/srv/irclogs.ubuntu.com/2012/05/24/#ubuntuone.txt

mmccJust sat down to write an email and ended up working on bug 1003692. If someone wants a nice morning review, here: https://code.launchpad.net/~mikemc/ubuntu-sso-client/fix-1003692/+merge/10715905:12
ubot5Launchpad bug 1003692 in Ubuntu Single Sign On Client "ubuntu-sso-login-qt crashed with TypeError in got_state() " [Undecided,Confirmed] https://launchpad.net/bugs/100369205:12
mmccgood night05:12
Chipacammcc: you are a machine of awesome :)09:19
mandelmmcc, is that due to the changes we made for the network state changes on OS X?10:04
mandelmmcc, 'cause it used to be ok, wasn't it?10:04
mattwj2002Good morning all10:11
mattwj2002:)10:12
mattwj2002with the Windows client for ubuntu one is there anyway to get a more verbose output on the syncing/10:14
mattwj2002?10:14
mattwj2002most cloud systems like Amazon, Google Music, and even the iCloud tell you what it is doing :P10:15
mandelmattwj2002, right now it is not possible, we used to have a very verbose output but most people hated it :(10:18
mattwj2002bummer10:18
mandelmattwj2002, I think is one of those things that you can file as a bug in lp:ubuntuone-control-panel so that it gets re-evaluated and better implemented10:19
ralsinamattwj2002: we will get a medium-verbose version in a few months though10:19
mattwj2002okay10:19
mandelralsina, morning!10:19
mandelralsina, are we? cool :)10:19
mattwj2002yeah like a progress bar with one file is uploading even would be nice10:19
mattwj2002:P10:19
ralsinamandel: well, we have "windows notifications" in the roadmap somewhere10:19
mattwj2002*with what file10:19
ralsinamandel: and good morning :-)10:19
mattwj2002especially when you are trying to upload 8 GB on a 896kps uplink10:20
mattwj2002:P10:20
ralsinamattwj2002: per-file progressbars are tricky because it does multiple simultaneou uploads, but at least a progressbar showing how many files are left to sync like we have in ubuntu10:20
mandelralsina, he, that roadmap should be called an autobahns instead10:20
mattwj2002that would good then ralsina10:21
mattwj2002:)10:21
mattwj2002or like utorrent10:21
mandelralsina, I was thinking about it, sending a signal per progress of a file might do the trick, but in terms of design would be hard10:21
mattwj2002a list of files it needs to upload and progress on each10:21
mattwj2002:)10:21
ralsinamandel: a port of the current notiications is straightforward10:21
ralsinamandel: that list may be a few thousnd items long ;-)10:22
mattwj2002true10:22
mattwj2002something10:22
mandelralsina, yes.. that is what the design part is the hard things10:22
mattwj2002a hey I am actually working! :P10:22
ralsinamandel: all we need is some IPC server on u1cp and add it to the menu, we *have* a design for windows "indicator"10:22
mattwj2002what about this?10:23
mattwj2002a percentage of the total size it is uploading?10:23
mattwj2002like right now I am doing 8 GB10:23
mattwj2002I am currently at 2.7 GB10:23
ralsinamattwj2002: something like that, yes10:23
ralsinamandel: https://docs.google.com/a/canonical.com/open?id=0B4l0pd7xzubjMTRmNDdkZTUtMzE4MS00MmU0LTkxYTAtMDBkMDJmMWZjZDAy10:24
mandelralsina, with the current state of the IPC (which I consider quite robust right now) doing that in twisted should be very easy10:24
mattwj2002and a transfer rate speed would be nice too10:24
ralsinamandel: we also have u1cp's "unique app" local socket open10:24
mandelralsina, AWESOME!!!10:24
mandelralsina, I mean the doc, lisettte is great :)10:24
ralsinamandel: but yes, syncdaemon and qt loops shall not mingle10:24
ralsinamandel: that's patricia's10:25
mandelralsina, the patricia :)10:25
mandels/the/then10:25
ralsinamandel: not that lisette is not :-)10:25
ralsinamandel: there is the problem that the "progressbar on taskbar" is not accessible from PyQt and that we currently minimize to the tray, but all that's just effort needed10:26
mattwj2002oh nevermind10:26
mattwj2002I just required access to that doc10:26
ralsinamattwj2002: long story short: we would like that too10:27
mattwj2002sorry I thought that was for everyone10:27
ralsinamattwj2002: sorry, design docs are just fr canonical employees :-(10:27
mandelmattwj2002, work is progress, but looks great10:27
mattwj2002no problem10:27
mattwj2002I thought you were giving that out to the whole room10:27
mattwj2002:)10:27
mattwj2002I wasn't trying to break in honest!10:29
mattwj2002:P10:29
mattwj2002though I am curious about the new design10:29
ralsinamattwj2002: no problem, I wish we did everything in public, too :-)10:37
mattwj2002it is no problem10:39
gatoxgood morning!11:06
mandelgatox, morning!11:06
gatoxmandel, hi11:07
gatoxmandel, how is it going?11:09
mandelgatox, back to fsevents and running tests with the unify implementation of tools11:10
gatoxmandel, cool...... yesterday i was able to see how my darwin implementation receive and process some events...... but because macfsevents is running in a different thread i'm having a race condition problem :S11:11
mandelgatox, do you want me to give you a hand after my lunch11:13
mandelgatox, we can do some pair programming and see if we can get it working11:13
gatoxmandel, yes no problem..... i'm also schedule a 1-1 with alecu11:13
gatoxmandel, if you can, it would be great11:14
gatoxis a stupid thing11:14
gatoxbut between twisted and another thread..... i'm not really an expert with that11:14
mandelgatox, sure, no problem, I'll ping you after my lunch and we can take a look at what is going on11:15
gatoxmandel, thx11:18
MoscherkoboldHello everyone, can I ask something about ubuntu one?11:58
MoscherkoboldI´ve created an account and sync my laptop running ubuntu 12.04 and my phone with android, it´s working fine so far... but if I log in on the website there are no files12:00
Moscherkoboldalthough on my windows PC using the ubuntu one client I can´t see any file12:00
ryeverterok: ping12:11
ryeMoscherkobold: are you able to see the files on android phone?12:12
Moscherkoboldyes with the ubuntu one client12:13
ryeMoscherkobold: but you are not able to see the files in the web interface ,am I correct? Are you sure you are using the same account?12:13
Moscherkoboldis it possible to have two account with the same email adress?12:14
ryeMoscherkobold: it should not be possible, could you please PM me the e-mail address you are using for Ubuntu SSO and I will look up the account info12:15
Moscherkobolddone12:16
ryetechnically web interface and rest api Ubuntu One client using are looking at the same database, so if one shows entries and another does not, then there is something wrong12:16
ryeMoscherkobold: hmmm you indeed have two accounts12:17
Moscherkobold?!?12:17
Moscherkoboldhow? :)12:17
Moscherkoboldcould you delete one?12:17
ryeMoscherkobold: i am now escalating this12:18
Moscherkoboldthx for your help12:19
mandelok, lunch time forme12:35
* mandel lunch12:35
=== mpt_ is now known as mpt
ryeMoscherkobold: do you happen to remember when you signed up originally?12:39
ralsinaalecu, dobey, mandel, mmcc, gatox, briancurtin, thisfred: no team call today because I have almost no voice13:08
gatoxralsina, ouch..... ok13:08
ralsinaalso, I will leave early today and work tomorrow morning (but don't tell chipaca because I have not loaded it in the admin)13:09
thisfredkk, hope you get better sii13:09
thisfredsoon13:09
ralsinabecause tomorrow is natl holiday in argentina and monday in the US13:09
ralsinaand I don't want to keep you guys unsupervised too long ;-)13:09
ralsinathanks thisfred!13:09
thisfredwe might get into all kinds of trouble113:09
thisfredwhat's this button do?13:10
ralsinaI do feel much better but all the coughing has been ... interesting13:10
ralsinathisfred: boom!13:14
thisfredOh. I guess we didn't really need that part of the data center13:15
Moscherkoboldrye: sry dont understand your question13:25
MoscherkoboldDo you ask me for the time when i have created my account?13:26
ryeMoscherkobold: yes, if you recall that. It looks like the account has been broken before the migration to new account system13:29
Moscherkoboldah ok13:29
Moscherkoboldi created my account yesterday13:30
ryewebm0nk3`: ^13:30
ryeMoscherkobold: hm, do you recall the sequence - i.e. have you created account via android phone first, website or ubuntuone control panel?13:31
ryeand in what sequence were you performing the registrations13:31
Moscherkoboldfirst ubuntu laptop, then android phone, worked so far13:32
Moscherkoboldthen windows...13:32
ryeok, great13:32
Moscherkoboldbut I noticed some problems while signing in after the installation on my windows pc13:32
* rye tries to replicate13:32
alecugatox, still need a hand with threads?13:33
gatoxalecu, yes13:33
gatoxalecu, do you have time for a mumble?.... it can be now or later.... i was fixing another bug in the meantime13:34
ralsinaanyone needs reviews? I have time!13:35
alecugatox, let me start the mumbling machine13:35
gatoxalecu, lp:~diegosarmentero/ubuntuone-client/darwin-fsevents13:44
mandelbriancurtin, so, how can I help you with jenkins?13:53
trubborIs there anyone here that might be able to look at Ubuntu One ticket #15507?13:54
briancurtinmandel: i'm not even sure right now, actually. i'm logged into the machine and i can make the build work perfectly when i run it myself in the command prompt (using the same commands), but when jenkins runs it i get an exception that should not be happening13:54
briancurtinmandel: for example: https://jenkins.errormessaging.com/job/ubuntuone-windows-installer/8/console13:55
briancurtinthat TypeError *used* to happen when we were looking in the wrong place (a Win7 path rather than Win2008 path), but that has been changed and the function in question returns the right value13:55
mandelbriancurtin, which project is that, ubuntuone-windows-installer?13:56
briancurtinyeah13:56
briancurtini'm just making that one work on its own right now. it'll be a once-a-day build13:56
mandelbriancurtin, ok, let me see if I can give you a hand..13:57
briancurtini'm about to reboot it partially out of frustration, partially out of "that's what you do on windows"13:57
mandelbriancurtin, hehe I know the feeling13:58
mandelbriancurtin, give me some mins before we reboot it, lets see if we can reproduce it13:58
briancurtinmandel: sounds good13:58
ralsinajoshuahoover: can you give a hand to trubbor? I don't have RT access14:00
joshuahoovertrubbor: i will look at that ticket soon...need to finish up one other thing...sorry for the delay14:01
trubborjoshuahoover: Thanks. It should just be an account delete (so that I can then recreate it).14:04
karnitrubbor: Welcome back.14:04
mandelbriancurtin, I wonder.. do you know in which case the output can be none?14:06
briancurtinmandel: it would be None if it can't find the "My Documents" folder. if you import build_installer manually and call that function, you can see that it comes back14:07
briancurtinmandel: hmm...actually, i wonder if when Jenkins runs it, it doesnt have the USERPROFILE environment variable14:08
mandelbriancurtin, exactly :)14:09
mandelbriancurtin, that will explain the issues with the cp tests14:09
trubborkarni: Hey14:09
mandelbriancurtin, lets do the following, we point to a +junk branch with extra print statements (so that we don't have to go via reviews) in the jenkins project and try to build it14:09
briancurtinmandel: sounds good, i'll make a few print changes right now and push it in a minute14:10
mandelbriancurtin, we log if we have user profiles + all we can think of and lets see what happens, if there is no user profile (which I would not be surprised) we can use something else of modify the jenkins user14:10
briancurtinmandel: i've never pushed a junk branch, do you have an example?14:13
mandelbriancurtin, simple, lp:~braincurtin/+junk/verbose-exe-building14:14
briancurtineasy enough14:14
mandel:)14:14
briancurtinlp:~brian.curtin/+junk/print-userprofile14:14
mandelbriancurtin, ok, let me change the jenkins config14:16
joshuahoovertrubbor: alright, i can't do that SSO account delete but i'm getting a hold of the person who does most of those14:16
mandelbriancurtin, we are pulling from that junk, want exactly should we be executingonn jenkins?14:18
mandelbriancurtin, 'cause right now it complains because we don't have a run-tests.bat which is ok14:18
briancurtinthe rest of the config shouldn't change14:18
briancurtinhm, which project are you looking at again?14:19
mandelbriancurtin, https://jenkins.errormessaging.com/view/Windows/job/ubuntuone-windows-installer-windows-test/14:19
mandelbriancurtin, or is it another one?14:19
briancurtinmandel: ah, that's a different one14:19
briancurtinit's just "ubuntuone-windows-installer", no test at the end14:19
mandelbriancurtin, oh! ok fixing14:20
dobeyhmm14:20
mandelbriancurtin, we should remove the other project then, makes no sense to have it14:21
briancurtinmandel: probably, since there are now no tests in there14:21
briancurtinrather, there's nothing to test14:22
ralsinawe could do a run-tests.bat that generates the installer ;-)14:22
mandelbriancurtin, seems to be blocked, lets give it a few mins and check what is going on14:23
alecugatox: ping14:25
alecugatox: https://developer.apple.com/library/mac/#documentation/Darwin/Reference/FSEvents_Ref/FSEvents_h/index.html#//apple_ref/c/func/FSEventStreamCreate14:25
gatoxalecu, pong14:25
alecugatox: check out the "latency" parameter14:25
* gatox looking....14:25
alecugatox, macfsevents is forcefully setting that to 0.114:25
gatoxalecu, ahhhhhh14:26
briancurtinmandel: fail. i just added another print, will try again14:26
alecugatox: that means that the FSEvent API will take that amount of seconds before passing the events from the kernel to your process.14:26
gatoxalecu, soo... wait that amount of time befure returning from the add_watch method should be consider or it's a crime? :P14:27
gatoxahhhhh nono14:27
gatoxi understand now14:27
gatoxalecu, the problem is not in the add_watch, rather than in every event being generated..... am i rright?14:28
alecugatox: also there's FSEventStreamFlushSync14:28
alecugatox: that's the FSEvents API function to make sure that all events were delivered.14:29
alecugatox: any delay we add in our code to SD is a crime14:29
gatoxalecu, mmmm so which are our options?? modify the lib?14:30
gatoxfork it?14:30
alecugatox: so, as a first step, I would consider recompiling macfsevents with that 0.1 turned into 0, and seeing if your issue goes away.14:30
mandelbriancurtin, do you know in jenkins how can we add that project to the windows tab?14:30
* mandel is a sorting freak14:31
gatoxalecu, okkkkk.... i'll try that14:31
briancurtinmandel: i'm not sure, i haven't looked at how to do that but i'll check it out14:31
alecugatox, then, as a second step we should do some serious stress testing of this library, to see if it will suit us.14:31
gatoxalecu, ack..... i'll start with the recompiling thing..... and i'll let you know14:31
mandelbriancurtin, that looks funny: C:\Windows\system32\config\systemprofile\My Documents\AutoUpdate\output exists14:31
alecugatox: I'm mostly interested in the number of folders we can watch; how much memory it takes to watch 10000 folders.14:32
briancurtinmandel: are you logged in to the machine right now?14:32
mandelbriancurtin, yes14:32
briancurtinthat path *does* look a bit weird, but maybe it's a link to C:\Users\Administrator14:32
alecugatox: so, I think we should do stress testing of this too before you move much forward with your branch.14:32
mmccmorning, folks14:33
mandelbriancurtin, that path does not exist.. I'll get out so that you can take a look14:33
mandeldone14:33
mandelmmcc, morning!14:33
gatoxalecu, ack14:33
mandelmmcc, question, the network changes on linux, are they due to the changes we made on darwin? cause it used to work, right?14:33
gatoxalecu, ok.... i'll start doing that.... and let you know what happend14:33
alecugatox: also: SD needs complete assurance that when a watch has been added by some external API, it has started "working" just before the API returns.14:34
mmccmandel, no - the bug is on trunk, and my darwin changes aren't in there yet14:34
alecugatox: even if that means that the event will arrive a bit later (like it seems to do in this case).14:34
ralsinamandel, briancurtin: according to MSKB that is probably a corrupt user profile14:34
briancurtinsweet14:34
alecugatox: if it takes a little bit to arrive, it should be ok. (we can work the tests to fix this).14:35
ralsinamandel, briancurtin: which is interesting because we had all those strange permission problems in the past, remember?14:35
mmccmandel: the is_machine_connected call didn't work before - it just wasn't being called and the tests were wrong14:35
alecugatox: *but* if it never arrives, then we cannot use this API.14:35
mandelmmcc, uh, really.. so we have never got it working? fuck, nice catch114:35
mandel!!!14:35
alecugatox: does the above makes sense?14:35
ralsinamandel, briancurtin: http://answers.microsoft.com/en-us/windows/forum/windows_vista-desktop/cwindowssystem32configsystemprofiledesktop-is-not/8adbc6ac-64c5-4418-9223-95c92d2dc95814:35
gatoxalecu, yes sir, ok.... good things to keep in mind, i'll start working on that! THANKS for your help!14:35
alecugatox: SD absolutely needs reliability when dealing with events, because of local rescans and all of that.14:35
mmccmandel: well, it might have worked at one point, and then broke after the code that called it was removed. I didn't dig through the history14:36
alecugatox: awesome! please let me know and we can mumble again anytime about this.14:36
mandelmmcc, I'll review it as soon as I'm done with jenkins :)14:36
gatoxalecu, roger that14:36
briancurtinralsina: thanks, looking14:36
mandelbriancurtin, ralsina so it is a matter of fixing the user profile, which means that we might be able to get cp to pass all tests and remove the skipIfJenkins (maybe not)14:37
mmccmandel: so the bug showed up after gatox put the network status check back in sso-client recently.14:37
mandelmmcc, ah.. I had no idea14:37
ralsinamandel: maybe, maybe not, this is "daemons flying out of your nose" behaviour14:37
ralsinamandel: as in "undefined"14:37
gatoxmmcc, but i only change the behaviour of the page..... not the network_detection module14:38
gatoxmandel, "14:38
gatox^14:38
mmccgatox - right! you didn't make a bug, you just uncovered an old one :)14:38
gatoxahhhhhh14:38
mmccgatox: and the tests were no help :\14:39
ralsinagatox: happens when you look under things noone moved in a long time ;-)14:39
Moscherkoboldrye: any news?14:39
gatoxalecu, this is the ninja-mug that change the color: http://youtu.be/wEm5fkM8RPo14:40
gatoxthat is perrito's one14:40
mandelalecu, gatox I think ninja has more merchandising that ubuntu.. hehehehe14:42
alecugatox: lol14:42
gatoxjejejeje14:43
alecumandel, I'm not sure if in units, but surely in variety of items!14:43
gatoxmandel, actually..... maybe14:43
alecuthisfred, ping14:43
mandelalecu, there are certainly way more fun :)14:43
ralsinagatox: that looks exactly like my kitchen, including the misligned cabinet door14:44
gatoxralsina, that's perrito's kitchen14:44
ralsinaI know, I have no ODEX14:44
alecuthisfred, I saw lucio yesterday at the asado at beuno's. He gave me a little box that will mean I'll have to buy quite a few beers next sprint...14:44
gatoxralsina, jejeje14:44
thisfredalecu: ah, cool!14:45
alecuthisfred, thanks a bunch!!!!14:45
thisfredyw!14:45
gatoxalecu, ok.... with the latency 0.... we can't be sure, sometimes it works, sometimes doesn't14:54
mandelI guess we have a standup, right?15:00
briancurtinmumble?15:00
dobeyno15:00
gatoxme15:01
alecume15:01
dobeyralsina said no team meeting15:01
ralsinasorry, I sid it early, should have repeated15:01
alecubriancurtin, we are doing standup today, because ralsina lost his voice.15:01
briancurtinme15:01
mandelme15:01
mmccme15:01
dobeybah15:02
alecugatox: "we can't be sure" will probably mean "we can't use it"15:02
ralsiname15:02
alecuthisfred, standup!15:02
thisfredme15:03
mandelalecu, uh, I don't like that last sentence..15:03
* joshuahoover watches from afar15:03
alecumandel, the "we can't use it" one?15:03
gatoxback15:03
alecugatox: go15:03
mandelalecu, yes, because you are right..15:03
gatoxDONE:15:04
gatoxWorking on SSO client bug, macfsevents debugging, mumble with alecu about fsevents.15:04
gatoxTODO:15:04
gatoxKeep testing macfsevents to see if suit our needs.15:04
gatoxBLOCKED:15:04
gatoxNo.15:04
gatoxalecu, go15:04
alecuDONE: worked on setting up brewy stuff on the mac, mumbled with mandel, mumbled and debugging with gatox15:04
alecuTODO: finish with osx setup, do varied stress testing.15:04
alecuBLOCKED: no15:04
alecuNEXT: briancurtin15:04
briancurtinDONE: fiddled with the jenkins machine to get it setup properly to run the tests, make a couple of last minute adjustments once it got on the machine, then debug it a bit15:04
briancurtinTODO: figure out this profile issue and watch jenkins produces a nice installer artifact for us15:04
briancurtinBLOCKED: none15:04
briancurtinNEXT: mandel15:04
mandelDONE: Merged tools.py for windows and darwin because they are the same in every detail but the transport used. Got back to fsevents and made tests run correctly. Added more tests.15:04
mandelTODO: Review for mmcc. More tests.15:04
mandelBLOCKED: no15:04
mandelmmcc, please15:04
mmccDONE: tech docs wiki, hacked at pyinstaller, sso bugfix15:04
mmccTODO: find next pyinstaller bug, repeat. OR revisit py2app15:04
mmccBLCK: NO15:04
mmccNEXT: ralsina15:04
ralsinaDONE: worked a bit more in the "get_rootdir" bug, reading & learning about lenses and stuff, worked on mac mini setup, coughed and coughed TODO: early out today, will put hours back tomorrow morning because of assorted national holidays BLOCKED: not, but uninspired for coding :-/ NEXT thisfred15:04
thisfredDONE: bug #1003610 TODO: finish bug #1003610 BLOCKED: no15:04
ubot5Launchpad bug 1003610 in U1DB "get_doc() and get_docs() should not return deleted documents by default" [Medium,In progress] https://launchpad.net/bugs/100361015:04
dobeyλ DONE: bug #711162, stable-4-0 series and branch setup15:05
dobeyλ TODO: set up milestones, triage, find someone to test installer fix, installer SRU15:05
ubot5Launchpad bug 711162 in Ubuntu One Client trunk "ubuntuone-login crashed with ValueError in call_async(): Unable to guess signature from an empty dict" [Undecided,Triaged] https://launchpad.net/bugs/71116215:05
dobeyλ BLCK: none.15:05
alecuoh, and NOTE: tomorrow is a national holiday in Argentina, so I'll not be working.15:05
ralsinaalso NOTE: monday is natl holiday in the US so only mandel works both days15:06
ralsinaand we need to be flexible about reviewers, since a lot of people will be missing one day or the other15:06
mandelherb, the world is going bananas when a Spaniard looks like a hard working man15:07
* mandel goes and takes a siesta..15:07
gatoxalecu, well... about macfsevents.... i'm thinking that in the tests we are generating the events........ but (correct me if i'm wrong)..... the normal use case of filesystem_notifications..... is that module to notify the other modules..... in that way..... it's ok..... we don't have any trouble there..... the problem is when we force some event and try to get the answer right after we push the event15:07
gatoxalecu, we can mumble about that if i'm not making myself clear15:08
dobeyralsina: simple solution. don't propose any branches tomorrow or monday :)15:09
ralsinadobey: I would rather have a less life-is-hard-let's-go-shopping solution ;-)15:10
dobeyralsina: i think we can survive it15:11
dobeypeople are allowed to do reviews when it's not their review day15:11
alecugatox: let's mumble15:11
gatoxalecu, ack15:12
mandelralsina, dobey, I can do reviews, I have not problem is a good way to rest from the objective-c code15:12
dobeyalecu: btw, were you going to review my branch, or leaving it for someone else to review, because it doesn't add tests? :)15:12
dobeyor well, doesn't add more tests, and doesn't break the current tests15:13
alecudobey, I should put a "needs fixing" on it. Hold a sec... :-)15:13
dobeynoooooo15:13
ralsinaalecu: and then go on vacation :-)15:13
briancurtini need coffee or i am going to fall face first onto the keyboard. brb15:14
dobeyand i need to get lunch15:18
dobeybbiab15:18
Moscherkoboldis there a way do delete an ubuntu one account?15:19
ralsinaMoscherkobold: it's tricky because we don't actually have ubuntu one accounts15:21
ralsinaMoscherkobold: so if you really want to delete all traces of it, you need to delete the ubuntu sso account15:21
ralsinaMoscherkobold: And I am not sure on how to do that, either15:22
Moscherkoboldhmm ok lets see first if there is a way how rye can help me15:22
ralsinaMoscherkobold: AFAICS, it has to be requested and done manually15:23
ralsinaMoscherkobold: so you could do a support request for it15:23
ralsinaMoscherkobold: it's bug #66422915:24
ubot5Launchpad bug 664229 in Canonical SSO provider "Offer a way for people to cancel their Ubuntu SSO account" [High,Confirmed] https://launchpad.net/bugs/66422915:24
Moscherkoboldok thank you, i will wait until rye has finished (somehow i have created 2 "accounts" with the same email)15:25
ralsinaMoscherkobold: ouch15:25
ralsinaMoscherkobold: OTOH, good to see you are not closing because you hated it :-)15:25
Moscherkoboldi think because of a bug while connecting a windows pc15:26
trubborralsina: I'm in the throws of the same thing - U1 ticket # 1550715:26
Moscherkoboldso my laptop with ubuntu and my android device are connecting to the "correct" account with files in it15:27
ralsinatrubbor, Moscherkobold: sorry to hear that :-(15:27
Moscherkoboldbut if i log in with a browers or a windows client the other/wrong/empty account appears...15:27
Moscherkobold*browser15:27
trubborralsina: Is there a proper was to remove U1 from an Ubuntu user account, so that when you re-connect with your new U1 account, things are "fresh". I tried to uninstall U1, and then re-install - but it seems to have found my old info (as if the uninstall didn't clear out everything)15:30
ralsinatrubbor: yes, the instructions are operating system specific though15:31
ralsinatrubbor: so, ubuntu or windows?15:31
trubborUbuntu15:31
ralsinatrubbor: ok, you don't need to uninstall15:31
ralsinatrubbor: first, in the control panel, remove your device so you are not authenticated anymore15:31
trubborDid that...15:31
ralsinathen remove the contents of ~/.local/share/ubuntuone/syncdaemon15:32
ralsinaAnd then login again15:32
trubborralsina: Didn't do that yet...:)15:32
ralsinabefore removing, do u1sdtool -q just to be paranoid about it ;-)15:32
trubborralsina: just looked for a man page for ulsdtool - didn't have one - what is that?15:34
ralsinatrubbor: just a command line tool to control u115:35
ralsinathat's u1sdtool (a one after the u) and should have a man page, too15:35
trubborralsine: yep - just saw that :)15:36
trubborralsine: That did it15:37
ralsinatrubbor: cool :-)15:38
ryeMoscherkobold: i was able to reproduce the issue with your instructions and contacting the responsible parties15:40
Moscherkoboldgreat, thx a lot15:41
* alecu has lunch15:41
Moscherkobolddid you find a quick solution already?15:41
* gatox lunch15:41
=== gatox is now known as gatox_lunch
trubborralsine: I'm worried that there is a "parallel" directory - with my old U1 profile sitting out there. For awhile, I had two separate U1 cloud syncs going - one with old accounts seeing old stuff under the "new" U1 login, and then new accounts (new local system accounts) were seeing the "fresh" U1 directory contents. Is this something I should worry about now? I'm wanting everything "clean" before I commit to U1 storage for important dat15:42
trubbora, etc15:42
mandelI need to take a rest, my head is about to explode, will be back later..15:43
ralsinatrubbor: sorry, didn't see that earlier because you mistyped me ;-)16:08
ralsinatrubbor: U1 is not meant to be used with two accounts from the same system user, so if you atry to do that, it's gling to end in tears16:08
ralsinatrubbor: what you *can* do, is, from account A create a folder and share it to the account B, then you can use only account B16:09
ralsinatrubbor: OTOH, maybe I am just misunderstanding you16:09
ralsinalunchtime. Will be back in a bit.16:20
trubborralsina: Had U1 support delete my account (U1 support ticket 15507) after re-creating, I get the "parallel" account behavior16:24
dobeytrubbor: https://one.ubuntu.com/help/faq/what-does-the-root_mismatch-error-mean/ this?16:26
trubbordobey: Yes..and no. U1 Staff deleted my account (at least my login info), because of a bug in the mobile name changing code. I re-created the account - same account name, same password. Did not "clean" my local systems before they reconnected. Using the same username, and same password - two systems had one set of cloud file sync, and another login which had never had U1 sync setup before, and a mobile device - had the same (but differe16:35
trubbornt from the first two). All 4 using same login credentials. The first two - still had sync'd files from the original iteration of the account16:35
dobeytrubbor: are you getting that error where you say there is a "parallel" directory somewhere? we don't understand what you mean by parallel there exactly16:36
trubborI followed instructions from *ralsina* and the problem no longer occurs on any of the systems that I have applied the "cleaning" procedure to. Before that - there were two separate "versions" of my U1 account that would run concurrently. My concern is that even though my login was deleted, that my user directory may not have, and that I may now be using a new U1 directory that has the same account info as the old one. This is probably m16:46
trubborore a procedural problem on my end. I should have "wiped" my local U1 info, before re-attaching to U116:46
trubbordobey: Sorry - Wife ACK going on here..16:50
mmccoh hey neat, tips n' tricks relevant to py2app can be found only on py2exe's page...17:11
dobeytrubbor: ok, so it sounds like it's all ok now17:22
dobeyoh launchpad, your bug management antics are so confusing17:23
trubbordobey: I think so - I'm going to use it for a few days and see how it goes. U1 is really great - I actually like it better than iCloud right now - just wnat to get Thunderbird contact sync back!17:23
mmccdeep in pyinstaller's test script, I learn that "support for eggs" means "only supports zipped eggs"17:23
dobeysigh17:32
dobeyalecu: do you have any idea how to write tests for that change, that doesn't involve me spending 3 days refactoring all the tests so i can?17:32
alecudobey, I think I do. I'll try to write a sample...17:49
dobeyalecu: thanks17:50
briancurtinughhhh. jenkins has a newer BitRock, which sneakily has been placing files in different spots18:02
mmccbriancurtin: computers, amirite?18:05
* mmcc going to lunch18:05
ralsinahello again, sorry, internet died :-/18:14
ralsinanot an easy day to work18:14
joshuahooverbriancurtin: i think you've learned that you can't trust that guy, jenkins...i suspicious by the name alone18:15
joshuahooveri'm, even18:15
ralsinammcc: py2app *is* py2exe so 90% of the tricks needed should already be in place18:15
briancurtinwhat we need is the jenkins plugin that changes the background from the little cartoonish butler guy to a picture of chuck norris18:15
ralsinabriancurtin: really? if the new bitrock puts things in logical places it may be worth upgrading ;-)18:16
briancurtinralsina: i didn't even notice that what i installed was a new version. i installed 8.1.0 on a machine a few weeks ago, then i installed (without noticing) 8.2.0 on jenkins. the failures mandel and i were seeing are because it looks in the old, weird place18:16
ralsinabriancurtin: sigh, it happens18:17
ralsinabriancurtin: maybe we should update the setup docs to point to copies of everything so we always get the same versions of everything in all the systems.18:17
ralsinabriancurtin: of course then we would need to update it every week for new versions so we don't get out of whack with ubuntu18:18
briancurtinerr...actually, it's not as easy as i thought and not specific to the version. because jenkins runs as a system process or something, not as a user, it puts the files in a different spot because it *can't* put them in taht weird, user specific spot18:19
ralsinammcc: in fact, at least oneo the workarounds and dirty hacks we use in py2exe we took from a py2app page18:20
ralsinabriancurtin: aha18:20
briancurtinso i'll check one place on jenkins, one place on our own machines...meh to that, but it's simple enough18:20
ralsinabriancurtin: maybe we can use runas?18:20
ralsinabriancurtin: and BTW that sounds more and more like the cause of the permission issues that caused SkipIfJenkins18:21
briancurtinralsina: it looks like jenkins runs as a windows service which wouldn't help there18:21
dobeywhat have i gotten myself in to18:22
briancurtinand yep, that does make sense about those permission problems. jenkins processes run in the context of a system process rather than a user process...or something like that, been a while since i fiddled with services18:22
ralsinabriancurtin: right, and UAC or however it's called is locking it out of stuff18:22
ralsinabriancurtin: http://jenkins.361315.n4.nabble.com/Slave-Won-t-Run-Under-Configured-User-td3617555.html18:23
ralsinabriancurtin: you can tell windows to run that service as a specific user, it seems18:23
ralsinaFYI  I just logged into the bar's wifi/adsl router and rebooted it. Goes to show I *want* to work guys.18:43
ralsinaIt's fun that the password of my home's router works at the bar as well. And it can't be changed.18:43
dobeyit's good that you broadcast that to the public18:44
ralsinadobey: you need to get into the wifi first to do it. So doesn't do anything at my home network18:44
dobeybetter change your password :P18:45
ralsinadobey: and is handy at bars, let me tell you18:45
* dobey needs to set up his new router, buy more cable, and run it18:46
ralsinadobey: that's the beauty of it: they give you the restricted admin site, which can't change the password. So all router have the same one, forever.18:46
ralsinadobey: OTOH, all you can do is reboot it.18:46
dobeys/reboot/DoS/18:48
ralsinadobey: pretty much. Or s/reboot/fix/ in this specific case.18:48
dobeyalso, i wish it were easier to find non-pvc ethernet cable18:50
ralsinadobey: why non-pvc? It's like 99.99% of the cable you will find18:52
dobeyto have less random toxic stuff laying around all over my house?18:53
ralsinaPVC is not toxic unless you... eat it?18:53
ralsinain fact, PVC is one of the least toxic plastics18:54
ralsinaunless you burn it of course18:54
ralsinaAND then eat it18:55
dobeymaybe i'll do that in 50000 years, when i'm old and senile and the pvc is still laying around18:57
ralsinayou won't notice18:57
ralsinadobey: any idea why bug #985484 is not fix-committed?18:59
ubot5Launchpad bug 985484 in ubuntuone-control-panel (Ubuntu) "Qt control panel says kilobits per second, means kibibytes per second" [High,Triaged] https://launchpad.net/bugs/98548418:59
ralsinadobey: or should I just blame launchpad?19:00
dobeyeh?19:00
ralsinadobey: it has a branch associated, branch is merged, bug is triaged19:00
dobeyralsina: blame yourself19:01
ralsinahmmmm19:01
ralsinahappy to blame myself, but what did I do wrong?19:01
ralsinajust to avoid repeats ;-)19:01
dobeyi don't know, i guess that isn't the one that wasn't linked properly19:01
ralsinaok, I'll set it manually19:01
dobeyi did19:02
ralsinaok19:02
dobeyoh19:02
dobeymaybe launchpad changed in some way19:02
ralsinaSo did I and it got back to triaged!19:02
ralsinaok, no, committed now19:02
ralsinaweeeeeeird19:02
dobeyand tarmac isn't picking up bugs that aren't targeted to a specific series properly19:03
dobeyralsina: the (Ubuntu) task isn't fix committed.19:03
dobeyit should be triaged19:03
ralsinadobey: ok, backing that one, or you do it?19:03
dobeyi did19:03
ralsinaok19:03
gatoxalecu, ping19:05
briancurtinralsina: https://jenkins.errormessaging.com/job/ubuntuone-windows-installer/14/19:10
gatoxmmcc, ping19:11
gatoxok..... now i feel ignored....... jejeje19:12
gatox_macgatox ping19:13
alecudobey, https://code.launchpad.net/~alecu/ubuntuone-client/signature-dict/+merge/10726519:13
alecugatox, pong19:13
gatoxalecu, can you take a look at the mmcc wiki..... i added some comments about macfsevents..... not very good19:14
alecugatox: "Doing a stress test (creating 100 folders with 100 files inside each), it seems to lose a lot of events"!!!!19:15
alecugatox, can you tell me a bit more on how you did that?19:15
gatoxalecu, and that is not very stressful19:15
gatoxalecu, creating a folder, with a file inside..... lose the file creatoin event too19:15
gatoxalecu, with this script: http://paste.ubuntu.com/1005257/19:16
gatoxreally simple19:16
alecugatox, also, please move that bit to the "Technical Decisions (Open)" part of the doc, since we will have to revisit it.19:16
gatoxack19:16
dobeyalecu: but that still doesn't test the situation where the bug happens19:16
alecudobey, right: this is not an integration test, it's only a unit test.19:17
ralsinabriancurtin: looking...19:17
gatoxalecu, done19:18
ralsinaalecu, dobey: unit tests are valid tests19:18
mmccgatox, pong - what's up?19:18
dobeyralsina: that's a statement of assumption, not fact, though :)19:19
gatoxmmcc, questoin: about the network_detection bug, do you know if there is someone working on that? because i need to fix something in sso.... and i could do that if noone is looking at that19:19
ralsinadobey: no, it  is a fact that unit tests are a good thing in general, and valid tests to add to our projects, although not the only kind f tests we ever want (detailed enough? ;-)19:19
gatoxalecu, sooooo..... really bad news.......19:20
ralsinabriancurtin: yay!19:20
mmccgatox: you mean is someone reviewing that? I think mandel was going to, but then he left19:20
dobeyralsina: but all tests which are unit tests are not necessarily valid tests, regardless of whether they pass or fail, i mean19:20
gatoxmmcc, ahhhh you already fix it? (for linux i mean)19:20
alecugatox: tell me19:20
mmccgatox: yeah, fixed last night19:20
gatoxmmcc, cool19:20
ralsinadobey: define valid in that context. Or rather, don't ;-)19:21
gatoxalecu, no, the bad news is what the document says..... even with few files or folders, if the events are really close together, we lost events19:21
alecubriancurtin, is that a real windows package built by jenkins itself? awesome!19:21
briancurtinyep19:21
alecugatox: wait.19:22
gatoxalecu, but only when we add files inside a folder that was just created19:22
alecugatox: the FSEvents API does "aggregation" of events.19:22
alecugatox: ah! files inside a folder that was just created is a slightly different case.19:22
alecugatox: we have that same issue with pyinotify. And code in sd to work around it.19:23
gatoxalecu, really?19:23
gatoxi didn't know that19:23
gatoxalecu, can you point me to that code..... so i can do the same for the stress test?19:23
alecugatox: when SD detects that a new folder was created, it might have passed some time before SD can put the watch on it.19:24
alecugatox: so SD forces a local rescan of that folder just after putting the watch.19:24
alecugatox: don't know where it is. I just know it's there :-)19:24
alecugatox: I can try finding it anyway :-)19:25
gatoxalecu, mmmm but i understand.....19:25
ralsinabriancurtin: great job, man, this should make our life 2% easier AT LEAST19:25
briancurtini hope so19:25
mmccFYI, I have to leave a bit early today (in ~30min)for a doc appt, will come back later for more night hacking19:25
gatoxalecu, in this case, we don't need to add the watch, but we know that we are not going to receive the events instantly after the folder is created19:25
dobeysigh19:26
gatoxalecu, but it's kind of the same problem.... we can implement sometihng similar19:26
gatoxalecu, i'm going to look for the code you mention19:27
dobeywhy is it so hard to test that things actually fix bugs19:28
ralsinahave to go for a few hours, will be back tonight & tomorrow morning19:28
ralsinamail me in you need me to know anything19:28
gatoxack19:28
alecugatox: ubuntuone/syncdaemon/sync.py -> new_local_dir19:44
alecugatox: that's the bit where the local rescan is called.19:45
alecu"self.m.lr.scan_dir(mdid, path)"19:45
gatoxalecu, thanks! i'm running more stress tests in this moment19:45
* mmcc has to run -> be back later19:54
alecubriancurtin, it's awesome to see the console output of jenkins while it's building. Great job, everybody that worked on it :-)19:55
briancurtini'm happy it finally got wrapped up19:55
alecubriancurtin, I've just forced jenkins to build a new package, then tested it in a pristine w7 vm, and it worked awesomely; it even synced everything back into the vm.20:02
alecubriancurtin, this is just awesome. Congratulations!20:02
joshuahooveryes, congrats briancurtin! this is a HUGE step forward for our windows client20:02
briancurtin:)20:02
gatoxEOD here!! see you tomorrow people!20:13
dobeywow. now this is a slow laptop.20:29
dobeywelcome to suse linux enterprise desktop 1020:32
alecummcc, do you know about this step in the setup doc?: $ source env-mac (located in: ubuntuone-windows-installer/scripts/devsetup)20:36
alecummcc, I can't find the env-mac in that u1-win-installer branch by brian (as shown on the doc)20:37
alecuonly a env.bat.20:37
alecuI see... it seems I should be using trunk20:44
dobeyalright, have a good evening all21:13
=== m_conley is now known as m_conley_away
alecuok, this is EOW for me.21:59
alecubye all!21:59

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