [18:49] nessita: " ralsina - was under the assumption that we were already picking up the theme, so I'm ok with your suggestion" [18:49] ralsina: may I see the question as well? :-) [18:49] nessita: where my suggestion was ": branding question! Can we make the font on the control panel follow the theme on Ubuntu? Currently it's hardcoded to ubuntu font, which makes sense on windows." [18:50] ralsina: of course, i think my branch also breaks that on windows? [18:50] dobey: yes, so we need to tweak that [18:50] dobey: I am happy to be the tweaker if you have other stuff to do [18:50] ralsina: ok then [18:51] I am happy to report that we look much nicer with the default font in a fresh account. [18:51] However, u1cp is slightly taller [18:51] ralsina: my cup overfloweth. but would be nice if i knew *how* the tweaking is supposed to work [18:51] ralsina: how did you make the big fonts work? [18:51] dobey: basically, we need to have a separate windows.qss that only loads on windows (before or after ubuntuone.qss) [18:52] dobey: have not fixed that yet [18:52] ah [18:52] dobey: it just happens to look decent at default font size :-) [18:53] does qt not have a tag or whatever for the text? [18:53] instead of 16pt can we not just say xx-large? [18:53] dobey: in the html? maybe [18:53] hey! [18:53] we can [18:54] * ralsina goes read the "this is the weird html+css2 subset we support in labels" doc again [18:54] although i am seeing a weird issue now [18:54] the top of the "1% used" bit is getting covered up [18:55] dobey: push? [18:56] ralsina: done [18:57] dobey, care for a very simple branch review? https://code.launchpad.net/~ralsina/ubuntuone-control-panel/lang-tested [18:58] dobey: not seing the % being cutoff [18:58] dobey: may be because you are in gnome3 and all that which caused different visuals on friday [18:59] ralsina: eh? that shouldn't matter. unity is mostly a special case of gnome [18:59] dobey: yes, but the qt theme is picking style hits from gtk [18:59] dobey: so, different styles in gtk can cause visual differences [19:01] well it could be a weird bug in qt [19:01] or a result of my font selection [19:01] dobey: can you merge trunk so I get a smaller diff in the qss? [19:01] dobey: yes, you are sort of a corner case on fonts [19:01] eh? did you change something in trunk recently? [19:01] dobey: define recently :-) [19:01] in the last few hours [19:02] dobey: well, a branch or two merged [19:02] dobey: not sure they touched the qss [19:02] oh, my truk was out of date. nevermind [19:02] trunk [19:02] yeah nothing changed in trunk qss :) [19:03] dobey: but I have a large diff on the qss of your branch not just fonts! [19:03] oh... lots of context [19:03] * ralsina hates diff now [19:05] ralsina: can you please propose the same lang-tested branch for ussoc? :-) [19:05] (when you have some slot) [19:05] nessita: happy to! [19:05] nessita: can you file the bug? I am kinda juggling things right now :-) [19:06] ralsina: should be the same bug also affecting ussoc [19:06] I can do the also affects bits [19:07] nessita: then it's ok, I'll do it [19:07] ralsina: already done! [19:07] nessita: cool, thanks! [19:08] now i just wish synergyc would stop crashing [19:09] dobey: synergic used to be real stable, I had to disable it because it's become crap the last 6 months :-( [19:09] Now I can't fix my wife's computer without standing up [19:10] #0 0xb757abf2 in std::_Rb_tree_increment (__x=0x8b0a75c0) at ../../../../src/libstdc++-v3/src/tree.cc:65 [19:11] lovely :-/ [19:11] "let's include our own copy of the c++ stdlib" [19:13] i wonder if i can just install an old version and have it work [19:13] https://bugs.launchpad.net/ubuntu/+source/synergy/+bug/926198 [19:13] Error: ubuntu bug 926198 not found [19:13] le evil [19:13] oh right [19:13] dobey: is the openssl linking a new thing? We have been using https and stuff since ever [19:15] ralsina: it's new in ubuntu-sso-client. we've been ignorant with ubuntuone-client forever [19:15] ok, so it's mostly old news but we just figured it out [19:16] dobey: it's just u1-client and ussoc? No u1cp? [19:17] ralsina: i'd say we should just add the exception to anything we make that is GPL, to be safe. [19:17] ack [19:17] ralsina: cp will need it for proxy stuff i'm sure, anyway [19:18] dobey: right [19:24] dobey: I am happy to approve system-fonts as it is now if you file a bug about forcing the ubuntu font on windows for me to tackle later [19:27] dobey: so, for client packaging dailies I need to edit the control.in, right? [19:28] ralsina: https://bugs.launchpad.net/ubuntuone-control-panel/+bug/953318 [19:28] Ubuntu bug 953318 in Ubuntu One Control Panel "Windows styling affects Ubuntu themes as well" [Undecided,Triaged] [19:28] nessita: yes [19:28] dobey: acxk [19:29] first missing build dep is python-qt4reactor, so will add that as a first step [19:30] nessita: it's the only thing that needs added afaik [19:30] nessita: we also need to add the new binary package for the proxy tunnel stuff, though [19:30] dobey: yes, but the tests will be run for the proxy module, no? [19:30] yes [19:31] (if the proper deps are there, squid y apache2-utils) [19:31] so adding the build-dep should fix that [19:31] so I guess the proxy tests will fail as well [19:31] I didn't add the python-openssl yet, I'm waiting to see the actual error [19:31] (sanity check) [19:31] or just get skipped [19:31] let's see [19:31] python-openssl is probably arleady in there [19:32] dobey: in u1client/packaging-dailies? naha [19:32] nessita@dali:~/canonical/client/packaging-dailies$ grep openssl * [19:32] nessita@dali:~/canonical/client/packaging-dailies$ [19:32] rly? that's odd [19:32] dobey: why would it be there already? [19:32] nessita: well it's been using it for years [19:32] and the tests haven't been failing [19:32] dobey: perhaps you're confused? [19:33] eh? [19:33] why would i be confused? [19:33] dobey: u1client was not using ssl directly so far, was just telling the reactor to connectSSL (but was never importing twisted.foo.ssl) [19:34] dobey: what triggers the IMportError on openssl is importing twisted.foo.ssl [19:34] dobey: OTOH, protocol does list python-openssl as dep :-) [19:34] ah [19:34] protocol imports it [19:34] yeah [19:36] nessita: but it's been importing that module from protocol, which imports ssl [19:36] nessita: but i think protocol already depends: python-openssl, which is why it worked [19:36] dobey: that's what I said :-) [19:36] dobey: OTOH, protocol does list python-openssl as dep :-) [19:37] (at (04:34:39 PM) ) [19:37] right [19:37] i didn't see that [19:37] ah [19:37] as i was on another workspace, looking at the source :) [19:37] ja [19:38] ralsina: so i should propose my system-font branch as-is? [19:39] dobey: hmmm maybe leave the ubuntu font family in [19:39] nessita: btw, can you upload patches to ubuntu for the broken 'link' things in sso and control-panel? [19:39] dobey: then I will remove it forUbuntu on the fix for the other bug [19:39] and we go flexible size on both platforms [19:39] dobey: yes, sure. But, what link is broken on controlpanel? [19:39] ralsina: hrmm, ok [19:39] nessita: 'sign in' ? [19:39] I know about the broken link for "password forgotten" only [19:40] dobey: where? [19:40] an no [19:40] it's in sso also [19:40] bug #951584 [19:40] sorry [19:40] Launchpad bug 951584 in ubuntuone-control-panel (Ubuntu) "Cannot sign in to Ubuntu One - "Sign in" button disabled" [Undecided,Confirmed] https://launchpad.net/bugs/951584 [19:40] dobey: that's another bug, more complex and un-debugged so far :-/ [19:41] nessita: are we setting the button insensitive until the form is filled? [19:41] dobey: yes, design spec [19:41] nessita: simple solution then: "don't do that" :) [19:41] dobey: guess my answer :-) [19:41] nessita: i'd rather have people using our software, than trying to be perfect with design doc [19:42] design is useless if people can't use it [19:42] we can fix it to be right later [19:42] for now, please make it work :) [19:42] dobey: that would just make having reports "clicking on sign in makes the app crash". Apparently there is a race condition when the backend is not ready in time [19:43] dobey: is not that simple, if you please read me, you may help come up with a solution that actually works (and not just moves the error somewhere else) [19:43] well i didn't know the code was broken in that way [19:43] dobey: the button is not enabled because something gets stucked in the backend [19:43] all i see is "some people can't log in, let's let them log in" [19:43] a friend of mine have the issue, but we have not debugged it yet [19:44] dobey: but, the button is working for a lot of people, so is not that simple as "is inactive" :-) [19:44] well, we could just use the gtk panel instead :) [19:44] dobey: those comments stopped being funny long time ago ;-) [19:45] nessita: could you at laest comment on the bug then, and provide some reassurance or something? [19:45] dobey: is on my queue [19:45] it's not supposed to be funny [19:46] dobey: then if you're serious... I would have to kindly ask you to please align to the team/code direction we're going, which is not GTK+ (despite our personal preferences) [19:46] so no, going back to GTK+ is not an option [19:50] nessita: I have a branch for the test fails in ussoc when LANG is set, but there seems to be a "real" bug in the spawner when LANG is set. Not really sure though. [19:50] ralsina: very likely... any chance you debug further? [19:50] (today or tomorrow) [19:50] nessita: sure. Here is the trace if it reminds you of something https://pastebin.canonical.com/62164/ [19:50] looking [19:51] ralsina: aaaahhhh maybe the thing we\ re using to detect failed to start is also lang'd [19:51] nessita: could be [19:51] ralsina: can you debug on runner/tx.py where the error is handled? [19:52] Oh, tx.py I was looking at qt.py! [19:52] you will see that 'No such file or directory" is used [19:52] Right [19:52] Let's remove the localised part [19:55] comparing results with string literals in tests, is always frightening :( [19:57] ralsina: https://code.launchpad.net/~dobey/ubuntuone-control-panel/system-font/+merge/97094 [19:57] dobey: yep [20:00] dobey: is not test, is live code! :-D [20:00] dobey: live code not used in production, though [20:00] even worse [20:00] and you wrote it!? bad teacher! [20:00] dobey, ralsina: the time I checked, twisted didn't let me much choice. [20:00] dobey: :-D [20:01] you should see how beautiful it is a piece of C code I wrote this weekend [20:01] nessita: there is a similar bug in the glib launcher, checking... [20:01] with asserts everywhere, ensuring prea dn post [20:01] nessita: want to fix memory errors in ubuntu? :) [20:02] nessita: check my last C code ;-) http://code.google.com/p/raspf/source/browse/trunk/raspf.c [20:02] ralsina: didn't you say you were reviewing https://code.launchpad.net/~dobey/ubuntuone-client/transient-notify/+merge/96815 this morning? [20:02] ralsina: I would very much appreciate the debugging of that. The main problem, when I tried to solve that, is that the errors comes directly from the "child" stdout/stderr, son string parsing is needed (as far as I recall) [20:02] dobey: I think I did! [20:02] ralsina: there are no votes still :) [20:02] dobey: no thanks, that's why I code python in my mainly job [20:02] dobey: slipped right by me! ;-) [20:03] ralsina: YOUR INCLUDES ARE NOT ALPHABETICALLY ORDERED :-D [20:03] nessita: but tell me that's not the most pythonic C *ever* ;-) [20:03] +1 to use bstring [20:04] the first thing every new C programmer should learn is "never ever ever use null-terminated strings if you canpossibly avoid it" [20:05] nessita: there are even tests! http://code.google.com/p/raspf/source/browse/trunk/testspf3.c [20:05] nessita: that library is a literal port of a python one of course :-) [20:07] never ever use strings which are not null terminated [20:07] dobey: +1 I found the terminal with the tests running in it ;-) [20:07] heh [20:08] dobey: yes, use strings that know their own length ;-) [20:09] dobey: you my C guy! Is errno 2 guaranteed to be ENOTFOUND and viceversa? [20:12] no? [20:12] ralsina: 'man errno' doesn't even list ENOTFOUND [20:13] u1client dailies are building! https://code.launchpad.net/~ubuntuone-hackers/+recipe/client-dailies [20:14] dobey: interesting... [20:14] dobey: I meant ENOENT [20:15] dobey: can you please also affect ussoc in https://bugs.launchpad.net/ubuntuone-control-panel/+bug/953062? and also propose a branch if you have the tim [20:15] Ubuntu bug 953062 in ubuntuone-control-panel (Ubuntu) "Font and sizes are hard-coded" [Undecided,New] [20:15] dobey: so we keep cnsistency between bopth [20:16] ralsina: ah right. i don't think it's guaranteed to be 2, but it is guaranteed to be ENOENT :) [20:16] dobey: grmbl [20:16] ralsina: particularly, i don't know what windows does for it [20:16] dobey: you also may wanna merge trunk in [20:16] nessita: did you change the qss? [20:17] well, what we have now really doesn't work in other languages anyway. [20:17] dobey: nopes, but in the LP diff there are some diffs against the tab tab tab branch [20:17] wtf [20:17] bad launchpad [20:18] eod for me! [20:18] see you tomorrow people [20:19] bye gatox [20:20] dobey: oh tabtabtab did not have a commit message [20:20] ... [20:20] set one now [20:21] nessita: sorry! [20:21] ralsina: but then why LP was showing that in the diff for dobey? [20:21] nessita: no idea, I had that too [20:21] what the hell [20:22] did I break tarmac? [20:22] no [20:22] ralsina: did the test-lang-fixed have the tabtabtab merged in? [20:22] lp is weird [20:23] nessita: ohhhhh hope not! [20:23] if it did, that was by mistake :-/ [20:23] it doesn't matter if it's in trunk or not [20:23] kindertime for me [20:23] see you laters, all. [20:23] bye alecu! [20:24] nessita: let me try something else [20:24] dobey: where? when? who? [20:24] nessita: regardless of what landed in trunk, lp shouldn't be showing that stuff in my branch [20:32] nessita, ralsina: ah, it seems like i messed up when i made the branch [20:32] dobey: ok [20:33] IT WAS NOT MY FAULT??? AMAZING! ;-) [20:33] fixing it [20:33] this week just keeps getting better! [20:34] i'm sure i could find some huge problem for you to have to deal with :) [20:34] dobey: killjoy! [20:44] nessita, ralsina: there i fixed it [20:45] dobey: ack! same MP? [20:46] nessita: yep [20:53] nessita, dobey: another fix-tests-with-lang-set branch, now for ussoc https://code.launchpad.net/~ralsina/ubuntu-sso-client/lang-tested/+merge/97097 includes fixes for internationalized spawners [20:57] ralsina: nice! [20:58] nessita: I am not terribly happy about the spawner fixes but they are what we can do, I think [21:02] we can probably do better [21:02] but i'm not fussed about it at the moment [21:11] dobey: if you suggest fixes on MP I can do them late tonight maybe [21:12] ralsina: i already gave it +1 [21:13] dobey: ack [21:14] i think we might have a problem with the proxy tunnel vs. web client in sso thing [21:16] because, circular deps are bad, mmkay [21:21] well, is past time for me to get off here [21:21] later all [21:21] bye dobey [21:22] EOD for me as well [21:29] * alecu is back [21:30] we will not have circular dependency problems with the webclient and the tunnel, because the webclient will be in sso, but will not use the tunnel_client. [21:31] txweb.Webclient instances will take a "reactor" optional parameter that will be set to "tunnel_client" only when being used by SD. [21:32] so, dobey: don't worry about that :-) [21:51] alecu: anda! [21:51] nessita, what anda? [21:52] alecu: your branch [21:52] náis! [21:52] alecu: have you tried testing it without having squid installed?> [21:52] (I do have squid) [21:52] nessita, it does not matter if squid is installed, right? it only matters if gsettings point to it. [21:53] nessita, I've tested with gsettings pointing at the squid, and with gsettings set to "none" [21:55] ack [21:55] alecu: ok, approving [21:55] good work! [21:55] and I need to run vbery fast to pilates now [21:55] bye! [21:55] byeeeeee === yofel_ is now known as yofel