[11:39] good morning everyone [11:40] good morning al [11:40] all* [11:42] good morning! [11:47] good morning everyone! [11:47] rye: do you know if anyone has set up U1 on a server edition and written a guide about it? [11:48] i seem to remember someone was going to do this... [11:48] hello karni [11:48] hi duanedesign [11:51] how are you today? [11:51] karni: ^ [11:51] in a lecture? [11:54] duanedesign: not really, in hometown since Mon/Tue night :) [11:54] duanedesign: helped out with huge shopping twice and finding some time for coding :) [11:54] duanedesign: feeling good, thanks! I feel it will be a good day for programming haha [11:54] duanedesign: how are you :)? [11:57] * karni can't understand how ant builds broken apk while eclipse does the job right :< [11:58] hello nessita! [11:58] nessita, hi [11:59] hello everyone! [11:59] nessita, regarding bug I talked about yesterday, I'm still investigating when exactly the problem occurs, I'll report it when I have more data [11:59] zyga: you mean the bug re: throttling? [11:59] no, xdg stuff [11:59] ah, right [12:00] zyga: did you tried new control panel? [12:00] nessita, I was going to ask about tha [12:00] nessita, is it in any PPA I should be aware of? I didn not get any new u1 packages on natty yesterday [12:01] ralsina: FYI, yesterday I released and packaged control panel v0.1.0 (without the services tab) [12:01] nessita: great :-) [12:01] zyga: do you have any control panel installed, even if it's the old version? [12:01] nessita: are you on track for the devices tab as we talked? I got dropped off the internet too early [12:02] nessita, what's the package name? [12:02] zyga: becasue the package is in the repo, universe for now until we fill the MIR [12:02] https://launchpad.net/ubuntu/+source/ubuntuone-control-panel [12:02] ah,that explains it [12:02] just a sec [12:03] ralsina: more or less, yesterday I took a pause in the afternoon and kept working after dinner. And Chipaca said he will be pairing with me today, helping, so that is good [12:03] nessita: cool [12:03] ralsina: I should be able to get the tab ready for eoding today [12:03] ralsina: so I may bug you tomorrow for a review [12:03] nessita: no problem [12:05] Chipaca: let me know when you're ready to sprint [12:05] * nessita grabs a glass of milk [12:06] nessita, got it [12:06] nessita, playing with it, the captcha failed early on, let's see what i can find [12:19] nessita, I got it working [12:19] nessita, should it list all my synced folders already? [12:19] zyga: yes, on the second tab, 'Folders' [12:20] nessita, on clean install it behaves strange [12:20] meaning? [12:20] nessita, right now I can see everything about my plan etc but since I'm not signed (not user, with network) it does not list the folders [12:21] nessita, on top of that when I first started it there were two popups for sign-in [12:21] one from the maverick cycle [12:21] and the second one from your new control panel after I clicked the button for existing users on the right hand side of the panel [12:21] (both popus were different) [12:21] ah, that was syncdaemon that didn't have credentials [12:21] anyway, now I'm signed in but not connected so it says there are no folders [12:22] yeah [12:22] is a known bug but not reported [12:22] I think so too [12:22] could you please report that 2 sign in dialog? is caused by activating syncdaemon before the creds are in place [12:22] sure [12:23] I --connected [12:23] zyga: regarding your folders, if it says there is no folders, then there is no folders. Let's confirm: [12:23] zyga: wait, connection has nothing to do with folders [12:23] nothing changed so far, description is with user with network and "ready to connect" [12:23] processing queues [12:23] I think it just needs to get to a stable metadata state to know there are folders [12:23] but the UI is wrong [12:23] it's not "no folders" [12:23] but "no data yet" [12:24] zyga: in the folders tab, what do you see? a spinner? a legend? [12:24] I made a screenshot, just the text sayin "no folders to show" [12:24] ok, now, in a terminal [12:24] nessita, I'll be taking screenshots as I go, I can post them somewhere later [12:24] what does u1sdtool --lidt-folders show? [12:25] zyga: wait, if the UI says no folders, that's the info that syncdaemon sens [12:25] sends* [12:25] it shows my folders correctly [12:25] (in u1sdtool) [12:25] ok, I need the logs then [12:26] could you please do the following: [12:26] I know where the log file is [12:26] just tell me where to put them [12:26] wait! :-) [12:26] :D [12:26] ok [12:29] in terminal : killall ubuntuone-control-panel-backend; DEBUG=True python /usr/lib/ubuntuone-control-panel/ubuntuone-control-panel-backend [12:29] in another terminal: killall ubuntuone-control-panel-gtk; DEBUG=True ubuntuone-control-panel-gtk [12:30] and paste the output from both terminals to two text files and please attach it to a new bug report saying something like "folder info mismatch" [12:30] zyga: so I can debug further [12:31] nessita, I did that but now folders show up correctly [12:31] nessita: let's do it [12:32] if you want I'm on mumble, It's quite annoying to switch betweens vms to talk over irc [12:32] Chipaca: yes! I assigned a bug report for you [12:32] nessita: very generous of you [12:32] zyga: honestly, I can't. I'm running a lot against a deadline [12:32] zyga: I'll be happy to first week of jan, or in the rally if you go there [12:33] Chipaca: looking for it [12:33] I'll be there [12:33] Chipaca: bug 690325 [12:33] Launchpad bug 690325 in ubuntuone-control-panel (Ubuntu) (and 1 other project) "Download banner from a fixed URL, on failure use a default banner (affects: 2) (heat: 431)" [High,Triaged] https://launchpad.net/bugs/690325 [12:33] so do you want those _new_ logs? that second run showed all folders correctly [12:33] I suspect it's the very first run that's borked [12:33] zyga: not really... [12:34] zyga: can be, I will try it myself [12:34] zyga: thanks for your feedback! [12:34] Chipaca: please branch u1cp and I'll give you pointers [12:34] thanks [12:34] nessita: done [12:35] Chipaca: so, in u1/cp/gtk/gui.py there is this ManagementPanel which holds, for now, a fixed image that acts as a banner [12:35] Chipaca: tests are located at u1/cp/gtk/tests/test_gui.py [12:36] nessita: why the frontend and not the backend doing the image grabbing? [12:36] Chipaca: glade file for that piece is located at data/management.ui [12:37] Chipaca: yeah, I was pointing what we have so far [12:37] nessita: so moving it to the backend makes sense? [12:37] Chipaca: but +1 to image grabbing on the backend [12:37] Chipaca: backend has kinda 3 layers: [12:38] lower layer can be either webclient (does api calls) or dbus_client (calls SD and SSO thru dbus) [12:38] then, there is a backend.py layer that uses the above [12:38] and on top of that, a dbus_service layer that provides all the dbus for the backend layer [12:39] lowest layer and backend use deferred to handle blocking ops [12:39] dbus make use of those deferred and dbus signals to provide asynchronism [12:40] Chipaca: I think this image grabbing should go into webclient and escalte to the dbus_service module, where it should provide something like: [12:41] download_banner(params may be: locale? size? anything else? -> so, a dict, at least empty for now) [12:41] return signals should be something like: [12:42] BannerDownloaded(path?) [12:43] nessita: question: what does the devices tab show if there is no network? [12:43] BannerDownloadError(error_dict) (error dict is a string-string dict of the form {utils.ERROR_TYPE: type, utils.ERROR_MESSAGE: message} [12:43] Chipaca: we don't handle that. Network access is only handled in the overview screen :-/ [12:44] it will be stay 'fetching' for ever [12:44] so, in the UI, a spinner that never stops, in the backend, not sure if alecu did anything to timeout [12:49] duanedesign, well, i have rebuilt my server so that i can continue testing that. There are two ways - first is to make ubuntuone-syncdaemon working on the server and second - provide a special commandline client that implements basic upload/download/list/etc [12:49] duanedesign, i am now testing the latter variant, since running dbus in ssh of some sort is really hacky [13:56] nessita, dobey standup in 6' [13:56] argh! /me forgot [13:57] nessita: dobey is away so maybe I should just call you on the phone :-) [13:57] huh? [13:57] dobey: well, you were away 2 minutes ago :-D [13:58] * nessita writes the summaty [14:00] ooh yeah [14:01] hrmm, latency went up 2-3 ms it seems [14:01] me [14:01] me [14:03] DONE: started PyQt/SSO-client coding [14:03] TODO: finish it, start PyQt preferences following nessita's doc and Ivanka's UI design [14:03] BLOCKED: no, just slow [14:03] nessita? [14:03] DONE: work on bug 673672 and bug 673673 (services tabd, basically). Answered honks on #u1. Follow up on somw ussoc bugs (found bug 693531) [14:03] TODO: finish services tab [14:03] BLOCKED: nopes [14:03] NEXT: dobey [14:03] me [14:03] Launchpad bug 673672 in ubuntuone-control-panel (Ubuntu) (and 1 other project) "Check with apt-get if the bookmark extension is already installed (affects: 1) (heat: 85)" [Medium,In progress] https://launchpad.net/bugs/673672 [14:03] Launchpad bug 673673 in ubuntuone-control-panel (Ubuntu) (and 1 other project) "Use apt-get with gksudo to install the extension (affects: 1) (heat: 85)" [Medium,In progress] https://launchpad.net/bugs/673673 [14:03] λ DONE: bug 693462, bug 660648, major headway on initial client backport work [14:03] Launchpad bug 693531 in ubuntu-sso-client "Failing to store a token in the keyring results in a verification email (affects: 1) (heat: 6)" [High,Confirmed] https://launchpad.net/bugs/693531 [14:03] Launchpad bug 693462 in ubuntuone-dev-tools "dbus config loading fails if not installed in system (affects: 1) (heat: 6)" [High,Fix committed] https://launchpad.net/bugs/693462 [14:03] λ TODO: 683351, finish initial client backport work for lucid, cp nightlies [14:03] Launchpad bug 660648 in rhythmbox-ubuntuone-music-store (Ubuntu) (and 2 other projects) "Non translated elements in the interface (affects: 1) (heat: 6)" [Low,New] https://launchpad.net/bugs/660648 [14:03] λ BLCK: None. [14:03] http://www.speedtest.net/result/1082145214.png [14:03] wheeee [14:04] dobey: ok, envy achieved ;-) [14:04] any comments about the standup? [14:05] dobey: and while we' re at it... http://twitpic.com/3iqf8a [14:06] ralsina: you're not working *on* the ferris wheel? :) [14:06] me! [14:06] comment: Chipaca is hacking on the cp! ueeeee [14:07] nessita: congratulations. I feel lonely, I never code with anyone :-) [14:07] and another comment, already said: u1cp v0.1.0 was released and packaged yesterday [14:07] dobey: Not yet. I may go there in the lunch break, though ;-) [14:07] natty universe repo was hit yesterday as well, now we need to fill the MIR [14:08] ralsina: would you help me with the MIR? https://wiki.ubuntu.com/UbuntuMainInclusionRequirements [14:08] ralsina: I started the bug at bug #690227 [14:08] nessita: sure! [14:08] Launchpad bug 690227 in ubuntuone-control-panel (Ubuntu) "[MIR] ubuntuone-control-panel (affects: 1) (heat: 488)" [High,New] https://launchpad.net/bugs/690227 [14:08] anything else or eom? [14:09] ralsina: you can read the wiki page and use all your super powers to convince the mir team why cp is so awesome that should go to the CD. Chipaca and dobey may help, I think [14:09] eom for me [14:09] I will try, I have not used THOSE superpowers yet :-) [14:10] it's so awesome it should go in ubuntuone-client ;) [14:10] dobey: you can let it go, can't you? is xmas! :-) [14:11] nessita: I should nominate it for natty, or is some other superpower that' s needed? [14:12] ralsina: what do you mean? the u1cp package is already on natty universe repo [14:13] nessita: then I have no superpower-related links on launchpad for that bug and I am not sure what you are asking of me [14:13] It's one of those things I just never did before and need to learn how to do [14:15] ralsina: let me explain myself a bit more. The wiki describes the procedure to nominate a package already in universe to be in the main repo. For this procedure, we need to fil a new bug which description should conform all the requirements from section "Requirements" [14:15] ralsina: what I would need help from you is grabbing the bug 690227 and completing the description so we can apply for inclusion in the main repo [14:15] Launchpad bug 690227 in ubuntuone-control-panel (Ubuntu) "[MIR] ubuntuone-control-panel (affects: 1) (heat: 429)" [High,New] https://launchpad.net/bugs/690227 [14:16] nessita: ooooohhh ok, will do immediately [14:16] s/completing/complete/ [14:16] ralsina: this is the description of the MIR process: https://wiki.ubuntu.com/MainInclusionProcess (sorry, I should have pointed this before) [14:17] nessita: ok then :-) [14:17] umm.. how do I share a folder from under Nautilus ? [14:17] ralsina: basically we're at point 2 (File a bug report for the package, titled "[MIR] sourcepackagename". Include the rationale and description of the violations of UbuntuMainInclusionRequirements, and a confirmation that you checked the requirements carefully. ) [14:18] karni: right click, context menu -> Ubuntu One -> share (or something similar) [14:18] so, I can do the confirmation, and then subscribe MIR [14:18] karni: right click on the folder [14:18] nessita: I don't have Ubuntu One in the contest. apart from Publish [14:18] ralsina: yes, we need to complete the bus description first, is not ready [14:18] no wait! it says [14:19] nessita: right [14:19] Synchronize on Ubuntu One [14:19] nessita: I'll make a pass at completing it and get back to you, ok? [14:19] karni: nopes, that is for making a folder. So the folder is not under Ubuntu One yet? [14:19] ralsina: yes! [14:19] karni: I mean, what you're trying to share is not in U1 yet? [14:19] nessita: I put it in my home. bah, it's probably because it's not under U1 folder? [14:19] right [14:20] it's just for testing, I'll make a new folder under U1 [14:20] karni: you want to share with other people or with other machines of yours? [14:20] nessita: I am still lost in some spots but I promise you won't have to walk me through the same things more than once, usually ;-) [14:20] nessita: one person (anyone actually ;D ) [14:20] nessita: there it is, Share option, thanks [14:20] prego [14:21] ralsina: :-) [14:21] ok, going away for coding for a while [14:24] beuno: I have shared a test folder with you. Please accept ^ ^ Upload works under ROOT and UDFs, time to test SHAREs [14:27] nessita, dobey: I won't be able to do the weekly meeting at 12:15. We should reschedule or just skip it, since nothing much changed since last week except "everyone worked", I think [14:27] ralsina: +1 to skip [14:28] skipping is good [14:28] unanymous skipping then. I feel we would just be later by having it :-) [14:29] nessita: would you mind if I shared a low-traffic test share with you? I need to have at lest one to test Ubuntu One Files upload to shares [14:29] and beuno is not here ATM [14:30] karni: sure! as long as you don't need much input from me, I'm kinda busy :-) [14:30] karni, woooo, share away! [14:30] nessita: no, not really. just a e-mail on priv, and that's it ^ ^ [14:30] hahaha [14:30] nessita: ok, no need for share, beuno's here, thanks! [14:30] karni: :-) [14:31] hola beuno [14:31] hola nessita! [14:31] beuno: enjoying the weather? :-D [14:31] * nessita is melting [14:31] nessita, I'm thinking of just opening the refrigerator and leaving it open [14:32] beuno: it may break, but that may not be an issue given this insanely hot [14:34] beuno: you can check 1 file uploaded from U1F to the share. I'd appreciate if you made another test folder and shared it with me (I know previous share is bi-directional, I want to check something else) [14:37] nessita: I did a first pass, still needs the background section, which I will do in ... 90 minutes. I have to go buy groceries and eat. [14:37] sure! [14:45] karni, sure [14:46] beuno: thanks [14:46] karni, done [14:48] * karni waits for U1 to process the share. not visible yet. [14:49] karni, I used the desktop, so let me know if it doesn't arrive [14:50] beuno: yes, I can't quite see it yet, not under webUI [14:50] nor an e-mail (should I get one?) [14:50] * Chipaca looks at beuno [14:51] beuno: there it is! [14:54] bah, stupid internets [14:55] nessita: I installed u1cp, and got no binary for my effort [14:56] Chipaca: what u1cp package did you install? there are 3: python-u1cp, u1cp, and u1cp-gtk [14:56] nessita: right, just realized taht [14:57] nessita: you need to make u1cp recommend u1cp-gui, and make u1cp-gtk provide it [14:58] Chipaca: would you please please file me a bug under the package tracker? https://bugs.launchpad.net/ubuntu/+source/ubuntuone-control-panel/+filebug [14:58] nessita: sugar [14:59] * nessita prefers honey [14:59] both for eat and for to be called [14:59] :-P [15:03] hiya Chipaca [15:05] beuno: hi [15:06] nessita: is the u1cp banner already on the servers? [15:10] nessita: bug 693798 [15:10] Launchpad bug 693798 in ubuntuone-control-panel (Ubuntu) "u1cp should recommend u1cp-gui, and u1cp-gtk should provide it (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/693798 [15:14] Chipaca: no, is not, we need that branch as well [15:14] nessita: ok, wfm [15:15] Chipaca: beuno said it should go to /media/images, and alecu suggested we should hit a url pointing to a view that delivers the image, so we can add any logic we need in between [15:15] Chipaca: if you feel like doing that branch as well, you will be awarded (?) [15:16] nessita: on this, i agree with beuno [15:16] Chipaca: me too, and with alecu, is mainly the same? we need to process the accept-language stuff, for example [15:19] nessita: apache handles that [15:20] Chipaca: but what if we want to send different images depending on the client revno? [15:20] version perhaps. revno not possible [15:20] dobey: right [15:20] nessita: we'd have to send the client revno for that [15:20] heh, no $revno$ in bzr [15:21] Chipaca: yes, can be done easily, right? not saying to do it today, but should be a small branch on bug fixing stage [15:21] sending the client version is easy [15:21] just a one line change probably [15:22] well, maybe more, since you need to store it somewhere [15:22] plus a few lines on tests, but yes [15:22] but as far as the http bit goes, it's one more call i think [15:22] one more? isn't it the same? [15:22] nessita: addHeader or whatever it's called [15:23] setUserAgent or something [15:23] http://hit.me/give-me-image/I'm-natty-0.582.5 [15:23] ah, I was planning on passing it as django view argument [15:23] no you should set the User-Agent header properly [15:23] dobey: why? [15:23] not passing http arguments [15:23] why not a django view argument? [15:23] because it's the correct way to do it [15:23] dobey: can you please explain why? [15:24] I find it cleaner passing an argument to the view, but I know little about http [15:24] nessita: RFC 2616 [15:24] dobey: I won't read that today, can you please share the interesting part? [15:25] nessita: how do you get the information to the server? [15:25] nessita: in the URL? [15:25] nessita: http://pastebin.ubuntu.com/546946/ [15:25] nessita: but then we're stuck with serving it from django [15:25] Chipaca: yeah... just like the remove phone api, let's say [15:26] nessita: and it's a static file, modulo language and revno or whatever [15:26] Chipaca: I would you with you and dobey think best, like I mentioned I'm not particular skilled in this area. I still think that having a django view gives us a lot of power, but maybe that's overkill at this stage [15:27] I would go* [15:27] dobey: understood, thanks [15:27] the django view is probably the easiest way to do it, but the language and user-agent should still be passed with the correct headers as per the RFC [15:28] dobey: right [15:28] dealing with the language might require a script anyway, to deal with falling back through the list [15:29] so that if someone is using turkish, we don't just fail [15:29] dobey: I thought that apache can handle fallbacks [15:30] nessita: not sure what you mean there. apache has no knowledge of what the appropriate content to serve is, outside of the url being requested [15:31] dobey: ah, I was confused [15:31] * nessita brbs [15:32] nessita: ping [15:32] * dobey prepeares to watch a million failing tests [15:33] dobey: got a second for a quick question? [15:33] mandel: what's up? [15:34] dobey: is there code to access the secrets api from python already available somewhere? [15:34] dobey: like the txsecrets cde in ubuntu-clent-sso [15:35] mandel: just the txsecertes in ubuntu-sso-client afaik [15:36] dobey: I'd like to use the code for destopcouch, so that we do not worry about kwallet etc.. [15:36] mandel: right [15:37] but adding a dependency to ubuntu-sso is not a great idea... [15:39] mandel: yeah, the txsecrets code needs to become a library on its own [15:40] dobey: indeed that would be the best idea, I'd try to do that [15:40] nessita: any objection? ^ [15:41] well i guess we need to coordinate that [15:41] because it means new packages in ubuntu and more red tape stuff [15:43] dobey: yes, I agree, but the sooner the better, also I've heard bad things about the python bindings for the gnome-keyring [15:43] ralsina: could we get this started at some point? [15:43] * mandel wonders if ralsina is here.... [15:44] mandel: ralsina left at 15:37 for 90 minutes, FYI [15:45] * karni could have been a bit less detailing :F [15:45] mandel: i'd say write the code and import from sso for now, and we can deal with splitting it after the holiday when we're all back [15:46] karni: hahaha cool, thx for the info :) [15:46] dobey: makes sense, at least I'll start working on it [15:46] mandel: if you really insist on working on it on your holiday anyway. at least make it work as-is, and then we'll make it landable first week of jan [15:47] dobey: heheh I should stay away for the holidays, that is true... [15:47] beuno: things looking good today. photo upload fully functional, I also know how to possibly speed up initial sync (writes to sqlite per file should be converted to batch or some other burst write ) [15:47] mandel: I was away for a few minutes, wanna repeat? [15:47] nessita: I was just talking about when should we move desktopcouch to use txsecrets, but dobey correctly pointed out that it could be done when people are not on holiday, to many pieces to move [15:48] mandel: so, uyou're splitting the tcxsecrets on its own library? we could just do a separated package [15:48] dobey: what about just another package? [15:49] not that I can do that today, but january I can. I also have to split the gtk code into a separate package [15:49] nessita: then I can wait 'itl then, not a huge issue [15:50] and I'm on holidays :) [15:50] nessita: i think it should be moved out of the ubuntu_sso namespace. there's no point in just putting it in another binary package, when the idea is to avoid depending on ubuntu-sso-client to use it. it would be better as a separate source at that point [15:51] dobey: like a whole new package and project, you say? [15:52] nessita: I believe dobey is right, it should be a diff tree so that e can use in diff projects with not heavy dependencies [15:52] yes i think it makes sense in this case, unfortunately [15:54] dobey: ok then! Let's do this. I'll do it on Jan. mandel: Can you please file a bug for me? on https://bugs.launchpad.net/ubuntu-sso-client/+filebug [15:54] nessita: sure [15:54] * mandel flies to launchpad [15:56] nessita: there you go bug 693821 [15:56] Launchpad bug 693821 in ubuntu-sso-client "Txsecrets must be reused (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/693821 [15:57] mandel: thanks, I'll tweak a bit the description [15:58] nessita: please do, I suck at those things... [15:58] :-) [16:36] * nessita -> lunch [16:39] nessita: anywhere in u1cp where the version is stored right now? [16:40] hrmm [16:41] Chipaca: i suppose the question is do we want to send the cp version or the file sync client version, or what [16:41] Chipaca: if the version is goign to be used to determine a "new features" image to use, we should probably figure out how to classify that properly [16:42] ISTM that we can safely not send anything now, and add the version info we want down the road [16:42] i.e. we'll know the one with no version info is the one in Natty :) [16:47] lunch time [17:00] mandel: ping [17:00] ralsina: pong [17:01] what should we get started? ;-) [17:03] ralsina: at some point, after xmas, to move txsecrets out so that it can be used by the other projects [17:03] mandel: ok [17:04] care to explain what txsecrets is, unless you prefer I go google it and look more competent ;-) [17:05] ralsina: txsecrets resides in ubuntu-sso-client (under utils) and is the code that uses the secrets api from dbus to access the keyring [17:05] oh, ok [17:05] ralsina: that way we get kwallet and gnomekeyring for 'free' on linux [17:05] so, it' s a binding to gnome-keyring via dbus? [17:06] oh, good [17:06] is it better than python-keyring? [17:06] because that one also supports windows and "nothing" [17:06] ralsina: is a binding to the secrets api via dbus http://www.freedesktop.org/wiki/Specifications/secret-storage-spec [17:06] mandel: cool, didn't know that API at freedesktop [17:07] So, yes, I agree it would be nice to have as a separate project someday [17:08] ralsina: I'd like to have it so we can get it in desktopcouch ( i was looking at thekeyring for windows when I though about it) [17:08] mandel: desktopcouch was migrating to python-keyring last I looked :-) [17:08] * ralsina may be wrong, of course [17:09] beuno: this is why initial sync is slower than it should http://code.google.com/p/android/issues/detail?id=681 -- we can walk this wround with a bit of hacking, but I don't think it's at the very top of TODO ATM. I'll shift my focus a bit to app Preferences. [17:09] * ralsina even has a vague memory of reporting that bug and/or sending a patch [17:09] ralsina: that was an option, but did not work due to the way we store things in gnome-keyring [17:09] ralsina: [17:10] ralsina: sil know more about the specifics [17:11] hey, I did report that bug and you answered it, a year ago :-) Too bad it didn' t work [17:11] 1 year+1 day ago to be exact [17:12] beuno: on the very low lever, this is how I'll improve speed a little -- http://paste.ubuntu.com/546979/ [17:12] but anyway, sure, that's ok by me. What are you going to use on windows? [17:12] *level [17:13] ralsina: we currently have a small code that uses the DPAPI from windows and the keyring, since vds and found issues with python-keyring on windows [17:13] mandel: ok === Mohan_chml is now known as IAmNotThatGuy [17:14] I'm getting segfaults with libsoup :( [17:26] Chipaca: is that when running soup + gtk? or libsoup alone? [17:26] nessita: soup alone [17:26] argh [17:26] Chipaca: so, not threads involved, and it crashes? [17:27] nessita: 1 sec and I'll pastebin my code [17:27] maybe it's just me (?) [17:28] ack. Anyways, we can download the image using plain urllib and if there is proxy or some other failure we just use the fallback image? [17:28] nessita: I'd hate to do that [17:28] hate, hate, hate [17:28] nessita: but, I can do it to get the feature working, hate myself, and then we fix it later [17:29] Chipaca: right. The most complicated part, that I know of, is detecting a problem since urlopen will not timeout on its own [17:35] nessita: http://pastebin.ubuntu.com/546984/ [17:35] nessita: you can make it timeout, let me check [17:35] Chipaca: missing f.close on line 12 [17:35] f.close() [17:36] nessita: in python 2.6 and later urllib.urlopen has a timeout parameter [17:36] Chipaca: that's why the full image is not completely stored (I think) [17:36]