[08:02] morning all! === urbanape` is now known as urbanape [10:44] mandel: mornings! [10:44] rye, morning ! [10:44] rye, how did the holidays go? [10:45] mandel: :-D [10:46] mandel: pretty well, thank you, refreshing :) [11:15] hmm, isnt secure by visa working as a payment method @ ubuntuone ? [11:16] used: 451% not nice :) [11:20] good morning [11:41] vds: morning, bug #989157 [11:41] Launchpad bug 989157 in Ubuntu One Servers "updown OAuth may fail with valid credentials" [High,In progress] https://launchpad.net/bugs/989157 [11:42] failure_: at the moment, no, we are working to switch to Ubuntu Pay infrastructure which supports paypal which in turn supports visa 3D security [11:42] gatox, morning! [11:42] mandel, hi [11:43] gatox, can you do give me a hand and run the stupid tests for this: lp:~mandel/ubuntu-sso-client/use-unix-domain/ in you evil win machine [11:44] gatox, there should be only one failing [11:44] mandel, ok [11:52] mandel, http://paste.ubuntu.com/964475/ [11:55] gatox, sweet, just what I wanted! [12:13] gatox, can you pull and run it again, it should pass ok now on windows (I have an issue on linux.. bloody twisted) [12:13] mandel, yes [12:16] mandel, http://paste.ubuntu.com/964521/ [12:16] rye, thx! [12:17] gatox, fuuu! [12:23] rye: ETA ? [12:26] webm0nk3y: ^ ? [12:26] also, joshuahoover ^ [12:26] rye, what is ./oauth_sign in the comment of the bug? [12:28] rye: is there a question? [12:36] gatox, I think I've fixed it, please pull and take a look (and add a +1 if you want too :) ) [12:36] ac [12:36] ack [12:37] gatox, would be nice to test sso on mac os x, linux and windows just in case but everything should wok [12:37] s/wok/work [12:37] now, lunch :) [12:37] * mandel lunch [12:37] mandel, no problem.... i'll run the tests everywhere [12:37] gatox, also IRL would be nice, I've done so, but double check [12:37] mandel, ack [12:51] webm0nk3y: i recall we were to migrate to Ubuntu Pay, the question was about ota [12:52] vds: http://people.canonical.com/~roman.yepishev/us/oauth-sign, described @ http://rtg.in.ua/blog/2012/03/upload-to-ubuntu-one-using-curl/ [12:52] rye, thx! you rock! :) [12:55] good morning! [12:58] ralsina, hi [12:58] hola gatox! === dobey_ is now known as dobey [12:58] hola [12:59] gatox: my son wants to see the Avengers movie too (big fan of the cartoons) so I may have to see it in spanish (yech) [12:59] aloha dobey [12:59] mandel, all the tests are ok (windows/linux/mac)..... but i'm not being able to login or create an account in mac [12:59] gatox: or is it too violent for a 5 year old? [12:59] ralsina, naaaaa.... it's awesome! and really funny too [12:59] gatox: and you don't mind seeing it in spanish this time? [13:00] * gatox thinks.... :P [13:00] gatox: then I'm inviting and paying for the popcorn [13:00] ralsina, i think i can live with it [13:00] jeje [13:00] ralsina, naaa.... no problem! [13:01] gatox: awesome, buying tickets [13:01] ralsina, a friend of mine will come to the cinema too [13:01] ralsina, you meet him at the pycon last yea [13:02] gatox: cool [13:02] gatox: so 4 tickets? [13:03] ralsina, yes please :D [13:03] gatox: is coming to zona norte a problem? [13:04] ralsina, i don't know anything in bsas...... but lets say no..... we are going to be just paseando :P.... so we can go everywhere [13:04] mas para pasear [13:04] gatox: cool [13:04] heh [13:04] it's been a longgggg time since i went to bsas to really enjoy being a turist :P [13:05] gatox: it's here: http://www.todoshowcase.com/index.php?option=com_showcase&view=cine&slug=showcase-norte&c=cine&Itemid=171 [13:06] ralsina, ack [13:06] brb..... need to buy some cat food across the street :P [13:10] gatox: there, 4 tickets at the DOT (closer to downtown) friday 7PM === yofel_ is now known as yofel [13:26] ralsina, awesome! [13:33] rye: ubuntu pay won't support 3dsecure - yet [13:33] vds: your personal message- was it before I added the trace of curl to the ticket or after? [13:33] joshuahoover: i know, but paypal will, i heard that [13:33] joshuahoover: right? [13:34] and paypal can be a backend for ubuntu pay === m_conley_away is now known as m_conley [13:34] rye: maybe, in a round about sort of way...you still need a paypal account and you'd need to associate your card with your paypal account [13:35] rye, after [13:35] rye, why? [13:35] dobey, ,gatox: any reviews needed? [13:36] not at the moment [13:36] ralsina, nop... working on filesystem watcher for mac here.... i'm going to be busy for a while [13:36] gatox: cool [13:36] gatox: try to get mmcc involved :-) [13:36] need to get some RAM back though, so brb. :) [13:36] ralsina, roger that [13:38] vds: ah, then i'll test this with dev env [13:43] duanedesign: joshuahoover - mint12 comes with u1 client 2.0.0 with bug #869920 - when we get it to -updates, it will work in oneiric as well as in mint 12 [13:43] Launchpad bug 869920 in ubuntuone-client (Ubuntu Oneiric) "Files in new UDFs are not uploaded due to filtering" [High,Fix committed] https://launchpad.net/bugs/869920 [13:43] rye: ah, ok, thanks [13:44] umm, verification-done, anything else needed? [13:44] rye: there are other bugs in the update [13:45] dobey: i don't see any verification-needed tags on ubuntuone package [13:48] aha [13:48] control panel bug [13:48] i see [13:48] huh? https://bugs.launchpad.net/ubuntuone-control-panel/+bugs?field.tag=verification-needed [13:49] well, protocol, sso, and the timestamp bug [13:49] i guess it uses Fix released as a filter and fix is released on one distro [13:49] gatox, ok, let me see if I can [13:53] dobey: erm... ubuntuone-control-panel does not seem to have a proposed branch attached for bug #692597 [13:53] Launchpad bug 692597 in Ubuntu One Client stable-1-6 "If the user has the date/time/timezone wrong, connection fails (Expired timestamp error)" [High,In progress] https://launchpad.net/bugs/692597 [13:53] so oneiric still fails [13:53] rye: indeed. why are you looking for control-panel? [13:54] dobey: because ubuntuone-control-panel also suffers from that bug [13:55] rye: yes, and we don't have a fix for it yet for the version in oneiric [13:55] dobey: and we don't have it in proposed, it is hard to say verification-done if syncdaemon works but control panel does not [13:55] gatox, I have no problem in login in or creating a new account [13:55] rye: well when control-panel gets SRUed, it will go back to verification-needed [13:55] dobey: oh, ok [13:55] rye: but for now, one needs to check that ubuntuone-storage-protocol, ubuntu-sso-client, and ubuntuone-client all work in that respect [13:57] dobey: so, syncdaemon works (connects -storageprotocl and publishes files - rest api in sd) [13:57] ok, i will verify it [13:57] rye: also, there is bug #825366 [13:57] Launchpad bug 825366 in ubuntuone-client (Ubuntu Oneiric) "Syncdaemon not running (ValuerError: mmap offset is greater than file size)" [Undecided,Fix committed] https://launchpad.net/bugs/825366 [13:58] mmmmmmm [13:58] mandel, weird [13:59] ralsina: btw, i will need to leave the meeting a little early today, for an appointment. about 10 mins early. [13:59] dobey: np, I'll make you talk first [13:59] ralsina: good valicert post :) [13:59] briancurtin: thanks, I am like, wtf at them. [14:00] briancurtin: and we still don't know where that certificate came from anyway [14:00] gatox, how did you launch it? [14:01] it came from cloud-based value-rich enterprise services [14:01] with extra enterprise [14:01] mandel, python bin/ubuntu-sso-login-qt --app_name="u2" (with and without) --login_only [14:03] gatox, set U1_DEBUG=1 and see what it outputs [14:03] ralsina, wtf is with our cert? [14:03] ralsina, just read the post.. [14:04] mandel: we have a user who is getting a cert from valicert on login.ubuntu.com [14:04] mandel: let me find the log [14:05] mandel: https://pastebin.canonical.com/65401/ [14:05] mandel, in the ui says: "canot import name backend" [14:06] mandel, also in the outputs [14:06] gatox, have you forgotten to set the python path? [14:06] mandel: good news are, SSL checks do what they are supposed to do. [14:07] ralsina, so is not a bug \o/ [14:07] mandel: well, we may be catching the wrong certificate [14:07] mandel: supposing there is a SSL tunnel to a router or something. But I don't see how that may happen. [14:07] mandel, pythonpath to u1-client? sso shouldn't need that or am i wrong? [14:08] hello, all! [14:08] alecu, hi [14:08] hello alecu! [14:08] gatox, try to run PYTHONPATH=. U1_DEBUG=1 python bin/ubuntu-sso-login-qt --app_name hola --login_only [14:08] alecu: How's the back? [14:08] alecu, buenas! [14:09] ralsina, a bit better. [14:09] alecu: good to hear [14:09] "a user who is getting a cert from valicert".... awesome! [14:10] ralsina, it would be interesting to know how is the network set up [14:10] I mean, not so good that it's a wrong certificate. But good to see that we are catching that. [14:10] alecu, yeah, ralsinas last post is very interesting [14:10] mandel, nop [14:10] mandel: yes [14:10] gatox: no por llevarte la contraria ;-) [14:11] :P [14:11] ralsina, jeje estaba respondiendo otra cosa yo [14:11] jeje [14:11] mandel: the post is not really about our problem but about the state of CAs but, really, yesterday was a pile of weirdness. [14:11] rmcbride has a VM where SD fails auth and u1cp doesn't [14:11] mandel, it has the same problem :S [14:12] ralsina: the "share this post" box on your blog, covers things [14:12] gatox, pastebin please :) [14:12] dobey: make the window wider [14:12] dobey: but yes, known bug :-/ [14:12] ralsina, you have to pay dobey as a QA for you css ;) [14:12] mandel: I need to pay someone who knows CSS for my CSS ;-) [14:12] mandel, http://paste.ubuntu.com/964720/ [14:13] gatox_mac, line 39 of what you just gave me [14:14] mandel, mmm it copy the wrong thing…. i'll show you the other paste… wait a sec [14:14] gatox_mac, ok :) [14:15] ralsina, "a VM where SD fails auth and u1cp doesn't" <- SD fails with SSL issues? [14:15] alecu: no, AUTH_FAILED [14:16] mandel, http://paste.ubuntu.com/964725/ [14:16] ralsina, rmcbride: does that VM run windows? has it a proxy configured? perhaps the time is skewed? [14:17] gatox_mac, can you do a bzr st or even branch from 0 lp:~mandel/ubuntu-sso-client/use-unix-domain ? [14:18] mandel, ok [14:18] gatox_mac, I'm doing the same just to be sure [14:18] alecu, mandel: a windows user now (the other one was ubuntu): https://pastebin.canonical.com/65460/ [14:18] alecu, mandel: same certificate issue [14:19] ralsina, so we found the problem!!! hurray! although, this is with qt 4.7.1 right, with 4.8 we get no cert details, is that correct? [14:20] gatox, mmcc_away, you guys fancy a catch up chat over mumble after the team meeting? [14:21] "Valicert SecureTransport is an enterprise file transfer system -- leading the industry in security, reliability, performance, scalability, and extensibility" [14:21] that's just awesome. [14:22] http://en.wikipedia.org/wiki/Tumbleweed_Communications [14:22] mandel: this is 4.7.1 yes [14:22] "Tumbleweed Communications merged with Axway in 2008" [14:22] mandel, yep [14:23] alecu: yes, it's an ancient certificate probably [14:23] gatox, I have branched from lp and works ok [14:23] alecu, got the mac mini around? [14:24] ralsina, not only that: they make a security suite used by big enterprises and governments. [14:24] ralsina, """"Tumbleweed products were used to block security threats, protect information, and conduct business online. Tumbleweed provided solutions for inbound and outbound email protection, secure file routing, and identity validation that allow organizations to conduct business over the Internet. Tumbleweed offered these solutions in three product suites: MailGate, SecureTransport, and Validation Authority. MailGate provides protection against spam, v [14:24] iruses, and attacks, and enables policy-based message filtering, encryption, and routing. SecureTransport enables customers to safely exchange large files and transactions without proprietary software. Validation Authority determines the validity of digital certificates. [14:24] """" [14:24] mandel, i don't know why.... but it's not working here :S [14:25] ralsina, so, it's very likely that this is happening in an environment where the internet is filtered with this weird security suite. [14:25] alecu: and they licensed stuf to trendmicro, so it may very well be some fancy gateway thing [14:25] mandel, are you using the environment from the buildout? [14:25] alecu: right [14:25] gatox, no [14:25] mandel, that might be [14:25] hm.. [14:25] gatox, but it did work for you last time, right? [14:26] mandel, never create and account.... just get the ui responsive [14:26] gatox, hm.. the deal is, that import error is complaining that it cannot find sso.backed which is there, right? [14:27] mandel, yes..... [14:27] gatox, so the problem is with your python path somehow [14:27] gatox, get in the term anc check if you can import it :) [14:27] gatox, we might find some info that way [14:27] mandel, which will be the import? [14:28] exactly [14:28] gatox, let me check [14:33] gatox, can you paste the logs from sso-login rather than just the ui ones? [14:36] mandel, i get this: http://paste.ubuntu.com/964760/ when i try to execute ubuntu-sso-login [14:37] gatox_mac, launc activity monitor and kill the python process that is already executing [14:38] mandel, done [14:38] gatox_mac, and try again :) === mmcc_away is now known as mmcc [14:39] mandel, same problem [14:39] mandel, let me check [14:39] gatox, what do you mean, already started error? [14:39] gatox_mac, ^ [14:40] gatox_mac, that does not make sense unless you do have another process [14:40] mandel, now is running… it was blocked somewhere [14:41] mandel, aja [14:42] mandel, ralsina: I've found the issue with the certificates. http://ubuntuone.com/6uWF1LvwTWpgnqDQhZzTTF [14:42] mandel, http://paste.ubuntu.com/964772/ this is what i get from sso-cliemt [14:42] ralsina, the certificate for one.ubuntu.com is signed by godaddy. And in very old times, the godaddy certificate was signed by valicert. [14:42] mandel, it seems to be a problem with qtnetwork and unicode.. the last part of the paste [14:43] ralsina, like in the unpatched ie6 on xp sp0 of the screenshot. [14:43] alecu: wtf??? [14:43] ralsina, and it seems that new qt is hating being run on old OS where that valicert certificate is still there. [14:44] ralsina, does that make sense? [14:44] alecu: maybe, I am a bit out of my depth [14:44] alecu: but we got this on UBuntu p [14:44] ralsina, I'll try running the installer in my VM with XP and no service packs. [14:44] gatox_mac, which line? [14:45] ralsina, but not so much as we get it on windows, right? [14:45] mandel, last one 109 [14:45] ralsina, are we sure it's P? perhaps it's somebody running nightlies on something older. [14:45] alecu: yes, on ubuntu, just a couple of users [14:45] alecu: was P, the nightlies made it show the error [14:45] mandel, gatox: mumble after meeting = AOK [14:46] ralsina, hmmm. Ok. [14:46] gatox_mac, so, funny thing is, that is there due to an error getting the timestap, right? [14:47] gatox_mac, can you add a print in line 202 of the qtnetwork webclient implementation and show the error and the content [14:47] mandel, ok [14:47] gatox_mac, I wonder why are you getting an error there in the handle_finish [14:48] mmcc, superb, I have a number of bugs I think you can help with (starting with some integration with the OS issues) :) [14:48] alecu, joshuahoover: one possible solution is to install the valicert root cert into the browser but it's scary [14:49] mandel: cool. I also want to talk about the packaging plan, I did some poking around and some people say 'pyinstaller' is better than py2app for macosx… [14:49] alecu: there is at least one FLOSS project that did that (clementine) [14:50] mmcc, I think is something we can investigate [14:50] mmcc: we could also use pyinstaller on windows, it's supposed to be nicer/easier/less insane [14:50] ralsina, or tell them to remove old certs! [14:50] mandel: or that, maybe, yes [14:50] ralsina, mmcc, if someone is willing to do the move from py2exe to pyinstaller I'm all for it (pido no!) [14:51] mandel: which is why we never switched ;-) [14:51] mandel: some of the users with this cert issue are using 12.04 [14:52] ralsina, mandel - maybe magic will happen and getting pyinstaller to work on mac will be 90% of the work for windows [14:52] * mmcc has never seen magic happen [14:52] mmcc: well, if it happens it happens ;-) [14:52] mmcc: look, a bunny! [14:52] mmcc, oh, can I have you help testing a branch (gatox has problem with it) [14:52] mandel, jejej this is f [14:52] f? [14:53] fucked!? [14:53] mandel: of course, what platform? [14:53] mandel, f$%&ing kidding me… i added the print. and now it works… [14:53] mmcc, your platform :) [14:54] gatox_mac, interesting, remove it and add a pdb ;) [14:54] but mandel, I love all platforms equally. what branch? [14:54] well i think this VM is shot [14:54] except AIX [14:54] mmcc, can you branch lp:~mandel/ubuntu-sso-client/use-unix-domain [14:55] mandel, i remove it.. and now is working… this is weirddd [14:55] mmcc, then set up the env and do what gatox was trying here: http://paste.ubuntu.com/964772/ [14:55] mandel, so….. it's working… :P [14:55] gatox_mac, so, it could be a problem in your network when we do the ping to the servers to get the date,and we should not have that error [14:56] i'm testing the create account now [14:56] gatox_mac, I'd say, add that as a bug (the error with the encoding) so that next time we get a decent message :) [14:57] mmcc, if that works, we might be done with sso on mac except for the bugs I have for you and packaging [14:58] mandel, ok… everything is working… login and create account [14:58] call in 2'? [14:58] gatox_mac, hurray! [14:58] gatox_mac, can you add the bug for the webclient error? [14:59] gatox, and do add a +1 to the branch ;) [14:59] has anyone else used mumble on macos? It ships two .apps, and no description of which one I should use :\ [15:00] team, mumble! [15:00] mmcc, I use it in mac os x [15:00] team blue! [15:01] mmcc: on windows it does the same - one is "regular" and one is "backwards compatible" [15:02] thisfred: mumble [15:02] alecu: mumble [15:02] a [15:05] mumble! [15:05] crap, my laptop is not connecting to mumble right :-( [15:06] uh-huh, we used all the mumbles with vd$ [15:06] :-P === mpt_ is now known as mpt [15:35] joshuahoover, ralsina: do we need a new RT to get these new binaries signed, or do i use the previous one (51590) [15:36] briancurtin: i was going to send a new request since they did the last one [15:36] cool [15:36] briancurtin: yes, new one [15:37] ralsina, briancurtin: k, ticket was filed [15:38] joshuahoover, alecu, mandel: http://forums.anandtech.com/archive/index.php/t-1961729.html [15:40] joshuahoover: so there is a chance the user can just click on the control panel, search for certificates and fix it as a workaround. I don't have details for it though. [15:42] ralsina: possibly, for windows...i'm not sure how that works on ubuntu [15:42] joshuahoover: me neither [15:42] :) [15:42] joshuahoover: sorry, i got a bit distracted with other things, what do you say about ubuntu - issue confirmed there too? [15:43] rye: we're seeing this on both windows and ubuntu (though only turkish users so far on ubuntu for whatever reason): http://paste.ubuntu.com/964896/ [15:45] wow [15:51] joshuahoover: can you add me to RT 52678? it seems i dont have permission to view it in order to upload the binaries to be signed [15:51] briancurtin: i copied you on the ticket request so in theory you should be able to see it...hmmm [15:54] ok, verification done on oneiric timestmp === salgado is now known as salgado-lunch [16:08] mandel, gatox, here's what I got from the ~mandel/ubuntu-sso-client/use-unix-domain branch [16:08] mandel, gatox: http://paste.ubuntu.com/964946/ [16:08] alecu, ralsina: 109.169.67.174 [16:08] mmcc, let me take a look [16:09] mmcc, kill that guy and try: python bin/ubuntu-sso-login-qt --app_name="u2" [16:09] mmcc, that one shows ui :) [16:09] mmcc, try to execute PYTHONPATH=.:$PYTHONPATH U1_DEBUG=True python_u1 bin/ubuntu-sso-login-qt --app_name="u2" ........ and create a new account to see if it works [16:10] mmcc, and then the same plus: --login_only...... to see if you can login with the account you just created [16:11] * gatox lunch === gatox is now known as gatox_lunch [16:11] gatox, mandel: no good for the UI test: http://paste.ubuntu.com/964960/ [16:11] mmcc, python setup.py build [16:11] first [16:11] did I need to build something first? setup.py? [16:12] ayup [16:12] mmcc, yes, qt things :) [16:12] * mmcc is new to qt [16:13] OK it shows UI but gets an import error for qt4reactor. hmmm. [16:13] http://paste.ubuntu.com/964966/ [16:14] OT: does paste.ubuntu.com have an API? I'd like to write a script to autopaste [16:14] or drag-to-paste or something [16:15] mmcc, so, you need to branch lp:qt4reactor and add it to the path [16:15] mmcc, I though the buildbot did that [16:15] .. [16:16] yeah I thought so too [16:17] * ralsina wonders if he can ask his wife's cousin in Istanbul to run a nslookup [16:18] ralsina, you might not be the first one to ask him lol [16:18] mandel, indeed it is in the path for python_u1: >>> [f for f in sys.path if "qt4" in f] ====> ['/Users/mmccrack/Documents/Canonical/Source/buildout-env/scripts/devsetup/eggs/qt4reactor-1.0-py2.7.egg'] [16:19] mandel, and I can import qt4reactor from the prompt [16:19] mmcc, maybe you have to do PYTHONPATH=.:$PYTHONPATH, right? [16:20] mandel, I am doing that, but even if I don't, if I run just 'python_u1', I can import qt4reactor… [16:21] mmcc: I use pastebinit to post to paste.ubuntu.com [16:21] mmcc, are you doing python or python_u1? [16:21] thisfred: hooray, thanks. [16:21] gatox_lunch, how did you run it ^ [16:21] mmap offset file verified [16:22] mandel, I am using python_u1 to run the qt test (as in gatox's last post) and also to do my import test [16:23] mmcc, lets try with python only... I wonder what is going on.. [16:24] I wonder if gatox is still using 'python_u1' . ubuntu-sso-login has #! /usr/bin/env python . [16:25] if gatox changed his setup so he has 'python' in his path instead of 'python_u1' that might explain things [16:26] a hacky tweak (changed the ubuntu-sso-login hashbang to python_u1) got much further but I'm going to test my theory by renaming python_u1 and changing my PATH [16:27] ok, looks like i verified the remaining 3 bugs on oneiric-proposed, dobey, should I do anything else or the upload will be performed automatically? [16:28] mmcc, ok, I think we might be using python, let me know if it works [16:29] mandel btw, what's the significance of the app_name="u2", is that a specific test name or just something arbitrary [16:29] mmcc, random 100% [16:30] mandel: what's the real app_name? [16:31] mandel, neat, I get all the way to captcha and see a PIL decoder error [16:32] why is it using the x86_64 PIL egg? 64-bitness was something I noticed pyinstaller wasn't well tested with (ie, it's in SVN but not a released version) [16:32] mandel, sorry pasted: http://paste.ubuntu.com/965006/ [16:33] mmcc, so, to fix that you have to do the following, remove PIL [16:33] mmcc, then brew install libjpg [16:33] mmcc, and reinstall PIL [16:34] mmcc, easy_install does not bring the libjpg with if for whatever reason.. no idea why [16:34] mandel OK [16:35] joshuahoover: ok, if everything is fine, oneiric will get a SRU in 7 days, according to SRU policy. yay! [16:35] rye: excellent, thanks! [16:36] rye: should be fine [16:40] alecu, may I have that c code you had? [16:40] mandel, sure, I'll mail you the link [16:40] mandel, (give me a few minutes, I need to start the macmini and find it) [16:40] alecu, sure [16:41] alecu, mandel, cc me? [16:48] mmcc, I guess that is a yes from alecu :) [16:48] mmcc, any luck with the test? [16:48] yes! [16:48] yes, working OK so far. just got a confirmation email [16:49] where are we putting UI bugs? I see a few picky things already :) [16:49] as in, is there a specific tag / asignee, etc [16:50] mmcc: for which UI? :) [16:50] this is the qt SSO gui [16:51] mandel, I'm seeing a zombie python rocket in my dock, is that something you guys also see? [16:51] mmcc, yes, he is working yet I don't want him to be there :) [16:52] mmcc, that is one of the bug I wanted to give you [16:52] mandel, ok cool [16:52] then against ubuntu-sso-client, and assign to ~ubuntuone-design-ux [16:52] mmcc, and add the u1-darwin tag [16:52] mmcc, related to what dobey said [16:52] mandel dobey ack [16:53] mmcc, ok, so it is working! :) [16:54] mandel, yup but the last part of gatox's instructions - use --login-only, ubuntu-sso-login-qt doesn't recognize that… [16:54] mandel, oof but it says it should [16:54] mmcc, is --login_only [16:54] mmcc, probably a typo [16:55] mandel, thanks :) no, gatox typed right. it was a read-o [16:55] mmcc, :) [16:55] mmcc, so, here https://bugs.launchpad.net/ubuntu-sso-client/+bugs?field.tag=u1-darwin you can find a list of bugs I think you can tackle [16:56] mandel, success logging into 'u2' with --login_only [16:56] mmcc, superb :) [16:56] \o/ <--- did I do that right [16:56] lol, yes [16:56] mandel, ack on the bugs [16:57] btw, I have a list of merges ready for review. whom do I ping [16:57] mmcc, let me check the calendar [16:58] mmcc, briancurtin and ralsina have a review day, if you send me a list to my canonical mail, I'll do a review tom morning before you are up :) [16:58] mandel: I have done no code this week, so it's unlikely I will have anything [16:58] ah, ok. but since we're all up, I should ping those guys? [16:59] mandel: hopefully tomorrow I will hack [16:59] mmcc: i can review whatever's needed [16:59] mmcc, yes :) [17:00] briancurtin, ralsina: https://code.launchpad.net/~mikemc/ubuntuone-client/refactor-launcher/+merge/104461 https://code.launchpad.net/~mikemc/ubuntuone-client/cleanup-platform-constants/+merge/104206 https://code.launchpad.net/~mikemc/ubuntuone-client/refactor-platform-event-logging/+merge/104301 [17:00] mmcc: will do one or two today [17:01] mmcc: thanks, will look soon [17:01] apologies in advance for "My First Merge™" mistakes… [17:05] brb.... in 15' === gatox_lunch is now known as gatox_away [17:14] ok, EOD for me, catch you all tom! [17:15] bye mandel [17:15] * mmcc heads to lunch === mmcc is now known as mmcc_lunch [17:20] man installing updates is boring === gatox_away is now known as gatox [17:24] doctor's appointment, will be back in a couple of hours [17:27] ack [17:36] aaaargh. iphone ubuntu one files client is (again) re-uploading all of the photos in my phone [17:36] all 654 of them [17:37] urbanape, ^ [17:37] Did you move them from the target UDF? [17:37] i did nothing since yesteday when photos previous uploaded [17:37] i also don't know what UDF means [17:38] the web folder "Pictures - " [17:38] i havent touched it [17:38] Hmm… [17:38] and as I recall, you're running recent iOS version, right? 5.1+? [17:38] yes [17:40] Anything done to the app itself on the device? [17:41] nope [17:43] beuno: do you know if anything's changed with generations in the JSON responses for the Files API? [17:43] urbanape, nothing should of changed, no [17:44] there have also not been any deployments today [17:46] I know this happens to mattgriffin, because he moves the photos out of the UDF they get uploaded to, and we know we need to update the local CoreData objects to know that they've already been uploaded. [17:46] But as it stands, we do check the target resource, and if it exists (with a non-zero generation), we don't upload. [17:47] this is some sort of bug where the local CoreData cache has been blown away. [17:47] popey: do you know if you're running low on space on the device? [17:49] urbanape: ah, yes I might be [17:50] So, as a guess, you were running low on space, and the device wiped out the CoreData cache, which, technically could be rebuilt. [17:50] However, we were mistakenly relying on it to determine whether or not to upload a photo. [17:53] ralsina: ping [17:54] mandel: ping? [17:55] Chipaca, mandel is eod..... and ralsina has a doctor appointment [17:55] http://www.youtube.com/watch?feature=player_detailpage&v=UEaKX9YYHiQ#t=8s [17:55] urbanape, so there's a bug here needing a fix? [17:55] gatox: i'm getting an import error on resources_rc in ubuntu sso [17:55] gatox: any ideas? [17:55] beuno: there is. I'm currently looking to see if there's a better way of annotating those photos (besides doing a HEAD on each one, which could also work) [17:56] Chipaca, executing from sources? you probably need to execute setup.py build [17:56] but if we can eliminate it before we get to the queueing stage, uploads will also go much faster [17:56] urbanape, there's magic uploads and things like that which could be useful [17:56] gatox: build fails with something about wintypes [17:56] yeah, but if we can recognize earlier, they don't come into the equation at all. [17:56] urbanape: thanks, have deleted some old games i don't use anymore [17:56] Chipaca, linux? can you show me the trace? [17:56] popey: really sorry about that. [17:57] np, i am at the sprint so not my bandwidth [17:57] just means ages before U1 is back in sync [17:57] urbanape, and there's no way to make sure iOS doesn't nuke our DB? [17:57] gatox: i'll try, gimme a sec [17:57] Chipaca, ack [17:57] maybe a sqilite db in the sahpe of steve jobs? [17:57] We might be able to annotate the DB to be non-nukable. [17:58] gatox: http://dpaste.org/D2nih/ [17:58] urbanape, can you file the bug? [17:58] yup, will do [17:58] * gatox looking.... [17:59] gatox: i might be missing some libraries -- pointers appreciated [17:59] Chipaca, :S weird..... [18:00] gatox: where's ubuntuone-devtools? [18:00] Chipaca, which version of python? [18:00] gatox: 2.6 [18:00] Chipaca, lp:ubuntuone-dev-tools ........ that's what you mean? [18:00] yah [18:00] was trying without the last hyphen [18:01] where's dbus.mainloop.qt from? [18:02] python-qt4-dbus [18:02] ah, was missing that [18:02] Chipaca: do you have nightlies ppa enabled? [18:03] dobey: still failing with the v error, though [18:03] dobey: in lucid? [18:03] dobey: the v error comment was for gatox :) [18:03] Chipaca: yeah. build-dep still works; and some of the packages are built on lucid [18:04] * Chipaca toddles off to see which [18:04] Chipaca, the v error thing is really weird :S..... i don't know what is happening there [18:04] like dirspec and ubuntuone-dev-tools :) [18:04] sso is not yet [18:04] beuno: it's pointed out to me that that db file should just live in Documents, since it's precious, on-device created content. [18:04] because the gtk+ bits won't work there now [18:04] that should be a very easy fix [18:04] dobey: i'm branching stable-3-0 of everything :) [18:05] dobey: right, but the qt bits should [18:05] beuno, popey: https://bugs.launchpad.net/ubuntuone-ios-files/+bug/994126 [18:05] Ubuntu bug 994126 in Ubuntu One Files App for iOS "Important information stored in CoreData can be deleted leading to re-uploaded photos." [Undecided,New] [18:05] right, but ./run-tests tests everything. and fixing the packaging to not package -gtk only on lucid is not trivial :) [18:05] and with that, I need to run an errand. [18:05] dobey: yup yup [18:06] briancurtin: any light you can shed on the traceback in http://dpaste.org/D2nih/ ? [18:06] Chipaca: looking [18:06] briancurtin: (note: 2.6) === lifeless_ is now known as lifeless [18:10] thanks urbanape [18:11] If you leave off auto-uploads until I can get an interim fix that stores the db in the right place and crawls your asset library, rebuilding that correlation db, it'll save you a minor headache. [18:12] popey: stop breaking things! :) [18:13] Chipaca: very weird...which project are you running that setup.py from? [18:13] briancurtin: ubuntu-sso-client [18:13] briancurtin: the stable-3-0 branch [18:13] Chipaca: oh oh, i know what that is [18:13] Chipaca: you need newer python-distutils-extra [18:14] Chipaca: we have a newer version in nightlies for lucid, for exactly this issue :) [18:14] ahhhh [18:14] also, aaaahh [18:14] and ah. [18:14] dobey: thank you [18:14] sure [18:21] dobey: you are the second person to say that to me today ☺ [18:21] heh [18:22] popey: did you break the cake trolley? [18:26] hah [18:27] :) === mmcc_lunch is now known as mmcc_away [18:38] I was back, now I have to run to the bank, ~30 min. === pfibiger` is now known as pfibiger === mmcc_away is now known as mmcc [19:42] * alecu runs to the doctor. [19:42] bye [19:49] gatox: https://code.launchpad.net/~mikemc/ubuntuone-client/refactor-launcher/+merge/104461 [19:49] dobey, yes..... i'm just running the test in linux for that [19:49] ok :) [20:17] briancurtin: just saw that your windows tests caught something I missed because I still can't run tests on windows. are you available to help me fix my windows test setup? [20:17] mmcc: yeah what's up [20:17] ok cool, let me fire it up and see what the problem was [20:18] eod here..... see you on monday! (i'm not working tomorrow.... swap) [20:19] bye gatox, have a great weekend [20:19] briancurtin, is there a way to get more scrollback on the win command prompt? [20:20] in the top left of the cmd prompt, right click on the icon and go to properties, then layout tab. in the "screen buffer size" height box, just enter 9999 [20:22] mmcc: and do that every time because it doesn't remember (at least here) [20:22] thanks! I saw that but it didn't map to "scrollback" in my head [20:24] briancurtin, for your tests do you run with a newer devtools or just use the buildout as-is? [20:25] after the recent developments around devtools, i now have a devtools trunk that i inserted into the path. i need to figure out a way to get that into the actual buildout functionality, though [20:25] ah, ok. I just took another look at my notes from tuesday and I was doing that too... [20:26] so what i did is just bzr branch lp:ubuntuone-dev-tools $somewhere, then in the bin\python-script.py file, i inserted that path into the sys.path it creates [20:26] s/that path/$somewhere/ [20:26] you shouldnt have to do that manually, just havent been able to add it into buildout yet [20:27] right. I put it in parts/ and just updated my PYTHONPATH env var [20:29] briancurtin: ok, so I get a lot of passing tests, then a few with this error: https://pastebin.canonical.com/65502 [20:30] mmcc: ah, taht is because the path is too long. i got stumped on taht for a while. i just moved everything to C:\u1, but i believe there's an environment variable you can set to tweak path names, let me look for it [20:31] TRIAL_TEMPDIR i think [20:31] will use a different temp dir for the tests [20:32] oh [20:32] TRIAL_TEMP_DIR actually [20:32] TRIAL_TEMP_DIR [20:32] looks like it's … yep [20:35] wheee [20:36] oh man now ralsina is following me on twitter. no more dumb jokes, gotta be all professional [20:37] mmcc have you read *my* tweets [20:37] or blog [20:37] or IRC [20:37] hahaha [20:37] * ralsina can't complain about others [20:38] ralsina :) [20:38] dumb jokes > twitter professionalism [20:39] briancurtin: don't you mean dumb jokes = twitter professionalism? [20:39] let the madness begin! pushing zipped-python-files-as-a-script to /us/ [20:39] or professional twitterism [20:40] rye: awesome! [20:41] rye: a new era in friction-free deployment [20:41] ralsina: joshua and duanne will not be amused [20:41] ralsina: no, i mean i am convertin all my support scripts to zip files with python scripts in them [20:42] rye: with the trick that lets people execute them? [20:44] ralsina: well, not really, python $filename is still required [20:44] joshuahoover, duanedesign, may I ask you to download and run this - http://people.canonical.com/~roman.yepishev/us/bin/ubuntuone-diff [20:45] ralsina: but a compatibility layer for 3 releases.... it is a nightmare [20:45] rye: sure [20:45] rye: no, you can put a shebang in the zipfile [20:46] rye: http://sayspy.blogspot.com.ar/2010/03/various-ways-of-distributing-python.html [20:46] rye: it does require "a" python though [20:47] ralsina: yes, but that will require chmod +x first [20:47] joshuahoover: i really hope it is not exploding. tested on precise and oneiric [20:48] rye: i get that one file was not hashed - ~/Documents/Canonical/Diagrams/ubuntu_one_bug_workflow.dia~ [20:48] rye: you could zip that zip into a zip, then it will expand with the right permission :-) [20:49] joshuahoover: interesting. u1sdtool --info ~/Documents/Canonical/Diagrams/ubuntu_one_bug_workflow.dia~ ? [20:49] joshuahoover: i guess you have a rule to ignore this file [20:50] ralsina: and attach sfx header :) [20:51] rye: no special rules, default settings...that command gives a keyerror [20:51] rye: right :-) [20:51] did you know that our minimal requirements for ubuntu state ubuntu works on 64Mb of RAM ? [20:51] rye: that's with no ui, right? ;) [20:51] joshuahoover: with desktop [20:51] joshuahoover: https://help.ubuntu.com/12.04/installation-guide/i386/minimum-hardware-reqts.html [20:52] rye: that's the alternate install, at least [20:52] rye: the regular media doesn't boot in less than 512 [20:56] ralsina: and indeed, it does not === m_conley is now known as m_conley_away [21:10] rye: you actually went and tried it? :-) [21:18] ralsina: precise boots on my laptop with 512M [21:19] dobey: yes, 512 is the lower limit I think. Remember that the install media requires more memory than the final installed product [21:20] well i upgraded it [21:20] no install media [21:20] I maybe remember not being able to install lucid in 256MB [21:32] ralsina: sure, i have a vm farm of all supported releases... and a few other things [21:32] 14 virtual machines in kvm, 4 lxc containers [21:33] rye: restarting and will run now [21:33] alright, need to go. have a good evening all! [21:37] bye dobey! [22:03] duanedesign: thanks [22:04] anyone know a convenient command to recursively kill all *.pyc files on windows? (in ubuntuone-client, in case there's a .bat for this already) [22:04] mmcc: i dont. is there a reason you need it? [22:05] briancurtin: it'd be nice to have. for now I know which ones to remove. I was running the tests on my branch that broke for you, and they didn't break for me, because i had an old .pyc in my branch :\ [22:08] mmcc: there could probably be some way to do "run-tests.bat clean" by adding parameter handling to the batch file, but that will be a pain in the ass. using hte standard windows tools i dont even know how you'd do that [22:08] for sure it'd be completely unreadable and usually not work [22:09] briancurtin: ouch. ok, well the right way is probably to start with a clean branch in cases like mine… I'll just make a note [22:09] mmcc: you could also just turn off python's bytecode generation to never write PYCs, at least while you're running tests [22:10] set PYTHONDONTWRITEBYTECODE=1 [22:10] perhaps that could even be set in the env.bat file [22:10] briancurtin: huh. I thought there was a way to do that! cool. yeah, might be useful to do that for test sanity... [22:12] i'm heading out of here. see everyone tomorrow [22:18] what do you mean, configuring Windows updates? the tests were almost done… [22:19] * mmcc head --> desk === salgado is now known as salgado-afk [22:36] time to go. I guess I have to check on these windows tests later. === mmcc is now known as mmcc_gone