/srv/irclogs.ubuntu.com/2012/07/17/#ubuntuone.txt

wi43any ubuntuone core teamers up?04:52
JamesTaitGood morning all! :)08:37
mandelmorning all!08:50
czajkowskimornign09:58
czajkowskimandel: have a good weekend :)09:58
mandelczajkowski, so so :)10:16
mandelczajkowski, ran a lot hehehe10:16
czajkowskiI'm sure!10:19
czajkowskimandel: do you happen to know if there are any plans to make the web U1 more useful?  Ie. Under files/photos being able to sort the pictures by dates or size/name ?10:20
mandelczajkowski, yes, there are works to change the web, when, no idea10:21
czajkowskicool wont report the bug so10:21
mandelczajkowski, but I know that you can ask beuno about it10:21
mandelczajkowski, do report the bug :)10:21
mandelczajkowski, maybe something was forgotten10:21
czajkowskiI have OODLES of pictures due to phone uploading them automatically which I love, but no way to sort them now10:21
czajkowskiaquarius: ^^10:22
mandelczajkowski, maybe JamesTait can give us some info ;)10:22
czajkowskimandel: I'll end up pinging everyone now at this rate :)10:22
mandelmaybe in #canonical though..10:22
mandelczajkowski, ask in #u1-internal more info is usually found there10:22
czajkowskiah well tis a public question I dont mind waiting to have the chat here. thanks though10:23
JamesTaitczajkowski: Hi! :)10:23
czajkowskiJamesTait: HELLO!10:23
JamesTaitczajkowski: There's always something happening in the web UI. :)10:25
czajkowskican it be made nicer :)10:25
JamesTaitczajkowski: I think the focus of our attention at the moment is more performance, though.10:25
czajkowskiJamesTait: http://twitpic.com/a8lkd0/full  thats only 1/4 of a page10:26
czajkowskiI love U1 on my phone for files! it's brilliant, and I love the way I now have a back up of my pics. But not being able to make sense of all them ther on the web is a bit frustrating as I delete some on my phone and spring clean it sometimes. I need to be able to sort through the colums so I can update it there also10:27
JamesTaitczajkowski: There's an awful lot going on behind the scenes at the moment which will make the web UI more responsive, and enable us to develop more "value-add" services around people's files and other synchronised data.10:27
czajkowskigreat stuff10:28
czajkowskithanks10:28
JamesTaitLike re-enabling Facebook contact sync, for example. ;)10:28
czajkowskiam still trying to work out how to get my phone to use the pictures from Fb it used to on old phone, but not this one. Which means I'e no clue when people ring who Im talking to as useless with faces/names10:34
gatoxgood morning10:56
gatoxralsina, ping! are you here?? we have a mumble with lisettte if you can11:02
ralsinahello gatox, I am starting mumble11:13
ralsinasorry I'm late, had to go to a bar11:14
gatoxralsina, no problem!11:14
mandelralsina, early drinking ;)11:47
ralsinahola mandel!11:48
ralsinamandel: computers are driving me to drink11:48
* ralsina waits for google's driverless cars for that joke to work11:48
mandelralsina, hehehe11:50
mandelralsina, I know the feeling :)11:50
mandelralsina, do you feel able to do a review?11:51
ralsinamandel: sure11:52
mandelralsina, superb! here it is: https://code.launchpad.net/~mandel/ubuntuone-client/unify-processors11:52
ralsinamandel: looking11:53
mandelralsina, I'm so close to get these things landing today :)11:53
ralsinamandel: said the terrorist in the jet11:53
ralsinaok, bad taste11:53
mandelralsina, lol11:53
mandelralsina, my taste11:54
mandelralsina, but yeah, proposing the changes for the code that allows to pass the monitor via command line using --fs_monitor=default or --fs_monitor=daemon11:54
ralsinaTo compensate http://icanhascheezburger.files.wordpress.com/2012/07/funny-pictures-cyoot-kitteh-of-teh-day-look-at-me-not-crying-over-spilt-milk.jpg11:54
ralsinayay11:54
ralsinamandel, use items insteadof iteritems for python3 compatibility11:56
ralsinaor at least for not-more-incompatibility11:56
mandelralsina, where did I use that?11:56
ralsinamandel: line 26 of the diff11:57
mandelralsina, can you add that in the comment? I did a move from one to another so is possible old code11:57
ralsinamandel: there11:58
ralsinais there something broken in your bzr mv? ;-)11:58
ralsinaoh, not a mv, just a move-out11:59
ralsinaok, other than that it looks good on a quick code review11:59
mandelralsina, yes, I was moving part of the code out of a file to make the import nicer11:59
mandelralsina, is later used to have more than one implementation on darwin and not re-implement the notifier12:00
mandelralsina, question, items is present in python2.6? and which is the latests python we support?12:09
ralsinamandel: items is in 2.4 or so12:09
mandelgreat!12:09
ralsinamandel: it's just that it's slightly less efficient because it's not a generator12:09
mandelack12:11
mandelralsina, I'm running the tests, I'll push asap and will let you know12:12
* mandel wonders why or tests are sooo slow12:17
gatoxalecu, please let me know when you are here12:18
ralsinahave to go back to the mothership12:20
ralsinawill be back online in 15' or so12:20
mandelralsina, change pushed12:32
ralsinamandel: re-looking12:47
dobeyholas12:47
mandeldobey, buenas!12:48
mandelI'm off to have lunch :)12:49
mandelbbl12:49
* mandel lunch12:49
dobeymandel: buen provecho12:50
mandeldobey, gracias! ha este paso hablaras espanol todo el rato :)12:51
dobeyheh12:52
dobeyhopefully can get all these releases done today, and done quickly13:00
gatoxbrb... need to restart!13:12
ralsinamandel: +113:31
=== zyga is now known as zyga-afk
dobeyhrmm, dailies broken for dirspec on natty/lucid now though, because of python3 stuff13:39
alecugood day, sunshine!13:49
alecugatox: ehlo!13:49
gatoxalecu, hi13:49
gatoxalecu, do you have 5 min to mumble?13:49
alecugatox: for you sir?13:49
alecugatox: always!13:49
gatoxalecu, :D thx13:49
gatoxon mumble!13:50
* mandel back13:51
mandelralsina, thx!13:51
ralsinadobey: it has BEEN. For over a WEEK13:54
dobeyweak.13:54
ralsinahe13:59
mandelalecu, do you think you will have time to finish the review for https://code.launchpad.net/~mandel/ubuntuone-client/fsevents-daemon/+merge/11483614:08
mandelalecu, sorry for begin a PITA but is blocking some other branches :)14:09
mandelralsina, last branch regarding fsevents on darwin proposed: https://code.launchpad.net/~mandel/ubuntuone-client/daemon-options/+merge/115354 if that pass we are in a situation in which we can run u1 using the fsevents work from gatox or the daemon14:29
mandelthe default is the fsevents one but we can talk on how to integrate this with the installer work14:29
ralsinamandel: am on call (for a change) but will look at it14:29
mandelralsina, thx14:30
dobeybrb, need to run an errand.14:30
mandelbriancurtin, are you dealing with the broken test on windows?14:30
ralsinamandel: can you write the commit message in the standard format ? Like " - Added command line option to choose filesystem monitor (LP:1234567)."14:30
gatoxralsina, do you have a couple of minutes to mumble?? alecu and i are already there14:30
mandelralsina, is there a bug number?14:31
mandelralsina, I'll create a bug14:32
briancurtinmandel: didnt know they were broken. i guess i need to set a notification to watch u1-internal14:32
mandelbriancurtin, you can add one in the irc client to tell you when jenkins cries :)14:33
mandelbriancurtin, if you are not I can take care of it, seems like a unicode error.. which is weird 'cause we should have not touched anything regarding the config14:33
ralsinagatox: I just finished 90 minutes of mumble, I'll be there in 2'14:34
briancurtinmandel: either way is fine with me. i'm working on making dirspec tests work with buildout but i can switch to this unicode thing14:34
mandelbriancurtin, I'll take care of it then, it should not be too much work and I don't like to have broken tests :)14:35
alecugatox: I14:40
alecu'm back14:41
ralsinaalecu, gatox: mumble then?14:41
briancurtinme15:02
ralsiname15:02
gatoxme15:03
thisfredme15:03
dobeymeh15:04
=== zyga-afk is now known as zyga
alecume-otp15:05
mmccme15:05
mandelme15:05
ralsinabriancurtin: go15:06
briancurtinDONE: trying to make sure i get unicode correct15:06
briancurtinTODO: finish up this dirspec/buildout thing, unicode15:06
briancurtinNEXT: ralsina15:06
ralsinaDONE: manuel 1-1 call, design call, pre-mgmt-call call, mgmt call, alecu+gatox call, gatox+lisette call, tech leads call, see a theme here? design reviews, reviews TODO: more 1-1s? more reviews, etc. BLOCKED: no NOTE: starting vacation tomorrow if big boss approves , will be around anyway in a limited role NEXT gatox15:06
gatoxDONE:15:07
gatoxMumble and mumble, reviews, read menu docs.15:07
gatoxTODO:15:07
gatoxFix branch with refactor. Start working on the indicator menu.15:07
gatoxBLOCKED:15:07
gatoxNo15:07
gatoxCOMMENT:15:07
gatoxvacation from July 23 to July 3115:07
gatoxthisfred, go15:07
thisfredDONE: investigation into parser generators that work with C and Python. Gave up on that for now TODO: implement combine mapping BLOCKED: no NEXT: dobey15:07
dobeyDONE: reviews, fix up dirspec python3 branch from reviews, SRU verification, holiday15:07
dobeyTODO: releases/uploads, investigate twisted-less dev-tools, work on some reorg to drop ubuntuone-installer for good, finish SRU verification poking15:07
dobeyBLCK: None.15:07
dobeyalecu: go15:07
alecuDONE: discussed dbusmenu with charles, and then with gatox and ralsina, worked on the fixes requested in review for the upstream twisted bug that opened spurious consoles on windows15:08
alecuTODO: walkthru statusaggregator with gatox, do mac reviews, more twisted, more py3k15:08
alecuBLOCKED: no15:08
alecuNEXT: mmcc15:08
mmccDONE: reviews, symlink packaging, config file debugging15:08
mmccTODO: more reviews, finish symlink .app, mb help fix dirspec jenkins15:08
mmccBLCK: none15:08
mmccNOTE: Still want to discuss ubuntu-sso-ssl-certificate-qt launching15:08
mmccNEXT: mandel15:08
mandelDONE: Proposed all branches to finish the events work.15:08
mandelTODO: fix bug 1025694. Get reviews.15:08
mandelBLOCKED: no yet reviews are a must15:08
ubot5Launchpad bug 1025694 in Ubuntu One Client "Config test failing on windows" [Medium,In progress] https://launchpad.net/bugs/102569415:08
alecummcc: let me know if you want to mumble re: ubuntu-sso-ssl-certificate-qt15:09
mandelmmcc, what is wrong with jenkins?15:09
mmccmandel, it needs to get a new dirspec - the changes we just made in u1-client require dirspec trunk15:10
mandelmmcc, told you last night I fixed it :)15:10
alecummcc: the tl;dr about ubuntu-sso-ssl-certificate-qt is that we are not using it right now, and we won't be using it in the foreseable future.15:10
mandelmmcc, should be in the backlog, around my 10 m or something like that15:10
mmccmandel: oh, great then. I thought you said you knew how, didn't realize it was fixed already :) thanks!15:11
mandelmmcc, no problem :)15:11
mmccalecu, ok - that explains how it could be broken and no one noticed :)15:11
mandelalecu, we might as well remove it, right?15:11
ralsinammcc: that's for allowing the user to use a broken cert. We should not provide broken certs ;-)15:12
mmccalecu: well, it should be unbroken by the sso-client branch I've put together, anyway, since it used the same API as the proxy-creds dialog15:12
mmccdobey: when you get a chance, the sso-client path finding branch needs your re-review: https://code.launchpad.net/~mikemc/ubuntu-sso-client/use-dirspec-get-program-path/+merge/11469015:13
alecummcc: why "unbroken"?15:14
dobeymmcc: thanks, i thought there was one of those i needed to re-poke :)15:14
gatoxalecu, i'm lookint at statusaggregator and it seems quite simple, so maybe we don't need the walkthrough15:16
mmccalecu, the way it is now, before my fix, utils/runner/qt.py spawn_program() would look at the sso-ssl-certificat-qt exe name and see that it doesn't end in .exe, and try to run it as 'python.exe ubuntu-sso-ssl-certificate-qt'. -- this was because there was code in ubuntu_sso/__init__ to add .exe to the executable names if sys.frozen=True, but that code left out the sso-ssl name.15:17
dobeyralsina: actually, it was for allowing the user to accept a broken cert from a proxy, iirc; and we don't provide those certs :)15:17
mmccalecu if you're really curious there's more detail in my comments here: https://code.launchpad.net/~mikemc/ubuntu-sso-client/use-dirspec-get-program-path/+merge/11469015:17
ralsinadobey: even worse, we don't support proxy certs ;-)15:17
gatoxalecu, so, i'm going to count until 5.... and if you don't answer i'll assume you agree with me and have lunch...15:18
gatox:P15:18
gatoxok then! :D lunch for gatox \o/15:18
dobeyralsina: right, it's a half complete feature, so i don't think that program ever actually gets launched, currently15:18
ralsinadobey: yes. We should disable it from the bundles15:19
dobeyralsina: eh, i don't think saving the 30 bytes or whatever it is for that script is a big enough gain for the trouble to do that :)15:19
ralsinadobey: on windows it's like 4k. And on mac before we do symlinks, it's probably 150MB :-)15:20
mmccralsina: I'm doing symlinks now. still, I'll leave it out anyway15:20
dobeywell, fixing the py2app problem on mac will fix that :P15:20
* gatox lunch15:21
=== gatox is now known as gatox_lunch
gr72(gr72) Hey guys, i downloaded the ubuntu one app for android, and it keeps trying to use my google account affiliated with my android tablet. And it wont allow me to use the account I made from my browser.15:21
dobeykarni: ^^15:21
karnigr72: Hi15:22
karnigr72: What do you mean by "it keeps tring to use your google account" ?15:22
karnigr72: or, should I say, what do you mean by "it wont allow you to use.." ?15:22
karnigr72: You can log in with any account you want. U1F > Menu > Settings > Preferences > Advanced settings > Remove from U1, then log in with whatever e-mail you have registered on Ubuntu One15:23
gr72When i try to login fom the app, it automatically fills in my gmail e,ail that i have associated with my android tablet, and when i try to use the login that i regestered from my browser, it bringa up a window saying that ubuntu one doesnt support multiple accounts amd that i need to delete one15:23
gr72Ok ill try that. One se.15:23
karnigr72: remove the Ubuntu One account from your Accnts & Sync, then upgrade Ubuntu One Files15:24
karnigr72: Where have you downloaded the apk from?15:24
gr72Google play.15:24
karnigr72: You have to remove your Ubuntu One account from the Accounts & Sync first. U1F does not support multiple accounts at that time.15:26
gr72Sorry,how do i upgrade an app? And yes i've removed it fom accounts and syn15:27
gr72This is my firat android.15:27
karnigr72: If you downloaded it today from google play, you have the most recent version :)15:27
gr72Kk15:27
karnigr72: I see, good for you! :)15:27
karnigr72: I'll have to go away now for a call, but I'll be around soon. I'll answer any of your questions.15:28
gr72Thanks, i dislike apple, and android is open source. And thanks, ill try to login now.15:28
mmccdobey, did you see that failed merge? it's just those lint errors stopping it right? looks like most of them are it not finding devtools...15:30
dobeymmcc: ah crap. your branch failed to land15:30
dobeymmcc: yes, i know what the problem is. will fix it immediately after lunch15:30
mmccdobey ok cool, thanks!. let me know if you need me to do anything.15:31
dobeyalso15:31
dobeyi hate pylint15:31
mmcc+115:31
dobeypylint is broken, and we depend on a patch to logilab-astng (i think that's the right one of the cacaphony of libraries logilab provides and pylint requires), which fixes our use case, but presumably breaks other use cases.15:32
dobeyi wish we could just move everything over to pyflakes15:32
dobeyanyway15:33
dobeygoing to lunch. bbiab15:33
gr72karni: im now getting an authntication error saying that the login is unauthorized.15:33
ralsinadobey: pyflakes has its own issues. I am using it in a personal project and when it gives you a bogus warning there's no way to disable it except twisting your code to its taste15:34
karnigr72: Please go to Date & time settings of your device and make sure "Automatic" is selected, as well as time zone is properly set.15:35
karnigr72: This "unauthorized" problem is almost always related to wrong time settings.15:35
dobeyralsina: so we can get those issues fixed; and much more easily than in pylint, which is way overly complex, and we have to patch anyway for it to work for us.15:35
dobeyanyway, really lunch15:35
ralsinadobey: bon appetit!15:36
gr72I dont have an automatic optin, btw this is samsung galaxy 10.1 tab.15:36
gr72*option15:36
karnigr72: Could you make sure your tablet time is close to your computer time (hopefully it syncs with NTP)15:37
gr72I made the ubuntu one acciunt from my tablet. I dont run ubuntu at home, I run arch and qubesOS.15:38
gr72*account15:38
karniah I see15:38
karnigr72: But you do have a clock there, don't you :)15:39
karnigr72: Your tablet date/time is most probably off. That is the problem.15:39
gr72O.k. One sec.15:40
gr72karni: thanks. That worked. I had the wrong timezone and it was off by a minute.15:43
karnigr72: hah :) the minute is nothing, but wrong timezone indeed doesn't help :) awesome! thanks15:44
gr72No, thank you. Another random question. Has ubuntu one thought about joining the 'cloud' computing area? Do you think adding an 'openstack' feature to ubuntu would help you any?15:46
=== salgado is now known as salgado-lunch
karnigr7.. No.15:53
mandeldobey, ping16:06
=== beuno is now known as beuno-lunch
=== gatox_lunch is now known as gatox
=== beuno-lunch is now known as beuno
mmccI have a style question - we have a few deep module hierarchies like a/b/c/d, and in a/__init__ we do things like foo = a.b.c.d.foo -- is this just because it's nicer to import a.foo elsewhere or what? I can understand if it's something like platform.darwin.foo and we want to access it as platform.foo, but there are other cases where there's only one implementation and we still have a 'shortcut' assignment...16:40
dobeymandel: hi16:41
mandeldobey, hello!16:41
dobeymmcc: i'd say we shouldn't do that, in general. the specific platform-spcific abstraction case is special though, of ocurse16:41
mandeldobey, question, when you made the move away from ubuntu_sso.xdg you removed a call to native_path that was doing a decode('utf8') to the xdg_path, was that intended?16:42
mandeldobey, that is the reason why tests fail on windows16:43
mandelcould just be a mistake very easy to solve :)16:43
mmccack dobey, that sounds like a good rule of thumb to me...16:43
dobeymandel: yes.16:43
mandeldobey, yes intended, yes re-add it?16:44
dobeymandel: yes, intended. can you point to the specific call that was doing that, and is breaking that test, so i can look at the code?16:44
mandeldobey, sure, atm the failing tests is https://jenkins.errormessaging.com/view/Windows/job/ubuntuone-client-windows-test/93/testReport/tests.syncdaemon.test_config/UnicodePathsTestCase/test_get_config_files_path_encoding/16:45
mandeldobey, the main issue is that if you don't do the decode it will be using the default platform one which is not utf8 but mbcs on windows16:46
mandeldobey, the native_path call was removed in rev no 1262.1.116:46
dobeymandel: right, i know where the test is. i'm more interested in which line of code changed16:46
dobeyand why it's breaking that test16:46
mandeldobey, line 168 in ubuntuone/syncdaemon/config.py use to have: xdg_config_dir = native_path(xdg_config_dir)16:47
mandeldobey, you removed that line and the decode is not longer used16:48
briancurtingatox: when you got dirspec tests to run on windows...uh, what did you do and how did you do it?16:48
dobeymandel: looking16:48
gatoxbriancurtin, install python-testtools and python-setuptools with easy_install, and then execute the test16:49
mandeldobey, is not a terrible problem, it was harder to find what changed :)16:49
mmccmandel, I left you a nice long comment about your daemon-options branch. two quick fixes and some discussion that you may decide to argue should be left to a different branch :)16:49
mandeldobey, I have a bug number for the issue: bug 102569416:50
ubot5Launchpad bug 1025694 in Ubuntu One Client "Config test failing on windows" [Medium,In progress] https://launchpad.net/bugs/102569416:50
mandelmmcc, great, I'll take a look tom in my morning :)16:50
briancurtingatox: after you installed those two things, you ran it from run-tests.bat or something else?16:50
gatoxrun-tests.bat16:50
mandelmmcc, I can fix all those in that branch with no problem, will be ready to a re-review tom :)16:51
mandelmmcc, better be consistent with the imports, and I have no problem in renaming the method16:52
mmccmandel: sounds good. I'll be sure to do the re-review tomorrow, since I'm gone for 3 days starting Thursday.16:52
mandelmmcc, ok, the I'm sure it will be ready :)16:53
mmcc(although if it were an emergency I could look at things from the road over the weekend)16:53
dobeymandel: i wonder why that test didn't fail on linux16:53
mandeldobey, it has a skip :)16:55
dobey:(16:56
mandeldobey, not big deal, we found it and is a super easy fix16:56
mandeldobey, do we re-add the decode or do something else?16:57
dobeymandel: well, i am guessing the test is wrong16:57
mandeldobey, we should ask nessita AFAIK she wrote it16:57
dobeymandel: right, but that was before we had to consider python3 for anything, and could do stuff wrong and it would "work"16:58
mandeldobey, nop, blame just pointed to gatox and alecu16:58
mandeldobey, is late here, can you take over it? if not, can you write your though in the bug and I'll fix it tom16:59
dobeymandel: i'll look at it16:59
mandeldobey, ok, thx!17:00
mandeldobey, is not a major blocking bug but is a PITA to have jenkins failing due to it17:00
mandelall, EOD here, see you tom :)17:02
ralsinabye mandel!17:03
mmccbye mandel17:03
gatoxbye Madkiss17:03
gatoxoh no17:03
gatoxbye mandel17:03
Madkisswtf17:03
gatoxwrong tab-completion17:03
mandelMadkiss, gatox has tabbing diarrhea :P17:03
gatoxjeje17:03
dobeyporting ubuntuone-client to python3 is going to hurt.17:09
=== salgado-lunch is now known as salgado
mandeldobey, a lot17:13
* mandel is not here..17:13
ralsinadobey: at least we have a reasonable schedule to spread the pain17:15
briancurtinheyyyy there we go. so weird that dirspec tests fail without testtools, but you get no indication that its because of a lack of testtools. ugh17:16
ralsinadobey, alecu: can I pass that list of twisted bits?17:16
alecuralsina: +117:16
ralsinadobey, alecu: I am not sure I get what we should ask about for qtreactor17:16
=== yofel_ is now known as yofel
dobeyralsina: i'm sure there's more, but it's probably fine17:17
dobeybriancurtin: you don't?17:17
briancurtindobey: i already cleared the screen that had that error, but i believe it was an AttributeError that there is nothing called "tests"17:17
briancurtini figured i would've gotten an importError or something more obvious17:18
dobeyralsina: qt4reactor is required to run the tests. make sure to not confuse it for the qtreactor thing that's in twisted, which is simply a very thin wrapper which requires extra stuff to be installed to use, and which only works with qt317:18
alecuralsina: I think that we qtreactor would be needed to run the control panel et all tests on ubuntu17:18
ralsinaso, I ask that we get a working qt4reactor?17:18
alecuralsina: right17:18
ralsinalet me do a draft then17:18
dobeywell, we either need a working qt4reactor, or tests not using twisted/qt4reactor17:18
ralsinadobey: is the second option feasible?17:19
dobeyi think so, but i haven't had time to really look into it yet17:19
ralsinaok, then17:19
dobeythings have been a bit overly hectic this weekend, unfortunately17:21
dobeybut we are moving closer to the target at least17:21
dobeyanyway17:24
ralsinahmmm endpoints and protocol is not for the windows IPC?17:25
dobeyhttps://code.launchpad.net/~dobey/ubuntuone-client/fix-unicode-path/+merge/115398 should fix the test failing in jenkins issue17:25
ralsinaalecu, dobey ^17:25
dobeyralsina: i don't know about those. but i just realized that you didn't list twisted.names, which we do need17:25
ralsinadobey: adding names17:26
dobeyi think endpoints/protocol are probably required by trial and reactor anyway17:29
ralsinadobey: reactors we need would be SelectReactor, PollReactor, gireactor, and qt4reactor then?17:34
dobeyralsina: and anything else they depend on, yes. qt4reactor is separate and upstream doesn't really maintain it any more, though17:35
ralsinayes17:36
ralsinathe "everything they depend on" is implicit :-)17:36
ralsinadobey, alecu, facundobatista (@canonical this time) sent updated list. If you +1 it I will get back to the contractors17:37
dobeyralsina: right, but it's good to be explicit, especially in contracts.17:39
ralsinadobey: will add it17:39
dobeyi'm not sure there's a good/easy way to magically print out a dependency report though17:40
dobeyyou'd think somoene would have written such a tool for python already17:40
ralsinadobey: there isn't one I know about17:40
ralsinadobey: the curse of dynamic languages in action17:40
ralsinadobey, alecu: could either one of you get me a rough estimate of how amazingly painful it would be to migrate out of qtreactor for tests?17:41
dobeyralsina: it's so easy to do some things, that nobody wants to solve the hard problems any more, even when they're easier to solve17:41
ralsinajust in case they refuse to port it17:41
dobeyralsina: i think it's something we should do anyway, and we can start doing in parallel even if qt4reactor does get ported17:41
dobeyoh joy17:44
dobeyso i tried to backport python-testtools to lucid; but it requires python-fixtures. so i tried to back port that; but it requires python-testtools…17:44
dobeydynamic languages are great for stabbing yourself in the face, recursively.17:45
dobeyralsina: if it helps any, i do have a branch of twisted on my machine right here, that has some of this stuff semi-ported already17:46
dobeyralsina: for instance, i have "python3 trial --help" working to print out the usage listing, but plug-ins aren't being loaded, so --help-reactors and such don't work (which means it doesn't run tests or anything)17:49
ralsinadobey: there is a cornucopia of branches of twisted that portbits17:50
ralsinadobey: I think yes, we should move away from qtreactor, but I am more concerned about timeframes17:51
mmccjust caught up after being neck deep in py2app for a bit - if we need a dependency report, would the modulefinder stuff in py2app/py2exe be useful?17:54
dobeyright17:54
alecuralsina: moving away from qt4reactor sounds hard...we would have to move our sso and u1cp UI tests out of trial too.17:55
ralsinammcc: yes except that I want a list excluding what we use only on windows nd not on ubuntu17:55
mmccralsina: ok, yes, not simple to do automatically17:56
alecuralsina: porting qt4reactor to python3 sounds much more easy, as it's only one file.17:56
ralsinaalecu: specially if it's mainly for tests17:56
ralsinait should not work any worse17:57
ralsinadobey, alecu: facundo just added a few modules, doublecheck?17:57
alecuralsina: qt4reactor depends on some twisted modules, so we may ask for them if they don't commit to porting qt4reactor.17:58
alecutwisted.internet import posixbase and twisted.python.runtime import platformType, platform17:58
ralsinaI'll just ask for qt4reactor and wait for them to refuse17:59
alecuperfect.17:59
ralsinaor not, if we are lucky17:59
dobeyalecu: those modules are required for the gireactor also18:00
alecugreat then.18:00
dobeyhrmm, actuall, i guess platformType and platform from runtime might not be18:00
dobeybut i think other stuff from runtime is18:00
* gatox hates when something break for a little change and have to start debugging something that was ready....... ¬¬18:14
mmcclunchtime, inching closer and closer to working bundle with symlinks to shared code... lots of fiddly bits18:15
mandeldobey, review done and approved branch because is a no brainer18:26
dobeymandel: thanks18:27
mandeldobey, no problem, also double checked that it work in the jenkins instance :)18:27
dobeymandel: well, when i enabled the test on linux, it failed with the same error as on win :)18:27
dobeythat test should never have been skipped on linux18:28
mandeldobey, he, I wonder why it was..18:28
dobeymandel: well, see the comment in the @skipIfOS line i removed there :)18:28
mandeldobey, yes, but it did not hurt to run it, right?18:29
dobeymandel: though i don't think that makes any sense either; and dirspec uses bytes everywhere now (thanks to the py3 port exposing various unicode issues)18:29
dobeymandel: right, running it wouldn't have added any significant overhead or extra time to the tests run18:29
mandeldobey, there must be a reason why that code was added, nevertheless when ever we do the py3 move everywhere we are going to have lots of 'stupid' code like that one18:30
dobeymandel: yeah, i was going to fix it the right way, but then realized that might just break more stuff elsewhere in syncdaemon right now18:31
dobeyso the python2 way will suffice for now18:31
mandeldobey, probably, it could leak a unicode file inside the state machine and we will have a number of failures sometimes18:31
mandelif we had only chosen to use unicode from the start..18:32
ralsinamandel, dobey: consider it as a chance to fix stupid code18:34
mandeldobey, I hate pylint, skipIfOs is not used, you will need to fix that18:34
mandelralsina, he, I'm not such a positive person :P18:35
dobeymandel: ah18:35
mandeldobey, happens for not fixing lint on windows..18:36
mandeldobey, I should do something about that18:36
mandelmaybe not in this life18:36
mandelok, I'm off again, see you all tom18:36
dobeywell, u1-client isn't using pylint18:37
dobeyit's using pyflakes :)18:37
dobeyso it should be usable on windows18:37
* briancurtin lunch18:41
ralsinaI am going to take a long break now. Will be back much later. Mail me for anything!18:46
gatoxalecu, this one is ready for review: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-tests-refactoring/+merge/11328919:01
* briancurtin back19:03
dobeywow. i didn't realize askubuntu.com was so political19:12
mmccback from lunch…19:22
dobeybrb, need to run for a bit.19:42
=== daker_ is now known as daker
=== technovi- is now known as technoviking
=== zyga is now known as zyga-afk
gatoxeod for me....... see you tomorrow people!20:05
briancurtin2need to run to the pharmacy, back shortly20:15
dobeymmcc: ah, you broke a pylint disable comment20:32
mmccdobey, ah, crap. where?20:32
dobey=== modified file 'ubuntu_sso/main/windows.py'20:32
dobeymmcc: it failed on win32process and win32security imports now20:32
mmccoh I see. shoot. I thought that only applied to winreg for some reason20:33
mmccso I understand - that's the 'unable to import' error, and win32* gives it because tarmac is running the tests on linux?20:34
mmcccan't really remember if I saw that much spewage when I ran the tests myself. if not, shame on me for not noticing those :\20:36
dobeyright; pylint is running on linux, so those can't be imported20:36
mmccif I did see that much lint spew, shame on me for assuming that was normal :\20:36
dobeywell, 2 lines isn't that much20:36
dobeythe TODOs aren't errors20:36
mmccoh, I was looking at the email with 100s of lines of errors20:36
dobeyand all the other stuff from earlier was due to a missing patch in logilab-common on quantal, because a newer version got pulled into the ubuntu archive20:37
mmccaha, ok20:37
mmccI'm re-running those tests on linux again just to see what I saw20:38
mmccthis will be a nice distraction from fighting py2app20:38
dobeywell, if you're running on precise with the nightlies PPA enabled, you wouldn't have seen the 100s. you should have seen the win32* errors though20:39
mmccyes, that's what I'm afraid of - I should've seen those :\20:39
mmccin other news, looks like I have to write code to merge two zipfiles :P <-- a little easier than figuring out how to get modulegraph to give me lists of dependencies separately from building an app20:40
dobeyheh20:44
mmcczipfile's API is giving me the unicode shakes21:06
mmccZipFile.writestr to write file bytes... will it get converted somewhere?21:07
mmcchmm, well, they always open in binary mode looks like.21:08
mmccdobey: I just pushed adding the pylint disables back on that sso branch. is it clear to flip the approved switch again?21:14
dobeymmcc: yeah, after lp rescans it, and the new revision shows up on the merge21:26
thisfredso, starting tomorrow, I will be asking for reviews on u1db here, as there is only one person not in desktop+ who can do them, and we're supposed to own it. Be warned ;)21:30
* mmcc is warned21:31
thisfredthere will be blood^WC21:31
mmccthisfred: I'm happy to review that stuff, I'm interested in u1db, and I've worked with my fair share of C21:31
thisfredmmcc, awesome!21:31
thisfredyou're on my list then ;)21:31
briancurtin2i have some C experience (and C++, but we don't talk about that)21:32
thisfredmost of the team has probably done more C than I have :)21:32
briancurtin2so count me in21:32
thisfredkewl21:32
dobeyi know nossing21:35
mmccI went to the irclogs to check which server the u1db channel was on... man you guys are chatty over there21:36
dobeymmcc: both networks :)21:38
thisfredmmcc, hehe, not so much on freenode, I admit, the internal one has been quite busy, but we really should be doing our discussing in the public one21:39
dobeyhopefully the public one will pick up more, once we get a release out and uploaded to quantal21:39
thisfredwell, we could just talk there by default, rather than the internal one21:39
mmcclooks like the public one isn't logged: http://irclogs.ubuntu.com/2012/06/17/21:39
thisfredto show we;re alive :)21:39
dobeywe really should have just used this channel for public u1db stuff21:39
thisfrednah21:40
dobeyyeah. irc channels are like e-mails. the fewer i have to deal with, the better. :)21:40
thisfreddobey, we'll just send you the logs of both in one big email everyday. Happy? :P21:41
thisfredl,21:41
thisfredoops21:41
thisfredok, dog needs walking21:41
thisfredlater21:42
mmccmmmm, moudle dependency soup!21:42
dobeyhave a good evening all. i'm out too!22:07

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!