[00:52] finally EOD, later anyone who's still on :) === JackyAlcine is now known as ThisCode === ThisCode is now known as JackyAlcine [09:35] omg is cold! [09:35] morning! [09:44] morning, nice and sunny here in South Africa :) === yofel_ is now known as yofel [09:50] psypher246, hi, i will mark bug #634013 as a dup of bug #720707 since the later contains the reason why the bandwidth throttling is not properly enforced [09:51] rye: Error: Bug #634013 not found. [09:51] Launchpad bug 720707 in ubuntuone-storage-protocol "Bandwidth limit is not correctly enforced: Transmission delays are inserted between data chunk writes (of arbitrary sizes) (affects: 31) (dups: 6) (heat: 150)" [High,Confirmed] https://launchpad.net/bugs/720707 [09:51] as we found out, syncdaemon bandwidth throttling works much like air conditioner [09:55] rye: thanks, trying to get the jist, so is there still a bug? or is it just jnot gonna work properly? [09:55] psypher246, there were no changes to the way storage protocol is handing the limits so this bug still exists [09:56] ok [09:56] btw I have now tried uploading 9000 files again [09:56] been about 5 minutes now and nothing is uploading [09:56] do you think I should still wait or log a bug? [10:01] rye: what does makefile mean? MakeFile(running=True, share_id='', path='/home/ruald/Ubuntu One/done/Pics03304.jpg', marker='marker:c1141b34-2043-45ec-8dc7-871a5c7ce751', name='Pics03304.jpg', parent_id='c810e06b-0934-45a3-bc1d-5b22ab251eb6') [10:02] psypher246, MakeFile is a call to the server to allocate the storage for new file, upload comes afterwards. Does u1sdtool --waiting work now? [10:02] psypher246, current clients send metadata queue item sequentially but content upload/download are sent simultaneously in chunks of 10 files [10:06] u1sdtool --waiting is working now after I waited about a day for it to action the "delete" of the 30000 files, had to constantly kill the syncdaemon process cos it klept getting stuck. then it still left the files on the web clinet and had to manually delete there too. After shutting down the daemon and staring again I could upload and delete small groups (300 files). [10:06] after deleteing thiose 300 I( uploading about 9000 [10:07] now it seem that the makefile count is decreasing slowly but nothing has uploaded yet [10:07] u1sdtool --waiting | grep Upload | wc -l [10:07] 9539 [10:07] u1sdtool --waiting | grep MakeFile | wc -l [10:07] 6051 [10:07] u1sdtool --current-transfers | grep path | wc -l [10:07] 10 [10:09] czajkowski, ping [10:09] mandel: pong [10:09] czajkowski, saying over g+ is lame, so congrats! [10:10] czajkowski, welcome to the getting constantly bigger family :) [10:11] mandel: hehe thanks, I start Monday :) [10:12] czajkowski, cool! [10:12] czajkowski, I want nicer lp! :P [10:12] have been doing the happy dance all week [10:13] czajkowski, haha I'm glad you got it, it was a matter of time [10:13] czajkowski, I'll talk to you later though, feature freeze is on friday and there is lots to do :P [10:14] Morning all! [10:15] psypher246, could you please pm me your e-mail to look up what shard you are on and test the performance there? [10:16] mandel: np enjoy try and not break things :) [11:34] 3eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeE [11:34] {NH>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>{{????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? [11:34] ='''''''///////////////////////////// [11:34] ='''''''///////////////////////////// [11:34] | [11:34] DC [11:34] good morning, duanedesign's cat! [11:36] ralsina, morning! [11:36] good last-morning-before-effective-feature-freeze mandel! [11:36] ralsina, yes, and things are finally comming together in the stupid dialog :) [11:37] mandel: cool! [11:38] ralsina, although, I have a question, the Keyring module on linux is using twisted.. which means that atm I'll have to use the qtreactor for that, which is a PITA [11:38] mandel: ouch [11:38] ralsina, can I use that before FF and then later change it to use the normal python-deferred module? [11:38] ralsina, I don't want to add a new dependency just yet [11:38] mandel: qt4reactor is not packaged [11:39] mandel: we could ship it broken, I suppose :-/ [11:39] ralsina, I wonder if using this is enough: http://packages.python.org/defer/defer.html [11:39] I don't know [11:40] I have no idea how it's using twisted there [11:40] ralsina, I'll ask alecu and nessita asap [11:40] mandel: yes please [11:40] ralsina, this is a huge PITA since we need to get that working today [11:40] mandel: I suppose they missed that one when they were clearing the dependency on qtreactor [11:41] ralsina, as soon as that is done I'll push the 5 branches with all the dialog work.. [11:41] mandel: ack [11:42] ralsina, and it seems that python-deferred is not packaged.. [11:42] puta puta puta.. [11:42] mandel: don't panic [11:43] ralsina, oh, it is.. phew! [11:43] ralsina, is python-defer :) [11:52] ah, cannot make bugpattern for this :-/ [11:55] good morning nessita! [11:56] hello everyone! [12:02] nessita, mandel has a problem where he may need a hand [12:02] ralsina: ok, tell me more [12:02] nessita: apparently, the keyring is using twisted and that makes it require qt4reactor, IIUC [12:03] nessita, more confirmation than anything else, can I use python-defer in the creds dialog? [12:03] nessita, do we use that already? or is it a PITA? [12:03] mandel: you could, but I'm not understanding why you need that [12:03] mandel: you can safely use twiste deferreds without needing a reactor [12:03] nessita, and they work without a twisted reactor? [12:04] mandel: of course [12:04] nessita, ok, then I'm happy :) [12:04] mandel: nice! that was the things that was blocking you? [12:04] nessita, more than blocking, scared me :) [12:07] mandel, ralsina: the "deferreds" mechanism is generic, and the only thing needed to progress is a mainloop (any mainloop) [12:08] nessita, ack [12:14] nessita, can you help me debug a problem that I have with this thing? [12:14] mandel: sure [12:14] hi [12:15] nessita, code is located at lp:~mandel/ubuntu-sso-client/keyring-integration and I use http://paste.ubuntu.com/835121/ to launch the dialog [12:15] nessita, I'm trying to irl test the dialog with a random domain [12:16] mandel: what do you mean with random? dummy, or truly random? [12:16] nessita, dummy [12:16] ah, great [12:16] branchinf now [12:16] nessita, I'm trying to see that IRL it works, but I never get pass the yield in the _on_save_clicked [12:17] hola gatox [12:17] ralsina, buenas [12:17] gatox: we missed the 1-1 yesterday want to do it really quick now? [12:17] gatox: 10 minutes tops [12:17] ralsina, i'm at nessita's home [12:18] gatox: it's ok, no problem [12:18] ralsina, do you prefer tomorrow? [12:18] gatox: sure [12:18] ralsina, great [12:20] mandel: first thing, not related: we can't have any Ubuntu One strings hardcoded in sso [12:21] mandel: what we have there is legacy and is specifically handling old U1 token, we'll remove that code soon [12:21] mandel: everything that you need to customize for your dialog (re: strings) should be UI creation params [12:21] nessita, hm.. I followed the present code, but I can pass it to the dialog, that is simple :) [12:23] mandel: right, but like I mentioned, that code is specific for handling old-style U1 tokens... those with the url in the name, remember? [12:23] rebooting because icons start disappearing [12:24] mandel: so, the UbuntuOneProxyKeyring should be a plain ProxyKeyring [12:24] nessita, yep, but I was looking at the U1_APP_NAME [12:24] nessita, ack, changing that [12:24] mandel: right, the token used to be called "UbuntuOne token for https://ubuntuone.com" and we needed a key to check if we should try with that, see the file: [12:25] try_old_credentials [12:25] (method, sorry) [12:26] nessita, yes, but there is U1_APP_NAME and U1_KEY_NAME, I wanted to have a ref to something similar to U1_APP_NAME but that uses the domain so that if you change of proxy I look for the proxy domain specific one [12:26] nessita, but I can remove that, is not a problem, I'm more interested in finding why keyring is not storing the creds [12:27] mandel: I know, but I also would like that you understand why that U1_ constants are there, and why we can't have Ubuntu One hardcoded for new stuff [12:28] nessita, sure, because other apps use sso, that I understand === Karmak23_ is now known as Karmak23 [12:28] mandel: and those U1_ constants are there to handle specific cases of users with tokens for U1 that were created with the old-style (no sso involved and a browser being opened for them) [12:29] ack [12:31] mandel: (I'm first quickly reviewing the code to see if anything pops up). So, this code will not return any icon on windows: QIcon.fromTheme('gtk-dialog-authentication') [12:32] nessita, true, and is intentional until we get a global qss, then I can add a fall back that uses that one in the resource file, which is passed as an optional parameter to the fromTheme method [12:32] mandel: ack [12:32] mandel: have you tried putting a try-except around the sentence that call the keyring? if that fails, you never know [12:32] nessita, lets say I'm doing mini branches and atm just focusing on linux regarding theme and ui [12:33] nessita, hm, I'll try that :) [12:33] mandel: is perfect, is what we agreed:-) [12:33] (regading leacing ui details for next week) [12:35] mandel: another detail: to avoid hundreds of conflicts with gatox's branch, and because is recommended to have each ui in it own file, would you please move your dialog to a separated file? and do the same for the tests for that dialog, please [12:35] nessita, I added the try-catch, no thing happens, I also added some print statements and got to yield self.service.open_session() in set_credentaisl [12:36] mandel: so we should have the proxy_dialog.py and test_proxy_dialog.py [12:36] nessita, sure, no problem what so ever, easy work :) [12:36] mandel: let me try it IRL [12:40] mandel: question, the is_error that is passed to the constructor will be removed when proposing for merge, no? [12:41] nessita, well, I need to think about it, I'm sure there is a nicer way to do it [12:42] nessita, atm just allows to show the error if you got a 401 from the proxy, but I'm sure there is a nicer way to put it all together in a nice package [12:42] package not refering to a python package, more like a simil :) [12:43] mandel: another question, any reason to pass a keyring_factory to the ui? [12:43] (instead of importing the Keyring class directly) [12:43] nessita, testing, that is all, that will be remove for sure [12:43] mandel: nice [12:44] nessita, but since it was not working I wanted to keep it dumb and stupid, then I'll clean up everything [12:44] mandel: another question, why do you need a specialization of the keyring for the proxy? [12:45] nessita, I simply followed the current code, I'll remove that since there is no real point [12:45] mandel: the UbuntuOneOauthKeyring is not what we use for auth, is a plain Keyring (the former handles the old-style tokens) [12:46] nessita, you told me that a few mins ago and I agreed that was going to be removed :) [12:46] mandel: did I? I just saw that clase, before I was talking about the U1_ constants... [12:46] nessita, is work in progress, imaging is not there, I'm interested in why I cannot open the session when using a QApplication and no twisted reactor :) [12:47] nessita, well, I uderstood that both things were deprecated, right? [12:47] mandel: yes [12:47] nessita, so we are in the same page :) [12:47] nessita, just ignore all that, is gone in the next revision [12:54] mandel: so, have you tried setting any key in the keyring usinf the plain d-bus service? [12:55] nessita, I did some tests, the keyring is working correctly [12:56] mandel: so, I just ran the sso service with U1_DEBUG=True PYTHONPATH=. bin/ubuntu-sso-login [12:56] mandel: went to d-feet, chose the Credentials interface, and executed store_credentials, passing as args: [12:56] 'foo bar baz', {'foo': 'bar'} [12:56] and that was stored in the keyring [12:57] let me try that using the qt mainloop (the former was using the glib plain mainloop) [12:57] nessita, I'm not using sso, I'm importing Keyring which uses txsecrets from the dialog [12:57] mandel: yes, and perhaps that's the issue? [12:58] using qt mainloop also works [12:58] nessita, could you try to do the same but not via dbus but importing the keyring? since txsecrets already does dbus, right? [12:59] nessita, and now I realize about the problem... I did not set the mainlopp to DBusQtMainLoop and txsecrets does not do it [13:00] * mandel facepalm [13:00] mandel: the thing is that our design does not expect that you access the keyring that way [13:00] mandel: you should go thru our backend [13:00] mandel: so, the use case would be something like: [13:00] (in my head) [13:01] control panel requires credentials for proxy domain '1.2.3.4' [13:01] control panel uses the same service it uses to gather oauth credentials, the sso. So it calls: [13:02] credentials_backend.find_credentials('Ubuntu One creds for proxy 1.2.3.4') [13:02] that will emit the CredentialsNotFound signal [13:02] sorry, if you use the CredentialsManagementTool is easier [13:04] tool = CredentialsManagementTool() [13:04] creds = yield tool.find_credentials() [13:04] if not creds: [13:04] creds = yield tool.login() [13:04] nessita, isn't that doing several calls over dbus? I mean dialog -> sso -> keyring? what does the sso step add to the overall design? [13:04] mandel: so, I think we need to provide a ProxyCredentialsManagementTool [13:04] mandel: the dbus part is hidden on the CredentialsManagementTool ... [13:05] but I'm not sure we can do all those changes before tomorrow [13:05] mandel: so, something in between would be that you use the sso backend (and not a tool in between) [13:07] mandel: or you can set the DBusQtMainLoop in the ui executable (not in the python modules) [13:07] nessita, I was typing just that :) [13:07] that's the faster option but is the dirtiest [13:07] nessita, I don't see that dirty, we just need a dialog that gets creds, stores them and tells webclient that he has them [13:08] nessita, then webclient uses them, or at least that is what I understood.. [13:08] mandel: yes, and we already have something similar in place for sso creds, no? and follows another design [13:08] mandel: but given the time constraints I guess we can go dirty [13:08] nessita, dirty, we meet FF then we improve === alecu_lunch_n_er is now known as alecu [13:12] hello girls and boys! [13:12] nessita, ha, that was simple, it was setting the correct dbus mainloop [13:12] hola alecu [13:18] alecu, buenos dias! [13:49] * mandel lunch [14:11] facundobatista: ping [14:11] nessita, ping [14:11] nessita, pong, I mean [14:11] facundobatista: I just the whole client suite in windows, and there are 2 test failures on test_offloadqueu, bug #929546 [14:11] Launchpad bug 929546 in ubuntuone-client "Test failure for offloadqueue (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/929546 [14:12] facundobatista: did you or any reviewer run your branches under windows? [14:13] I just run* I meant [14:14] nessita, nop [14:14] nessita, bah, I don't think so [14:14] nessita, why? [14:15] facundobatista: there are 2 tests failures in windows in that code [14:15] facundobatista: saw the link I pasted above? [14:15] nessita, let me see [14:16] nessita, facundobatista: here is where they started failing: https://jenkins.errormessaging.com/job/ubuntuone-client-windows-test/72/ [14:16] We really need to get jenkins messaging when these thigns happen :-) [14:16] facundobatista: for future branches, could you please ask me, or any other desktop dev to run the suite for you in windows? I understand you don't have a windows env setup, so we can run them for you (but ideally before they land) [14:17] nessita, ok [14:18] briancurtin2: hola! your branch is reviewed, a couple of trivial fixes and we can land it! [14:18] nessita: cool, i'll take a look [14:19] oh hi classic gnome no effects [14:19] i've missed you [14:19] nessita, did you see the test that fails? [14:20] nessita, there may be a strange situation where the log check fails because of the name... the whole test is [14:20] self.assertTrue(self.handler.check_debug("Using temporary file", [14:20] self.oq._tempfile_name)) [14:20] so, maybe the "temporary file name" is a strange thingie? [14:21] (in windows, I mean) [14:21] facundobatista: path length? [14:21] facundobatista: I can try reproducing in isolation and giving you the log [14:21] nessita, please [14:21] nessita, maybe put the handler in debug mode? [14:21] ack [14:22] (though is another price ;-)) [14:23] :) [14:25] facundobatista: strange, there is no log file under _trial_temp/xdg_cache [14:26] nessita, what stdout says? (regarding handler in debug) [14:27] facundobatista: added as a comment to the bug report [14:28] nessita, ah! is that "you can not compare file names in debug" in windows issue! [14:28] nessita, right? [14:29] /1/5 [14:29] facundobatista: perhaps you're using %r for logging? if so you need to pass to the memento check repr(path) [14:30] nessita, yes, I am [14:30] facundobatista: doing both? [14:30] nessita, nop, just the former :) [14:30] ajá! [14:30] facundobatista: remember in windows, %r != %s for a path [14:31] because of the backslashes [14:31] facundobatista: very likely [14:38] nessita, https://code.launchpad.net/~facundo/ubuntuone-client/compare-repr-path/+register-merge [14:39] you mean https://code.launchpad.net/~facundo/ubuntuone-client/compare-repr-path/+merge/92281 ? :-) [14:40] nessita, eso, sorry [14:40] ack! [14:41] nessita, also, how do I ask for client reviews? https://code.launchpad.net/~facundo/ubuntuone-client/different-dir-deletion-check/+merge/92282 [14:42] facundobatista: you mean for us to run the windows suite? [14:45] nessita, yes, please [14:46] facundobatista: to ask, you can ping directly anyone (me, mandel, briancurtin2, gatox, etc). I will run them now [14:46] nessita, gracias === briancurtin2 is now known as briancurtin [14:47] yep, i can (and probably should) do windows reviews/test runs to get more acquainted with things [14:53] briancurtin: wanna start with this one? https://code.launchpad.net/~facundo/ubuntuone-client/different-dir-deletion-check/+merge/92282 :-) [14:53] nessita: will do [14:53] * mandel back [14:54] facundobatista, need help? [14:55] mandel, I hope not! thanks [14:55] facundobatista, ok :) [14:56] facundobatista, if you need a window vm I can give you the details of my vm server, it might be slow but you can run the tests there and is ready for development :) [14:56] alecu: quick question re: webclient api [14:56] alecu: if wc.request does not throw any exceptions, does that mean the request finished with code 200? [14:56] alecu: or there are some greys in between? [14:58] nessita, I think no exceptions means only 200 [14:58] nessita, lemme check [14:58] nessita, is there a way to talk with syncdaemon in linux? [14:58] facundobatista: besides dbus you mean? [14:59] nessita, dbus is not working [14:59] facundobatista: it works from trunk, nightlies are failing to build [15:00] nessita, a trunk started SD, and using u1sdtool from trunk? [15:00] facundobatista: nopes, just a sd started from trunk and the system's u1sdtool [15:00] nessita, gatox, briancurtin, alecu, mandel, joshuahoover, dobey: mumble? [15:01] ohh, logging in [15:01] ralsina, nessita is turning on the other computer to connect to mumble [15:01] ralsina: apparently our wifi will not connect to mumble (wtf?) [15:01] dobey: I seem tobe having mumble issues, can you hear me? [15:02] nessita: wtf indeed [15:02] oh, mumble [15:02] ralsina, launching it [15:02] ralsina: you sounded like one of those dubstep songs [15:02] oh damn, I hate dubstep [15:02] no [15:02] i don't hear anything [15:03] ralsina: new alsa probably broke the world [15:03] hmm [15:03] oh, actually my problem might be simple [15:03] now i can hear [15:04] nessita, the webclient is only showing errors via the exception type. The http code is not exported (yet, let me know if you need it). [15:13] nessita, system's u1sdtool hangs talking to trunk's SD [15:14] (it hungs after doing the job, though) [15:16] nessita, mmm... u1sdtool -q doesn't work [15:23] all: https://bugs.launchpad.net/ubuntuone-client/+bugs?field.tag=u1-proxy [15:27] nessita, btw, ctrl-c doesn't work anymore to cut a running client (started from terminal), I need to ctrl-\ it [15:51] facundobatista, that is on linux or windows? [15:51] facundobatista, please tell me is on linux, otherwise there is a stupid thread being evil.. [15:52] mandel, linux [15:53] hmm [15:53] facundobatista, These aren't the droids you're looking for. [15:53] thisfred: now account loosks more fun, doesnt it? ;-) [15:53] mandel: the ^C issue is some weirdness with new glib i think, combined with twisted. [15:54] facundobatista: ^^ [15:54] dobey, hm.. beautiful.. [15:54] or maybe not even with twisted [15:54] facundobatista: yes [15:54] iirc, sso was doing the same, without twisted [15:54] but oddly, ^C works fine for me for the gwibber bits [15:54] so basically it's a "wtf" at the moment :-/ [15:56] dobey, and u1sdtool not finishing? [15:57] i don't know that it's the same issue [15:57] u1sdtool should be finishing [15:58] ralsina: hehe, I was just thinking... accounts maybe, u1db definitely ;) [15:58] * thisfred updated the desktop+ channel picture in mumble [16:00] thisfred: best thing would be to be the guy that does the 1st half of every project. But that's futures, right? ;-) [16:00] ralsina: yeah, they have a sweet gig === alecu is now known as alecu_lunch [16:00] can you guys browse LP urls? [16:01] mine is connecting indefinitely [16:01] nessita: wfm [16:03] hum, firefox is not loadin any page for me [16:03] but I do have dns [16:03] @ping [16:03] pong === facundobatista_ is now known as facundobatista [16:09] ralsina: so alecu said these bugs need FFE, right? https://bugs.launchpad.net/ubuntuone-client/+bugs?field.tag=u1-proxy [16:09] joshuahoover: yes [16:10] ralsina, nessita: and then we for sure need a FFE for the wizard and maybe 1 more for qt sso? did i miss any? [16:10] ok, i need to get lunch [16:11] joshuahoover: is not for the qt sso but for being able to spawn UI from the sso service. I'm searching the bug but firefox is not working [16:11] but i am not sure we are all on the same page about what various freezes actually mean :) [16:11] dobey: heh, always a possibility...enjoy your lunch, we can chat about this later [16:11] thanks :) [16:12] bbiab [16:20] I will have lunch as well [16:22] joshuahoover: the bug is https://bugs.launchpad.net/ubuntu-sso-client/+bug/919330 [16:22] Launchpad bug 919330 in ubuntu-sso-client "Execute UI as a separated process, not inside the same mainloop (affects: 1) (heat: 6)" [High,In progress] [16:22] nessita: thanks! [16:24] prego [16:26] alecu_lunch, let me know when you are around :) [16:31] ok, lunchtime here as well [16:31] beb [16:31] brb [16:51] ralsina, dobey I'm off to get some fresh air and an aspirin [17:02] facundobatista: I approved your branch (the first one), but I noticed that it got test failures on tarmac [17:02] facundobatista: can those be related to the latest changes you addeD? [17:03] nessita, I rebooted three times, since that... which branch? [17:04] facundobatista: the one that fixes the repr() [17:05] nessita, I'm off to walk the dog and EOD in theory but I'll be working later to night so that tom we have a full day for reviews.. [17:06] nessita, I'll be back close to your EOD (got rugby and i really need to get off some stressing hurting others) [17:06] mandel: ack! [17:06] nessita, anything you need send me an email :) [17:07] sure [17:07] nessita, or tweet ;) [17:07] nessita, >, 'HQ_HASH_ERROR', mdid='mdid')> [17:08] nessita, GIReactor is in the mix, there, in another test... I don't think it's the repr() [17:08] facundobatista: yes, already saw that, was wondering if it may have something to do with your latest branches [17:08] facundobatista: right, I was not considering repr() to be the caused there [17:18] facundobatista: so, can any of the changes you have been landing these days cause that? [17:21] nessita, no [17:22] nessita, hash_queue wasn't touched at all, directly nor indirectly [17:22] ok [17:30] ok, unity crashed === alecu_lunch is now known as alecu [17:32] mandel, ping [17:34] Chipaca: in the remote video scope, all the strings are coming from the server. Could a "lang" parameter be added for future-proofing i18n? [17:35] ralsina: I've thought about that, and am on the fence [17:35] Chipaca: ok. [17:36] ralsina: the way I see this evolving, you'll get results in the main language of the place you're search *from* [17:36] well bugger. [17:36] dobey: bugger what? [17:36] Chipaca: everyone hates when google does that, right? [17:36] Chipaca: the nvidia fiasco [17:36] Chipaca: remember budapest and everyone getting google in hungarian [17:36] dobey: heh [17:36] ralsina: well, but the alternative is serving us-only content to the whole world, for example [17:37] Chipaca: it's hard to enable the second monitor that just popped up on your doorstep, when all the tools dump core :( [17:37] Chipaca: google solves taht by having two parameters, a lang and a locale [17:37] lang is the language, locale is where you are [17:37] dobey: i'd feel for you, if my notebook with intel video had been able to use xrandr at any point during P without the panel going dead [17:38] ralsina: yep. [17:38] ralsina: where would you get the lang from, ie what would it look like? [17:38] heh [17:38] Chipaca: from $LANG and look like es_ES or en_US [17:39] Chipaca: and geoip for the location [17:39] ralsina: the location i'll be geoip'ing on the server, of course [17:39] I am not sure what we would be translating though. Genres? [17:40] ralsina: are you sure dbus-fired services like this have an accurate LANG? [17:40] ralsina: not in this version of the protocol, no [17:40] Chipaca: LANG should be set long before dbus is started, but I have not checked [17:40] (i.e. not with urls that start with /v0/) [17:40] right === dduffey is now known as dduffey_afk [17:41] ralsina: and, tbh, genres is going to happen before locale-specific content [17:41] ralsina: so you know what? forget it. no language. [17:41] What language? ;-) === dduffey_afk is now known as dduffey [17:41] ralsina: french [17:41] everything will be in french [17:41] I could use some fries [17:41] or neederlands [17:43] ralsina: the dutch are more into the fries than the french :) === dduffey is now known as dduffey_afk [17:44] in other news, http://copyranter.blogspot.com/2012/02/abandoned-hobbiton-from-lord-of-rings.html [17:44] Chipaca: they are called french fries because a guy called French sold them. Maybe he was dutch! :-) [17:44] Cobblers [17:44]