[07:50] morning all!! [07:52] morning mandel [07:54] fagan, morning ;) [07:54] fagan, have you heard the news? [07:54] mandel: steve dying yeah? [07:54] sad :-/ [07:55] really sad.. [08:41] Brrrrr! Good moaning! [08:47] Morning! [08:47] I'm testing some code which will use openid (specifically Ubuntu SSO) for auth. [08:47] is there a staging server for SSO I can use? [08:48] * popey discovers https://staging.one.ubuntu.com/ [08:51] popey, hm.. good queestion, what are you actually doing? [08:51] I want to make a simple website which uses launchpad/ubuntu SSO as the auth provider [08:52] popey, when I worked porting sso to windows I used the actual server using the + sign in a gmail account.. [08:52] in the same way that (for example) askubuntu.com does [08:52] popey, ok, seems reasonable, if there is no u1 involved you can ignore our code :) [08:52] right, okay, if that's acceptable [08:53] popey, yes, it is more thant acceptable, the service is there for you to use :) [08:53] popey what lang are you using, we migh have some example there :) [08:53] popey, and let me a sec and I'll as the people that manage sso to see if there are any issues using staging :) [08:53] php [08:53] shit, I use lots of smilies, I must be happy today hehe [08:54] ☺ [08:55] popey, just asked but we might need to wait a little, I think most people from SSO are in the other side of the athlantic, I'll ping you as soon as I get an asnwer [08:56] ok [08:56] seems most code I'd find would be python based [08:56] figured someone must have done an openid client library in php [08:57] popey, there should be one, I did some work with js, does that help you? [08:57] popey, we can also ping aqurius whenever he is up, he might now about some crazy lib [08:58] not sure, I'm not the actual coder ☺ [08:58] will go and have a play and come back later :D [09:00] popey, sure, I'll try to get you the info asap [09:01] thanks! I appreciate the help [09:03] popey, I'm here for that :) [09:03] heh [09:22] good morning. Mandel ping [09:30] ralsina, buenos dias! [09:30] buen dia mandel! [09:30] Sorry I am late with that review, ding it now [09:30] doing it [09:30] ralsina, I was about to have a coffee, so you have a couple of mins hehe [09:30] ralsina, and I moves to the migratin data bug which is nearly done [09:31] * mandel coffee break [09:58] mandel: I get test failures, but couldn't capture the log, so am running again [10:01] ralsina, weird, please do pate them somewhere [10:02] ralsina, there are 4 tests that fail in trunk already [10:02] mandel: could be those [10:05] ralsina, yeah, nevertheless please paste and I'll take a look [10:05] ralsina, better land that branch late than sorry :) [10:05] mandel: 2 fails: http://pastebin.ubuntu.com/703295/ [10:05] and 13 errors [10:06] let me paste more :-/ [10:07] ralsina, exceptions.IOError: [Errno 2] No such file or directory: 'C:\\Users\\ROBERTO\\canonical\\is_path\\data\\logging.conf' [10:07] mandel: yes... [10:07] ralsina, you have to do ./autogen.sh & make to get that file :) [10:07] mandel: windows [10:07] if it has to be there for tests, then run-tests should put it there [10:08] ralsina, because using autotool for a python project is a good idea :P [10:08] ralsina, it has to be done for the tests, so far we have being doing make froma shared folder in a unix system that has the tools [10:08] ralsina, which I agree is less that desirable [10:08] mandel: I haven't :-) [10:08] ralsina, hahaha [10:09] ralsina, the ones that touch the code did hehe [10:09] mandel: there is a logging.cnf in windows folder [10:09] mandel: make run-tests.bat move it if needed [10:09] ralsina, how is the progress of removing the c code form ubuntu one client source? [10:09] ralsina, sure, I can add that there [10:09] let me check if it is correct first [10:09] mandel: stalled, I think [10:10] mandel: it's what we are shipping so it is correct ;-) [10:10] same for clientdefs if needed [10:11] ralsina, clientdefs are copied, of that I'm 100% sure [10:11] cool, so same thing :-) [10:11] ralsina, give me 2 mins and I'll get that done for you [10:11] awesome [10:13] ralsina, so about the c code, are we moving it out, 'cause doing all this autogen crap is a pain [10:13] ralsina, I got it working for mac os x, but installing gnome-common for that on a mac just feels stupid [10:13] mandel: dobey was moving it, I think, before he became the linux team [10:13] but yes, using autogen to generate logging.conf is overkill [10:14] specially since we always generate the exact same one anyway [10:15] ralsina, indeed, and having all this in a shell script or the setup.py would be great [10:15] I'd love to do python setup.py test or something like that hehe [10:16] ralsina, I'm testing the changes to run_tests.bat atm [10:16] will push as soon as I see that everything is ok [10:16] ack [10:18] ralsina, in the main time, can you take a look at https://code.launchpad.net/~mandel/ubuntuone-windows-installer/uninstall-old-app [10:18] mandel: sure [10:18] ralsina, that code will uninstall the old msi for the beta in case someone has it, will propose the code that does the data migration in 5/10 min [10:18] mandel: ack [10:18] after that I just have one last branch performing the 3 steps adn will do some IRL test with the old-beta and new one [10:20] mandel: that branch cdefines the code but is not called from aywhere? [10:22] ralsina, yes, I have done it in small steps, find old app/remove msi/migrate data (next branch)/ put things together (last one) [10:22] ok [10:22] ralsina, I wanted to make smaller merge proposals with clear tests so that the reader does not have a 1000 lines to look at :) [10:22] and we all appreciate it ;-) [10:25] ralsina, see, I listen to my users hehe [10:27] ralsina, tests finished and runtest.bat works as expected, please pull (my finger) when ever you want [10:46] mpt: ping [11:04] czajkowski, I think he is here now [11:08] mandel: where are you ? [11:12] czajkowski, here :) [11:12] but if you mean in the world, would be Madrid :) [11:12] ralsina, other branch: https://code.launchpad.net/~mandel/ubuntuone-windows-installer/migrate-data/+merge/78387 [11:13] ralsina, that one contains just the migration of the data from the old beta to the new one [11:13] mandel: ack. Gotta take the kid to school thouhg [11:13] well, new client not new beta :) [11:13] ralsina, no probelm [11:13] hi everyone [11:14] gatox, hola! [11:14] gatox, got internet? [11:15] mandel, 3g right now... but in a few few hours i should have a proper connection and i'm going to be REALLY HAPPY! :D [11:16] gatox, congrats!! [11:16] mandel, thanks...... this situation was driving me crazy!!! :P [11:16] gatox, I was like a month like that, I understand the pain [11:17] mandel, true [11:17] Morning all - just signed up for ubuntu one music. nice - now to hack an upload tool to run on OSX )() [11:18] talios, are you looking to port the daemon that syncs the files, or just an uploader, cause if you want to hack I know the people that can help you for that :) [11:18] talios, if it is the daemon, I'm also interesting in doing some work there, but I don't want to have to do the UI hehe [11:19] mandel: ahhh thought you meant london [11:19] czajkowski, oh, no hehe I meant here as in the channel right now [11:19] mandel: a bit of both :) I actually started working on a personal cloud music tool the other day and am thinking of exploiting Ubuntu One for it. [11:21] talios, oh my!!! that sounds great! [11:21] talios, so, the one that can help you with the rest api would be urbanape, he did the iOS client [11:21] talios, so you might well be able to take advantage of his work [11:21] talios, regarding the daemon, I can give you a hand, but only on fridays :( [11:22] mandel: call me crazy ( every one does ) but I have a map reduce style config for all my iTunes playlists ( heavy smart playlist setup, along with scripts which hack the iTunes meta data ) - I'm trying to write something to ween myself of iTunes :) [11:24] talios, I dont see anything crazy about that, I see coolness, do you mind if I get some time (walk the dog) and chat about it in a couple of hours? [11:24] talios, we can get urbanape too and see if there is something nice to hack :) [11:24] mandel: I'd love to but its 12:24 am and I should be heading to bed :) [11:25] I'm sure I'll be loitering in here a bit as I explore the APIs [11:25] talios, superb, please do so, I'm in CET time so you can ping me earlier than this :) [11:25] talios, cath you later! [11:25] * mandel walks dog [11:26] now if the android player gave me ID3 from AAC files I'd be immediately happy too :) [11:39] good morning everyone! [11:54] in less than 2 hours i'm going to have internet!!!! please be patient for 2 hours more :P === yofel_ is now known as yofel [12:17] good morning gatox [12:17] ralsina, good morning.... i'm here about an hour ago...... but with internet problems...... in less than 2 hours i should have a real connection :D [12:17] good :-) [12:20] * nessita -> quick errands [12:29] * mandel lunch [12:42] * nessita is back [12:46] hola nessita [12:46] hi ralsina [12:47] ralsina, can you see me? [12:47] mandel: now I can [12:47] mandel: is run-tests.bat updated in that branch? [12:48] ralsina, yes, i guess you did not get my msg [12:48] it seems I didn; t [12:48] ok, trying [12:49] ralsina, not a. big deal, I also sent you an other message with another branch for re view, did you see that one? [12:49] the one about data migration? Yes, got that one [12:50] ralsina, weird cause they were more or less at the same time.. well who cares :) [12:51] mandel, gatox, alecu, nessita, dobey: standup in 9' [12:51] mandel: maybe I just didn't read it. I am very sleepy for some reason [12:51] ralsina, or an xchat fail.. :P [12:52] brb....... internet dude is here :P [12:52] that is what she said!! [12:52] jeje [12:52] gatox, would have been funnier with a pizza delivery guy... or a plummer :P [12:52] jejejjeje [12:53] ralsina: lies [12:53] dobey: I am sleepy! [12:55] ralsina: no standup [12:56] ralsina: mumble instead, no? [12:56] mumble is in 1 hou [12:56] hour [12:56] ralsina, no standup? [12:56] oh [12:56] so, yes standup, but do it desultorily [12:56] with a slouching posture, maybe [12:56] say meh instead of me [12:57] nay [12:57] sig [12:57] meh [12:58] mandel: still no joy: http://pastebin.ubuntu.com/703365/ [12:59] gatox: you reading me in PM? [12:59] ralsina, I'll take a closer look aft the stand up, can you put that in the mp? [12:59] me [12:59] mandel: sure [13:00] me [13:00] gatox: ping [13:00] me [13:01] We [13:02] nessita: gatox is with the internet guy [13:02] So, mandel, go [13:02] DONE: Prosed merges for lp:~mandel/ubuntuone-windows-installer/migrate-data for the code that migrates the data from the old beta. Fixed run-tests for lp:~mandel/ubuntuone-client/is-path-dir-fix so that there is no need to do make on windows. That branch fixes bug 867567 TODO: autoupdate bugs. BLOCK: nope [13:02] Launchpad bug 867567 in ubuntuone-client "The _path_is_dir function in file system notifications on windows is broken (affects: 1) (heat: 6)" [High,New] https://launchpad.net/bugs/867567 [13:02] nessita go [13:02] DONE: administrivia, arranging things for mgmt sprint trip, reading on u1db, some bug triage, some reviewing. TODO: close my damn bug I have been poking for 3 days, bunch'o'reviews, call with you guys BLOCKED: no NEXT: nessita [13:02] DONE: finished and landed bug #862991, bug #862540, bug triaging, lots of debugging for Lisette's bug about empty folders not being deleted in the web. Found bug #868661, started a branch for it [13:02] TODO: finish bug #868661, lots of reviews I didn't do yesterday, special meeting [13:02] BLOCKED: nopes [13:02] NEXT: all the dobeys [13:02] Launchpad bug 862991 in ubuntuone-control-panel "Set a better title for the error dialogs (affects: 1) (heat: 94)" [Low,Fix committed] https://launchpad.net/bugs/862991 [13:02] λ DONE: e-mail, more magic, tarmac config stuff [13:02] λ TODO: tarmac config updates, more of #865593, call, SRU work [13:02] Launchpad bug 862540 in ubuntuone-control-panel "Do provide a nice error message when there are issues (affects: 1) (heat: 6)" [Medium,Fix committed] https://launchpad.net/bugs/862540 [13:02] λ BLCK: None. [13:02] Launchpad bug 868661 in ubuntuone-client "u1sdtool is not multiplatform (affects: 1) (heat: 6)" [High,In progress] https://launchpad.net/bugs/868661 [13:03] cest fin? [13:03] I guess so [13:03] oui oui [13:03] good morning! [13:03] dobey: would you be able to review https://code.launchpad.net/~nataliabidart/ubuntuone-client/update-sso/+merge/77370 today? [13:03] hi alecu [13:03] alecu: you! say me! [13:03] me [13:03] alecu, go [13:05] alecu.standup.addCallback(print("EOM")) [13:06] ralsina, twisted does not work on python 3 yet! [13:06] alecu: from __future__ import print [13:06] ok, I'm off for lunch, wii be back in 30 min more or less [13:06] ;-) [13:07] buen provecho! [13:07] ok, from __future__ import print_function, I was close [13:08] talios: I'd love to see what you come up with. I was working on a simple upload/publish tool, but work keeps getting in the way. [13:08] DONE: took morning off, then worked a bit more on server-time sync, on the sd and some fixes on cp. [13:08] TODO: finish fixing the tests, propose for reviews [13:08] BLOCKED: find out what if we want to use a specific url for this [13:08] in fact, let's ping karni [13:08] EOM [13:09] alecu: what specific url you need? [13:09] alecu: Yes? [13:10] Ah [13:10] alecu: Let me find it [13:11] nessita, right now I'm using http://one.ubuntu.com/ but perhaps we want to use *some* other url. [13:12] nessita, that's why I'm asking karni, that has already been thru this. [13:12] karni, also: what are the non-caching headers that you are using? [13:12] alecu: I do HEAD https://files.one.ubuntu.com/?status=1 with the header Cache-Control: no-cache [13:12] alecu: and parse the Date header :) [13:13] alecu: sorry I'm a little out of context, I think you discuss this while I was on vacations. You need to hit a url to achieve what exactly? [13:13] karni, right. [13:13] nessita: Find the server time to avoid unauthorized response to OAuth signed calls that used invalid timestamp (invalid, as in: to far from server time) [13:14] karni: and then how do you "overwrite" the timestamp? [13:14] nessita, I'm getting the server time, so I can tell it the local time is skewed. And then I use that "skew" to adjust the oauth_timestamp parameter. [13:14] nessita: a bit of review done [13:14] nessita: I extended CommonsHttpOAuthConsumer class and override the generateTimestamp method with return Long.toString((System.currentTimeMillis() + timeDriftMillis) / 1000L); [13:15] alecu: ah, tricky :-). Do we need to "fix" to many places or (ideally) just ussoc? [13:15] alecu: Exactly. [13:15] karni, just "Cache-Control: no-cache" is enough? I remember that cache headers were a bit more complicated when the browsers had to go thru different transparent proxies and such. [13:15] alecu: Cache-control should override everything IIRC [13:15] nessita, I've got five branches [13:16] alecu: To my best efforts I have confirmed this solved the issue. I still have a bunch of users with the same problem, but it's most probably due to timeouts on /sso-done-so-get-tokens [13:16] nessita, storage protocol, sso, control panel, syncdaemon, windows installer. [13:16] alecu: So I would say, yes. [13:16] karni, ralsina: thanks. [13:17] alecu: guh, when you have a moment, would you share why? I mean, I would like if all the oauth stuff was isolated in ussoc, so perhaps we can move the logic back to ussoc and re-use from the other locations? [13:17] alecu: You're welcome. [13:18] nessita, those are all the places that are doing oauth signing. (as opposed to getting the tokens used to sign) [13:18] dobey: ack, thanks [13:18] alecu: in ussoc there is a method that returns signed headers for any http request, can you use that in the other places that sign the urls? [13:19] mandel: just for you, We will make everything use imake to build/install [13:19] xmkmf -f was it? [13:19] nessita: alecu: as to what nessita just wrote - please keep in mind that SSO time may be different than U1 time. I'm fixing the issue in two places. [13:20] nessita: alecu: With different server URLs [13:20] karni: which 2 places? [13:20] nessita, I'm taking a look at using the same bit from sso. [13:20] nessita: When I talk to SSO, I get the Date header from https://login.ubuntu.com . When I talk to U1, I get the Date hear from https://files.one.ubuntu.com/?status=1 [13:21] alecu: maybe protocol will have to duplicate code (it does not depend on SSO, and I'm not sure if we want it to depend on that) [13:21] karni: are you doing that because you know for a fact that those 2 sites differ in their time? [13:21] ralsina: xmkmf -ftw [13:21] nessita: We had a chat with pindonga and beuno that they need not to check the timestamp (only use it to track the order of requests, as a monotonically increasing integer) [13:23] nessita: I know for fact I had problems with both. Does that answer your question? [13:23] nessita: It was the same "UNAUTHORISED" problem. [13:24] if we have two servers with time more that 15' off, right now the desktop client would fail for one of those all the time [13:24] nessita: And fixing the device time drit/skew did solve the issue for both. [13:24] karni: yes, that answers my question, thanks [13:24] nessita: You are most welcome :) I am happy to help. [13:24] :-) [13:26] really, all We want to do today, is get results from this call. [13:26] ralsina, yes: that would be really weird: we would have that problem on many desktops too if the servers were more than 15' off [13:26] karni, ^ [13:26] alecu: nessita: I also know that SSO has a cluster of servers that are time-synced (with one NTP I believe). We had huge differences in time, because it turned out we were receiving cached responses. So, the Cache-Control: no-cache header solved that for us. In case of U1, I know this is the URL we need to hit. [13:27] karni, great. [13:27] it makes sense now, that some servers are behind a caching proxy, and some are not. [13:28] and also probably the ?status=1 invalidates some proxies [13:28] alecu: ralsina: I had many users with SSO issues, some claimed they do not want automatic (synced) time set on their Androids, some said they just changed to daylight saving, etc etc. That really sucks. [13:28] alecu: Most probably yes :) I got that url from beun'o \o/ [13:28] karni: yes, we have the same problem on windows, thus the need for this fix ;-) [13:28] karni, yeah, time on cellphones can be set manually very often. And on windows is the same. [13:29] ralsina, exactly. [13:30] * JamesTait pokes talios [13:31] nessita, I'm taking a look at using oauth_headers from ubuntu_sso/utils/__init__.py. It's something that can be done, but I won't be doing it in my current branches. The main issue is that different parts of the code are using a different web client. [13:31] alecu: ok, I see. Thanks for checking though! [13:32] alecu: that's the kind of cleanup we should set time apart to do [13:32] alecu: as in "stop using different web clients" :-) [13:33] ralsina: beuno: Seeing how much resources we've already put into this and will now be fixing for Windows, I'm thinking if they shouldn't drop the timestamp verification (as the specs say they CAN) for request invalidation purposes, and use it only for tracking the order of requests. [13:34] ralsina, yup: and we were discussing with mandel a way to replace them all with just one, so proxy support would be easier. I mean feasible. [13:34] alecu, ralsina: when implementing proxy support, unifying that will be a must [13:34] nessita, exactly. [13:34] ralsina: beuno: Oh.. I just said "they" thinking of SSO, but we have the very same issue with OAuth on U1. [13:34] 'rewrite everything in vala so we can use it from EVERY language without having to write the same code N times' [13:34] mandel, gatox: I made all the pending review from you, let me know if I forgot something [13:34] karni: they also use it to discard nonces, I think. jdo gave an explanation of why it couldn't be discarded on that email thread. [13:34] dobey, but not on windows, right? [13:35] alecu: also on windows. [13:35] dobey: fixing update-sso now [13:35] ralsina: Interesting. Fine, can use to discard nonces, but drop the time lag check. [13:35] dobey: we are not using it from multiple languages. We are just using multiple python libraries to do the same thing on different places [13:35] dobey, we'll probably just use qtnet + python. It's a much smaller change. [13:35] karni: yes, yahoo and google don't check it so it can be done, but he... [13:35] ralsina: i'm pretty sure we are doing authentication from C/C# apps [13:36] ralsina: we should also set time aside so we all migrate to O (I know I didn't do it yet :-/) [13:36] nessita: indeed [13:36] nessita: these are some of the things we need to talk about in ... 24 minutes :-) [13:36] crap. [13:36] oh, meeting. [13:36] * ralsina is terrified of O. Last attempt went bad really quick. [13:36] * alecu grabs a bite [13:36] nessita: migrate to O ? [13:37] is that like a 15 times better C? [13:37] karni: all of us are supposed to be on Oneiric since beta at least [13:37] karni: our desktops, upgrade to oneiric [13:37] * alecu will wait for C++ [13:37] * alecu will wait for O++ [13:37] ralsina: nessita: oh [13:37] alecu: you are thinking of C++++++++++++++ [13:37] * dobey throws a O(n) at alecu [13:38] * JamesTait tried Oneiric from a Live SD Card on an Asus EeePC 701 last night - very nice, I have to say. [13:38] JamesTait: yes it's pretty! Then I upgraded and it stopped booting. [13:38] ouch [13:38] ralsina: :( [13:39] ralsina: does it sit "Waiting for network blah blah" ? [13:39] dobey: no, it kernel-panicked on boot [13:40] dobey: I will try again pinning drivers+kernel first, I suppose [13:40] oh [13:41] ralsina: hold shift immediately after the BIOS goes away, and boot an older kernel [13:42] ralsina: And I can only imagine how much more work this will require for you guys to fix that OAuth issue. Can't we really fix it on the server side.. ? Then you would only have to fix the sso part [13:42] dobey: killed the sucker already, will try again on saturday [13:43] karni: that's why I did that email thread. Since server side told us no... [13:43] ralsina: I'll re-read it in a spare moment. [13:43] dobey: question, you mentioned that g_cclosure_marshal_VOID__OBJECT should be VOID__VOID for the CredentialsNotFound signal. I'm not sure if you're suggesting that I should change the marshaller declaration, or if I should add a new marshaller. I would guess you meant the second one (since we need the VOID__OBJECT for the credentials found and credentials error signal), and I did not add it since I read there are a couple of "builting" marshallers alrea [13:44] dobey: so, since previously we had no marshaller for STRING_VOID (which was the signature of the auth_denied signal), I would assume there is a builting marshaller for VOID__VOID [13:44] nessita: We mean that signal should use a _VOID__VOID marshaller as it is not expecting any return values, nor is it passing any arguments to the callback [13:45] dobey: right, but since I removed the signal, I don't have to change that, no? [13:46] nessita: right, if you removed the new signal addition, then that line should no longer be in the diff [13:46] dobey: ok, I now understand what you meant, thanks [13:49] YES! HABEMUS INTERNET! [13:49] gatox, WELCOMUS! [13:49] alecu, thankus! [13:49] jeje [13:49] gatox: WHAT A JOY! [13:50] nessita, you have no idea :') [13:50] I DON'T KNOW WHY WE'RE YELLING! [13:50] diez minutos [13:50] nessita, ugh, manually marshalling. my condolences. [13:50] CardinalFang, jeje [13:50] alecu: we needed a firewoman in the team (?) [13:51] * alecu thinks of firewomans and of poles and of series of really bad jokes [13:54] dobey: fixes pushed to revno 1149 [13:55] nessita: cheers [13:59] oh meeting! [14:00] ugh, 33ms ping [14:00] alecu, nessita, dobey, gatox, thisfred mumble? [14:00] and slow upload [14:00] mandel too [14:00] ralsina, ack [14:01] ralsina: waiting for you [14:01] on my way [14:01] shit. laptop's upstairs, be there shortly [14:03] ralsina, what? [14:04] mandel: mumble! [14:04] mandel: team mumble [14:04] yes, launching it [14:04] mandel: there was mail, and invitation, and it's in your calendar :-) [14:04] ralsina, I know, bad phone call at the same time === m_conley_away is now known as m_conley === zyga_ is now known as zyga-afk [15:07] nessita, ping [15:07] mandel: pong [15:07] regarding https://code.launchpad.net/~mandel/ubuntuone-windows-installer/migrate-data/+merge/78387 I had a lint error when importing from platform, shall I just add an ignore comment? [15:08] nessita, ^ [15:08] mandel: what lint error? [15:09] nessita, let me un do the change and get that for you one sec [15:10] mandel: sure [15:11] lunch time, bbiab [15:13] nessita, the following is when executing runtest in O for that branch: http://paste.ubuntu.com/703438/ [15:13] lunch...... brb [15:13] nessita, once I have fixed the import, that is [15:16] mandel: if you open a python terminal, can you do the import? [15:18] nessita, yes, is just lint being an ass on O [15:18] nessita, I worry that we will start getting false possitives.. [15:19] nessita, http://paste.ubuntu.com/703442/ [15:19] including typo hehe [15:20] mandel: hum, can you try with the ignore? [15:20] mandel: bad news is that I think the pyflakes do not listen to the ignores :-/ [15:20] nessita, this is the installer, so we are ok with that, right? [15:20] nessita, I'll add the ignores [15:20] mandel: eh! wait [15:20] mandel: this is installer? [15:21] mandel: we certainly should not be using os_help platform stuff outside syncdaemon [15:21] mandel: SyncDaemonTool, yes. os_helper, no. [15:22] mandel: since this is windows code, you should always use unicode paths, and use the builting functions (eg shutil.move) [15:22] nessita, and add the long path prefix too [15:22] mandel: well, it depends, the old beta handled long paths? (I would guess no)( [15:22] nessita, yes, it did [15:23] mandel: but... question [15:23] nessita, also, tools comes from platform, why can't we import the rest? [15:23] mandel: we won't leaking os_helper outside syncdaemon, otherwise it gets difficult for us to fix/move stuff there [15:24] mandel: and outside syncdaemon, the app knows in which paltform it is, so multiplatform FS handling is not needed [15:24] mandel: question about log path prefixes: [15:25] mandel: hum, no, nothing, just use the builting function passing unicode to them and we'll be fine [15:26] nessita, the default funtions you mean, sure I can change that, the installer does not touch that many times the file system [15:27] mandel: default functions? [15:28] nessita, I'm lost here: 'builting function' I supose you mean the stdlib ones, right? [15:29] mandel: yes, sorry for not being clear [15:30] nessita, no worries, is irc, no one is ever clear :) [15:35] nessita, on question, in ubuntuone-windows-installer, why do we have linux code? [15:35] or at least a linux.py? [15:35] mandel: because the wizard must work on both platform (is not working yet, but it will) [15:35] mandel: the wizard code will be moved to controlpanel, eventually [15:36] nessita, an the project will be renamed to unbutone-installer, right? 'cause is confusing :) [15:36] mandel: nopes, the project will have only the windows specific bits, like the building script and the migration code [15:37] nessita, ah.. ok === beuno is now known as beuno-lunch [15:51]