TLEGoodmorning everyone09:00
TLEdpm: can you give me an update on the build of natty lang packs at some point during today?09:01
dpmTLE, I certainly can, even straight away: the full langpack was made available last night -> https://translations.launchpad.net/ubuntu/natty/+language-packs09:02
TLEnice, will you ask pitti to copy them to proposed or should I?09:03
dpmTLE, for full langpacks, they need to be built manually, not copied. I'm a bit swamped these days, would you mind asking him? (feel free to CC me if you send an e-mail)09:05
TLEnp, will do09:06
dpmthanks a lot TLE09:06
dpmTLE, I've just pulled r27 from your docs screenshot viewer branch and I'm still getting the same problem. Could it be that the database needs to be populated or something?09:34
TLEI don't think so, the only thing that is in the database is information on the paths, the rest is pulled from the filstructure09:34
TLEdo you have a link?09:35
TLEit has to somehow be tied in with the new datastructure, because you can see that there is the right number of languages09:36
TLEall right, all the missing information on the project page is pulled by quiring a method from a object09:39
dpmI wonder if the docs branches are ok?09:41
dpmlet me check09:41
TLEsorry I have to go untill after lunch, I'll be happy to help, but I can't untill at that point09:45
dpmTLE, no worries. Here's something for when you come back. I've tracked it down to be the .get_*() functions in the ImageCollection object. While the DB returns a correct language set, it seems that calling e.g. get_common_name from the template on any of the languages returns an empty string10:22
dpmah, here's some debugging output:
dpmTLE, it seems get_common randomly fails. As in randomly, I mean most of the time, but it sometimes does return the right value10:41
=== greyback is now known as greyback|lunch
TLEdpm: the fact that it happens every once in a while is very odd, since it means that something must fail everyonce in a while11:56
TLEbut all the stuff I do, should re reproducable11:56
dpmTLE, yeah, I cannot reproduce it locally on django 1.3, and I still haven't figured out how to best debug a django app on the server. 'python manage.py shell' on the server return the right results, it's just the template that doesn't11:58
TLEme neither on django 1.311:59
TLEcan you get print statements into a log or something11:59
TLEthe only thing I do which I can think off, that I'm not sure is totally all right in all versions of pyton and so on is that I save all the ImageCollections in Paths local name space dict: http://bazaar.launchpad.net/~k-nielsen81/translated-documentation-screenshot-viewer/trunk/view/head:/image_projects/paths.py#L7612:03
TLEbut I really don't that's a problem, and I don't see why it would make it fail randomly12:03
dpmTLE, I kind of get the print statements on the template itself -> (btw, this time it worked for me, but I'm sure that if I refresh the page, the output will be empty again). I guess next step would probably be to get a log of the http requests12:04
TLEhmm, just read an interesting thread, if the method call in the template generates an exception, django will catch it, but we could just try an call that method in the view function in stead, to see if the method fails12:08
TLEso do something like: for l in langs:l.get_common_name()    at line 28 in views.py12:09
dpmTLE, regardless of this issue, have you thought about making more use of the DB capabilities and models in Django? I.e. use models instead of collections, that is, making it less pythonish and more djangoish, if that makes sense :)12:13
TLEyes, I have thought about it12:16
TLEbut have leaned against not doing it12:16
TLEI think, despite for the fact that this is not working right now, it seems like a reasonable solution12:18
TLEit saves me the problem of having to write a script that update the database after the local archives are updated12:19
TLEand, since it is so small amount of data that has to be pulled, i don't see it as a large problem to hit the filesystem for it12:20
=== greyback|lunch is now known as greyback
TLEespecially not since I have to hit it for the figure anyway12:21
TLEso unless there is some reason that I'm missing I don't see a large drive for it, but I am open to other arguments :)12:25
dpmit was just a thought, I haven't been diving deep into the code yet12:28
TLEWell, at the core I get the path (components) from the database and get the list languages and the list of images for a language by means of the filesystem12:30
dpmI think we'll need a command script to do a bzr pull every now and then. The database could be updated in that script (again, I'm not familiar enough with the code yet to argue for one option or the other, just throwing in some ideas)12:33
TLEyes, it is definitely doable, what I'm struggling with are the advantages, especiall because the data amounts are so small, but I'll give it some thought12:37
dpmI think it's more like diving into Django's MVC model and philosophy for the long term, rather than any e.g. performance advantages12:43
dpmunrelated to that, this bug is really annoying me now. /me puts on hacking hat.12:44
TLEyes ok :: do you have time for it?12:45
TLEwhat color is it btw?12:46
dpmTLE, yeah, it's my 20% day today, although I haven't been able to do much of it today, due to other tasks. What do you mean? The hat? Certainly not red ;-)12:49
TLEahh yes :: yes the hat12:49
dpmTLE, I've just reset the database in my local system and I can no longer see the list of project in the index page. Do I need to add anything manually to the DB to see them?12:54
=== bulldog98_ is now known as bulldog98
TLEwell on that note, while the bug is annoying, and pretty wierd, if we are thinking about changing the datastructure again maybe it is not worth it to spend to much time on it, it just means that people wont have pixels sizes on the images untill I find time to do the hacking12:55
TLEyes you do12:55
TLEhold on12:56
TLEis the data12:59
TLEor you can just pull the vldi file from the repo13:00
dpmok, cool, thanks13:03
TLEdpm: I couldn't provoke an error now, did you do something?13:17
TLEahh damn, there it was13:18
dpmTLE, no, I've noticed sometimes it fails less often. Try refreshing the page a couple of times, and it should trigger the error13:18
dpmyeah :/13:18
TLEwell I did refresh it a lot of time, but obviously not enough13:18
=== bulldog98_ is now known as bulldog98
dpmkelemengabor, TLE, all set for the call?15:58
kelemengabordpm: uh, not yet15:59
dpmkelemengabor, no worries, I'm going to start it for the others, feel free to join when all set16:00
dpmkelemengabor, is this your user? If so, I'll add it to the translations board https://trello.com/gaborkelemen17:00
kelemengaboryup, it has my email :)17:01
kelemengabordpm: now, I know what did I wanted to bring up: translating universe packages. What's up with that project?17:04
dpmkelemengabor, yeah, I couldn't see the e-mail, just the user. Ok, added to you to the translations board17:04
dpmyou now simply need to accept the invite17:04
kelemengaboralso the https://wiki.ubuntu.com/Translations/Meetings/ToDo is updated17:05
dpmcool, I'll move the actions to the board, then17:05
dpmkelemengabor, re: universe packages, it's still the same status as discussed a while back - the feature is implemented in LP, it seems to work (the arkose package is using it). I don't want to make it a big switch and include all universe packages at once, so I suggested starting with a selected set of packages17:07
dpmandrejz was interested in driving the project, and started a list of packages17:07
dpmbut we haven't talked about it in a while. I can't remember whether there was any blocker, I don't think there was17:08
dpmkelemengabor, ok, I've added a bunch of actions: https://trello.com/board/translations-team/4f621c87861db54230b9ca3917:43
kelemengaborlooks good :)17:45
jokerdinowooh trello board17:55

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