mandel_morning all!07:18
JamesTaitGood morning all! :-D08:42
ryecompiz desaturated firefox window during high i/o and dimmed in. Upon i/o load drop it undimmed the window but forgot to draw colors09:38
ryenow I have black and white firefox09:38
=== yofel is now known as Guest57033
gatoxgood morning!11:04
alecuI think I forgot to say "hello" today.12:09
alecuthere you go.12:09
alecugatox: ping!12:09
gatoxalecu, pong12:10
gatoxand hello12:10
alecuhi gatox!12:10
alecugatox, I heard that you were cursing twisted before leaving on friday...12:10
alecugatox: is there anything I can give a hand on?12:10
gatoxalecu, yes, i don't know exactly if the problem is twisted, i "finish" refactoring everything and now everything is working except the tests that involved: _perform_operations function..... which is exactly the same..... so i'm trying to figure that out12:11
gatoxbut i like to curse twisted :P jeje12:11
alecugatox: if you want me to take a look, please give me an url to the branch.12:12
gatoxalecu, ok, just a sec..... i'll upload my changes12:13
gatoxalecu, here: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin4-fsevents12:16
gatoxalecu, the problem is running tests/platform/filesystem_notifications/test_windows.py12:17
gatoxalecu, but you can test it with u1trial and run just test_dir_create for example, which is one of the tests that fail12:17
gatoxfor what i see, it looks like that the event is not being generated.... so i'm debugging that12:18
alecugatox: congrats on the release, btw!12:20
gatoxalecu, thanks!! :D i truly believe that is an awesome version with really cool features!12:21
alecugatox: and you are quite the salesman :-)12:21
gatoxalecu, jejejeje12:21
ralsinagood morning!12:34
gatoxralsina, hi12:34
mandel_gatox, hello :)12:35
mandel_gatox, you own me reviews :P12:35
gatoxmandel_, hi.... looking....12:35
mandel_gatox, hehe12:36
mandel_gatox, how was fgirday?12:36
ralsinay felicitaciones mias tambien  gatox!12:36
mandel_uh, friday?12:36
mandel_ralsina, what? what?12:36
mandel_ralsina, why 'felicitations'?12:36
ralsinamandel_: Ninja release12:36
mandel_ralsina, ohh gatox kudos!12:37
gatoxralsina, gracias! :D12:37
gatoxmandel_, thx!12:37
mandel_gatox, does it have vim support yet?12:37
gatoxmandel_, friday wass lot of tests refactoring.... and asking for tests in reviews (like a bad ass)12:37
gatoxmandel_, ¬¬ jeje12:37
mandel_ralsina, did we get something to try on friday? package or something of the kind12:38
ralsinamandel_: not before I EODd12:38
ralsinamandel_: we'll have to wait for mmcc12:38
mandel_ralsina, we need to talk about the fart that the shebang does not longer uses the env python but the one from /usr/bin/python12:39
ralsinamandel_: hope that was "fact"12:39
mandel_ralsina, which means that right now we cannot launch the diff processes correctly from trunk if you followed the instructs to set the machine12:39
mandel_ralsina, yes, sorry, stupid autocomplete from mac os x...12:39
ralsinamandel_: that was because of ubuntu packaging guidelines12:39
mandel_ralsina, but fart sounds funny enough :P12:40
ralsinamandel_: so, the only alternative I can think of is to patch the shebang on install, which is fragile12:40
mandel_ralsina, yes, I wonder how are we suppose to deal with that with py2app12:40
ralsinamandel_: patching it before packaging?12:40
mandel_ralsina, I was thinking of making the code that creates the new process to call it with the correct python directly12:41
mandel_ralsina, like, python ubuntu-sso-login instead of ubuntu-sso-login12:41
ralsinamandel_: why not. On windows brian had once done something like it IIRC12:41
ralsinapatching the command line to insert a python12:41
mandel_ralsina, seems like a little nicer that patching the shebang al time, right?12:42
ralsinamandel_: also, after py2app that doesn't matter because we have .app files that don't have a shebang, right?12:42
ralsinamandel_: also, objectives12:42
mandel_ralsina, no idea, is something I wanted you to know while I'm not 'here'12:43
ralsinamandel_: haha12:43
mandel_ralsina, fuuuu objectives, I forgot!12:43
alecugatox: perhaps I'm not looking right, but it seems that -4 is over 4k lines long...12:44
gatoxohhhhhhhhhh god....... what happend here!!!!112:44
gatoxit tooks common as a new file..... i did bzr mv12:45
mandel_gatox, was it present in the previous branch?12:45
mandel_gatox, http://wiki.bazaar.canonical.com/BzrPipeline12:46
gatoxmandel_, no.... but and in this branch i did a bzr mv to that file12:46
gatoxalecu, i'll try to revert and do a bzr mv again12:46
mandel_gatox, so if it was not present in the previous file it makes sense that the mv is like a new file12:46
mandel_gatox, try bzr diff --old lp:gatox-darwin-3-branch12:47
gatoxmandel_, but bzr mv, does a rename, so you don't see this kind of changes, i've use it a lot of times12:49
gatoxalecu, i'll try to fix the tests first.... and then fight with bzr12:49
mandel_gatox, yes, unless the file was not present, that is, bzr add and later bzr mv is a simple bzr add in the diff12:53
mandel_gatox, that is why I'm saying it would be nice to know what bzr thinks it happened12:53
gatoxmandel_, now it says just added.... so, i'll check doing add and the mv.... but i never did it that way..... :S12:54
gatoxusually it works just with mv without the file being there12:54
mandel_gatox, well, maybe there was a problem in the merge12:54
mandel_gatox, lets hope you don't have to divide it again … :(12:55
gatoxanyway..... i'm going to fix the tests that i don't know why they are failing.... and then clean up the diff12:55
mandel_gatox, alecu, let me know if you need reviews, I'll be 'available'13:03
gatoxmandel_, wow... that pycon should be really boring jejeje13:03
mandel_gatox, no, is just that I like to code while I listen, also, I'm easily distracted :)13:04
ralsinamandel_: if you run into a guy called Kay Hayen, say I said hi13:12
mandel_ralsina, ok, will do13:12
=== mhall119_ is now known as mhall119
ralsinaEveryone, your objectives are approved, please countersign ASAP13:28
gatoxralsina, ack13:29
alecuralsina: already done13:29
dobeyralsina: done. and disturbed by the grammar on the screen after doing so :)13:31
ralsinadobey: ha13:32
ralsinadobey: allhands is much nicer in the original klingon13:32
dobey"can't be anymore updated"13:32
joshuahooverralsina: any update on bug #1017019? we have a workaround but do we have a new build in the works to fix this?13:32
ubot5Launchpad bug 1017019 in Ubuntu One Windows Installer "3.0.2 not run on XP and Vista" [High,Confirmed] https://launchpad.net/bugs/101701913:32
ralsinajoshuahoover: we have an updated installer uploaded already :-)13:33
dobeywell, i /am/ using th_QS as my language13:33
ralsinajoshuahoover: should just redownload (since if you have the problem by definition you can't autoupdate)13:33
joshuahooverralsina: ok, the bug never showed any progress/status update, thus my question :)13:34
ralsinajoshuahoover: adding a comment and marking fixed now13:34
ralsinajoshuahoover: we tend to rely on launchpad to notice things and this was not done through launchpad ;-)13:34
dobeyralsina: does installing the new one actually remove the old un-needed files, or does uninstall happen first?13:34
ralsinadobey: ooooooh good question. It's supposed to uninstall first.13:35
joshuahooverdobey: good question13:35
dobeyheh. i'm pretty sure lp will never notice things about windows :)13:35
ralsinadobey: but I don't think we actually *tried* it.13:35
dobeysomeone should try it :)13:35
=== Guest57033 is now known as yofel
ryealecu: do you happen to know if it is possible to find out whether SD is currently hashing something via some sort of api, not by peeking at the log file?13:50
ryejoshuahoover: so, re: cpu usage/ecryptfs - ecryptfs access is itself slower than raw fs access but we are writing the compressed file to /tmp at the speed of disk (if the CPU is fast enough) thus preventing any other I/O. I am now compressing a 20Gb file and can only IRC only because I am on a remote server13:52
joshuahooverrye: ah, ok13:52
ralsinarye: would that be alleviated if we ionice'd and nice'd our processes?13:54
=== zyga is now known as zyga-food
ryeralsina: ioniced to 2 (best effort) and it does not appear to change much, there are spikes in activity though, renicing to 19 makes the computer still be hot but no noticeable delays, so my machine is disk bound, if there is enough to write - all processes will wait until that gets written14:06
dobeyrye: there should be events for hashing things in the event_queue which should be monitorable internally; and i think there's a way to get them over dbus as well, but enabling that will result in dbus getting flooded with sd events stuff, iirc14:10
diogobaederHi, guys, I'm new here. About the Windows installer, I'm not sure if this is what you needed, but I downloaded the installer from a Windows a VM and installed it successfully. No problems so far, files synced normally.14:11
ryediogobaeder: hello!14:12
diogobaederrye: hi :-)14:12
dobeywow. my hard drives are surprisingly cool at the moment14:12
ryeCPU is at +45, compressing the file14:13
ralsinadiogobaeder: thanks, the thing is,we had a broken installer before. And welcome aboard :-)14:13
diogobaederralsina: nice, if it was broken, now it's not anymore. At least in Windows XP.14:14
ryejoshuahoover: also, compressing 20Gb file took around 20 minutes, the user claimed that it took several hours, it looks like there is a real issue with the hard drive, nevermind14:16
ryeralsina: you deleted the .dlls ?14:16
ralsinarye: the new package doesn't have them14:17
ralsinarye: and when we build it again, it doesn't include them14:17
ralsinarye: so how they got there is a bit of a mistery14:17
ryeralsina: uhm, why is it required to uninstall the previous version before installing the new one?14:18
=== zyga-food is now known as zyga
alecurye: btw: I can't find any current way to get the state of the hash queue only. Though I can see that "u1sdtool -w" waits for the hash queue to be clean (among a few other things, like all transfers done).14:34
gatoxalecu, can you please take a look at this branch? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin3-fsevents/+merge/11166614:36
gatoxalecu, i think i'm going to revert darwin4 (it hasn't any review yet), and merge carefully with what i did..... there's something really screw up with the last push14:37
alecugatox: a revert sounds reasonable14:38
alecugatox: make sure to set the MP status to "work in progress"14:38
mmcchi guys14:38
gatoxalecu, i was going to delete it..... it doesn't has any review14:39
gatoxmmcc, hi14:39
alecugatox: don't delete it, just set it to wip, then repush, then set to needs review again.14:39
gatoxalecu, ack14:39
mmcccatching up on the backlog, looks like mandel was talking about the /bin/env thing - I fixed that in this mp: https://code.launchpad.net/~mikemc/ubuntu-sso-client/fix-1018924-use-buildout-python/+merge/112813 -- which just needs some discussion apparently14:41
gatoxmmcc, hi..... i leave yesterday and you were fixing your branches.... are they ready for re-review?14:43
mmcc(that was the bit about using 'python' when launching subprocesses when running from trunk because they don't have 'env' in their shebang anymore)14:43
mmccgatox, let me check. I have a lot of merges laying around...14:44
mmccand gatox, dobey disagreed with your comment on https://code.launchpad.net/~mikemc/ubuntuone-client/fix-1018125-darwin-no-gireactor/+merge/112702 - instead of installing gireactor only for platform == 'linux2' , he liked it the original way14:45
alecugatox: and by "yesterday" you probably meant "friday"14:45
gatoxalecu, yes..... that :P14:45
mmccgatox, yes I added the tests you suggested here: https://code.launchpad.net/~mikemc/ubuntu-sso-client/fix-992593-backend-path-darwin-pkgd/+merge/11270914:46
gatoxmmcc, ok with the [darwin, win32]14:46
mmccand as a bonus, made it do something saner when running from source on windows - use the sso client in the source tree not the installed one14:47
dobeymmcc: i also had a concern on your other branch14:47
gatoxalecu, now darwing4 is 886 lines again jeej14:48
mmccdobey, yes I'm looking at that now. I understand your comment, just thinking through how we'd make it independent again.14:48
alecugatox: awesome14:49
gatoxalecu, would it be too painful for you to review/merge this and do the refactor in the next branch? ..... i accept NO obviusly14:49
alecugatox: in -3: what's the issue with "IN_MODIFY as in_modify"14:50
gatoxalecu, flakes14:50
alecugatox: how is it complaining?14:50
alecugatox: and why?14:50
gatoxalecu, it says that IN_MODIFY is not being used, and it's right, but as we have windows and darwin, using common, and we have all the rest of the IN_SOMETHING in common, i though it would be best to have everything there, in other case, windows (for example), will need to use IN_CREATE for common (as the rest)..... except for IN_MODIFY which has to be imported locally.....14:52
gatoxi can change it if you want, i just thought it would be best to have everything in one place14:52
alecugatox: oh, I see.14:53
gatoxalecu, what do you prefer?14:53
alecugatox: this is a bit ugly... but after hearing the issue, it seems like a good solution.14:54
gatoxalecu, yes.... that the thing with flakes, that we can't ignore issues14:54
alecuyup, I hate that. But not as much as I hate pylint :-)14:55
gatoxjejej agree14:55
gatoxalecu, i'll try to send a patch to pyflakes for that.... i already fix that in ninja14:55
alecugatox: is this fixed in -4 ?14:56
alecu# TODO: Implement this decorators to fix some encoding issues in darwin14:56
mmccdobey - now that I'm looking, get_activation_cmdline in SSO is called from platform.ipc in client, but it really doesn't need to be. it just uses it to get a path that it uses to create an ActivationConfig, which takes the cmdline as a param. So I can move the part of get_activation_cmdline that needs to know about client into client and make the SSO branch independent again pretty cleanly14:56
gatoxalecu, oh yes.... i always forgot about that... i thought mandel was working on that... but i don't remember his answer.....14:57
gatoxwe already had this conversation14:57
gatoxbut i can't remember the outcome14:57
dobeymmcc: right, i wasn't sure why you put that stuff in sso. thanks :)14:58
alecugatox: I can't either. Two days of weekend is too much!14:58
alecugatox: let's shorten it!14:58
gatoxalecu, jejeje agree.... i keep coding during the weekend, so it doesn't feel to much like a weekend :P14:59
gatoxalecu, i'll ask mandel via pm on twitter14:59
mmccdobey, I didn't put it in there, I just updated it without thinking too hard about where it should be :)14:59
dobeypotato, buffalo. :)15:00
mmccralsina, any significance to the fact that my objectives were in reverse order upon review? :)15:00
dobeymmcc: they're probably in date-due order. so if you set different dates due, then that could be why15:01
alecummcc: mine got .reversed() too15:02
alecummcc: so it's probably the allhands site doing it's funky thing.15:03
mmccI figured it was a hard-coded philosophical imperative to reconsider the order of my priorities15:03
ryetomdroid seems not to like our datetime format now :-/15:03
ralsinammcc: nah15:04
ralsinammcc: allhands just trains us to expect the unexpected15:04
ralsinammcc: like, objectives from people who don't work for me, or objectives from years before I joined the company appearing in my task list15:05
mmccralsina: nice!15:05
gatoxshall i start?15:05
alecugatox: you shall15:07
gatoxReviews, and refactor (this is taking so long!), revert a branch because of some merging problems.15:07
gatoxApply again for the tests refactoring.15:07
gatoxbriancurtin, go15:07
briancurtinDONE: finished the first of what will be several unicode branches15:07
briancurtinTODO: look at the StringIO branch for dobey's comments, propose the first unicode branch15:07
briancurtinBLOCKED: none15:07
briancurtinNOTE: i'm only here today and tomorrow. US holiday on wednesday, personal holiday the rest of the week through monday15:07
briancurtinNEXT: dobey15:07
dobeyDONE: objectives, sso 3.0.2 SRU, reviews15:07
dobeyTODO: finish 3.0.2 SRUs, poke at some bugs15:07
dobeyBLCK: None.15:07
dobeymmcc: gord15:07
dobeyerr, go15:07
mmcc DONE: dev path fixes, packaging fixes, tests, lotsa branches15:07
mmcc TODO: tweak path fixes, land fixes, send a .app to ralsina for real this time15:07
mmccBLOCK: no15:07
mmcc NEXT: thisfred15:07
thisfredDONE: objectives, bug #999029, bug #999562 TODO: bug #1019333, keep dogs from killing one another BLOCKED: no NEXT: alecu15:07
ubot5Launchpad bug 999029 in U1DB "make test suite cleaner for no c_backend_wrapper" [Low,Fix committed] https://launchpad.net/bugs/99902915:07
ubot5Launchpad bug 999562 in U1DB "retry logic on 503 " [High,Fix committed] https://launchpad.net/bugs/99956215:07
ubot5Launchpad bug 1019333 in U1DB "When InvalidGeneration is raised proceed with sync anyway" [High,Confirmed] https://launchpad.net/bugs/101933315:07
alecuDONE: objectives, reviews, some py3k reading, played around with t3k15:07
alecuTODO: techleads mumble, reviews, more t3k15:07
alecuBLOCKED: no15:07
alecu*SPECIAL NOTE*: I'll be out from friday - tuesday15:07
alecuNEXT: ralsina15:07
gorddobey: this is your three monthly reminder to stop trying to tab complete real words :P15:08
alecubtw, all: let's sing "happy happy, joy joy" for briancurtin15:08
ralsinaDONE: cmake tweaking, some reviews, objectives discussions and reviews TODO: my own objectives, maybe, cmake tweaking, canonicaladmin, askubuntu, other stuff BLOCKED: no15:08
alecubriancurtin: 27 now?15:08
mmccSO OLD15:08
thisfredspring chicken15:09
gatoxbriancurtin, hey! happy birthday! :D15:09
thisfredhappy birthday Brian15:09
ralsinabriancurtin: you are so old, you couldn't be my son. Think about it.15:09
gatoxBTW: i have the same special note than alecu! PyCamp! \o/15:09
dobeygord: i demand the feature! :)15:09
briancurtinhaha, thanks all15:09
alecuI demand the future!15:09
ralsinaI demand the facturas15:09
ralsinasorry, argentinian joke15:09
* dobey got it15:10
* mmcc googled it15:10
thisfredI demand the fracturas15:10
ralsinammcc: http://lateral.netmanagers.com.ar/tr/es/weblog/posts/BB896.html15:10
mmccooh, that *does* sound good. I'll join ralsina in demanding some pastries15:11
ralsinammcc: yes, facturas is either pastries or bills. I prefer pastries :-)15:11
ralsinaOne comment: if you haven't yet, please go to allhands again and countersign your objectives15:11
ralsinaI have no way of knowing if you did, so consider this the first last and only reminder15:12
thisfredarghh, thunderbird will not stop reminding me of the standup15:12
mmccthisfred, you need to actually stand up for it to stup15:12
* thisfred tries15:12
ralsinaat 11:50 I get a popup window, a chime from my phone and an email, and still I usually forget in the following 7 minutes15:13
dobeyralsina: second, last, and secondary?15:13
ralsinadobey: that too15:13
thisfredralsina, that's why I moved it to 5 mins in advance, which is better15:13
ralsinathisfred: will try it15:13
dobeyi just hope everyone else forgets it15:14
ralsinawe could cancel friday standups since we have the long call on thursdays15:14
ralsinabut I would feel like I forgot something all weekend15:14
thisfredThe objective sheet can't be anymore updated15:14
ralsinathisfred: not can it15:15
thisfredwhich is to say: I countersigned15:15
dobeythe standup can't be anymore stood up15:15
mmccit is so updated, it can't be any more updated. completely updated15:15
ralsinadobey: you can't standup when you are standing, so yes15:15
thisfredallhand web ui developed by yoda, it is15:15
ralsinadobey: unless you have extra fold-out legs or something15:15
ralsinathisfred: yoda does the copy, curious george does the coding15:16
dobeyralsina: i can stand-up the standup. you know, in the "don't show up for your date" sense :)15:16
ralsinaok, that was our daily 5 minutes of weird, EOM15:17
* briancurtin coffee, back in a few mins15:24
* gatox lunch15:31
=== gatox is now known as gatox_lunch
dobeyalright, lunch and such. bbiab15:35
* briancurtin back15:40
mmccdobey, when you get back from the such, please revisit https://code.launchpad.net/~mikemc/ubuntuone-client/fix-1018125-darwin-no-gireactor/+merge/11270215:43
ralsinaOh fun another possible turkish locale bug.15:57
ralsinaAtaturk Y you make locale strange?15:57
ryeralsina: 2012-06-28 13:42:40,690 - ubuntuone.SyncDaemon.Pb - WARNING - Could not emit signal 'on_status_changed' to <twisted.spread.pb.RemoteReference instance at 0x04357148> due to 'underlying C/C++ object has been deleted' - have you ever seen this?16:02
ralsinarye: nothing like it in a long time16:02
ralsinarye: in fact, I don't know how that would appear in the sd logs, do you have a matching exception on u1cp?16:04
ralsinaalecu: can you take http://askubuntu.com/questions/156811/proxy-settings-for-ubuntu-one-on-windows please?16:06
alecuralsina: sure16:06
ryeralsina: nope, does not look like control panel was up at that time16:08
ryeralsina: continuing poking...16:08
ralsinarye: then SSO, because that is a PyQt error :-)16:08
mmccwhoa, Qt built in only 64 minutes on this new macbook air16:09
=== salgado is now known as salgado-lunch
mmccbrew still doesn't use bottles like it claims, :\ but, SSD good!16:10
ryeralsina: does SSO subscribe to on_status_changed ? o_O16:12
ralsinarye: shouldn't16:12
ralsinarye: but hey, those are our two Qt apps :-)16:12
ryeaha, pyqt... that narrows the search, but ralsina, why is that a syncdaemon's issue that it cannot emit the signal - fire and forget I thought....16:13
ryebeuno:  GetPublicFiles(running='False', _iri="u'https://one.ubuntu.com/files/api/public_files'") failure: ('Moved Permanently to https://media.one.ubuntu.com/offline.html', None) - i guess we will want to have a 302 redirect in case of maintenance, not permanent.16:16
beunorye, yes, RT that16:16
=== zyga is now known as zyga-afk
ralsinarye: it's probably getting an exception and has a logging errback16:25
ralsinarye: so it's very pssible that this is not a problem at all16:25
ralsinarye: like "u1cp started, connected to the signal, crashed, syncdaemon gets an error"16:25
=== gatox_lunch is now known as gatox
gatoxalecu, i talked with mandel, he say i should do the decorator thing16:34
alecugatox: in what branch will you be doing it?16:35
gatoxalecu,  if you want i can remove the empty implementations of darwin3 and do it there.... or add it in a future branch16:36
gatoxi need to take a look what each decorator involves and why they should be there16:36
gatoxalecu, so? darwin3 or future?? (to know if i should return to darwin3 or keep working in the tests refactoring)16:39
alecugatox: a future branch sounds fine.17:05
gatoxalecu, ack17:05
=== salgado-lunch is now known as salgado
dobeymmcc: approved17:17
mmccarg, lots of code sharing, name sharing and comment sharing between invoking sso using credentials proxy and using ipc to invoke syncdaemon in u1-client17:27
mmccseparate code paths look similar but only actually share one function - get_activation_cmdline.17:29
* mmcc is done complaining17:29
alecugatox: ping17:33
alecugatox: one last thing regarding -317:33
alecugatox: I remember that mandel sent you some tips on how to make some tests not need "time.sleep" anymore.17:33
alecugatox: I saw that the branch has just the removal of the sleeps, but I can't find the code to wait on those events arriving.17:34
alecugatox: am I missing something?17:34
gatoxalecu, it was an unnecesary use of time.sleep17:34
alecuok. I'll run the tests and approve.17:36
gatoxalecu, great17:36
* briancurtin lunch17:38
mmccargh, can't test changes to ipc on darwin without a working filesystem_notifications :|17:41
mmccthat is, u1-client platform.ipc17:41
dobeysigh, i rebooted earlier and my audio is all screwed up again because the on-board sound isn't loading properly all the time for some reasaon17:43
* ralsina lunches17:45
dobeyand this raid enclosure is a bit more internally complex than i was expecting :-/17:49
lduroshello,  I have 4 songs that appear as "queued" for a good 30 minutes17:58
ldurosthat I just purchased17:58
ldurosit doesn't look like there are any issue with the music store in the status17:58
lduroswhen I login from the web, I don't see those songs either17:59
* briancurtin back18:05
dobeyjoshuahoover, rye: ^^ help lduros please?18:08
ldurosdobey: thx :-)18:09
joshuahooverlduros: we're working on this right now...it's a server side issue...i just had an engineer look into it about 15 min. ago and it's getting fixed18:09
ldurosjoshuahoover: ok cool :-) Thanks for the update!! :-)18:09
gatoxbrb..... need to buy some cookiesssssss18:23
mmccwondering why we are testing perspective_broker ipc on linux when we only use it for windows and darwin, and it used to be named 'windows'... I think the platform refactoring a while back missed some stuff...18:33
dobeymmcc: i think we should run tests for everything on every platform that we can18:35
mmccit used to be skipped - it had been in the test/platform/windows/ subdir but got moved out into test/platform/tools/ , and so now it gets run on linux, which only worked because the cmdline code was in sso, not client18:35
mmccdobey I can certainly test it on linux but I'd need to write code to help it find the executable for syncdaemon just so I can test that, which we'll never use18:36
dobeywell, any skipping we do i think should also be done via @skipIf() decorators or such, rather than having them in files that we ignore on some platforms18:37
dobeyhaving the N different ways to run tests is annoying18:37
elopiohey desktop team.18:40
elopioany idea what this could mean? ubuntuone.controlpanel.qt.filesyncstatus - ERROR - on_sync_status_button_clicked: backend method is None!18:40
mmccso I'll have to put skipIf(linux) on every method in test_tools, since you can only get a meaningful perpective_broker proxy for darwin and windows... ok18:40
mmccback, my emacs died. having one of those days18:46
elopiohey mmcc. Why is this bug assigned to the windows installer? bug #101914218:48
ubot5Launchpad bug 1019142 in Ubuntu One Windows Installer "setup-mac.py does not package syncdaemon" [Critical,In progress] https://launchpad.net/bugs/101914218:48
mmccelopio, the windows installer is a temporarily misnamed project. we're using it to house both mac and windows packaging tools18:48
mmccI think we're going to rename it but it's a low priority18:49
elopiommcc: ah, got it.18:49
mmccgoing to lunch now18:58
alecugatox: I'm trying to run the tests for your branch, but it seems I'm missing some bits of u1-devtools, and the tests are not running right.19:10
alecugatox: I'll take a better look later today, I need to run to kinder.19:10
gatoxalecu, ok...... if you are using the buildout19:15
gatoxalecu, you will probably need to update ubuntuone-dev-tools19:15
gatoxin the eggs folder of the buildout19:15
gatoxit happens to me sometimes19:15
gatoxi'm off for today..... see you tomorrow20:13
dobeyalecu: you might need a new buildout tree. i think briancurtin fixed it to use the trunk dev-tools now.20:22
briancurtineh, buildout kind of sucks and i couldnt make that work20:22
briancurtinit works on the assumption that you deleted the ubuntuone_dev_tools egg from the eggs folder, then did a bzr branch of ubuntuone-dev-tools yourself and keep it up to date20:23
mmccthe way I've been making it work is to just have a branch of dev-tools trunk in parts/ and just setting PYTHONPATH20:23
briancurtinthat would also work20:23
mmccit's not ideal but you can make forward progress20:24
dobeytoo bad we can't just ship updates to everyone in a PPA for that20:25
mmccdobey, I suppose we could make a homebrew 'recipe' for dev-tools20:25
dobeywell, that doesn't fix it on windows20:26
dobeybut yeah, it would be nice to have auto-updating magic for all the things20:26
mmccwell, homebrew isn't that. there's never going to be anything really nice for os x open source code20:30
mmccso I want to just avoid running ipc/test_tools (tests perspective_broker) on linux - even skipping all its tests still means it imports perpective_broker, which would then have to have a dummy implementation of get_command_line for linux just so I could skip its tests... I don't like that, so I figure, skip the whole file from the makefile. anyone disagree? dobey?20:33
mmccoh, never mind. it isn't that simple because of course there are two files named test_tools.py20:35
dobeyare the tests specific to the pb ipc?20:36
mmccyes, superficially at least - they create a perspective_broker.SyncDaemonToolProxy() in setUp20:37
mmccbut IIRC there are tests of linux SyncDaemonToolProxy. I wonder if they test the same thing20:38
dobeynot sure20:38
mmccboy, nothing will make you want to put your head through a wall faster than a crashy text editor20:40
mmccI think bzr is giving emacs vc mode heartburn or something, wtf20:40
mmccok, yes they are specific to perspective_broker. they test internal methods not used anywhere else20:48
mmcc^ answering dobey20:48
dobeytest_tools is probably the wrong name exactly then.20:49
mmccyeah, I was thinking the same thing. I could rename it test_perspective_broker and that'd let me skip it from the makefile20:49
mmccor - what if it was 'test_not_linux' so it runs on both win/darwin and the makefile only has to add one more filename to skip20:50
mmccwe're already ignoring files named test_windows and test_darwin, this is just test_windows_AND_darwing20:51
mmccblast from the past20:53
mmccok, unless anyone yells at me in the next couple minutes I'm renaming tests/platform/tools/test_tools.py to test_not_linux.py and adding that a path to exclude that in the linux makefile, since it just tests perspective_broker, we don't use that on linux, and we'd have to add unused linux-specific code just to be able to run the tests on linux.20:56
ralsinammcc: why not test_pb.py20:58
ralsinammcc: not that I much care one way or the other, just want the filenames to be descriptive20:58
mmccralsina: understood. the appeal of test_not_linux is that I add that exclusion once to the makefile and that file name might come in handy elsewhere for things that are mac+win only.20:59
thumperwho looks after file sync these days?20:59
mmccralsina: I'd say the full path to the filename is as descriptive as the test_darwin and test_windows that are already there20:59
mmccralsina: tests/platform/tools/test_$platform21:00
ralsinammcc: I say go ahead then21:00
mmccralsina: ok, done21:00
ralsinaand I am exiting stage left, will be back later for a little bit21:00
thumperbeuno: how's things?21:02
mmccwaaaitaminit - there was already a function over in platform.tools.darwin to get the path to syncdaemon - so why was platform.ipc using the function from sso !?21:03
dobeymmcc: further down the spiral?21:07
ralsinammcc: we had a bit of organic growth in that area21:08
mmccdobey, so it goes21:08
mmcc"organic growth" : http://www.roofrepair.net/layout/roofrepair/userFiles/images/left-black-mold.jpg21:10
mmccBut I kid. btw, DO NOT google for images of black mold21:10
thumperoh, why not...21:13
* thumper is tempted21:13
ralsinablack mold... good name for a metal band21:13
diogobaederor a computer virus :-P21:14
ralsinaBlack mold and their first single "don't google us!"21:14
mmccthumper: apparently sometimes it looks like a swarm of snails. gets all 3-d off the wall and starts coming for you.21:14
beunothumper, heya21:14
thumperbeuno: do you know who looks after file sync?21:15
thumperbeuno: my machine is having problems21:15
thumperbeuno: and I'd like to help debug WTF is going on21:15
dobeythumper: what sort of problems?21:16
thumperdobey: it isn't syncing, and the U1 dialog shows "File Sync starting..." forever21:17
thumperbeuno: btw, I have lots of space now, thanks :)21:17
beunothumper, yeah, dobey  is your man21:17
dobeythumper: anything in ~/.cache/ubuntuone/log/syncdaemon-exceptions.log ?21:18
thumperdobey: http://paste.ubuntu.com/1072047/21:20
excelsiorSo if I move one directory inside of another in my Ubuntu One Documents Directory, will UbuntuOne wind up with 2 copies of the files?21:21
dobeythumper: hrmm, seems like your metadata might be corrupt and syncdaemon isn't starting. there's probably a little more info in syncdaemon.log itself. should be reasonably obvious if it's exiting or not21:22
dobeyexcelsior: not if you move them; if i understand your question correctly21:22
excelsiorok, so the move will be reflected in UbuntuOne as a move as well, and it will not simply copy them?21:24
thumperdobey: well there are lots of events in the syncdaemon.log21:25
excelsiorI just don't want to wind up with redundant files21:25
thumperdobey: recently just --mark entries21:25
thumperbut nothing is syncing21:25
dobeyexcelsior: a move is a move, not a copy.21:25
dobeythumper: what does "u1sdtool -s" say the status is?21:26
excelsiorok, cool, I just guess I fundamentally don't understand how ubuntuone works. Is it based on any technology Unix guys already know about?21:26
thumperdobey: http://paste.ubuntu.com/1072055/21:27
dobeythumper: well it's not connected, so that explains the not syncing, and it probably won't connect because the metadata is corrupt and it keeps working on that21:28
dobeythumper: do u1sdtool -q and tell me if it keeps running, or stops after that21:28
thumperdobey: how do I know if it keeps running?21:29
dobeythumper: check "ps fx" output21:30
dobey"ps fx|grep ubu" even to limit it21:31
thumperdobey: seems like it has stopped21:31
dobeythumper: ok. while it isn't running do "mv ~/.local/share/ubuntuone/syncdaemon ~/.local/share/ubuntuone/syncdaemon.old"21:32
dobeythumper: then start it back up with u1sdtool -c21:32
thumperdobey: done21:33
dobeythumper: might take a few minutes for it it rescan everything, but it *should* start working now21:34
thumperdobey: ok thanks21:34
thumperI'll let you know :)21:34
mmccNow I have a windows question. there seem to be two different registry keys that might be used to get the installed path to syncdaemon. one is in the sso client in ubuntu_sso/main/windows.py - it uses the key 'path-ubuntuone-syncdaemon', which exists on my machine (i've run tests but not installed U1 on there)21:39
mmccthe second is in u1client, platform/tools/windows.py, which uses the key 'SyncDaemonInstallPath', which doesn't seem to exist21:39
mmcc(on my system)21:39
mmccwhich one is the right one?21:40
dobeyno idea21:46
dobeythe one that works, i would guess :)21:46
dobeyhrmm, just need one more upvote on askubuntu today21:47
mmccso -- currently, the one in sso is used in code that tests if we're running syncdaemon (but the command line it gives is ignored) and the one in u1client is actually used to launch the syncdaemon21:48
mmccso the one that looks like it won't work on my system is the one that actually appears to need to work when installed :/21:48
dobeywell, have you installed u1 on that system?21:49
mmccno. so I guess when it's installed, the other key gets inserted21:49
mmccbut I couldn't find that string anywhere in the source21:49
mmccin the tree, that is, I grepped everything21:49
dobeywell, there shouldn't be any registry keys existing if it's not installed21:50
mmccadding keys to the registry was part of the test setup21:50
mmccwhich reminds me, I was going to make it so that wasn't necessary21:50
mmccas a side effect of this branch.21:50
dobeytests shouldn't do anything outside the branch21:51
mmccthis epic spelunk21:51
mmccok now I remember. in the sso client tests, we have a registry file that includes (fake) paths to the sso client and the syncdaemon21:56
=== salgado is now known as salgado-afk
mmccwindows-installer/scripts/devsetup/env.bat checks for one of the test keys and installs that .reg file if the keys don't exist21:59
mmccin summary, the registry key "SyncDaemonInstallPath" looks like the real one although I can't tell how it gets set, and the other one is only for tests, and happens to be used in the real code that tcpactivation uses, but maybe it's OK if it's bogus because we don't actually use the command line it returns in that case22:01
mmccat least we don't for syncdaemon22:02
dobeysounds like a shark circling the meat22:04
mmcc...no , I don't follow22:09
mmccanyway I just need to confirm that SyncDaemonInstallPath is actually the value we use in production. anyone with a windows install listening? briancurtin ?22:10
briancurtinmmcc: i'm about to head out the door. no idea about SyncDaemonInstallPath. sounds familiar...which im sure is helpful22:10
mmccbriancurtin: heh, ok - can you tell me where to look to see what reg keys we set up when we install it?22:11
briancurtinmmcc: ubuntuone.xml in ubuntuone-windows-installer/scripts is what creates our installer. i know it sets a few things up22:12
briancurtini dont know if it sets that particular one up, though22:12
mmccOH NO it sets the ones I thought were just for testing and now my whole theory is shot22:13
mmccthanks briancurtin  :)22:13
alecummcc: do you know which branch I should be using to create a fresh buildout?22:24
mmccwhich branch of the windows-installer? I think trunk should work, but I see the instructions still say briancurtin's branch... hmmm22:25
mmccalecu, yes, the branch the instructions refer to was merged on 4/1822:27
mmccso I'll edit the instructions22:27
mmccoh i see you are22:27
alecummcc: so ubuntuone-windows-installer trunk should do?22:27
mmccyes, it should22:28
alecummcc: go ahead and change!22:28
alecugreat, thanks.22:28
mmccok, I have to run. I'll be working later, might not have internet though. see you all tomorrow22:32

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