/srv/irclogs.ubuntu.com/2011/07/07/#ubuntuone.txt

mandelmorning all07:52
mandelfagan: happy bday07:52
faganmorning mandel thanks :)07:52
* mandel is really tired...07:59
* fagan takes a quick break to get some milk 08:45
faganconnection issues mandel? :)09:24
mandelfagan: 3G, so it sometimes decides to go… also adium is a bit shit regarding irc...09:27
faganmandel: my ma has one of those yokes and IRC doesnt even work on it09:28
mandelfagan: in my case, I don't know why it just decides to crash...09:30
faganmandel: weird09:32
mandelfagan: yes.. is a PITA but I can deal with it, 3G so far has been working very well, can browse etc… and Windows has updated several times so far...09:37
czajkowskimandel: how is your brother???09:50
mandelczajkowski: he is ok, apparently he decided to slid downhill with a plastic and water or something like that… went to hospital had an MIR and got told he was lucky :)09:51
mandelczajkowski: is in our blood we are dangerously stupid :)09:52
czajkowskimandel: spanish :p09:58
mandelczajkowski: yes hehe10:00
lalejandHi, I have ubuntuone-sync and ubuntuone-indic both oscilating between 65 an 95% CPU :/10:22
lalejandand u1sdtool is "processing the command pool"10:23
lalejandIs ti normal taht it takes so much CPU ? Nautilus is barely usable10:24
mandellalejand: no, it is not normal at all10:24
mandellalejand: if you fill like taking a look on what is going on I recomend you use https://launchpad.net/magicicada10:25
ralsinamorning!10:26
mandellalejand: is a front end for sdtool that should make more sense and can show you WTF is syncdaemon doing...10:26
mandelralsina: morning!10:26
mandelralsina: fancy a nice mumble about utf8, file systems and their different mothers...10:26
mandelralsina: at least before I take a swim/walk to think about it...10:26
mandelactually better skype10:26
ralsinaMy son is asleep 2meters from here10:27
ralsinaso no talking for at least 90 miutes sorry :-(10:27
mandelralsina: then we type10:27
ralsinabasically, file names are bytes, andnot guaranteed to decode correctly in any way10:27
mandelralsina: exactly, so we can use unicode in all the windows operations with no problem10:28
mandelralsina: but sd wants bytes.. so we can deal with the diff paths but when we encode, we are fucked...10:28
ralsinawell...10:29
mandelralsina: we could use url encoding, is ugly but will fix that issue but then users can type paths like that10:29
ralsinawe could be bytes all the way10:29
mandelwhich is when I'd tell them to stop screwing up with me :)10:29
faganoh hi ralsina10:29
ralsinaas in, taking the file names as-is10:29
mandelralsina: waht do you mean?10:30
ralsinathe problem you have is that you take a str from the filename, convert it to unicode, modify, and then you need to re-convert to str to use it on the fs again, right?10:30
ralsinaOr maybe I am understading the problem itself wrong10:31
ralsinabecause if sd wants bytes, and windows gives you bytes... let's just use bytes10:31
mandelralsina: no, what I do is, convert to unicode, modify it and write it to the fs with no problem10:31
mandeland windows can always give be utf8 with no problem10:32
ralsinaso, the problemis converting to bytes to give to sd?10:32
mandelralsina: actually know that I think about it I'm stupid… I know how to sort it out...10:32
lalejandmandel : no deb for magicicada ?10:32
ralsinaa utf-8 string *is* bytes10:32
mandelralsina: actually, I'm lying, I do not know how to fix it… we can talk in 90 min, over irc is hard to explain the issue10:35
mandelor solution for that matter10:35
ralsinamandel: ok10:35
mandellalejand: let me look for it10:36
ralsinaI don't understand what you mean by "sd" for example, sice your code *is* in sd ;-)10:36
mandellalejand: which letter? M, N or O?10:36
lalejandmandel : wich letter ? I don't understand10:37
lalejandmandel : I downloaded the tgz and lauched it10:37
mandellalejand: Maverik, Natty or Oneric :)10:37
lalejandwhat should I check in it ?10:37
lalejandmandel : ah sorry, N10:38
mandellalejand: :)10:38
mandellalejand: looks like it is in universe: http://packages.ubuntu.com/natty/magicicada10:38
mandellalejand: if you launch it and connect to sd you should be able to see the diff operations going on10:38
mandelralsina: the problem I have is the following, if you tell os_helper to give the abspath it must return byte with the illegal chars, and next os_helper operation should take those, change back to utf-8 and perform the operation.. and I think that would solve the issues10:40
mandelralsina: the annoying part is that we will move to unicode in the future, lo que me toca los huevos un poco...10:41
lalejandmandel : yes I see them. meta queue --> a loooot of unlink, content queue --> a looot of upload10:41
ralsinamandel: you know who knows a lot about unicode and such? facundobatista. And he also knows a lot about syncdaemon. Win win! ;-)10:41
mandelralsina: which is the person I was arguing with yesterday :)10:42
mandelin a friendly manner, of course, you know me10:42
lalejandmandel : the U1 indicator shows the same since almos 1h now : uploading 3 files : 8%, 0% and 0%10:42
mandellalejand: is that what you expected to see?10:42
mandellalejand: from magicicada you can tell to stop with those uploads (I think)10:43
lalejandmandel : I suppose, yes. But it looks like it is stuck now. I am generating and regenerating a lot of files10:43
mandellalejand: lets see if the debugging god is around, rye are you here?10:43
ryelalejand, mandel yep, looking10:43
mandellalejand: rye is way more intelligent than me, I'm just a code monkey, rye you take care of it? :)10:44
ryelalejand, try killing ubuntuone-indicator first - that might be causing excessive SD interrupts in some weird cases10:44
mandelralsina: I'm of for 30 min to think about the utf8 but I think I got it10:45
ryelalejand, are you uploading/downloading a large number of files?10:45
lalejandrye : ok I will disconnect U1, but I was leaving it in case you needed some information about its actual state10:45
* rye needs to rate-limit some operations in indicator10:45
ryelalejand, no no no10:45
ryelalejand, killall ubuntuone-indicator10:45
ryelalejand, it is an external process so it will not affect anything if it is killed10:45
lalejandrye : I think so, yes. And worst than that : generating and regenerating a large number of files that are being uploaded10:46
lalejandrye : u1-indicator ok10:46
lalejandrye : U1-indic killed10:47
mandelralsina: we will fix it, but on Windows we are doing a stupid amount of extra work I do not like… and I wonder how it works in other OS...10:47
mandelanyway I go, will be back soon10:47
lalejandrye :u1sync is at 46% cpu right now10:47
=== mandel is now known as mandel_mediterra
=== mandel_mediterra is now known as mande_afk
lalejandrye : now 3% :)10:48
lalejandrye : indic was the problem10:48
* fagan calls mandel mande from now on :)10:49
lalejandrye : oh no, 65% now10:49
ralsinafagan, you don't want to10:49
ralsinafagan: "mande" means "give me an order" :-)10:49
ralsinafagan: btw, happy birthday10:49
faganralsina: thanks :010:50
* fagan even learns a bit of spanish10:50
lalejandrye : 90% buuu :(10:51
ryelalejand, okay, i think i know what can be causing this in indicator, sd is scanning the folders/queueing up things. Every single item being added causes SD to emit notification that something has changed. Indicator listens on these notifications and queries sd about the changes.10:52
ryelalejand, could you please check how crazy is ~/.cache/ubuntuone/log/syncdaemon.log is being written? (tail ~/.cache/ubuntuone/log/syncdaemon.log)10:53
ryelalejand,  tail -f10:53
* fagan break10:58
lalejandrye : quite calm now10:59
ryelalejand, and still using the cpu ?11:00
lalejandrye : nope, not now11:03
lalejandrye : but using 42% mem11:04
lalejandrye : nothing happens in magicicada, no move11:06
lalejandrye : 52% mem now11:06
ryelalejand, could you please try tunning u1sdtool --waiting | wc -l11:10
lalejandrye : 90% cpu, 69% mem now11:11
lalejandrye : ok11:11
ryelalejand, it should print out a number11:13
lalejandrye : magicicada shows same list o meta & content queue11:13
ryelalejand, ah, could you please try shutting down magicicada too?11:14
lalejandrye : Traceback (most recent call last): Failure: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken. 211:14
ryelalejand, and run u1sdtool --waiting | wc -l in terminal11:14
lalejandrye : it says what I copi-pasted above11:16
lalejandrye : 99%, 51%. Nautilus totally stuck11:16
facundobatistaMuy buenos días a todos!11:16
ryelalejand, so it is too busy doing something and something is not right. Could you please pastebin the output of grep MARK ~/.cache/ubuntuone/log/syncdaemon.log* ?11:17
lalejandrye : it returns me nothing11:17
lalejandrye : 7% 56%11:18
ryelalejand, okay, so now neither ubuntuone-indicator nor magicicada is running, am I correct?11:18
lalejandrye : correct11:19
lalejandrye : nautilus still frozen11:19
ryelalejand, hm, could you please pastebin the last 10 lines of ~/.cache/ubuntuone/log/syncdaemon.log - they may contain filenames (tail -n 10 ~/.cache/ubuntuone/log/syncdaemon.log) - feel free to remove the filenames from there if they are sensitive11:20
lalejandrye : http://pastebin.com/0LU9d7gM11:23
lalejandrye : nautilus is usable now11:24
ryelalejand, is syncdaemon idle now (u1sdtool --status)11:25
lalejandrye : can't check it : Oops, an error ocurred: Traceback (most recent call last): Failure: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.11:27
=== mande_afk is now known as mandel
duanedesignmandel: that is one thing that really annoys me about Windows is the updates. I only boot my windows box once a month or so and it always wants to spend 30 minutes doing updates11:40
ryeduanedesign, the same would have happened if we booted ubuntu once a month... not 30 minutes though... probably11:41
mandelduanedesign: yeah.. is a major PITA, yesterday I wasted a stupid amount of time due to that11:42
ryefacundobatista, how often does sd react on dbus requests? I mean is this a separate thread for serving dbus or just an idle handler?11:42
facundobatistarye, for the deepest detail I'll need to check code or ask verterok, but it's immediate, as it's integrated in twisted's reactor11:47
ryefacundobatista, ah, ok, so if e.g. u1sdtool --waiting results in NoReply dbus exception it may mean that the amount of data being collected is quite large, right?11:48
facundobatistarye, yes11:49
facundobatistarye, see #754050 that will prevent that11:50
facundobatistarye, feel free to upgrade it to "high" if you think it's important11:50
nessitagood morning everyone!11:53
facundobatistaHola nessita11:53
ralsinagood morning nessita!11:53
=== teknico is now known as teknico_away
nessitahola facundobatista, ralsina11:54
ryefacundobatista, well, yes. I'd still like the remaining upload/download amount too though11:54
nessitafacundobatista, mandel: did you guys have sync regarding branches?11:54
lalejandrye : sorry I can't see the status. Trying, but I walways get an error message. And everything is very slow11:55
ryelalejand, ok, let's try this - "tail -f ~/.cache/ubuntuone/log/syncdaemon.log* | grep MARK" and keep it running for some minutes (it won't affect the performance more than SD)11:55
ryelalejand, i am trying to find out the number of files in the queue11:55
facundobatistanessita, I answered the mail yesterday evening11:56
ryefacundobatista, should --status be less memory-heavy ^11:56
nessitamandel: were you able to appply changes?11:56
facundobatistarye, status is not memory-heavy at all11:56
facundobatistarye, if SD is not answering then it may be working really too hard doing something else (maybe processing and trashing because the previous --waiting?)11:57
ryefacundobatista, ah, that can be it, and waiting request blocks all other operations, right?11:57
mandelnessita: I'm on the middle of a big branch that ensures that all paths are byte and with the correct chars in every os_helper method + pyinotify, that should fix the listdir, stat etc errors… but I did not talk with facundobatista yet11:58
facundobatistarye, not block, but make it to work too much to answer dbus in a timely manner11:58
nessitamandel: ok, When you can context swicth, please try to have those 2 branches moving in the land queue (were those 2 or only 1?)12:00
mandelnessita: there are 2, yet if the fix I'm doing know works correctly one of them will be supeseeded by it since the solution is cleaner, the other is regarding the test complexity12:04
nessitamandel: yeah, and we talked about that yesterday in the meeting, so if you have any doubts about how to fix the test, please ping me12:04
mandelnessita: yep, will do :)12:05
nessitamandel: basically, when you need to replace something with a no-op, do not use mock but patch, yes?12:06
mandelnessita: yep, got it12:06
nessitamandel: for the tests you're doing mock gives you nothing better than using self.patch12:06
mandelnessita: in that test precisely you are right.. also I wonder why removing 'b' did not brake the test correctly12:07
nessitamandel: so, when you finish this branch you're currently working on, I will suggest removing all the mock calls and replace that by patchs (In the branch you need to fix)12:08
mandelyep12:08
lalejandrye : 2011-07-07 14:08:58,328 - ubuntuone.SyncDaemon.Main - NOTE - ---- MARK (state: <State: 'QUEUE_MANAGER'  (queues WORKING  connection 'With User With Network')>; queue: 59481; hash: 0)12:14
ryeomg12:14
nessitaralsina: hey there! can I ask if sent your report yesterday? I'm worried since I only got alecu's, and maybe thunderbird is acting weird on me12:14
rye60k queue items. lalejand were these files originally in Ubuntu One or you have added them recently?12:14
ralsinanessita: no, didn't, felt horrible so stayed in bed :-(12:15
nessitaralsina: oh, you feeling better now?12:15
ryefacundobatista, do we have any benchmarks regarding cpu usage/upload eta of a large number of files (where large is 60k) ?12:15
ralsinanessita: not really, but I have stuff to do :-)12:15
facundobatistarye, we did a couple like 10 months ago12:16
nessitaralsina: any chance you send your report now?12:16
ralsinanessita: sure, in 2'12:17
nessitathanks12:17
ryelalejand, could you please wait for about 2 minutes more so that we see how the queue is progressing and can plot an approximate eta for files upload?12:18
ralsinaalecu, nessita, mandel: no hurry, this is for any time today, can I get reviews for https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_805290/+merge/66739 and https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_803929/+merge/66827 please?12:22
nessitaralsina: queueing that up by opening tabs in FF12:23
ralsinanessita: thanks!12:23
nessitaralsina: that cancel message should be shown every single time the user clicks cancel, right?12:24
ralsinanessita: yes12:24
nessitaok12:24
nessitaralsina: can you please add tests for side_widget? basically everything not in an xml file or a generated ui files should have tests12:29
ralsinanessita: it has no logic. It instantiates the .ui and sets the window caption12:29
nessitaralsina: you can have typos, someone can screw something up. All the paths should be exercised by tests.12:30
nessitaralsina: also, I can argue you have logic :-)12:30
ralsinaok, I'll add a test for it. Checking that the title is right?12:30
nessitaralsina: stage should 0 when started12:30
nessitathe stage should be assigned when calling set_stage12:30
ralsinanessita: actually, no, because the wizard will reset it12:30
ralsinaok, will check12:31
nessitaralsina: when you call set_stage, you're doing 6 stuff, and each one should have have it's own test12:31
ralsinanessita: ok, will add tests for that too12:32
nessitaralsina: also, is there any chance you make test_stage_progression be a unit test? that means, splitting that test in several tests trying to test a single change12:32
ralsinaok, will do that too12:32
nessitathanks!!!12:32
ralsinaI think splitting test_stage_progression in 13 tests is overkill, but anything to make you happy!12:33
nessitaralsina: well, do you have 13 different options when calling on_currentIdChanged?12:34
ralsinanessita: yes12:34
nessitaralsina: then is expected to have 13 tests :-)12:35
nessitawhen reading that code as is, using "new" eyes, is hard to understand (just happened to me)12:35
ralsinanessita: ok12:41
nessitaralsina: for the are you sure branch, I added a needs fixing (requested tests) and a question12:42
ralsinaok, checking12:42
alecuhello Ubuntu One!12:42
nessitahi alecu!12:42
faganhey alecu12:43
alecuhi there all :-)12:44
fagan15 mins to standup12:45
ralsinanessita: I can set the text from python, sure. It will have a bit of HTML, but will be just a <br>12:48
nessitaralsina: a '\n' will not work?12:48
=== teknico_away is now known as teknico
ralsinanessita: no, because there is a link: https://launchpadlibrarian.net/74511342/are_you_sure.png12:49
ralsinanessita: in fact, a <h3> or similar is needed too.12:49
faganralsina: could you not do that in 2 labels?12:50
nessitaralsina: can we split that in a title, that the string definition has no tag, and when setting in the dialog you add the h3?12:50
ralsinanessita: yes, I can do that, too. I can split it in paragraphs, too.12:50
nessitaand then the rest of the message without the link, and then you add the link, which is not translatable, at the end?12:50
nessitasomething like:12:50
nessita"If you need more go to %(ubuntu_one_link)s."12:51
nessitaand then you use that passing named parameters12:51
ralsinayeah yeah12:51
nessitathat will ease the job for translators a lot12:51
ralsinaOk12:52
ralsinapushed the other one with 13 tests, and now it checks all the way up to the right labels being enabled for each possible value of page_id on a separate test.12:52
nessitaralsina: could it be that you forgot to add test_side_widget?12:54
thisfredstandup in 6!12:54
faganthisfred: 5 :D12:55
nessitaall: Diego is taking an exam, so he sent me his standup, I will pasting that soon12:55
ralsinanessita: the side widget is tested completely with 100% coverage over each path (I checked!) by the test_stage_progression tests.12:56
nessitaralsina: since we're developing independent widgets, we need to have unit tests for that independent widget12:56
ralsinabut sure, I can do the exact same tests passing 1 through 6 for set_stage12:56
nessitalet me see the test_stage_progression in detail, we should no be crossing boundaries12:57
ralsinanessita: sideWidget is not an independent widget. It can't be used anywhere else except as a side widget in this wizard, but don't worry, I'll repeat the tests.12:57
nessitaralsina: ok12:58
lalejandrye : "60k queue items. lalejand were these files originally in Ubuntu One or you have added them recently?" --> I think they have been added13:00
faganmoi13:00
nessitame13:00
thisfredme13:00
alecume13:00
faganmandel ralsina dobey13:00
nessitaDiego says:13:00
nessitaDONE13:00
nessitaSome corrections in the branch has been applied.13:00
nessitaTODO13:00
nessitaCreate Tests, Modified Loading animation with the new specification,13:00
nessitapolish minor ui details.13:00
nessitaBLOCKED13:00
nessitano13:00
ralsiname (no notes yet)13:00
dobeyme13:00
ryelalejand, any more MARK items?13:00
nessitamandel: ?13:01
nessitafagan: go13:01
faganlets just go he can catch up13:01
=== m_conley_away is now known as m_conley
faganDONE13:01
fagan* Got some more finished but the refactoring still is taking ages.13:01
faganTODO13:01
fagan* Finish off the branch (id say it will be done monday)13:01
faganBlocked13:01
fagan* nope13:01
faganNOTE: Birthday today so started work a little earlier so I could head off an hour early for errands for a party im having you all are invited but the plane journey would be a little expensive for even the best of parties :)13:01
fagannessita: go13:02
nessitaDONE: IRL reviews @ windows, meetings, tested UDF creation noticing that is not working, need to debug further. Reported lots of bugs in different projects.13:02
nessitaTODO: more IRL testing, meeting with design team13:02
nessitaBLOCKED: nopes13:02
nessitaNEXT: thisfred13:02
thisfredDONE: Bug #762722 Bug #806668 Bug #806500 TODO: Bug #762722 BLOCKED: No NEXT: alecu13:02
ubot4Launchpad bug 762722 in ubuntuone-control-panel (and 1 other project) "Opening the Control Panel during sync removes progress indicator (affects: 1) (heat: 9)" [High,In progress] https://launchpad.net/bugs/76272213:02
ubot4Launchpad bug 806668 in ubuntuone-client "Tests inheriting from DBusTwistedTestCase must yield setUp when calling super() (affects: 1) (heat: 6)" [Medium,Fix committed] https://launchpad.net/bugs/80666813:02
ubot4Launchpad bug 806500 in ubuntuone-client "logging in status/aggregator.py is broken (affects: 1) (heat: 6)" [Low,Fix committed] https://launchpad.net/bugs/80650013:02
alecuDONE: started a branch for TCP activation for SSO and SD; will take care of inhibiting multiple activations as well; objectives; some reviews13:02
alecuTODO: keep working on activation13:02
alecuBLOCKED: no13:02
alecuNEXT: ralsina13:02
ralsinaDONE: windows call, call with beuno about embedding pages in u1cp and wizard, bug #805405 started working on bundles, objectives with nessita, alecu13:02
ralsinaTODO: push with bundles, merge two branches I have pending on installer, calls, approve objectives.13:02
ubot4Launchpad bug 805405 in ubuntu-sso-client "In the Qt UI, errors are shown as dictionaries in popups. (affects: 1) (heat: 46)" [Undecided,Fix committed] https://launchpad.net/bugs/80540513:02
ralsinaBLOCKED: no13:02
ralsinanext: dobey13:02
dobeyλ DONE: triage, discussion on #80616313:02
dobeyλ TODO: start the shiv^Hm?, more fixes/triage/etc13:02
dobeyλ BLCK: None.13:02
dobeymandel: donde es?13:03
nessitaany comments anyone?13:03
alecuI've got two turn tables and a microphone!13:03
nessitaalecu: yey!13:03
faganalecu: nice13:03
nessitafagan: happy birthday13:03
thisfredalecu: w00t!13:03
mandelme13:03
fagannessita: thanks13:03
mandelsorry13:03
nessitamandel: go!13:04
mandelDONE: Looked and the branches that need fixing. Though about how to deal with unicode and byte path representations so that sd deal with byte but we use unicode when dealing with the Windows fs. Implemented most of it in os_helper13:04
mandelTODO: Fix test in branch refarding open_file. Finish implementation of previously mentioned solution.13:04
mandelBLOCKED: no, should have unicode, byte support this afternoon13:04
alecuthisfred, nessita, fagan: sorry, bad joke. It was a reply to "donde es?"13:04
thisfredI thought I heard a beck quote in there :)13:04
faganalecu: :D13:04
nessitaChipaca: can you meet a las y media?13:04
ChipacaNeoti: yes13:04
nessitalol13:04
ChipacaNeoti: o hi.13:04
Chipacanessita: yes13:04
nessitamandel, Chipaca, ralsina, alecu: meeting in 25 minutes (as scheduled on the calendar, I re-added the event so we don't need to sync the time every day :-))13:05
ralsinanessita: ack13:05
alecuack13:05
nessitaralsina: so, I got no ack from the deisgn crowd, did you?13:06
ralsinano13:06
thisfredalecu: at least it wasn't: "Get crazy with the Cheese Wizz"13:06
nessitapinging people will not work, they may be having lunch I guess13:06
nessitaok, mate time before mumble!13:06
mandelnessita: ok, I'm going for lunch now then13:06
nessitamandel: ejoy!13:06
nessitan*13:06
lalejandrye : "lalejand, any more MARK items?" --> didn't move since a long time13:06
mandelthx13:07
ryelalejand, could you please ctrl+c that command and then run again13:07
lalejandrye : I just deleted almost all the new files (it was cache files for some packaging operations I'm doing)13:07
ralsinanessita: pushed with separate tests for side_widget13:10
lalejandrye : for the moment it returns nothing13:10
ryelalejand, the log files are being rotated and so they disappear. could you please retry with "tail --follow=name  ~/.cache/ubuntuone/log/syncdaemon.log | grep MARK13:12
lalejandrye : nothing neither for the moment13:22
alecunessita, ping. Again: what's the convention for the message in assertTrue?13:22
alecunessita, the message should say what should happen? or what was the error?13:23
nessitaalecu: "This condition must hold"13:23
alecuack13:23
alecunessita, thanks!13:23
nessitaalecu: :-)13:23
ralsinanessita: pushed the "are you sure" branch with translator-friendly strings13:23
nessitaralsina: great! what about the tests for the are_you_sure dialog?13:24
dobeynessita: did diego fix all the issues you had in his branch?13:24
alecunessita, so the message rather than "the connection is dropped" should read "the connection must be dropped" instead. The "must" makes it clearer, right?13:24
nessitadobey: not all, he will finish today, he's taking an exam now13:24
dobeyah ok13:24
nessitaralsina: usually we define all the constants to be translated on a separated file, and import and use that from the code and from the tests. That approach has the advantage of gathering all string in one place, and ease re-use and testing13:25
* ralsina reads backlog to figure out what tests he missed on "are you sure"13:25
nessitaalecu: exactly13:25
nessitaralsina: is on the MP13:25
alecugreat13:25
ralsinanessita: adding the tests now.13:26
lalejandrye : still nothing13:26
nessitaalecu: so, if the assertion blows up, you get: AssertionError: the connection must be dropped <- since is an assertion-error, the assertion (which uses must) is not holding13:26
nessitathe water is about to boil!!!!13:27
ralsinanessita: please file a bug for moving the strings to a separate file, it's not in the scope of this branch13:30
nessitaralsina: yeah, I was saying for the new string, since you need to re use it for building the tests13:30
ryelalejand, and u1sdtool --waiting errors out, right?13:31
ralsinanessita: makes sense, will do it for the "are you sure" strings at least13:31
nessitathanks13:31
nessitaralsina, mandel, alecu, Chipaca, mumble?13:32
alecunessita, I'm coming13:33
nessitamandel: ping13:34
lalejandrye : yes. Error13:35
mandelNeoti: sorry, I'm back13:37
nessitamandel: mumble!13:37
mandelnessita: booting13:38
mandelI mean luanching13:38
lalejandrye : I'm going to stop and restart u1, hoping that it will be ok. My PC is to slow13:59
dobeyoh, i need to take a couple days off14:15
nessitadobey: we all do! :-)14:17
dobeywell, need to take several days off at some point too. :P14:18
nessitaralsina: ping14:20
ralsinanessita: pong14:20
nessitaralsina: any reason why the images are not in the installer? I'm referring to U1logo, and the icons for each "step" in the left14:20
ralsinanessita: I was expecting Diego to add those after he was finished with the u1cp14:21
nessitaralsina: ok, is that added programatically or with css?14:21
dobeyugh @ duplicating locations of things14:21
ralsinanessita: mostly CSS14:22
ralsinaor designer in some cases14:22
ralsinabut I don't think he will have to touch code anywhere14:22
nessitaralsina: ok, fair enough. Can you please file a bug to add those images? And assign to him :-)14:22
ralsinanessita: cool, on it14:22
ralsinanessita: BTW, both branches should be up to testing standards now ;-)14:23
nessitaralsina: great, I'm already IRL testing the side_widget now14:23
ralsinaThere is no bug for the bundles, is there?14:38
nessitaralsina: nopes, can you make some please?14:45
nessitaralsina: ah, I forgot to ask, any feedback on my goals?14:47
ralsinanessita: +1 on them from me14:48
ralsinaSo load them up and I will approve them14:48
nessitaack14:48
dobeyhmm14:51
ralsinadobey: yours too14:52
* ralsina wants to minimize cntext switches ;-)14:52
* fagan EOD/EOW14:54
fagansee you guys monday14:54
fagangtg go party14:54
dobeyralsina: i already put mine in allhands last week :)14:56
ralsinadobey: I meant that I will approve them14:56
dobeyok14:57
DiegoSarmenteroralsina, nessita hi! I'm back... I will finish the ui for ubuntuone-control-panel today, make the tests and propose a new merge15:04
ralsinaDiegoSarmentero: cool15:04
nessitaDiegoSarmentero: hey there! how did the exam go?15:04
DiegoSarmenteronessita, 9 :D15:05
nessitaDiegoSarmentero: groso! what subject?15:05
DiegoSarmenteronessita, Tecnologia de Software de Base15:05
DiegoSarmenteronessita, java... *escalofrios*15:05
DiegoSarmenterojejej15:05
nessitaAAAAAHHHHH15:05
nessitaDiegoSarmentero: did you shower after that? :-P15:05
DiegoSarmenteronessita, not yet jeje15:05
ralsinanessita: we all n mumble now!15:06
dobeyDiegoSarmentero: you can just fix in the existing merge, you don't need to propose it again, btw15:17
dobeyunless you are making massive other changes, anyway15:17
DiegoSarmenterodobey, ah ok.... yes, i was planning to use the same that i've created15:18
dobeyok, i'm off to get some lunch. bbiab15:21
facundobatistageneric question: how do I share a folder using nautilus? I chosen "Ubuntu One" -> "Compartir", but then I get a dialog that tells me to search something (I don't know what), it always says "found 0" nomatter what I search, and don't know how to continue15:38
=== JanC_ is now known as JanC
* mandel walking dog16:06
facundobatistaralsina, a friend of mine has the issue described in #57640816:08
facundobatistaralsina, is it really "in progress"?16:08
ralsinabug #57640816:08
ubot4Launchpad bug 576408 in ubuntuone-client ""Share on Ubuntu One" and "Stop synchronizing on Ubuntu One" menu items disabled (affects: 10) (heat: 44)" [High,In progress] https://launchpad.net/bugs/57640816:08
ralsinafacundobatista: I have half-a-branch on it16:09
ralsinafacundobatista: then got sidetracked for a few months16:09
ralsinafacundobatista: but the honest answer is "no"16:10
dobeyfacundobatista: hrmm, have we always allowed sharing from a UDF?16:13
dobeyhrmm, i guess so16:19
dobeyoh i see what's wrong16:20
dobeyfacundobatista: does your friend have the UDF that folder is in, subscribed, on the machine where this bug is happening?16:28
ralsinanessita: remember the 3PM thing for the bundles? Reschedule it for 6PM because I have to have lunch and have not advanced much.16:32
nessitaralsina: ack.16:32
nessitaLUNCH!16:32
nessita:-)16:32
alecufacundobatista, in the dialog that tells you to search for something, the list of contacts from evolution-data-server are shown.16:38
alecufacundobatista, you can just enter an email address, and it will be used and added to e-d-s.16:38
facundobatistadobey, asking him right now16:53
facundobatistaalecu, I don't use evolution16:54
alecufacundobatista, Me neither! You can just enter an email address.16:54
facundobatistaalecu, where? in the search field?16:55
alecuright16:56
dobeyyeah, the contact picker UI sucks. and i think one of the requirements for TB being default in Oneiric, was integrating with EDS16:57
* dobey wonders why anyone is still running maverick anyway :)16:57
facundobatistaalecu, oh, "magically" appears one button that wasn't there before16:58
alecufacundobatista, yeah, it's awful. And I've been doing some tests connecting e-d-s with canonical's ldap, and now that dialog is painfully slow.16:59
lalejandHi, U1 began to upload some large files. I deleted them, but U1 indicator still shows that U1 is uploading those files17:00
alecuI'll open a bug for that.17:00
lalejandalecu : which package ?17:00
aleculalejand, sorry, was talking about a different issue.17:01
dobeylalejand: i think he's talking about something else17:01
dobeylalejand: if you want to open a bug against u1 indicator, i think you need to file against one-indicator17:02
alecufacundobatista, bug #807111. Perhaps you want to open a bug for the "magical" button.17:04
ubot4Launchpad bug 807111 in ubuntuone-client "The share contact picker is slow for 400 contacts (affects: 1) (heat: 6)" [Medium,Confirmed] https://launchpad.net/bugs/80711117:04
nessitamandel: ping17:04
lalejanddobey : I don't know if it is u1-indicator's fault or not17:05
mandelnessita: pong?17:06
dobeylalejand: no, but filing a bug against the thing where you see the problem is a good start. if it's not the indicator applet's fault, then the bug can be moved. it's more about documenting it, than solving it17:07
nessitamandel: how are branches going? (I'm catching up since you said 2 hours after the meeting)17:07
mandelnessita: changes have been done but I have a failing test in test_vm due to having no access rights for some strange reason, I'm debugging right now17:08
nessitamandel: that is for the unicode branch, right? any news on the in-review branch?17:09
mandelnessita: I'm stuck with the unicode one, is more important since it should fix lots of issues17:09
=== beuno is now known as beuno-lunch
lalejanddobey : okay17:10
Chipacanessita: ping-u17:16
nessitaChipaca: pong17:21
Chipacanessita: mumble plz?17:21
nessitaChipaca: I'm about to have lunch (nuddles almost ready). Is it talkable in less than 3 min? :-)17:21
Chipacanessita: have lunch. try noodles instead nuddles next time.17:22
nessitaChipaca: :-)17:22
* nessita -> lunch, brb17:22
ralsinahttp://www.nuddleblanket.com/Celebs.html17:24
ralsinaLUNCH!17:31
nessitaChipaca: I'm back, logging in...17:42
nessitaChipaca: ready when you are17:44
Chipacanessita: ok, let's do this17:45
nessitalet's!17:45
dobeyhttps://code.launchpad.net/~dobey/ubuntuone-client/no-pycentral/+merge/6723117:57
dobeypretty trivial review if anyone wants to do it :)17:57
Chipacaralsina: when you return, mumble plz17:59
ralsinachipaca, now is ok?18:01
Chipacaralsina: depends, are you still chewing?18:01
ralsinaChipaca: no, just having coffee18:01
ralsinaSo don't worry!18:02
nessitaralsina: when you finish your mumbling and chewing, I have a question/suggestion re: bandwith settings18:02
ralsinanessita: ok!18:02
facundobatistaalecu, ping18:02
alecufacundobatista, pong18:02
facundobatistaalecu, maybe you saw something like this when setting up your windoze vm?18:03
facundobatistapywintypes.error: (1332, 'LookupAccountName', 'No se ha efectuado ninguna asignaci\xf3n entre los nombres de cuenta y los identificadores de seguridad.')18:03
alecufacundobatista, while running the tests?18:03
facundobatistaalecu, I get that when starting ubuntuone-client, mandel has no idea of what could be happening18:03
mandelalecu: first time I see it, sounds like chinese too me… or an exchange thing, but I have no idea18:04
alecumandel, isn't that related to the pywin32 module you've recompiled?18:04
ralsinanessita: finished chewing and mumbling18:05
alecufacundobatista, I have no further idea than the pywin32 module that mandel fixed.18:05
ralsinadobey: checking!18:05
nessitaralsina: that was fast :-)18:05
ralsinanessita: yes it was!18:06
ralsinaOr "that's what she said" to preempt dobey & mandel18:06
nessitaralsina: so, I was looking at the installer wireframes (https://docs.google.com/a/canonical.com/drawings/d/10Io1nRr0mGr_EW5mHTKErS6s2oxVnF7uCDdIzuqYB50/edit?hl=en_US), in particular wireframe 1918:06
dobeyheh18:07
ralsinanessita: give me 1' to reach the right thingy18:07
ralsina19, yes!18:07
nessitaralsina: if we could agree with design crowd to change the button "Check bandwidth settings" to "Check file sync settings", and show as frame 19 the same screen as the Preferences tab (from https://docs.google.com/a/canonical.com/drawings/d/1O_Q9EIJNbvnimroBTvoxvpFXuawEp8pA01iyfW-lHn4/edit?hl=en_US), we could save A  LOT of work18:07
nessitaralsina: otherwise, separating the BW settings gets hairy due to the "restore default settings" and "apply" buttons18:08
nessitaralsina: you think proposing that is a good idea?18:08
ralsinanessita: I can always override the text on the wizard, if needed18:08
nessitaralsina: override what?18:08
ralsinaBecause in this case, "bandwidth" makes more sense (to me at least)18:08
mandel1dobey: I think we are doing quite well when the n-1 ihas started with the jokes too :)18:08
nessitaralsina: not sure if you understood my proposal (given that comment)18:09
dobey:)18:09
ralsinanessita: wait a sec, haven't checked the other wireframe you mentioned, o I may nort make sense right now18:09
nessitaright :-)18:09
ralsinaI am "the n-1"?18:09
ralsinanessita: I get it now. With the extra options, it makes sense to keep the text18:10
* ralsina of course, has a perfectly hackish solution for this problem that would work perfectly18:10
nessitaralsina: to keep or to change the text? (I'm proposing changing the text)18:10
ralsinakeep control panel's, change the wizard's18:10
nessitaralsina: wait, I still think you are not following me :-)18:11
ralsinaYou want to replace wireframe 19 with your perferences tab widget18:11
nessitaralsina: well, we embed the preferences tab into that step18:11
ralsinaAnd the button in 18 with another text18:11
nessitaralsina: right, so no change of text needed in the control panel18:12
nessitaralsina: and within 18, you remove the checkbox under folders18:12
nessitasince we provide that in 1918:12
* mandel1 has to go to the roof.. dont ask18:12
ralsinaI am ok with both changes. If, however, design says it's not good because it's too geeky or whatever, there is another way to do it, which is "use the widget from u1cp, and hide the parts you don't want visible "18:12
* ralsina ducks18:13
nessitaralsina: that can work, yes, but each you click on Apply, the whole settings will be changed (hidden and not hidden)18:13
nessitaralsina: so I will strongly discourage that18:14
ralsinabut the hidden settings will not change anything, since they will contain the value from the config file anyway?18:14
ralsinaBut I only propose that as a last resource if we can't make the change18:14
nessitaralsina: it depends, are you offering to change settings somewhere else? like in frame 18?18:14
nessitaif so, we have a mismatch there18:14
ralsinanessita: good question18:14
mandel1facundobatista: ping18:14
ralsinaI agree that the setting in page 18 should move to 19 and the button's wording change18:15
ralsina*at least*18:15
mandel1verterok: ping18:15
facundobatistamandel1, pong18:15
ralsinaAnd I don't think design is going to have problems with that18:15
nessitaralsina: great. Can you please email/talk this with design people? (they already left, I think)18:15
mandel1facundobatista: one question, the share.path is that unicode or byte?18:15
ralsinaafk, brb18:15
nessitaralsina: or better, can you implement it that way (is very easy) and see what they say by looking at that?>18:16
facundobatistamandel1, it should be bytes, IIRC (verterok?)18:17
mandelfacundobatista: also, udf.ancestors, are those meant to be using local paths, look at test_udf_ancestors in test_vm (line 1333)18:17
facundobatistamandel, stop being schyzofrenic18:17
mandelfacundobatista: the expected result uses / is that correct for windows too?18:17
mandelfacundobatista: ein?18:18
facundobatistamandel / mandel1 :)18:18
mandelfacundobatista: yes, stupid irc client seems to have a bad day...18:18
facundobatistamandel, the expert in that area is verterok, but I think that udf.ancestors is all local path, so we really should be using os.path.sep, no '/'18:19
mandelfacundobatista: I though so too.. verterok is that correct?18:19
mandelI know suggested path is correct and should be using /18:19
ralsinanessita: you mean maybe they won't otice? ;-)18:20
ralsinas/otice/notice/18:20
nessitano! :-)18:20
nessitaI meant they may like better by looking at it18:20
nessitanot by trying to imagine it18:20
ralsinaagreed. Since it's much MUCH easier, we can hack if it is not good enough18:20
ralsinaIn fact, I think that page is meaningless as part of the first setup, but hey, what do I know18:21
facundobatistamandel, he was having lunch18:21
mandelfacundobatista:  np :)18:21
mandelfacundobatista: so, share.path is that byte of unicode, I was let down my irc just a second after I asked you the question...18:22
nessitaalecu: ping18:25
nessitaalecu: does this ring any bell? bug #807146 (I've assigned it to you for now, we'll see later who has an empty slot to work on that)18:26
ubot4Launchpad bug 807146 in ubuntu-sso-client "Windows: twisted.spread.pb.DeadReferenceError: Calling Stale Broker (affects: 1) (heat: 6)" [Medium,Triaged] https://launchpad.net/bugs/80714618:26
alecunessita, pong. looking18:26
alecunessita, yes, it rings18:26
nessitaalecu: is that already reported?18:27
alecunessita, I think that happens when a client disconnects and the server still wants to call the signals.18:27
alecunessita, I think it's reported, yes, but let me check.18:27
nessitamakes sense, given what I described happened18:27
ralsinanessita: I get that all the time with SSO, it answers only one cnnection18:28
nessitaralsina: you reported it somewhere? (is not good, it should answer to several)18:28
ralsinanessita: no, because I thought it was related to the reactor problems we were having and later I forgot18:29
nessitaok18:29
mandelverterok: ping18:31
mandelfacundobatista: ping18:31
* mandel hates his irc client...18:31
alecunessita, I'm not sure again if it's the same bug, but it happens under similar conditions: bug #80653918:31
ralsinawhat version of ubuntu-client are we to be shipping on windows?18:31
ubot4Launchpad bug 806539 in ubuntuone-client (and 1 other project) "Cleanup API to connect to services (affects: 1) (heat: 6)" [Medium,Triaged] https://launchpad.net/bugs/80653918:31
ralsinaubuntuone-client that is18:31
alecunessita, for now let's keep this as two separated bugs.18:31
nessitaalecu: right, the error does not look the same, so yes18:32
nessitaalecu: FYI, the last report is 100% reproducible (at least here(18:34
nessita))18:34
alecugreat18:35
nessitaralsina: still reviewing your side-widget branch. Question: right after I logged in, the installer skipped the "Choose services" section marked in the left518:36
nessitaand went thru directly to Select sync folders18:36
nessitaralsina: also, the "add folder" does nothing, is that correct?18:37
alecunessita, ralsina: can I get reviews on this first branch of tcp activation? https://code.launchpad.net/~alecu/ubuntu-sso-client/tcp-activation/+merge/6724018:37
ralsinaChoose services is the one where we will embed the browser18:37
alecuthis branch has all of the functionality; in a coming branch I'll make sso use it.18:37
ralsinaAnd "add folder" is waiting for the branch that embeds the actual u1cp widgets18:37
alecubut first, lunch.18:37
ralsinabranch which I should finish now that u1cp can talk to syncdaemon :-)18:37
nessitaralsina: ok, so I'm lost :-D both "issues" are expected and "normal"?18:38
ralsinaAs of that branch ,yes, expected and normal18:38
=== beuno-lunch is now known as beuno
nessitaralsina: and the folders tab is also outdated, in that branch, right?18:41
ralsinanessita: yes18:42
ralsinaDamn distiutils-extra doesn't work if your setup.py is not called "setup.py"!!!!18:42
nessitawhat?18:42
ralsinanessita: http://pastebin.ubuntu.com/639665/18:43
ralsinaI will just do it without distutils-extra18:43
nessitaralsina: in the tests for the sidewidget branch, in all the test_stage_progression_X, I'm not why you have all the assertion over the state of self.ui.sideWidget(). From my POV, all the test_stage_progression_X shoul assert only over the sideWidget.stage value, and the rest is ensure by the sideWidget API and functionality18:47
nessitaotehrwise you're duplicating assertions that, if we sightly change what the sidewidget does, we need to modify them all18:48
ralsinanessita: because those were done first, before the other ones. I can remove that, no problem.18:48
nessitaralsina: thanks! it will make that code much easier to read and understand (and maintain)18:49
ralsinanessita: ok, changing and pushing18:49
facundobatistamandel, pong18:49
facundobatistamandel, which irc client is yours?18:49
verterokmandel: pong18:51
dobeyralsina: were you reviewing my branch btw?18:53
ralsinadobey: sorry, got sidetracked18:53
ralsinadobey: reviewing now18:53
dobeyralsina: hrmm, why would you not call the setup.py "setup.py"?18:54
ralsinanessita: pushed without all those asserts18:54
nessitaralsina: great!18:54
ralsinadobey: because I don't want someone running it on linux18:54
ralsinaso I was calling it setup-windows.py18:54
dobeyif sys.platform == 'win32': setup(); else: print "GTFO."18:55
dobey:)18:55
dobeyralsina: hrmm, that's in ubuntuone-client? :(18:56
ralsinadobey: yes18:56
ralsinadobey: it's basically to build the exe18:56
czajkowskimandel: how do18:57
dobeyralsina: yeah, i know. but still. :(18:57
dobeytrying to think of a solution that makes sense, but alas18:57
nessitaralsina: the branch looks great. Is pretty much approved, I just listed some lint and pep8 issues19:02
ralsinanessita: I will try to get pep8 0.5 in windows because I got no pep8 issues here, again :-(19:02
nessitaralsina: yeah, I imagine. I ran it in linux though19:02
ralsinanessita: ok, will fix them soonish19:03
dobeynessita, ralsina: i wonder what the acceptable requirements are for building u1client stuff on windows.19:03
ralsinadobey: don't understand the question19:03
nessitadobey: not sure what you mean. The acceptable requirements would be that "it builds and installs" :-)19:03
dobeyralsina, nessita: i mean as far as fixing the build system issue. there are some things there which i think we still need to do on windows, but which aren't so easily doable from bzr checkout, but aren't an issue when installing from tarball, for example19:05
ralsinadobey: such as?19:05
nessitaDiegoSarmentero: ping19:05
dobeyralsina: PNG icon generation from the SVG19:06
ralsinadobey: how's that converted now?19:07
mandelverterok: ping again, sorry I missed the pong19:07
dobeyralsina: running inkscape during 'make' to generate the icons inside the build tree19:07
mandelfacundobatista: ping again..19:07
verterokmandel: hi19:07
mandelczajkowski: how do?19:07
facundobatistamandel, pong19:07
facundobatistamandel, which irc client is yours?19:08
ralsinadobey: we are making inkscape a build dependency? Whoa.19:08
mandelverterok: question, the ancestors from udf, do the use local paths (that is os.path.sep) or do they always use /19:08
mandelfacundobatista: adium… terrible client19:08
dobeyralsina: it has been for a long time19:08
ralsinastill19:08
dobeyralsina: well, to build from trunk, not from tarballs19:09
verterokmandel: I think the ancestors are calculated from the local path19:09
mandelfacundobatista: share.path is that meant to be unicode?19:09
mandelverterok: ok, so in the test_ancestors in test_vm we should be using os.path.sep, right?19:09
verterokmandel: no idea, need to look at the code :)19:10
mandelverterok: hehe, want the file and line number?19:10
verterokmandel: no need to, gimme a few minutes to wrap what I'm doing19:10
mandelverterok: gracias!19:11
nessitaralsina: a couple of extra needs fixing in the areyousure branch, after code review19:14
ralsinanessita: ok19:14
nessitaralsina: also, maybe you forgot to add the test_are_you_sure file? I don't see it19:15
ralsinanessita: turns out I did forget it. Pushed it.19:16
nessitagreat19:16
verterokmandel: the path should be the "native" path, as it's used during local rescan to add pyinotify watches to the udf ancestors19:23
mandelverterok: ok, so I fixed the test :)19:23
mandellooks like this branch is looking good :)19:24
* mandel dinner, will be back later19:24
nessitaralsina: SyntaxError: invalid syntax (test_are_you_sure.py: 38)19:24
verterokmandel: and probably you need to check the ancestors property in volume_manager.py @ line 31619:25
ralsinanessita: uh? Let me check19:25
dobeyok, gotta run to an appt. be back in a few19:26
nessitaralsina: also, when running that branch and clicking on "Yes I want to cancel", I get an error in the terminal saying "QWaitCondition: Destroyed while threads are still waiting"19:26
ralsinanessita: so sorry, seems I had not actually finished that branch before I started eith the phone calls19:26
nessitaralsina: is ok, let me know when is done :-)19:27
ralsinanessita: pushed with the fix and tests pass now19:28
ralsinanessita: but let's try again that branch tomorrow ;-)19:28
nessitaralsina: not sure what that means :-)19:30
ralsinanessita: that I dn't want to waste your time before double-checking the branch again19:30
nessitaralsina: I'm trying to have a screenshot of the latest installer since DiegoSarmentero will start working on that tomorrow19:30
ralsinathe changes should not affect screenshots :-)19:31
nessitaralsina: well, having the latest may be better, specially the side widget19:31
nessitabut yes, is not mandatory19:31
ralsinanessita: ok, pushed , passes pep8, pylint, pyflakes and tests ;-)19:32
nessitaDiegoSarmentero: https://wiki.ubuntu.com/UbuntuOne/Contribute/WindowsTesting <- a new section for the installer was added, running it is extremely similar to running the control panel19:32
nessitaralsina: is that for side widget or are you sure? :-D19:32
ralsinano, the areyousure19:32
nessitaralsina: are you sure? :-P19:32
nessitaack19:32
ralsinaI'll fix the sidewidget one now so diego can have that one, which is more important19:33
nessitaralsina: ack19:33
DiegoSarmenteroralsina, ack19:33
nessitaralsina: is tarmac landing your installer branches?19:34
nessitaralsina: in the are you sure branch... did you apply the latest needs fixing from https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_805290/+merge/66739/comments/145569 ? (I don't have them)19:36
ralsinanessita: had not seen them. Will fix and push. The problem with using clientdefs is that I basically import all of control panel to get a string, because of how the package is organized. But I'll do it anyway.19:40
nessitaralsina: all the control panel? wow, why?19:44
nessita(clientdefs is inside client)19:44
ralsinanessita: because it imports the ubuntuone namespace, which imports controlpanel, sso, client and whatever else is in there because it has no __ALL__19:44
nessitaralsina: I see. Anyway, installer already depend on control panel, right?19:45
ralsinanessita: yes, which is why I did it anyway ;-)19:45
nessita:-)19:45
ralsinanessita: fixed the needs_fixing in https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_805290/+merge/66739/19:50
ralsinamoving onto the other one19:50
nessitaack19:50
ralsinanessita: you are using pylint? We moved to pyflakes for this one19:55
nessitaralsina: why?19:55
ralsinadobey suggested it19:55
nessitaralsina: errors from pylint makes sense here19:55
ralsinapylint ignores don't work on windows19:55
nessitaI mean, those errors are not false positives19:55
ralsinawhich means I have to re-check all the sources in another OS19:55
ralsinabut ok, I can do it19:55
nessitaralsina: well, checking the sources in all platform is a requirement for all of us19:56
nessitaanyways, I wasn't aware pylint ignores the ignore19:56
ralsinathe style checks should not be platform dependent?19:56
nessitashould not be19:56
nessitaralsina: running the tests in all paltform is the requirement, I mean. But if the ignore switch is not working, I see the problem there19:57
ralsinaexactly. pylint gives about 1200 lines of output on windows19:57
ralsinacomplaining about gtk, dbus, how QWidget is not a valid parent class, how all tests are using super in old style classes and all sorts of wacky stuff19:58
nessitaralsina: pylint --help does not list the ignore swicth :-/19:59
ralsinait's on the pylintrc19:59
nessitaso seems like is not implemented, at least in the version I'm suing19:59
nessitausing*19:59
nessitaralsina: what is in the pylintrc?20:00
ralsinawhich is, of course, undocumented everywhere20:00
ralsinathe ignore option20:00
nessitaralsina: right, but the pylint --help do not list that as a valid option, so seems like is not implemented in w :-/20:01
ralsinait's not implemented as an option in the CLI, only from the pylintrc config file20:01
ralsinain Linux, too20:01
nessitaralsina: I beg to differ, you can use pylint --ignore=something, that is what we use in the control panel20:02
ralsinanessita: did you try pylint --help on linux?20:03
nessitayes20:03
nessita    --ignore=<file>     Add <file or directory> to the black list. It should20:03
nessita                        be a base name, not a path. You may set this option20:03
nessita                        multiple times. [current: CVS]20:03
ralsinathe help in windows says exactly the same thing: http://pastebin.ubuntu.com/639703/20:04
ralsinaExcept it didn't work when I tried it, and that's why I talked with dobey and switched to pyflakes20:05
ralsinaor rather, added USE_PYFLAKES for u1lint20:05
nessitaralsina: and it seems to work on windows: I just ran pylint --ignore="ui,reactor" ubuntuone_installer and I got only failures/errors for the rest of the project20:05
ralsina(maybe it was a u1lint bug, not a pylint one)20:05
nessitaralsina: maybe you're mixing u1lint with pylint?20:05
nessitaright20:07
nessitaralsina: ok, we can leave using pylint and fixing the errors in another branch20:07
ralsinanessita: this is what I get running pylint directly, with --ignore="ui,reactor"20:07
ralsinahttp://pastebin.ubuntu.com/639705/20:07
alecuBTW: if anybody wants review a nice branch... https://code.launchpad.net/~alecu/ubuntu-sso-client/tcp-activation/+merge/6724020:07
nessitaalecu: right, sorry, I'll do it20:07
alecunessita, thanks!20:08
ralsinanessita: I think you'll agree it's not a very useful output :-)20:08
nessitaralsina: is very useful right below the PyQT4 import, and we alrady fixed that (we need to import QtCore before QTGui or something similar)20:09
nessitaI mean, a lot of errors there make sense, o no?20:09
ralsinanessita: from line 60 until 265 they don't20:10
ralsinaactually from 60 until the end20:10
nessitaralsina: I'm not sure why you have that, look at my output http://pastebin.ubuntu.com/639707/20:10
nessitamaybe some installation is a bit busted?20:11
ralsinanessita: well, that's the problem, how do I figure that out? Should I fight pylint for a couple of hours?20:11
nessitaanyways, that can be done later. Can you at least fix the unused variables I reported?20:11
nessitaW0612: 57:CalculateSize.run: Unused variable 'dirnames'20:12
nessitaW0612:149:LocalFoldersPage.stop_threads: Unused variable 'path'20:12
nessitaand all the missing docstrings20:12
ralsinaI added all the missing docstrings. What's a variable name I can use that doesn't give that warning besides _?20:12
nessitaralsina: to not using it? no other... are those unused variables, needed?20:13
ralsinafor example, when you use listdir() it returns tuples.20:13
ralsinaI am doing for dirpath, dirname, filenames in os.walk(self.path_name):20:13
nessitaright, so you can use _ there20:14
nessitafor _ , _, filenames in ...:20:14
ralsinabut then it complains I am overriding _ from the outer scope, which I can probably fix by moving strings to another file.20:14
nessitaright, which is even better! :-D (so we have all the to-translate-variables together)20:15
ralsina Invalid name "bad" .... pylint I hate you ;-)20:22
alecuguys and gals: got to run to kinder. See you laters.20:25
ralsinanessita: pushed sidewidget with the style fixes, pylint seems clean to me now, but I may be missing something in the noise20:34
nessitaralsina: great20:35
dobeynessita: because pylint is a royal PIA, and I'd like us to switch everything over to pyflakes if possible20:36
ralsinadobey: OTOH pyflakes doesn't detect some real problems (like missing docstrings)20:37
dobeyralsina: missing docstrings isn't a functional one, but a style issue.20:37
ralsinadobey: yes, but we were using pylint for style checks20:37
ralsina(too)20:38
dobeywell we were using pylint because it's what we started with 3 years ago20:38
dobeywe weren't exactly using *for* style checks per se, it just also does them20:38
dobeyralsina: is the docstring thing the only additional thing here? i'd really like to get us a solution that moves us completely off of pylint, because it's a pain20:40
ralsinadobey: bad top-level variable names, at least20:41
ralsinaI don't have a full list, as you may imagine :-)20:41
dobeyralsina: well the toplevel variable names aren't bad necessarily either. it is just a style thing we're imposing via pylint :-/20:41
dobeyralsina: also, pep8 is extensible, so we could presumably add these additional style checks to our pep8 checks20:42
ralsinadobey: yes, but consistent style in a team is a good thing. I don't have strong feelings. I was happy with pyflakes because it never complained ;-)20:42
ralsinaDiegoSarmentero: in case it's not merged whe you want to start working on styling the wizard: you should really use trunk merged with https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_803929/+merge/6682720:44
ralsinaso you should branch from that, and use it as a prerequisite for your (eventual) merge proposal20:44
DiegoSarmenteroralsina, ok! I will clone that repo20:45
ralsinaThere are two or three other branches that will affect you but that one is a big thing visible on all pages :-)20:45
nessitaralsina: let's merge those babies!20:48
nessitaralsina: I'll re-review20:48
ralsinanessita: do sidewidget, the other one is going to conflict like crazy now20:49
nessitaoooohhhoook20:49
ralsinabecause sidewidget grew tentacles with the style fixes :-)20:49
nessitaralsina: looks great, approving!20:50
ralsinanessita: should I merge?20:50
nessitaralsina: I guess so, yes20:50
nessitaI think I meeds with your head enough :-D20:50
nessitamessed*20:50
ralsinahehehe20:50
ralsinaMy excuse is that we were not using the same stylechecks20:51
ralsinaafter this one merges, I should have areyousure de-conflicted quickly20:51
nessitaralsina: ping me when that's done20:52
ralsinanessita: sure20:52
ralsinawhere's ubuntuone.storageprotocol supposed to come from? is it generated?20:53
* ralsina misses grep so much20:53
nessitaralsina: from the project ubuntuone-storage-protocol. See https://wiki.ubuntu.com/UbuntuOne/Contribute/WindowsTesting21:00
nessitaralsina: instructions are there21:00
ralsinanessita: cool, fetching21:00
dobeyfacundobatista: did you ever get an answer from your friend?21:05
facundobatistadobey, it was in an UDF, yes, a locally created UDF21:05
facundobatistadobey, which should be subscribed, but he didn't have access to that computer at that time to check that21:05
dobeyfacundobatista: ok, weird21:06
ralsinanessita: the syncdaemon bundle is close, i think.21:14
=== m_conley is now known as m_conley_away
nessitaralsina: you got your setup-windows.py working? :-)21:14
ralsinanessita: it's mostly building, checking what module fails, add, retry21:14
ralsinaI am having problems with storageprotocol, yet21:14
nessitaah, you need protoc21:15
nessitaand the protobuf module21:15
ralsinait works, it doesn't bundle21:15
ralsinabut it's probably because of zipped eggs21:16
nessitaah21:17
ralsinaneed to hack some more, will try to have it by tomorrow early21:17
* nessita no idea21:17
nessitaralsina: good luck21:17
ralsinathx!21:17
ralsinawill EOD for a little bit21:17
thisfredfacundobatista: not much luck yet with bug #807005, but I'm pretty sure that the problem originated with the original upload (i.e. '2011-07-06.191137-0300ART.txt'). You did not see a notification for that one right? (I know you may have missed it, but if you did happen to see it, it will disprove my current theory)21:18
ubot4Launchpad bug 807005 in ubuntuone-client "Filename in notification does not reset. (affects: 1) (heat: 6)" [Medium,Confirmed] https://launchpad.net/bugs/80700521:18
thisfredThe actual showing of the notification resets everything, so my guess is, sometimes notifications are not shown (you indicated that that happened as well) and then the state is never reset21:19
thisfredeven if that is the case, no idea why that would happen yet. I'm digging through a forest of Twisted callback timeouts21:20
facundobatistathisfred, it was too early (those files are cron-moved into u1)21:21
facundobatistathisfred, also, not all notifications go to a bubble21:21
thisfredIf you happen to have the part of the log that deals with that file, it may help21:21
facundobatistathisfred, I do21:21
thisfredfacundobatista: I know, but if they don't, the reset should still happen21:22
thisfredthis could be where the bug is21:22
facundobatistathisfred, the part of the syncdaemon's log, or status'?21:22
thisfredfacundobatista: both, if there is anything in status, but I doubt that21:23
thisfredI think I'm going to add errbacks to everything, as I don't think they are there21:24
facundobatistathisfred, I'm using your branch since midday, not early morning, so I have only syncdaemon's log21:24
thisfredand have them log21:24
thisfredfacundobatista: that's fine21:24
nessitaok, I'm off for the day21:25
nessitabye all!21:25
facundobatistathisfred, no, I have both logs21:27
facundobatistathisfred, I have to tell you... status log lines are not very informative :|21:27
thisfredfacundobatista: I know, working on that too21:27
thisfredlike adding a metric shitton more of them21:28
facundobatistathisfred, logs sent21:29
dobeyadorilson: hi21:29
adorilsonhi, dobey21:30
thisfredfacundobatista: thanks! will continue investigating tonight or tomorrow, have to walk the dog first21:30
dobeyadorilson: can you change your diff in bug #801938 into a branch committed with --fixes=lp:801938 and propose it for merging please?21:30
ubot4Launchpad bug 801938 in ubuntuone-client (Ubuntu Oneiric) (and 3 other projects) "Make command gets error: variable 'result' set but not used (affects: 1) (heat: 8)" [Medium,Triaged] https://launchpad.net/bugs/80193821:30
facundobatistathisfred, thank you!21:30
adorilsondobey: of course21:33
dobeyadorilson: thanks!21:33
dobeyalright, i am out of here. good evening all22:01
karnihuh. why is this channel not listed on https://wiki.ubuntu.com/IRC/ChannelList ?22:21
karniIf you guys haven't yet, I strongly encourage to "Preview Gmail's new look", I love it.22:57

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