=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== salgado [~salgado@200-206-134-238.async.com.br] has joined #launchpad
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: changed bug assignment statuses to the more clearly-named New, Accepted, Rejected and Fixed to overcome usability problems experienced while testing (patch-814)12:37
=== salgado_ [~salgado@200-206-134-238.async.com.br] has joined #launchpad
=== kiko_ [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== salgado_ [~salgado@200-206-134-238.async.com.br] has joined #launchpad
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
BradBOuch, that <pre> change on bug messages is horrific12:56
=== kiko_ [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== kiko_ [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== kiko_ [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== stub [~stub@dsl-] has joined #launchpad
BradBstub: dude, did you look at the <pre> change? :)01:25
stubI merged that into Rocketfuel, didn't I?01:26
BradByeah, but i mean did you look at it in the UI? it's nasty dude.01:27
BradBbecause, of course, lines aren't wrapping anymore.01:27
stubDid some limited tests here. Hmm...01:27
BradBthat's why i suggested only for code snippets.01:29
stubStill got the same problem. I need more CSS-fu I think :-(01:30
BradBstub: with code-only though, you only get the problem when people are stupid, which will happen from time to time, but i don't expect it to be the common case01:30
BradBe.g. i don't think pasting 100-column wide code snippets is the common case01:31
stubI feel certain that the common case will be forgetting to put in (or not even knowing) the special markup required to flag a code snippet.01:31
BradBin any case, i'm about the land a change to the default search criteria, so that it's New/Accepted. i'm leaving the other widgets alone until i've discussed them more with Mark, because he has some concern about them being difficult to handle when all the packages, products and people come into it.01:32
BradBstub: eh, you can only help the user so much, of course...01:32
BradBif you say "use <pre> for code snippets" and they don't, well...01:33
BradBit could even be a little note nearby the textarea01:33
stubIf Malone does not do the right thing, people will use other systems. If we start accepting certain HTML codes, and making people escape them somehow when they want to include them, it will be yet-another-markup language and piss people off.01:33
stubAnd we can't fix this with documentation on the web, because of the email interface.01:34
BradBstub: it's not YAML though! it's just only allowing them to use HTML tags that we /can/ allow them to use whilst not creating a security compromise.01:34
BradBstub: email users will already be familiar with the web UI, and those that aren't should expect (and will have to) read a little tutorial on how to format their email correctly anyway.01:35
stubIt is - people will be composing HTML in their text email clients, and god help the poor sods who are sending HTML mail.01:35
BradBand again, we're thinking too far ahead right now. we need some little bit of markup for code no matter which way you look at it, but we don't much care about the incoming mail UI atm, because (i hope) it won't be part of the first release anyway.01:36
stubOh - I think I see. For the nice_pre, space should only be converted to &nbsp; at the start of the line.01:36
BradBwhat problem does that solve?01:38
stubI'm thinking ahead to the email interface, because I don't want all our comments to break when we switch it on (because they were assuming fixed width display, and suddenly the system chooses to render them proportional with collapsed whitespace). (Hmm... I guess at least in this case we can run through and add the 'code' marker around the entire existing messages)01:38
stubIt will allow the browser to break lines I think.01:38
BradBstub: dude, i think you're assuming everybody that uses malone uses mutt.01:39
BradBstub: rest assured, this isn't the case. think of all the gnome developers, kde developers, etc.01:39
stubAt the moment, it is not breaking them because I stupidly told it to *not* break them in about the most explicit way possible.01:39
stubI don't mean text clients (oh... so 90's!).  I mean text mode. All developers I know only send plain text emails (if for no other reason than the flamage they receive on public forums)01:41
BradBstub: i'd be delighted if, when writing an email, all i had to do to highlight some code was:01:42
BradBx = 101:42
BradBthe font we had seemed pretty decent, as proportional fonts go01:42
BradBmixing prop with pre-for-code would look fairly professional, like, heck, our wiki.01:43
stubYes, And I'd be pissed off at the system if i sent 'You need to do something like '<pre><span tal:contents="foo/bar" />...</pre>' and the system decided to interpret y markup, or worse still, only *some* of my markup.01:43
BradBstub: "interpret your markup"??! huh? we wouldn't touch anything inside the <pre>, obviously01:44
BradBit's verbatim, which you'd expect.01:45
BradBand you'd understand why we do the same thing that other sites do in restricting html tags used for security reasons.01:45
BradBbut anyway, gah, whatever, i'm not too arsed. i just hope we can do another rollout today, so that the statuses become more sane and the default search params become slightly more sane too.01:46
BradBstub: if sabdfl actually *likes* the look of all <pre> messages though, i'd boggle. :)01:47
stubRoundup looks like it is formatting the text server side to 80 columns (and preserving whitespace and using a monospace font)01:48
stubActually, it looks quite fine here. You need to use a different monospace font that courier ;)01:48
BradBstub: will you be able to push up a new dogfood today?02:15
BradBgreat...my patch for default search params should land any minute.02:15
BradBif i implement "Quick Searches" like at http://plone.org/collector, we can basically drop the bugs assigned to, and bug submitted by reports.02:16
BradBimplement tomorrow, that is02:16
stubSounds good.02:29
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
stubGet a good collection of default ones in and we won't have to bother with making used-defined quick searches either ;)02:30
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
=== BradB heads out
=== BradB is now known as BradB|away
stublifeless: Does arch do anything sucky with line endings like subversion, or does it just leave them the way they were saved?03:15
stubkiko: So - Malone has this 'select a sourcepackage' requirement03:53
stubkiko: Unfortunately, there appears to be no way a user can select the correct sourcepackage without knowing the primary key id03:53
stubkiko: I suspect I either need to make UNIQUE(Sourcepackage.distro, Sourcepackage.sourcepackagename) or somehow only display a 'sane' list of sourcepackages (perhaps only displaying the sourcepackage with a given distro and name that has the latest creation date (which needs to be added))03:56
kikoThere could be a problem with UNIQUE distro/name because of inherited packages, I think.03:58
kikothis part of the workflow really hasn't been considered yet, so it sucks to be you the first one to have the problem.03:58
kikoyou *could* make a select distinct on them -- how do you feel about that?03:59
stubI can have multiple identical rows in the sourcepackage table, only differing by id. If I only display distinct rows, I need to know which one is 'current' and assign bugs to that, or perhaps I should always use the first one? I'm really not sure04:01
kikocan't people file bugs against old releases?04:02
stubI personally think that we either need UNIQUE(Sourcepackage.distro, Sourcepackage.sourcepackagename) or refactor the database schema - the situation seems wrong.04:02
stubkiko: If they can, they have no way of telling the old releases from the new ones at the moment.04:02
stubIt isn't a database issue, it is a UI issue.04:02
stub(but I suspect the UI issue is caused by the database design, and that this design is causing problems elsewhere too)04:03
stubDo you know who would know if UNIQUE(distro, sourcepackagename) would break the model? People I've asked all suspect, but nobody can actually tell me a use case.04:04
stub(we might have to move sourcepackage.maintainer into sourcepackagerelease to snapshot this historical information, but apart from that...)04:06
kikosourcepackagename is merely a normalization table, avoiding duplicating the package name in the sourcepackage table.04:06
stubYer - pointless optimization IMO but I havn't pushed it too hard ;)04:07
kikothe issue is that you can have multiple sourcepackages with the same name but representing different packages 04:07
kikoso you could have a web browser called firebird and a database called firebird04:08
kikohowever I do think that you can't have firebird the browser and firebird the database in the same distribution as it would, well, not work, right?04:08
stubThat is what I think. If anything, the issue would be that if Ubuntu had firebird-the-browser in warty, and firebird-the-database in hoary.04:09
stubThat might be our pathalogical use case?04:09
stub(If so, it indicates to me that sourcepackage should be related to distrorelease rather than distro)04:10
stubI'm not sure if the Manifest causes trouble - it isn't documented so I'm not sure about it.04:10
stubThe other Malone alternative might be to assign bugs to a (distro, sourcepackagename) or a (distrorelease,sourcepackagename) instead of to a sourcepackage.04:12
stubHmm... if that *is* our pathalogical use case, we might need to prevent it from happening anyway because an Ubuntu user upgrading that package might get a nasty surprise ;)04:13
kikoI think over a whole distro you need the sourcepackagename to be unique.04:48
kikoyou'll need mark to confirm, though, because none of this is complicated unless we consider derivative distros, and that's where my knowledge endgs04:48
kikoI really don't like sourcepackagename very much and I've said it often -- I don't think it brings us anything at all 04:49
stubI need to dig up a database design book which explains why sourcepackagename is wrong - I've gotten too used to relying on my intuition and not having to rationalize my decisions to anyone else ;)04:50
stubMark seemed to think the existing design on sourcepackage was required for some reason but was unable to come up with a use case - suggested it was discussed at the conference but I was hoping to have the UI fixed before then ;)04:53
kikoso the use case you have is the user is trying to say in what package he's encountering the bug?04:55
stublifeless: pqm hung08:01
stubelmo: pqm hung08:01
=== lifeless [~robertc@dsl-] has joined #launchpad
stublifeless: pqm is hung09:12
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: made New and Accepted the default statuses in the bug listing search (patch-815)09:37
sabdflhey stub09:47
sabdflstub: sourcepackagename is partly because a lot of the actual pointers in the real world are to sourcepackagename, not to a specific sourcepackage09:48
sabdflthe debian bug system, for example, assigns bugs to a sourcepackage name09:48
sabdfleven if we move sourcepackagename back into sourcepackage, you still have the basic problem that a sourcepackage with a given name is not unique09:49
sabdflstub: can i merge the corrected BugMessage refactoring?09:50
sabdflI added:09:50
sabdflALTER TABLE Message09:50
sabdfl    ALTER COLUMN id SET DEFAULT nextval('message_id_seq');09:50
stubyes - go for it.09:51
stubKeeping sourcepackagename in a seperate table makes sense when you factor in those references I guess (the arguments for become stronger and against less clear anyway).09:52
sabdflit's a shitty situation, but we're trying to model a difficult problem09:52
sabdfli'm thinking of refactoring Malone to assign bugs to (distro, sourcepackagename) rather than sourcepackage09:53
stubYup - that would be a solution. I've just been poking people trying to workout the best solution, and if there is an underlying problem with the data model.09:54
stubI think we need to come up with the most pathalogical case we plan on supporting - the firebird one above seems pretty good, but I don't know if we want to support that behaviour or prevent it.09:56
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: nice_pre CSS implementation for dogfooding (patch-816)09:59
=== Kinnison [~dsilvers@host81-153-126-219.range81-153.btcentralplus.com] has joined #launchpad
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
sabdflmorning Kinnison10:35
Kinnisonhey sabdfl 10:36
Kinnisonsabdfl: Assuming kiko and cprov agree; I'd like to have cprov working with me from Monday. Does that sound okay to you?10:37
sabdflstretch cprov a little10:38
sabdflstub: ok, i'm using patch 4-10-010:38
=== Kinnison puts in the order for the rack
=== Kinnison oils the rope
sabdfleasy on the quartering10:41
KinnisonCan I at least draw him?10:43
sabdflsteady on10:44
sabdflhe might take that personally10:44
KinnisonTrue; I'd best stick to keeping this purely professional10:44
stubNow PQM is awake again, I'll do a dogfood update. Yell now if people have patches they want rolled out today that aren't yet committed.10:47
=== elmo [~james@] has joined #launchpad
Kinnisonmorning elmo10:54
Kinnisonsabdfl: I did a very rough time-plan last night; I'm going through it now making sure it seems sane10:54
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: bugmessage refactoring (patch-817)11:15
Kinnisonsabdfl: Unfortunately every time I think I'm done; I think of another thing to get gina to do :-)11:20
sabdflstub: did that bugmessage refactoring make it into the dogfood update?11:23
stubIt will - just shutdown the existing services when I saw the dillys message11:23
=== cprov [~cprov@] has joined #launchpad
Kinnisoncprov: is Kiko around?11:55
sabdflSteveA: congrats on getting the ZODBAnnotations service up and running11:58
=== Kinnison sighs as his time-estimates balloon :-(
SteveAwith the ftp server, I think we might want to put a try:except: around the callbacks12:03
SteveAand log errors at that point12:03
KinnisonSteveA: sounds like a sensible precaution12:03
SteveAstub: I just read the email about a simpler way to restore backups.  Is the "wisdom of the dba" being recorded anywhere other than the mailing list?12:07
stubNot at present. Something like an FAQ may be appropriate.12:09
stubProcedures that exist are documented in launchpad/database/schema/README, but that is more for acting-dba's.12:10
cprovKinnison: no, I think 12:10
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: bugassignment renumbering (patch-818)12:13
Kinnisoncprov: Any idea when he normally turns up?12:13
cprovKinnison: sorry, but today I have no idea, maybe within this hour . may I help you ?12:15
Kinnisoncprov: I need to talk with you and him ;-)12:16
Kinnisoncprov: sabdfl has agreed to let me ask for your time on lucille :-)12:17
Kinnisoncprov: Which will give you a chance to see how things work from the other end of the package process in soyuz ;-)12:17
sabdflcprov: would you like to take on some of the deeper challenges in the package management system behind soyuz?12:29
cprovsabdfl: yes12:30
=== Kinnison adds the scary-shit to the list of things to give cprov ;-)
sabdflcprov: excelle nt12:31
cprovsabdfl: I'm just getting a kind of lost helping SteveA and Brad to have "better output on failed tests" <wink>12:31
sabdflyou've done a great job on the front end so far12:31
sabdflremember all those questions about "how uploads happen" and "what happens when a new package arrives" that I said not to worry about?12:32
sabdflnow's the time to start thinking about them :-)12:32
sabdflKinnison has made a good start and will lead  this work12:32
sabdflyou will work for Kinnison on tasks he assigns12:32
sabdflhe knows exctly what he wants done, and can help you when you get stuck12:33
cprovsabdfl: thanks, but I feel I can do more on it ... just need more "time" (48 hours day would be nice) 12:33
sabdfltogether you two need to get the back-end stuff working for the end of es-conf12:33
cprovsabdfl: great !12:33
Kinnisoncprov: My plan involves getting you into the lucille stuff starting on Monday. That gives you today and tomorrow to finish of the stuff you're working on and hand off anything to other people. Does that sound enough or do you want more time than that? (sabdfl: Do you want cprov to be exclusively on lucille or partially on lucille until es-conf?)12:35
sabdflKinnison: i think lucille should remain your baby12:36
sabdflgive cprove related tasks with a lower learning curve12:36
sabdflto start with12:36
Kinnisonsabdfl: Yeah; my plan had that kind of ramp-up12:36
sabdflso, for example, the build system integration12:37
Kinnisonsabdfl: I have a identified quite a bit to do on gina for now ;-)12:37
cprovKinnison: 2 days to finish my current tasks will be enough12:37
Kinnisonsabdfl: I think the build-system integration is actually quite a deep thing. I have a curve mostly planned out for now ;-)12:37
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
cprovKinnison: sabdfl: I'm very happy to work on this, let's book a meeting soon to talk about your plans and the targets12:48
Kinnisoncprov: Yes; I'd like to make sure kiko is happy with all this too12:48
cprovKinnison: let me help you with a simple phone call :)12:52
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
cprovKinnison: done, in few minutes he will arrive12:53
Kinnisonhi kiko12:53
kikohey everybody12:53
kikotoday is "I hate SSL certificates" day!12:53
Kinnisonkiko: To bring you up to speed... sabdfl and I had a meeting yesterday and we decided that it would be good to get cprov on board with me as a lucille developer. cprov is interested in this and we've basically agreed that he will start working with me on Monday. Is this okay with you?12:54
cprovkiko: sorry I should say "seconds" ...12:54
kikoKinnison, sounds good to me -- cprov has been wanting to get his hands dirty on it for a while already I suspect.12:54
kikoI was actually curious because yesterday you said an extra pair of hands would be good12:55
Kinnisonkiko: Excellent ;-)12:55
kikoI was going to offer but since the topic died out I thought you had reconsidered.12:55
=== Kinnison puts a tick-mark in kiko's b33r column
sabdflkiko: ssl certificates are paying for this little gig :-)01:04
kikoKinnison, tell me more about where lucille is going01:05
kikosabdfl, jesus, good point, you could give me some hints, openssl is kicking my ass.01:05
sabdflhold on, this is really soyuz, not lucille01:05
sabdfllucille is one component of soyuz01:05
stubMmm.... SSL certificates. Although I did get a $25 amazon gift certificate in thanks for calling Thawte extortionists in a survey. Sorry Mark :-)01:05
sabdfllucille manages APT archives for distros - ubuntu and derivatives01:05
Kinnisonsabdfl: Lucille manages a tad more than that to be fair01:06
sabdflstub: dude, i think they started that after i left01:06
Kinnisonsabdfl: She also manages the queues. She's the *archive manager*01:06
kikoexcuses, excuses01:06
sabdflwhat we are talking about is extending cprov's work to include more of the workflow of package management01:06
Kinnisonsabdfl: Indeed; so queue stuff; upload handling etc.01:06
kikoincluding the web-presented side of it, I expect?01:06
Kinnisonkiko: Yes; I hope so01:07
=== Kinnison doesn't want to have to learn zope
sabdflfrom upload, through build, through publishing, and including offering the package to derivatives01:07
sabdflkiko: especially :-)01:07
sabdflKinnison has a really good understanding of how that workflow should look01:07
sabdflwe need to give him a second pair of hands, and also someone with web / zope flair, so cprov sprung to mind01:08
kikoif this means we can finally fix the "Pending" feature in source package, I want to donate to the effort01:08
sabdflkiko: yes it does01:08
kikosabdfl, today's our first day running ubuntu diskless at async01:09
sabdflkiko: ltsp?01:09
kikowe encountered a few tricky issues but all in all it's excellent.01:09
kikono, ltsp runs apps on the server -- we run apps on the client. it's pure NFS.01:10
sabdflcan we sanely fold that into hoary as a standard feature?01:10
kikoI'd need to ask salgado, but there are a few package tweaks that would need to be done.01:10
kikothe issue is that what we do is install ubuntu into a directory on the server and export that via NFS01:11
kikowe use a custom-built kernel, and some init.d hacks01:11
kikoI wonder if this would be a "package" to install on the server or something else.01:11
kikoit's currently a tarball.01:11
KinnisonSounds like ideally it'd be a debootstrap following by a chroot and an install of a few choice patching packages01:12
kiko(it would be an 800MB package)01:12
kikoKinnison, let's talk a bit about this in spain, I'd like to package it nicely01:12
Kinnisonkiko: make sure you note down to talk about it or I'll forget01:12
kikosure thing01:12
=== carlos leaves to get lunch
SteveAcarlos: we have a meeting in 5 minutes01:24
stubsabdfl: Should everyone with a chinstrap account be allowed to access a launchpad_dogfood backup?01:31
elmostub: no01:35
stubelmo: Is there a group I can use? There is a request on launchpad@ to get a copy of the dumps for development work and I either need a drop point for people without accounts on mawson or deny it.01:37
carlosSteveA: I'm here01:37
elmostub: who needs it who doesn't have an account on mawson?01:37
carloselmo: I01:37
elmoanyone else?01:38
stubcarlos requested it. I have no idea if there are others.01:38
elmocarlos: congratulations you now have an account01:38
elmothanks for playing "Lucky Account Lotto"01:39
stubTa elmo.01:39
carloselmo: thanks01:39
=== debonzi [~debonzi@] has joined #launchpad
SteveAdaf: ???01:49
SteveAcarlos: I have another meeting later, and other things to do today.  Let's talk anyway.  Daf can read later.01:57
carloshere or #canonical-meeting?01:57
stubDo bzipped files rsync well?02:00
SteveAok.  Is the rosetta alpha running on the dogfood system?02:01
kikostub, well, not efficiently, but it works02:02
carlosSteveA: seems like some data is still pending of upload, but I'm not sure02:03
carlosSteveA: I know daf was uploading files already02:03
SteveAwhen we talked about it yesterday, we all thought that it could be done by the end of yesterday.02:04
SteveAwhat happened?02:04
carlosSteveA: I'm not sure, but I think the scripts we have to import the files were not updated with latest changes we hace when moving to the common infrastructure02:06
SteveAwhat, exactly, is left to do?02:06
carlosSteveA: guessing here (I was not able to talk with daf, I left to sleep before he finished). Import the files into launchpad02:07
carlosI think I got an account in mawson today, so I could help now daf 02:07
carloson this task02:07
SteveAok.  I wish daf were here.02:08
SteveAdaf: buy yourself a big alarm clock that has no "snooze" button.02:08
SteveAlet's assume that the dogfood stuff will be done a bit later today.02:09
SteveAnext, we have two important things02:09
SteveA1. get more people using rosetta02:09
SteveA2. the rosetta team needs to lead the rest of launchpad in internationalizing the application02:10
SteveAbefore I went on vacation, I emailed a list of resources on how to internationalize applications in python and zope02:10
SteveAdid you read these?02:10
carlosnot in depth02:11
carlosbut I got the idea02:11
SteveAdid you start to apply what you had learned to the rosetta code?02:11
carlosnot yet02:12
carlosI was working fixing bugs more than adding new feature02:12
carlosyesterday I started adding new features02:12
SteveAthis is an important feature for dogfooding rosetta02:12
carlosso I will do it now02:12
carlosI mean, from now02:12
SteveAok, great.  there are three parts to this.02:13
SteveA1. start internationalizing rosetta02:13
SteveA2. get the pot files into the rosetta dogfood system for translation02:13
SteveA3. write up how to do this so that others on the launchpad team can see what they need to do to existing code, and what they need to do with new code.02:14
SteveAthe rosetta team are the "internationalization csars" for launchpad.02:14
carlosSteveA: should we add projects to translate as we decide? (outside the list we already were talking about)02:15
SteveAyou mean in addition to stuff from ubuntu, zope3, plone, zwiki ?02:16
carloshmm, with all ubuntu resources we have already a big tasks02:16
carlosI was talking about which ubuntu resources should we add02:17
carlosuntil we could automatice the import02:17
carlos /s/automatice/automate/02:18
SteveAthe thing to consider is whether you can get more people using rosetta by adding a particular project02:19
carlosand how are we going to translate ubuntu's website?02:20
=== carlos is talking about the content
carlosnot sure if it will be handled by rosetta02:21
SteveAthat doesn't have anything to do with rosetta at this point02:21
carlosSteveA: anything else?02:26
SteveAthat's it for now, at least until daf turns up and we can talk about the dogfood02:27
SteveAkeep me informed about how the internationalization stuff is going02:27
SteveAI can help out if you have questions02:27
carloswill do02:28
carlosSteveA: sorry, but I'm not able to find the links you sent about zope i18n, I read them but seems like I forgot to add a bookmark02:41
carlosdo you have them around?02:41
carlosI think I found it: http://zope.org/Wikis/DevSite/Projects/ComponentArchitecture/HowToDoI18nAndL10nForZope3/Zope3Internationalization02:44
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: No more unused imports in soyuz app components. BugAssignmentStatus ajustment. (patch-819)02:51
SteveAcarlos: I have sent you another copy03:01
carlosSteveA: thanks03:01
=== BradB|away is now known as BradB
KinnisonHey brad03:07
carlosSteveA: ok, I know why I didn't found it, it's inside my inbox folder instead of the folder I have with your mails ..03:11
cprovBradB: please, tell me the RIGHT way to set a customized "optionflags" for a set of python tests 03:13
BradBcprov: canonical.functional.FunctionalDocFileSuite has an example of how to do it.03:37
carlosBradB: how is your css knowledge?04:04
BradBcarlos: limited04:05
carloswho knows css in launchpad team?04:05
carlosBradB: we have a problem with css in rosetta04:05
=== Kinnison knows some css
carlosKinnison: could you look at http://localhost:8086/rosetta/projects/gnome/evolution/evolution-2.0/translate ?04:06
carlosKinnison: from your local launchpad installation?04:06
Kinnisonis that part of the default data?04:06
carloswe are showing check buttons and they are expanded because some css magic...04:06
carlosKinnison: yes04:07
carlosyou need default data04:07
Kinnisonhow do I load default data?04:07
carlosa make inside database/schema should be enough04:07
=== Kinnison starts a launchpad
carlosor foo.bar@canonical.com and same password04:09
Kinnisonokay; I can see the page04:09
carloslook under the language name04:09
carlosthat's a check button04:10
carlosthat has the same size of the entry text04:10
Kinnisonand you want to know why yes?04:10
carloswell, know why and fix it :-)04:10
=== Kinnison looks
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
carlosKinnison: my X server died04:14
KinnisonThe problem appears to be that the checkbox is inside a <label> tag04:15
carloslet me check it...04:16
sabdfli need some charset help / fu04:16
sabdflany experts out there? stevea?04:17
carlosKinnison: same problem moving the checkbox out of the label04:19
carlosKinnison: the template is potemplate-translate.pt04:20
Kinnisonwhere is that?04:20
carlosKinnison: http://gollum.pemas.net:8085/rosetta/projects/gnome/evolution/evolution-2.0/translate <- My local launchpad server04:21
sabdflhow do i make sure a string is unicode before sending it to the db?04:23
sabdfli'm trying to put a bunch of emails into the db, and i hit one with 8-bit characters... BOOM04:24
kikosabdfl, heh, same problem as in gina.04:24
kikoyou can use my horrible-hack from it04:24
sabdflok, where's the code?04:24
Kinnisoncarlos: Okay, I can fix it with an explicit style rule04:25
Kinnisoncarlos: add to your css the following:04:25
SteveAsabdfl: when you say "putting a bunch of emails in the database", is this a script?04:25
carlosKinnison: ok04:26
Kinnisonfieldset.message-strings input[type=checkbox]  { width: auto !important; }04:26
Kinnisoncarlos: try that04:26
Kinnisondid that do it?04:30
carlosKinnison: I added it to the end of launchpad.css04:30
carlosand I don't see any change04:30
KinnisonIn case mozilla has it cached04:30
carlosok, that's the problem04:31
carlosit worked now04:31
KinnisonYou're welcome04:31
=== Kinnison takes off his "I read the CSS spec -- and what's more I can remember a lot of it" hat ;-)
sabdflSteveA: yes it is a script04:31
carlosKinnison: ;-)04:31
sabdfli'm rewriting the debbugs-bugzilla syncer that mdz wrote, for malone-debbugs sync04:32
=== debonzi [~debonzi@] has joined #launchpad
sabdflKinnison: wow, that's quite a memory04:32
Kinnisonsabdfl: I needed to understand the CSS box model a while ago; so I forced myself to remember as much of CSS as I could04:33
Kinnisonsabdfl: Unfortunately I'm reasonably sure that I forgot how to make spaghetti bolognese in the process04:33
Kinnisonblasted limited-capacity poorly-designed brain04:33
BradBsabdfl: got a moment to chat about the search widgets?04:35
=== kiko is now known as kiko-fud
BradBsabdfl: namely, i want to know if it's worth the effort to set defaults on the result of the widgets, based on the user's context.04:35
SteveAsabdfl: the general rule for writing code that interacts with things outside the system is, convert textual data into unicode as soon as you get it.04:35
BradBbecause if they're going away, or are going to be modified significantly, it may be wasted effort.04:35
SteveAsabdfl: or, decide that it is "opaque data", and store it as "bytes" in the database not as text.04:36
sabdflSteveA: here's the challenge04:36
sabdfl - take several hundred thousand emails in the debbugs system, and import them to the Message table04:36
sabdflwe have to be able to do this04:36
sabdflelmo: what's the ETA on a genuine upload facility that does not require a chinstrap account for hoary uploads?04:37
=== SteveA wonders if the email modules in the standard library can help
sabdflBradB: if we used wizard-style forms this would be easier, right?04:38
sabdflbecause we get to develop the context over a few page loads04:38
BradBsabdfl: heck no. that would take too much effort.04:39
=== SteveA -> lunch
elmosabdfl: if I go home tonight, I'll should be able to do it tomorrow04:41
sabdflelmo: great04:41
BradBsabdfl: when do we get to see a serious go at a combo box widget that can handle, say, 20,000 items?04:41
sabdflSteveA: i'm using the std library stuff to parse the emails fine04:41
sabdflBradB: it's not going to be possible client-side04:42
sabdflBradB: we need either a popup, or a behind-the-scenes query with JS04:42
BradBsabdfl: ok. it doesn't sound like it's worth it for me to do much more with setting things on the existing widgets then. this widget will be critical for launch. when do we get someone to have a serious go at making a popup that can work for the bug listing screen?04:44
cprovBradB: thanks04:45
=== BradB should really learn more about Sources in Zope 3, which are meant to deal with huge vocabularies, but I'm not yet sure just how much help they give (e.g. a really cool widget? not sure.)
BradBsabdfl: heck, i think a killer MonsterSelect widget would really help sell launchpad to developers. i for one get a kick out of using apps that do little things that make me go "whoa, now /that/ is neat."04:47
BradBwhich is what i did the other day when i saw the combo box thing in bugzilla04:48
sabdflBradB: ok, want to have a stab at it?04:48
sabdflmy preferred approach is a behind-the-scenes query that returns a JS array04:49
sabdfland keep this all inside the form, in an expanding fieldset04:49
sabdflinitial state shows search form04:49
sabdflenter text and press "find"04:49
sabdflfieldset expands to show text search field and list of up to 10 matches04:50
sabdflclick on a match, and fieldset collapses to show just that selected item, with a "change selection" option to search again04:50
=== daf [daf@muse.19inch.net] has joined #launchpad
BradBit needs to handle multi-selects too, of course (a common case in the bug listing)04:51
Kinnisondaf: sleep well?04:52
dafyes, thank you04:53
=== daf was having some connection trouble
=== Kinnison looks into heading home to finish his planning and todo-lists
carlosdaf: hey!04:57
=== Kinnison waves
Kinnisonc'y'all in about 45 mins again04:58
sabdflblush... what's the login for foo again?05:01
sabdflfoo.bar@~canonical.com, test?05:01
carlossabdfl: yes05:01
carloswithout the ~05:01
sabdflcarlos: doesn't work for me05:02
carlossabdfl: I think it's only available with sample data05:03
carlosthe dogfood server does not have it05:03
=== kiko-fud is now known as kiko
sabdflcarlos: ah, thanks :-)05:17
BradBum, why in the HELL am i seeing a bunch of messages like "rm: /mach_kernel: Operation not permitted" when i try to restore a dogfood backup with: bunzip2 -c /Users/bradb/launchpad_dogfood.20041118.sql.bz2 | psql -d launchpad_dev05:20
BradBrm: /dev: Operation not permitted. um, jesus fuck.05:22
kikodude, that's not good.05:23
dafBradB: you want bunzip2 -cd, surely?05:25
dafoh, no, never mind05:25
=== daf notices the "un" in the command name
kikoyou redundant welshie05:25
BradBi saw lots of garbage like "invalid command \n" or "invalid command \N" earlier in the run05:26
BradBand then this thing starts running launchpad ftests and such, which is thoroughly unexpected.05:26
carlosBradB: MacOSX is not unix ;-)05:26
carlosBradB: you just need to execut "ls -l /" to see it05:27
=== carlos hides
carlosdaf: I'm leaving to go to the university in some minutes, should we talk about anything that cannot wait?05:28
dafwe should05:28
dafbut I don't know of anything that can't wait05:28
kikoSteveA, yo?05:28
carlosdaf: btw, elmo gave me an account in mawson today so I suppose I could help you with the data import05:29
carlosdaf: ok, I will be back in about 3 hours05:29
dafcarlos: ok, I've been having some problems with the import scripts05:29
carlosdaf: anything else than recent code changes?05:30
dafmostly database stuff05:31
=== debonzi [~debonzi@] has joined #launchpad
carlosdaf: ok, will see it when I'm back05:31
kikowhere's the SteveA 05:32
debonziSteveA, ping05:34
kikodude, it's RAINING.05:36
BradBeh, didn't i have a .txt file somewhere explaining how to use batching? I wonder if someone removed that.05:42
BradBsabdfl: hm, i wonder. is it reasonable to require that a launchpad user's browser supports DOM mutation?05:46
BradBwhich i guess just means supports DOM...05:47
sabdflBradB: what's the coverage of that?05:47
=== Kinnison [~dsilvers@haddenham.pepperfish.net] has joined #launchpad
sabdfli think LP in lynx would be painful :-)05:47
sabdflkiko: dude, it's POURING over here05:47
sabdfland it's dark too05:47
sabdfland 16:4805:47
BradBhm, no idea which browsers don't support DOM, but i guess most do05:50
kikowinters in europe.05:52
kikoBradB, all modern browsers do, but text browsers don't. does that mean no filing bugs from lynx?05:52
BradBi think that's acceptable.05:54
dafI thought there was at least one text-based browser that support JavaScript05:55
kikolinks does somewhat.05:55
kikoI don't know -- Bugzilla has always supported lynx as policy and I've used it on occasion.05:56
=== BradB ignores lynx for now
elmonot supporting lynx is CRACK06:03
BradBi didn't say it wasn't going to be supported, but that i'm ignoring it for now.06:04
kikoignoring it for now means ignoring it forever, no?06:04
dafso it't not going to be supported?06:05
SteveAhi kiko, debonzi, cprov06:22
SteveAsorry I'm later06:22
kikoI'm about to skip out, too06:22
kikohow about we zap through?06:23
debonziHi SteveA06:23
SteveAin soyuz/distroapp.py there are examples of what is arguably database code06:25
SteveAfor exampe06:25
SteveA        self.roles=DistroReleaseRole.selectBy(distroreleaseID=self.release.id) 06:25
SteveAthat is, rather than use what this method / instance has06:25
SteveA(in this case, a release)06:25
SteveAor use stuff got via a utility06:25
SteveAit is using the database stuff directly.  rather than using operations specified on an interface, it is calling "selectBy" directly.06:26
SteveAso, it is great that it is there and getting the job done06:26
SteveAbut, I'd like stuff that is this closely linked to the database to be within canonical/launchpad/database/06:26
kikolet's look at that code06:27
SteveAIt may be that a DistroRelease object needs to grow certain extra methods06:27
SteveAto allow what this class we're looking at needs06:27
SteveAthat would be one way to fix it06:27
SteveAI'm looking at class DistroReleaseApp06:27
SteveAaround line 75 in distroapp.py06:27
SteveAwhat do you guys think about all this?06:27
kikosounds like the right way to go about it06:29
kikolet me get my tree updated, one moment.06:30
debonziSteveA, you mean: I should be a method inside DistroRelease SQLObject that return the selectBy result. right ?06:30
SteveAone of the reasons I want all the database stuff to be in one place is because I'm working on the advanced security stuff right now.06:30
SteveAwhen this is checked in, the most important parts to consider the security of are the database classes06:30
kikoSteveA, we should not use SQLObject-specific API outside of database.py, IMO. Agreed there?06:31
SteveAdebonzi: yes.  so, the DistroReleaseApp code has a reference to a particular DistroRelease object.  It is called 'release'06:31
SteveAIt needs to get the result of calling "DistroReleaseRole.selectBy(distroreleaseID=self.release.id)"06:31
SteveAso, DistroRelease can grow a method "getRoles" perhaps06:32
SteveAkiko: I agree.  Not database.py of course, but the database package06:32
kikodoh, yes, database/*.py.06:32
SteveAthe exception is in stand-alone scripts06:32
kikosuch as gina.06:32
debonziSteveA, Is classmethods ok for it too?06:33
kikothere's not much in the way of a security framework gina can work with anyway, I suspect.06:33
SteveAdebonzi: classmethods are a problem, because they are not really represented in the interface to an object06:33
SteveAyou have to make an assumption about what the object's class is06:33
SteveAand that breaks the encapsulation06:33
SteveAso, the public api should be instance methods, described by interfaces06:34
kikoI dislike classmethods personally06:34
kikoin this case there is a good reason not to use one too (self.release.id)06:34
SteveAnote that we're working with interfaces06:34
SteveAan interface doesn't really say whether something is implemented by an instance method or a class method or what06:34
SteveAbut, seeing as we have a release object, and that object is an instance06:34
SteveAit is natural to use instance methods to implement the operations needed by the interface06:35
SteveAdoes that make sense?06:35
kikoyes, though there are some weird exceptions in SourcePackage. I'd need to look at that code carefully.06:35
SteveAif it makes enough sense for the majority of the code, we can convert that to the new way, and look again at any weird things06:36
kikomy tree is bonkers :-/06:36
kikoyes, indeed.06:36
SteveAThere is something odd with my tree at the moment I think.  I'm getting a test that always fails, and a test that always hangs.06:36
kikothis is the way we should go, for sure06:36
SteveAbut I can still merge into RF06:36
=== SteveA wonders if any libraries are out of date
kikoI can't star-merge, I get bucketloads of conflicts and no undo love.06:37
kikoSteveA, did you get my Zope patch from salgado?06:37
SteveAI did06:37
SteveAI read the email, but I haven't thought about it yet06:37
kikookay. it's really trivial -- and follows what the code does in that part anyway.06:37
SteveAI'll think about it soon06:39
SteveAso, who is going to do what with the database code in the soyuz package?06:39
kikodebonzi and I can handle it06:39
kiko(if my tree unfuxors itself)06:40
SteveAok, cool.  when can you start?06:40
SteveAdaf: around?06:41
=== SteveA gets out a theremin to accompany the hit song "daf gets around"
dafSteveA: yes06:42
kikoSteveA, lemme check current statuspile06:43
SteveAcan we have a chat about rosetta stuff?  did you read what carlos and I talked about earlier today?06:43
dafI did06:43
SteveAany comments?06:43
kikoSteveA, starting monday we can do a review and rehack of these bits06:43
kikoSteveA, if you want some of it done tomorrow, we can drop the ball on other things06:44
dafSteveA: I've been working on the import, but have been having problems with the import script which mean it is not done yet06:45
SteveAkiko: it can be done on monday.06:45
kikookay, scheduled06:45
debonziSteveA, nice06:45
SteveAthanks guys06:45
SteveAdaf: what's up?06:45
kikodebonzi, can you file a bug for me while I unfutz my tree06:45
debonzikiko, yep06:45
dafwell, there were some database changes which the scripts hadn't been updated to reflect06:45
dafand now I am getting error messages which I am still in the process of diagnosing06:46
SteveAcan anyone help you with this?06:47
dafCarlos might be able to, but he isn't around at the moment06:48
SteveAcan you describe what the first problem is in some more detail?06:49
cprovBradB: finally something good, do you have time to analize the results of my task ?06:50
dafTraceback (most recent call last):06:50
daf  File "./poimport.py", line 168, in ?06:50
daf    options.potemplate, options.language, True)06:50
daf  File "./poimport.py", line 95, in update_stats06:50
daf    poFile = poTemplate.poFile(languageCode)06:50
daf  File "/home/daf/launchpad-devel/launchpad/lib/canonical/launchpad/database/pofile.py", line 263, in poFile06:50
daf    raise KeyError, language_code06:50
dafKeyError: "Couldn't find record in database"06:50
=== Signon time : Fri Nov 5 12:41:16 2004
=== Signoff time : Thu Nov 18 19:05:42 2004
=== Total uptime : 13d 6h 24m 26s
=== warthylog [~warthylog@port49.ds1-van.adsl.cybercity.dk] has joined #launchpad
=== Topic for #launchpad: lunchpad: home of the sandwich artists | fogo na bomba | "qorking along happily, with SteveA egging us on"
=== Topic (#launchpad): set by sabdfl at Tue Oct 19 11:05:38 2004
=== Your nick [ubuntulog] is owned by ~warthylog@port49.ds1-van.adsl.cybercity.dk
=== Your nick [ubuntulog] is owned by ~warthylog@port49.ds1-van.adsl.cybercity.dk
(SteveA/#launchpad) daf: do what you think will work best. but definitely use the debugger07:08
(SteveA/#launchpad) is the language table in dogfood the same as the one in rosetta alpha07:09
(SteveA/#launchpad) ?07:09
(daf/#launchpad) yes, probably07:09
(daf/#launchpad) I'll try reporoducing it on my machine07:09
(SteveA/#launchpad) why not do a query, and check that they are in fact the same?07:10
(SteveA/#launchpad) you have access to both databases07:10
=== fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has joined #launchpad
fabbionesorry for the introdusion :-)07:12
=== fabbione [~fabbione@port49.ds1-van.adsl.cybercity.dk] has left #launchpad []
(BradB/#launchpad) damn, looks like i'll be reinstalling. the db restore done verbatim as describe in the email to lp@ blew away all my applications.07:26
(kiko/#launchpad) that's crazy07:26
(kiko/#launchpad) how did that happen?07:27
(BradB/#launchpad) not the faintest clue...07:27
(daf/#launchpad) crazy!07:27
=== salgado [~salgado@200-206-134-238.async.com.br] has joined #launchpad
(lulu/#launchpad) night all :o)07:50
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has left #launchpad []
(elmo/#launchpad) quick hands up, who's still working?07:51
(elmo/#launchpad) seriously, I need to reboot some of the remaining boxes,including mawson, macquarie and emperor - any objections if I do it now? 07:51
(kiko/#launchpad) me07:51
(kiko/#launchpad) not me07:51
(daf/#launchpad) elmo: hmm, should be OK, I think07:59
(kiko/#launchpad) daf, what does it mean when tla update sits for hours writing to unlinked files in /tmp?08:06
(daf/#launchpad) I don't know08:07
(daf/#launchpad) sounds pretty broken, though08:07
(kiko/#launchpad) time to nuke my tree?08:07
(daf/#launchpad) hmm, perhaps08:07
(daf/#launchpad) try asking somebody who knows tla better first, maybe08:08
(kiko/#launchpad) hmmm08:10
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
(elmo/#launchpad) mawson rebooting08:21
(kiko/#launchpad) * searching ancestor revision in library in archive kiko@async.com.br08:27
(kiko/#launchpad) corrupt library (failed inode signature validation)08:27
(kiko/#launchpad) archive: daniel.debonzi@canonical.com08:27
(kiko/#launchpad) revision: launchpad--devel--0--patch-22908:27
(kiko/#launchpad) this is lots of fun!08:27
(daf/#launchpad) lies! lies!08:27
carloskiko: :-P08:28
carlosdaf: ;-)08:29
(kiko/#launchpad) hmmm08:29
carlosdaf: I will be available to talk in about 15 minutes, is it ok for you?08:29
(kiko/#launchpad) interesting, it's refueling my revlibs08:29
(daf/#launchpad) carlos: ok08:29
(elmo/#launchpad) macquarie/emperor 08:34
(elmo/#launchpad) + done08:34
(kiko/#launchpad) whee08:35
(kiko/#launchpad) SteveA, ping?08:38
(kiko/#launchpad) daf, do you know why psycopgda and sqlos aren't links to sourcecode?08:38
(kiko/#launchpad) or SteveA?08:39
(kiko/#launchpad) or carlos 08:39
(kiko/#launchpad) or anyone else08:39
(kiko/#launchpad) Kinni!08:40
(daf/#launchpad) kiko: because they are not compiled?08:40
(kiko/#launchpad) hmmm08:41
(kiko/#launchpad) is that the only reason? is it a good one?08:41
(daf/#launchpad) I think it's reasonable08:42
(daf/#launchpad) I don't see any major disadvantages08:42
(elmo/#launchpad) it's CRACK08:43
(daf/#launchpad) elmo: ?08:43
(elmo/#launchpad) it's the same CRACK as people who copy library source into their own projects rather than just linking with it08:44
(kiko/#launchpad) it's bizarre to me -- I always need to special-case psycopgda and sqlos into the directory.08:44
(elmo/#launchpad) CRACK CRACK CRACK08:44
(daf/#launchpad) elmo: they are in separate arch categories08:44
(elmo/#launchpad) daf: is psycopg modified at all?08:45
(elmo/#launchpad) if not - WTF is it in there?08:46
(daf/#launchpad) not psycopg08:46
(daf/#launchpad) psycopgda08:46
(kiko/#launchpad) I'm more curious about the consistency issue that arises from having bits in sourcecode and bits in lib08:47
(kiko/#launchpad) linked to sourcecode08:47
(kiko/#launchpad) and other bits not linked08:47
(kiko/#launchpad) ugh08:47
(daf/#launchpad) consistency is good08:47
(sabdfl/#launchpad) BradB|lunch: how do you mean, blew away your applications?08:47
(sabdfl/#launchpad) i've found that working with a backup of the dogfood db on my local machine i need to make sure08:47
(sabdfl/#launchpad) that i apply the patches to bring the db schema in line with my code08:48
(sabdfl/#launchpad) SELECT * FROM launchpaddatabaserevision;08:48
(sabdfl/#launchpad) then apply the patches needed to bring the revision up to the level of your code08:48
(kiko/#launchpad) daf, in this case, apparent lack of consistency :)08:49
(daf/#launchpad) well, if you can fix it, I say go ahead08:49
(kiko/#launchpad) I'd move sqlos and psycopgda into sourcecode and link it too. wouldn't you?08:50
(kiko/#launchpad) corrupt library (failed inode signature validation)08:53
(kiko/#launchpad) archive: daniel.debonzi@canonical.com08:53
(kiko/#launchpad) revision: launchpad--devel--0--patch-30208:53
(kiko/#launchpad) You should remove this revision from your library.08:53
(kiko/#launchpad) whee!08:53
(sabdfl/#launchpad) wow#08:54
(sabdfl/#launchpad) did you know you can tla add a file that does not exist?08:54
(sabdfl/#launchpad) and it works?08:54
(daf/#launchpad) which definiton of "works" are we using here?08:54
(elmo/#launchpad) guys - chinstrap's going down - anyone need it desperately right now ?08:54
carloselmo: as soon as it's back soon... ;-)09:00
carlosdaf: I'm ready to talk09:00
(elmo/#launchpad) carlos: it's back...09:01
carloselmo: perfect, thanks09:01
(kiko-fud/#launchpad) lifeless, the email to launchpad is probably one of those you'll say "bah, that's easy" I suspect09:07
(SteveA/#launchpad) kiko-fud: we agreed in a launchpad meeting a while ago that these things were going to move into "sourcecode" and out of "launchpad/lib"09:16
(SteveA/#launchpad) I never got around to it. Thanks for bringing it up now.09:17
(lifeless/#launchpad) sabdfl: can't do that in baz :)09:17
(lifeless/#launchpad) sabdfl: why are u not using baz ?09:17
(lifeless/#launchpad) :)09:17
(SteveA/#launchpad) daf: how are the imports going?09:18
(daf/#launchpad) still trying to reproduce the error on my machine09:19
(daf/#launchpad) I have an import running right now09:19
(SteveA/#launchpad) did you investigate the error on the server, so that you know exactly what the problem is?09:22
(daf/#launchpad) no09:23
(SteveA/#launchpad) it would be good to do that -- work out exactly what is wrong on the server. fix and test on your machine. that's my suggestion.09:23
(SteveA/#launchpad) you can't expect to solve a problem until you know what is wrong.09:23
(SteveA/#launchpad) is this something where I can help you, via screen ?09:24
(daf/#launchpad) I dont't know how I can test it on my machine before I can get it to fail on my machine09:24
(SteveA/#launchpad) it will be easy to get it to fail on your machine once we know exactly what is failing09:25
(BradB/#launchpad) sabdfl: in OS X, each user has some kind of copy in their home dir (in ~/Library/Applications or ~/Applications or something) of the apps they install. those all got blown away.09:26
(sabdfl/#launchpad) bradb: by what?09:26
(BradB/#launchpad) elmo: do netgear 802.11g cards work in powerbooks?09:26
(BradB/#launchpad) sabdfl: rm :)09:26
(sabdfl/#launchpad) ah09:26
(sabdfl/#launchpad) by one of our scripts?09:26
(BradB/#launchpad) sabdfl: from following the instructions verbatim to restore a dogfood db09:26
(sabdfl/#launchpad) oops09:26
(elmo/#launchpad) BradB: I don't know - it's just whether they work on Linux tho - not powerbook specific09:27
(BradB/#launchpad) elmo: what do you use?09:27
(elmo/#launchpad) cisco aironet09:27
(SteveA/#launchpad) elmo: have you prodded the ftp server yet?09:27
(elmo/#launchpad) $$$ tho, and it's not 100% on Linux/PowerPC09:28
(elmo/#launchpad) SteveA: no, been at the DC all day09:28
(elmo/#launchpad) I'll look at it tomorrow, when I'm home09:28
(SteveA/#launchpad) ok. let me know if you have questions about it.09:28
(SteveA/#launchpad) daf: I really want the rosetta imports fixed soon. consider me to be at your disposal to help get it sorted, until I go to sleep tonight.09:28
(elmo/#launchpad) BradB: looks like modern WG511 uses Atheros chipset, which is the same as the onboard one on modern powerbook's, which means BZZT 09:29
(elmo/#launchpad) bradb: (btw, by not 100% don't get me wrong, it works just fine for me, but some features are missing, like the ability to hot plug the card :)09:29
(elmo/#launchpad) stevea: will do09:30
(BradB/#launchpad) elmo: ah09:30
(BradB/#launchpad) hmph09:30
(elmo/#launchpad) BradB: mark has some card that limi got working that isn't cisco09:30
(elmo/#launchpad) (s/got working/'just works in warty' apparently)09:31
(BradB/#launchpad) the place i bought it from mentioned sommet09:31
(elmo/#launchpad) I don't recall what it is tho09:31
(BradB/#launchpad) s/bought it/bought my powerbook/09:31
(BradB/#launchpad) sabdfl: did you try the store instructions in linux?09:31
(daf/#launchpad) SteveA: if I get stuck, I will ask for help09:32
(BradB/#launchpad) sabdfl: see, what was really fux0red is that in doing this restore, it ran the test suite more than once too...did all kinds of bizarre things, including blowing away my Applications folder on OS X. I don't even have a bloody terminal anymore, and sshd ain't running, so i can't come in from the outside.09:32
(SteveA/#launchpad) daf: I'd like to get involved, and understand what's going on, so I'm ready to help when needed. what do you think of setting up a screen session?09:33
(daf/#launchpad) well, I think I just found the cause of the problem09:33
(SteveA/#launchpad) cool!09:35
(daf/#launchpad) carlos: around?09:37
carlosdaf: yes09:43
(daf/#launchpad) can we discuss POFile.updateStatistics?09:45
(BradB/#launchpad) elmo: can i use my pbg4's not-wireless internet card fine in ubuntu?09:47
carlosdaf: what happens there?09:48
carlosBradB: the ethernet ?09:48
carlosBradB: it works here09:48
(elmo/#launchpad) BradB: the wired port? sure09:48
carlosBradB: the only problem with new powerbooks is the bradcom wireless 09:48
(elmo/#launchpad) carlos: and the modem09:49
carloselmo: well, I never remember the modem...09:49
(elmo/#launchpad) but yeah they got suspend/resume now \o/09:49
carloselmo: isn't there a driver to use it?09:49
carlosa propietary one, I know09:49
(BradB/#launchpad) well, the machine i'm on is completely hosed. if i close this instance of x-chat i won't be able to reopen it.09:49
(elmo/#launchpad) carlos: only for i386 AFAIK09:49
(BradB/#launchpad) everything blown away, can't even send bloody mail anymore.09:49
(BradB/#launchpad) going over to other machine to burn warty, which i downloaded while i was out for lunch09:50
carloselmo: oh, I thought they are for ppc machines...09:50
(daf/#launchpad) carlos: first of all, I'm not clear on the newImport flag09:51
(daf/#launchpad) carlos: can you try explaining it to me?09:51
carlosdaf: ?09:53
(daf/#launchpad) s/flag/parameter/09:53
carlosdaf: without it, the "current" count will be broken09:54
carlosbecause it will take in account the msgsets that where moved from fuzzy to "complete"09:54
carlosor the ones that were moved from untranslated to fuzzy or  complete09:55
(daf/#launchpad) hmm, you're saying things, but my brain is not working09:55
carloswith it we can control that it will only change when we are importing a new pofile09:55
carlosdaf: X-)09:55
carlosblah, blah, blah, blah...09:55
=== daf nods and smiles
carlosdaf: should I improve the explanation?09:56
carlosor just wait for your brain?09:56
(daf/#launchpad) what does "current" mean?09:56
carlosthe number of translated strings in last .po import09:57
(daf/#launchpad) biab, workraving09:57
(daf/#launchpad) let's try again10:15
(daf/#launchpad) so, we have thre states?10:16
(daf/#launchpad) current, updated, rosetta?10:16
carlosand updated is not working correctly with current schema (or I don't know how to fix it)10:17
carloscurrent == translated messages from latest .po imported10:17
(daf/#launchpad) oh, is there a bug open about that?10:17
carlosupdated == messages that were already translated but were updated from rosetta10:17
carlosrosetta == messages that were fuzzy or untranslated and we finished them from rosetta (and they were not imported again with a new po import)10:18
carlosdaf: no, I just noticied after read my comment in the code10:18
(daf/#launchpad) and these states are mutually exclusive?10:18
carloscurrent and rosetta are10:19
carloscurrent and updated are not10:19
carlosforget that10:19
carlosyes, the states are mutually excusive10:19
(daf/#launchpad) ok, could you add a comment saying so?10:19
carlosa updated string cannot be in rosetta10:19
carlosdaf: in the fields declaration or in the updateStatistics method?10:20
(daf/#launchpad) either, both10:20
(daf/#launchpad) the code, for sure10:20
carlosI think it's defined in the interface10:20
carloslet me check10:20
(daf/#launchpad) or perhaps we should have more comprehensive documentation about statistics separately10:21
carlosdef currentCount(language=None):10:23
carlos        """Returns the number of msgsets matched to a potemplate for this10:23
carlos        project that have a non-fuzzy translation in its PO file for this10:23
carlos        language when we last parsed it."""10:23
carlos    def updatesCount(language=None):10:23
carlos        """Returns the number of msgsets for this project where we have a10:23
carlos        newer translation in rosetta than the one in the PO file for this10:23
carlos        language, when we last parsed it."""10:23
carlos    def rosettaCount(language=None):10:23
carlos        """Returns the number of msgsets where we have a translation in rosetta10:23
carlos        but there was no translation in the PO file for this language when we10:23
carlos        last parsed it."""10:23
(daf/#launchpad) ok, I'm still trying to grasp this parameter10:24
carlosthe main idea is10:25
(daf/#launchpad) when do you use it, when do you not use it?10:25
carloswhen you import a new pofile, the numbers could change a lot10:25
carlosit's used only when you import a new pofile10:25
(daf/#launchpad) as opposed to?10:26
carloswhen you translate from rosetta and add a new string10:26
(daf/#launchpad) is it an optimisation?10:26
carlosso you only need to recalculate updates and rosetta10:26
carlosin fact I don't think it should be called if it's not after an import 10:27
carlosI mean10:27
carlosthat method should only called after a new import10:27
carlosor if the statisticis cache could be out of sync10:27
(daf/#launchpad) what happens if you call it and you haven't done an import?10:27
carlosin which case the argument you don't like comes in the game10:27
carlosif the argument tells you that it's not an import10:28
(daf/#launchpad) well, I'm not sure whether I don't like it yet, I need to understand it first :)10:28
carlosthe updates and rosetta fields will be updated10:28
carlosthe current field will remain the same because there were not new imports10:28
(daf/#launchpad) so it is an optimisation?10:28
carlosthe method itself no10:29
carlosthe fields, yes10:29
(daf/#launchpad) I mean the parameter10:29
carlosif you execute it without a poimport the values will be broken10:29
(daf/#launchpad) well, you are trying to save some work by not always recalculating the current field, right?10:29
(daf/#launchpad) oh10:30
carlosyou cannot calculate the current field when you get a new translation from rosetta10:30
(daf/#launchpad) why not?10:30
carlosbecause the information is lost10:30
carlos(the fuzzy flag)10:30
carlosyou cannot know if it was fuzzy or untranslated10:31
carlosthat's why I suggested to move the fuzzy flag from msgset to the sighting 10:31
(kiko/#launchpad) SteveA, should I file a bug on you?10:31
carlosso we don't lose that information10:31
(daf/#launchpad) ok, I didn't realise that10:31
carloswell, we could lose it that why also10:31
carlosbut it's less usual10:32
carlos /s/why/way/10:32
(daf/#launchpad) don't we have information on the origin of a translation?10:32
carlosnot really10:32
carlosyou add a new translation from rosetta10:32
carlosthen export a po file10:32
carlosand reimport it again10:32
carlosthe origin will be rosetta10:32
carlosbut it was in the po file10:33
(daf/#launchpad) argh10:33
carlosdaf: funny, isn't it?10:33
(daf/#launchpad) this is annoying :(10:33
(SteveA/#launchpad) kiko: sure, thanks10:33
=== salgado [~salgado@200-206-134-238.async.com.br] has joined #launchpad
carlosdaf: do you want I work on a proposal so we don't lose any information?10:34
(daf/#launchpad) hmmm10:34
(daf/#launchpad) I don't like the fact that I have a hard time understanding our database code10:38
(daf/#launchpad) I would like to fix that10:38
=== BradB_ [~bradb@modemcable165.196-131-66.mc.videotron.ca] has joined #launchpad
(daf/#launchpad) not just the code, but the schema also10:38
carlosdaf: the dia image is a good start point10:39
(daf/#launchpad) well10:39
=== BradB_ is now known as BradB|osx
carlosbut I think it's obsoleted, we should update it10:39
(daf/#launchpad) that's a good conceptual view10:39
(daf/#launchpad) but if I want to know "how are statistics updated?", it's not much use10:39
salgadolifeless, rsyncing worktrees will trigger that "corrupt library" error because of the pristine tree located inside it, right?10:40
(daf/#launchpad) carlos: ok, let's sidetrack for a moment10:40
(daf/#launchpad) when you import a PO template, how are statistics updated?10:40
(lifeless/#launchpad) you will only have pristines inside a working tree if you don't have a greedy revlib, and everyone should have a greedy, sparse, revlib.10:40
carlosdaf: well, we don't have a policy about how to update the statistics...10:40
(daf/#launchpad) what do you mean by "policy"?10:41
carlosdaf: we need to define that procedure10:41
(daf/#launchpad) I mean10:41
carlosdaf: I just finished the script lalo started 10:41
(daf/#launchpad) is there a function somewhere that updates the statistics for a template?10:41
carlosthat does not mean we have documented or defined a way to update the statistics10:41
carlosdaf: no10:42
(daf/#launchpad) ok10:42
carlosdaf: and now that you talk about it...10:42
carlosdaf: is not doable with the current schema10:42
salgadolifeless, ok, so i'll have no problems. I already have sparse and greedy revlibs10:43
(daf/#launchpad) lifeless: "{tla,baz} get" doesn't create a pristine if the revision you're getting is in the revlib?10:44
(lifeless/#launchpad) right10:44
(daf/#launchpad) carlos: why not?10:44
(lifeless/#launchpad) and if the revlib is greedy, it will add it to the revlib rather than making a pristine in the project tree.10:45
(daf/#launchpad) carlos: same sort of problem as before?10:45
carlosdaf: yes, I just noticied it10:45
carlosdaf: if we don't have the original data that was imported, the current value cannot be updated10:46
(daf/#launchpad) hmm10:46
(daf/#launchpad) hmm, rosetta-states.dia is a good idea, but I think it needs updating10:47
carloswhat's that?10:47
(daf/#launchpad) lib/canonical/rosetta/docs/rosetta-states.dia10:48
carlosyep, it should be updated with more info10:49
(daf/#launchpad) suggestion: we have an RST or DocBook document in docs/ which has information on this sort of thing10:49
(daf/#launchpad) of course, it'll only work if we keep it up to date :)10:49
(daf/#launchpad) but we have some documentation in the wiki, and we have some documentation in docs/10:50
(daf/#launchpad) it would be good to bring it together10:50
(daf/#launchpad) so we can have cross-references10:50
carlosI don't know anything about RST10:51
(daf/#launchpad) DocBook?10:51
carlosso we move from wiki to docs/?10:51
carlosdaf: yes, I know docbook10:51
(daf/#launchpad) yeah, me too10:51
(daf/#launchpad) RST is very simple, though10:51
(daf/#launchpad) yeah, I like the idea of keeping the docs and the code close together10:52
carlosdaf: what do you think that fits better to our needs10:52
(daf/#launchpad) it's easier to update both at the same time, then10:52
carlosI don't have any problem learning new formats10:52
(daf/#launchpad) I think either format would work10:52
(daf/#launchpad) my personal preference is for DocBook10:52
carlossame here10:52
carlosdocbook with xml templates10:52
(daf/#launchpad) ok, let's open a bug on this10:52
=== ddaa [~ddaa@deep-thought.xlii.org] has joined #launchpad
(daf/#launchpad) ah, except Malone is down10:53
(daf/#launchpad) I'll do it later10:53
carlosis down?10:53
(daf/#launchpad) yeah, maybe it didn't come back after the reboot10:53
carlosright, the reboot10:54
=== daf wonders if that was a strategic exit by elmo there :)
salgadolifeless, is it possible to share a revlib between multiple arch users, using baz?10:55
(daf/#launchpad) BradB|osx: who can bring the dogfood back up?10:55
(lifeless/#launchpad) yes and no. if the permissions on your files allows it, and you don't hit race conditions, them yes. but its not guaranteed, and not supported.10:56
(kiko/#launchpad) BAZ ROX10:56
(kiko/#launchpad) hmmm10:56
salgadokiko, so, I think it's not a good idea10:56
(kiko/#launchpad) lifeless, will it ever be supported? it would mean significant savings for us..10:57
(lifeless/#launchpad) kiko: its not possible in general. So supporting it is hard.10:57
(lifeless/#launchpad) its not possible because...10:57
(lifeless/#launchpad) say I have a file thats mode 0700.10:57
(lifeless/#launchpad) its in my revlib with that mode. Now, HTF can that be shared ?10:57
(kiko/#launchpad) I see.10:58
(daf/#launchpad) hmm, so it would only be possible if the on-disk permissions were separate from the revision permissions10:59
(kiko/#launchpad) you'd need some intelligent revlib service.10:59
(daf/#launchpad) well, it could be quite dumb10:59
carlosdaf: ok, returning to more urgent things... can I help you with the pot/po imports?10:59
(kiko/#launchpad) from which you requested the files and then synthesized them into the correct perms.10:59
(daf/#launchpad) carlos: once I fix this bug in poimport.py, probably10:59
(daf/#launchpad) carlos: the bug is: it calls update_stats() when language is None11:00
(daf/#launchpad) yeah11:00
carloslet me check I don't forgot to deactivate some "debug" changes I did...11:00
(lifeless/#launchpad) kiko: its doable, but it makes it a lot more complex, even for the dumb case. That said, we have to do some stuff in this area anyway, I'll certainly try not to make it worse than it is.11:01
carlosdaf: ok, my fault, I forgot to commit a fix for that11:01
(daf/#launchpad) yeah?11:01
(daf/#launchpad) there are a few changes I'd like to make to that script11:02
carlosdaf: could you check it this way?:11:02
carlosprint "Importing %s ..." % options.file11:02
carlos            bridge.imports(person, in_f, options.project, options.product,11:02
carlos                           options.potemplate, options.language)11:02
carlos            if options.language is not None:11:02
carlos                print "Updating %s pofile for '%s'..." % (11:02
carlos                    options.potemplate, options.language)11:02
carlos                bridge.update_stats(options.project, options.product,11:02
carlos                                    options.potemplate, options.language, True)11:02
(daf/#launchpad) ah, looks good11:02
carlosdaf: will you commit that change or should I do it?11:03
(daf/#launchpad) well, easier for you to do it, I think11:03
(daf/#launchpad) so go ahead11:04
BradB|osxdaf: stub11:04
(daf/#launchpad) garh11:04
(daf/#launchpad) we need to serviceify launchpad11:04
BradB|osxi've already filed a bug on that (at least the part about making it run gracefully as a deamon)11:05
(daf/#launchpad) groovy11:06
carlosdaf: merge request sent11:11
(daf/#launchpad) thanks11:11
(daf/#launchpad) SteveA: there's a class here that inherits from PlacelessSetup -- what does that mean?11:13
(SteveA/#launchpad) where is "here" ?11:22
(SteveA/#launchpad) PlacelessSetup is a mixin class used in zope3 unit tests11:22
(daf/#launchpad) in lib/canonical/rosetta/scripts/poimport.py11:22
(SteveA/#launchpad) it sets up and tears down the framework such as the adapters service and utilities services11:23
(daf/#launchpad) this class is not calling its superclass's __init_ or anything11:23
(SteveA/#launchpad) that code is a mess11:23
(daf/#launchpad) yes :(11:24
(SteveA/#launchpad) at first glance, the PlacelessSetup is a decoy11:24
(daf/#launchpad) ok, I'll try removing it and seeing if it still works11:25
(SteveA/#launchpad) also,11:25
(SteveA/#launchpad) move all of the stuff in the if __name__ == '__main__': section at the end into a main() function11:25
(daf/#launchpad) I don't see the point of the exception munging either11:25
(SteveA/#launchpad) then split out the option parsing part into a parse_options() part11:25
(daf/#launchpad) yes, good idea11:25
(daf/#launchpad) I don't want to make too many changes before I've merged carlos's changes to this file11:25
(SteveA/#launchpad) ok11:26
(daf/#launchpad) (which should allow imports to work again)11:26
(SteveA/#launchpad) and it should be formatted more like pop811:26
carlosdaf: it's a trivial change, so it should not be too difficult11:26
(SteveA/#launchpad) pep811:26
(daf/#launchpad) what formatting is lax?11:26
(SteveA/#launchpad) line 42 is too long11:27
(SteveA/#launchpad) line 46 is a mess11:27
(daf/#launchpad) right11:27
(SteveA/#launchpad) another issue is the try:except: around line 4611:27
(SteveA/#launchpad) it is around way too much11:27
(SteveA/#launchpad) it should be around the absolute minimum11:28
(daf/#launchpad) I'm thinking of getting rid of the try/except completely11:28
(SteveA/#launchpad) and even better, it shoudl use a check of the length of the results11:28
(SteveA/#launchpad) rather than half expecting an index error11:28
(SteveA/#launchpad) just sloppy programming11:28
(SteveA/#launchpad) lines 61 onwards -- bad formatting11:29
(SteveA/#launchpad) not sure what it's doing in the except: clause anyway11:29
(SteveA/#launchpad) and incoprehensible XXX comments11:29
(SteveA/#launchpad) there's more, but I'll stop there11:30
(SteveA/#launchpad) ok, so you reckon that when you're done with fixing this code, you can do the imports11:31
(SteveA/#launchpad) and the dogfood stuff will be done11:31
(daf/#launchpad) and the virtual hosting11:31
(SteveA/#launchpad) ok11:31
(SteveA/#launchpad) great11:32
(SteveA/#launchpad) I must go to sleep now. please send me an email to let me know how it goes11:33
carlosSteveA: night11:33
carlosdaf: if you are going to improve the script, could you do it in a way I could use importing it from other scripts?11:33
(daf/#launchpad) SteveA: sure, thanks11:34
(daf/#launchpad) carlos: yeah, I like that11:34
carlosdaf: thanks11:34
(daf/#launchpad) hmm, the merge seems to be taking a long time...11:38
carlosit was just a line change...11:38
(daf/#launchpad) lifeless: did PQM come back up after the reboot?11:39
carlosdaf: the merge is done here11:40
carlosI got the confirmation mail11:40
(daf/#launchpad) ah11:40
carlosI thought you were talking about the star-merge11:40
(daf/#launchpad) did dilys announce it?11:41
carlosdaf: I don't think so11:41
(daf/#launchpad) grr11:41
(lifeless/#launchpad) daf: pqm is cron based, it always comes up11:42
carlosdaf: dilys needs some love, he miss you since you went to NY :-P11:42
(daf/#launchpad) carlos: heh11:42
(daf/#launchpad) I'll spend some time with her later11:42
(daf/#launchpad) I'm going out right now11:42
(daf/#launchpad) I'll carry on working on the imports later11:43
carlosdaf: could you CC: to me the mail about the import status?11:43
(daf/#launchpad) sure11:43
(daf/#launchpad) later!11:43
carlosSteve always asks me but I don't know what to answer because I go to sleep 11:43
carlosdaf: thanks11:43
carlosdaf: later!11:43

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