=== jblack_ [~jblack@mercury.merconline.com] has joined #launchpad
=== spiv_ [~andrew@fuchsia.puzzling.org] has joined #launchpad
=== debonzi [~debonzi@] has joined #launchpad
jblack_debonzi: ping12:39
lalomy bug disappeared :-(12:42
laloah, no, here it is. :-)12:42
lalowhoop, there it is...12:49
=== lalo [~lalo@200-096-070-061.paemt7014.dsl.brasiltelecom.net.br] has joined #launchpad
dafSteveA: is it a Soyuz or a Launchpad meeting that's happening at 12:00 UTC?01:02
kikoon monday? soyuz afaik.01:08
carlosdaf: tomorrow is launchpad, on monday, soyuz01:09
dafoh, I see01:12
=== kiko is now known as kiko-afk
lalodaf: is there any way yet to run the importer from the web interface?01:26
laloI have reason to believe #1935 would only happen in the script and ftest - it seems to be tied to the "zopeless" code in canonical.database01:26
dafno, you can't import over the web yet, but it will be imported01:27
lalopity - I would like to test this hypothesis, but I don't want to spend an hour hacking traversal just for this :-P01:28
dafyou might well be right01:30
laloI'm right :-P01:34
lalozopeless is simply discarding values01:34
lalothe only thing that acutally *happens* on a zopeless setup is an insert, or so it seems01:34
debonzijblack_, pong01:34
=== lalo takes a big breath and starts trying to figure it out
dafI traced the code01:35
dafwhat happens is this:01:35
dafin the adapter code, something like "msgset.sequence = self.len" is executed01:36
jblack_debonzi! Lets fix your problem. :) 01:36
dafthis invokes an SQLObject function01:36
dafwhich sets a value somewhere and also sets a dirty flag on the msgset object01:36
dafbut it seems that the value is never actually written to the database01:37
dafSQLBase is also involved somewhere01:37
SteveAI think I know what might be happening01:37
SteveAstub and I looked at problems with transactions / writing to the database01:37
SteveAthere's code in canonical.publishing that deals with this01:38
SteveAby clearing an sqlobject or sqlos connection cache01:38
SteveAand explicitly starting transactions etc.01:38
SteveAThis code won't be invoked for a zopeless setup01:38
SteveAperhaps it is needed in this casE?01:39
laloyou're talking about it? sorry :-P was looking at the code01:39
lalohere's what happens01:39
lalothe "zopeful" setup uses sqlos transaction objects01:40
SteveA        name = getUtility(IConnectionName).name01:40
SteveA        key = (thread.get_ident(), name)01:40
SteveA        cache = sqlos.connection.connCache01:40
SteveA        if cache.has_key(key):01:40
SteveA            del cache[key] 01:40
laloso when an object is marked "dirty", it's inserted in a queue of objects that have to be flushed01:40
SteveAdo that at the start of the script01:40
SteveAand get_transaction().commit() at the end01:41
lalothe zopeless setup doesn't have any similar logic - there is no queue of objects to flush, so when the transaction is committed, there is no way to know what objects must be sync'ed01:41
lalo(the "queue flush" thing of sqlos basically calls the sync() method of each object in the queue)01:41
dafSteveA: I tried putting in a get_transaction().commit() at the end01:42
lalothere seems to be, on first read, two possible fixes - either do make a queue of objects to sync(), or don't set lazyUpdates at all01:43
laloI think for a script or ftest the second approach may be better suited01:43
SteveAlalo: I need to go to sleep now.  Can you send a mail to the launchpad list with what you've found out, after you've done a bit more trying things?01:45
dafSteveA: wouldn't adding comments to the bug make sense?01:45
SteveAdaf: perhaps.  I want spiv and stub to see this.01:46
dafthen perhaps they should be CC'd to the bug01:46
=== carlos goes to bed and tries to recover the hours he wasted programming on Powerbuilder instead of sleep since last week
carlossee you tomorrow01:46
laloI think I'll post to the list with a summary of what's going on and an URL to the bug01:47
lalocarlos: 'nite01:47
lalosuppose we need a component for this kind of stuff (database, sqlo, sqlos, canonical.database)01:49
=== lalo does a dance
daflalo: yes, a component would make sense01:53
laloremoving SQLBase._lazyUpdates makes my tests pass :-D01:55
laloyay for being right01:55
lalojust have to be sure to commit at the end of the script01:56
lalo...which we're already doing, so it's ok01:56
lalolaunchpad_test=# select * from pomsgset where sequence = 0;02:00
lalo id | primemsgid | sequence | potemplate | pofile | iscomplete | obsolete | fuzzy | commenttext | filereferences | sourcecomment | flagscomment02:00
lalo(0 registros)02:00
=== kiko-afk is now known as kiko
daflalo: hurrah!02:02
lalohmm. canonical.database is in launchpad, so technically I can commit02:03
lalonot very sure how to proceed in this case :-/02:03
lalowho is the right person to poke about the bugzilla component?02:06
=== lalo decides the list is it :-P
dafSteve or Dave can do it02:07
lalothen the list is it indeed02:10
debonziSteveA, Hi02:16
debonzispiv_, ping02:17
debonzispiv_, I saw that some links in sourcepackage still broken.. I'm gonna to fix it tonight I sent the changes to rocketfuel tomorrow morning.. see u02:21
debonzis/I sent/and send02:21
=== debonzi go dinner
kikojblack_, you missed him again :-(02:22
jblack_no I didn't.02:23
jblack_that's why he disapeared for awhile right after he came back. I hooked 'em02:23
kikodid you fix it? :)02:26
=== jblack_ grins
=== lalo is off for the night - be back in a few hours
=== lalo [~lalo@200-102-075-177.paemt7014.dsl.brasiltelecom.net.br] has joined #launchpad
=== kiko is now known as kiko-afk
=== stub [~zen@dialup-] has joined #launchpad
=== lalo [~lalo@200-180-185-073.paemt7003.dsl.brasiltelecom.net.br] has joined #launchpad
=== stubish [~zen@dialup-] has joined #launchpad
=== justdave_ [~dave@] has joined #launchpad
=== justdave_ is now known as justdave
=== jblack_ is now known as jblack
=== sabdfl [~mark@wblv-228-55.telkomadsl.co.za] has joined #launchpad
=== Signon time : Tue Sep 7 09:17:35 2004
=== Signoff time : Thu Sep 9 09:51:49 2004
=== Total uptime : 2d 0h 34m 14s
=== warthylog [~warthylog@port1845.ds1-khk.adsl.cybercity.dk] has joined #warthogs
=== Topic for #warthogs: /join #ubuntu
=== Topic (#warthogs): set by Keybuk at Tue Sep 7 20:30:22 2004
=== #warthogs [freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify: http://freenode.net/faq.shtml#nicksetup
=== warthylog [~warthylog@port1845.ds1-khk.adsl.cybercity.dk] has joined #launchpad
=== Topic for #launchpad: <kiko> how's england? <daf> wet <daf> but soon, I will go back to Wales <kiko> and how will wales be? <daf> probably wet, too
=== Topic (#launchpad): set by kiko at Fri Aug 27 16:07:21 2004
=== #launchpad [freenode-info] why register and identify? your IRC nick is how people know you. http://freenode.net/faq.shtml#nicksetup
=== Signon time : Thu Sep 9 09:51:53 2004
=== Signoff time : Thu Sep 9 09:53:00 2004
=== Total uptime : 0d 0h 1m 7s
=== warthylog [~warthylog@port1845.ds1-khk.adsl.cybercity.dk] has joined #launchpad
=== Topic for #launchpad: <kiko> how's england? <daf> wet <daf> but soon, I will go back to Wales <kiko> and how will wales be? <daf> probably wet, too
=== Topic (#launchpad): set by kiko at Fri Aug 27 16:07:21 2004
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
luluHi guys - anyone know how the Rosetta logo came about? Who made it? and is it from clip art?11:01
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
luluhi carlos12:15
=== doko [doko@dsl-082-082-140-133.arcor-ip.net] has joined #launchpad
lulucarlos: the Rosetta server is down - do you know how to fix it?12:51
carlosno, sorry. I don't have access to it12:52
carlosdaf is the only one with an account in that server and perhaps lalo, but I'm not sure12:52
luluI am sure spiv and Stu also have.12:54
luluspiv: u around?12:54
spiv_lulu: Just woke up :)12:54
lulumorning hon! sleep well?12:54
luluspiv: would u mind checking the rosetta server mate - it's down and I'd like it available if at all possible. thanks in advance :o)12:55
spiv_I don't think I have an account on it.12:55
spiv_I just tried, and I don't seem to :/12:56
spiv_lulu: I always sleep well, it's the waking up that's the problem ;)12:56
lulummm - this isn't good - we need to have more than one person who can kickstart it when down....can't rely on only one person. All launchpad team leaders should have access. I'll chat to Stevo.12:57
luluspiv: :o)12:57
luluSteveA: ping12:57
luludaf: ping12:58
luluCarlos: how are your graphs looking?12:58
=== spiv_ is now known as spiv
sabdflmeeting in 60 mins right?01:00
carloslulu: the graphs are in place, waiting for the changes from Limi01:00
carlossabdfl: yes01:00
lulucarlos: what changes are required from limi?01:02
carloslulu: png and css changes01:02
carloswe are using still my images01:02
lulucarlos: ok. I'll add that to the to do list. Anything else you need from limi next week on Rosetta?01:03
carloslulu: he should do a global review of all UI, It's the bug #193401:04
lulucarlos: ok - does he have an account on Bugzilla?01:05
lulucarlos: good - thank you.01:05
lulucarlos: as the server is down - I can't see if the warty applications have been added to the home page of Rosetta.01:07
luluhave they?01:07
carloslulu: limi added some code to the templates01:07
carlosbut it's not done01:07
carloswait I will take a screenshot of my local installation01:08
luludoes that need limi, or can you guys just hard code the links?01:08
carlosIf that's the final UI it will have, we don't need limi01:09
carlosthat's the bug report related with that feature01:09
carloslulu: and this is the metabug https://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=1915 that list all pending things for the Alpha release01:10
carlosany bug report you add should block that one01:10
carlosthat way we are sure we don't forget anything01:10
=== lulu just creating a bugzilla account...
carloslulu: http://gollum.pemas.net/~carlos/Screenshot-Rosetta.png01:12
carlosthat's the main page now01:12
luluok - the translatable packages in Ubuntu - do we have all of those in the database?01:13
luluif not, we need to get them in, with the correct title, descriptions and package names.01:14
lului haven't looked at the bug, but is inputting the data part of it?01:14
carloslulu: we are creating the scripts to do it01:14
lulucarlos: who is? and when will they be done?01:14
carlosand we need to talk with arch people to fetch the projects from the arch repository01:14
luluok - who's responsible for this task?01:15
carlosI'm creating the script to create persons into launchpad01:15
carlosso I suppose I will create also that script01:16
lulucarlos: it's assigned to Daf I see....depends on how many bugs you have, Daf has, Lalo has.....and if he needs to delegate it to you.01:18
luluhow many rosetta bugs do we have outstanding?01:18
carloslulu: all bug reports are assigned by default to Daf01:19
carlosas soon as we are free, we pick any bug with the NEW tag and change the assgned field01:19
luluok - if you want to take it on, then great. we go live next week.....01:20
carloslulu: https://bugzilla.warthogs.hbd.com/bugzilla/showdependencygraph.cgi?id=191501:20
carlosthere you will see the status with a graph01:20
carlosthe green ones are the open ones, the grey ones are fixed already01:21
luluok - lots to do....I'll let you get on with it Carlos...hope Daf wakes up soon.01:22
carloslulu: ok, but don't worry, some bugs are trivial to fix01:24
lulucarlos:then lets get them done....running out of time! :o(01:25
carlosdaf: morning01:34
luludaf: i tried to call u this morning. jane wants to know where the rosetta logo came from - did u design it?01:35
daflulu: hmm, I think my phone was downstairs01:36
daflulu: I went to bed at 7am, so I might not have heard it annyhow :)01:36
daflulu: yes, I designed the logo01:36
luludaf: cool - did u use clipart?01:37
daflulu: nope01:37
luludaf: drew it?01:37
luludaf: good - will you get in touch with Jane then - she sent an email to Warthogs. cheers!01:38
luludaf: why bed so late/early?01:38
daflulu: it happens to my sleep patterns... I was fixing bugs :)01:38
daflulu: right, hadn't read my mail yet today -- I'll reply now01:39
luludaf: good to hear u fixing bugs :o) we have 4 days to go, so we need to squash as many as possible asap. hows; the schedule looking today?01:40
luludaf: thanks :o)01:40
carlosdaf: wow... no more plural forms missing error :-P01:50
dafcarlos: yeah, I finally got around to fixing it01:52
carlosdaf: the po traversal is working now.01:53
dafoh, great!01:54
carloswhen we don't have a pofile, I suppose we could export the .pot file, right?01:54
carlosit fails now01:54
carlosor show only a message?01:54
dafwhat's the URL?01:54
carlosit's not yet at rocketfuel01:54
carlosbut for instance:01:54
dafI'm not sure which is best01:55
carlosthat will fail because we don't have the cy pofile (yet)01:55
dafby the way: what do you think on having the PO file as an attachment vs. having it inline?01:55
carloshmm, in fact, http://localhost:8085/++skin++Debug/rosetta/projects/gnome/evolution/evolution-2.0/cy also fails so we should fix it also01:55
carlosdaf: GTP people prefers them as inline01:55
carloslike you01:56
dafwhat's your opinion?01:56
carlosdaf: that's an easy change01:56
carlosI think that you should download it and open with an editor01:56
carlosbut it's long ago since my last active translation, so perhaps I don't have in mind the real user case01:57
dafI'm not 100% sure about -- I prefer inline myself, but I don't know what other people like01:57
carlosI suppose we will know it with the alpha :-)01:58
dafcarlos: once you have finished the traversal, do you have other things to work on?01:59
carlosChoose any bug from the list of pending things02:00
SteveAhi folks.  Launchpad meeting here in 1 minute!02:00
carlosSteveA: hi02:00
carlosdaf: I was thinking on implement the project/product/resource creation script02:00
sabdflhi all02:00
=== cprov [~cprov@200-206-134-238.async.com.br] has joined #launchpad
spivcprov: Hello :)02:01
cprovspiv: hi 02:01
=== debonzi [~debonzi@200-206-134-238.async.com.br] has joined #launchpad
spivJust in time :)02:02
dafhi mark02:02
carloshi * 02:03
SteveAcprov, debonzi: hi.  Can kiko make it to this meeting?02:03
cprovSteveA: of course, just one minute more :)02:03
SteveAok, thanks02:03
debonziSteveA, Hi02:03
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
sabdflok, should we get going?02:04
kikoby me, bombs away02:04
sabdflstevea, could you set the stage, then let's have each team describe status and issues02:05
sabdflgo ahead steve02:05
SteveAwhat I'd like to do today is to talk with each team about exactly what the current short-term and medium-term goals for their part of launchpad is02:07
=== sabdfl [~mark@wblv-228-55.telkomadsl.co.za] has joined #launchpad
SteveAand how they are progressing towards those goals02:07
SteveAwhat things are standing in the way of progres02:08
SteveACan we start with the soyuz team?02:08
SteveAI spoke with kiko yesterday about how that is going02:09
SteveAand we're having a meeting on monday to go through the system, as checked in02:09
SteveAand go through all of the soyuz functionality02:09
SteveAbut first,02:09
SteveAwhat are the current short-term and medium-term goals of the soyuz project?02:09
kiko(short-term meaning pre-sprint, I assume)02:10
kikowe've got the functionality we discussed at the oxford hackathon laid down02:10
kikoand we have been producing mockups and implementing them02:11
sabdflhow are we doing on the db-to-html mapping?02:11
kikosomething like fire and motion02:11
sabdflfire and motion?02:11
kikoall pages are intended to be db-backed, and there are very few things left that are stubbed out02:11
sabdflok, is this accessible on rosetta.warthogs.hbd.com yet?02:12
SteveAkiko: am I right in recalling that you think that all will be DB-backed by the meeting on monday?02:12
kikomost of the stubbed out information remaining is stuff that hasn't been really discussed or doesn't have db backage yet.02:12
kikoSteveA, well, certainly not *all* of the content in the pages, but all the content which can be displayed from the database (given our existing database infrastructure)02:13
kikoif you look at the checklist and at the pages you'll see there's not a lot left.02:13
SteveAso, on monday, I'd like to go through each relevant database entity, and we'll look at the page(s) that present it02:13
sabdflin that case well done02:13
kikoas I said yesterday, by the middle of next week, unless we have a trusted domain expert that can guide us into moving the database forward together with the pages, we'd like to be assigned some extra work.02:14
SteveAkiko: so, there are some changes to the database that you want to discuss.  Can these be discussed at the sprint?02:14
kikothey can, but that leaves us with ~ 10 days that could be invested somewhere that needed it.02:14
SteveAsabdfl: how about I get at least one of them to help me with infrastructure?02:15
sabdflgood idea02:15
SteveA(actually, kiko suggested it in our meeting yesterday)02:15
SteveA(thanks kiko)02:16
sabdflwe need a consolidation / cleanup process that spans the infrastructure02:16
SteveAI can lead that, but I don't have time to execute it all02:16
kikonow, medium and long-term we are *really* in need of someone to guide us more closely02:17
sabdflsteve, can you map out the areas that need that work for me this afternoon, then I can give the go-ahead on async time for that02:17
SteveAyou mean, a package maintainer, a user of souyz02:17
kikowell, a soyuz user, yes, but someone with a reasonable vision of how everything should fit together and time to spend with us02:18
SteveAsabdfl: I can send you a short bullet-pointed list.  Will that be enough?  I already have a lot of work planned for this afternoon.02:18
sabdflyes, that's fine02:18
kikoso far we've been winging it based on what we know about distributions and what we discussed at oxford, but we're running low of fuel already02:18
SteveAsabdfl: ok, willdo02:19
sabdflkiko: we can cover a lot of ground during the sprint02:19
sabdfldaf: rosetta passwd?02:19
kikosabdfl: great, but we would appreciate someone after that that could be in close contact, do you think we can arrange someone with the cycles for that?02:20
dafsabdfl: wartyhoarygrumpy, I think02:20
spivkiko: I should be more available from now on.  The US leg of my travelling was more disruptive than I'd anticipated.02:20
SteveAspiv: even so, you're not a domain expert.  You don't know what a package maintainer needds.02:21
kikospiv, problem is we need someone who really can make decisions based on domain knowledge02:21
sabdflthat did the trick, thanks02:22
spivGood point.02:22
SteveAkiko: the immediate goal is presenting what's in the database02:22
kikothat's already in its final steps02:22
SteveAwe'll look over progress towards that on monday02:22
sabdflelmo and kinnison will get going on archive management stuff at high speed02:22
kikosoyuz team needs a more distant horizon to work with from there on :-)02:23
sabdflas soon as we are using their new code for archive management we can dogfood ubuntu, which will mean LOTS of users02:23
SteveAkiko: are there any database changes that you need?02:23
sabdfli believe our db is already a nice superset of the existing db management stuff, so that should go quickly02:23
kikoSteveA, there are some, yes. 02:23
SteveAkiko: is there a wiki page, or something like that, listing them?02:24
SteveAsabdfl: when will elmo and kinnison start that?02:24
cprovdaf: is https://rosetta../ working ? I'm just getting blank pages 02:24
SteveAdaf: same here02:25
dafoh, me too02:25
dafit was working a minute ago02:25
sabdflSteveA: at the sprint02:25
kikoSteveA, not summarizing DB changes by themselves, we have a checklist, some wiki bullet items and a couple of email requests. 02:25
kikoSteveA, we'll summarize them.02:25
SteveAkiko: ok, we can go through those on monday too.02:25
dafSteveA: I'm getting some interesting Zope errors on the server02:26
SteveAdaf: let's look at it after this meeting02:26
SteveAis that everything for soyuz so far?02:26
sabdfli have a question02:27
sabdflwhat sort of sample data are you guys using?02:27
spivCompletely hand-crafted.02:27
dafSteveA: hmm, restarting the server made it go away -- perhaps it was demo-bogons02:27
SteveAdaf: keep the logs02:27
spivIt would be good to start basing it on actual real data.02:27
SteveAis there anything stopping us from using real data?02:28
kikoSteveA, well.. is there real data?02:28
dafSteveA: logs are always kept02:28
spivDo we already have tools for extracting metadata from deb packages, and that sort of thing?02:28
SteveAdaf: mail the relevant ones to me please02:28
spivI'd guess sourcerer has some of this stuff.02:28
kikothe sample data is built from what we *think* would be reasonable02:29
kikobut you have to realize that some of the stuff -- release team, for instance -- doesn't exist anywhere that I know of02:29
kikothe package side of things could be extracted from ubuntu debs, definitely02:30
SteveAso, a good task starting next week is to get real sample data in there, augmented with made up data where needed02:30
kikomaybe that's a task for 50% of the soyuz team post-next-week.02:30
kikoyeah, agreement02:30
sabdfli'm trying to figure out if we can get the system to monitor our archive and automatically try to represent the data in there02:30
sabdflelmo: listening in?02:31
SteveAdo you mean monitor, as in "real time", or as in "let's import/sync with the archive from time to time" ?02:32
kikothat would be awesome, but does keybuk's team have the tools up to speed to give us this data?02:32
sabdflsync from time to time02:32
debonzidaf, could you please merge into rosetta.warthogs.hdb.com the fix I've sent to rocketfuel02:32
sabdflthere's a database used to manage the existing archive, it's part of a set of tools called "katie" that james mostly wrote for debian02:32
elmosabdfl: yeah02:32
dafdebonzi: sure02:32
debonzidaf, thanks02:33
sabdfland james and daniel (silverstone) will write a new set for us02:33
dafdebonzi: hmm, it says that everything is up to date02:33
sabdflelmo, is the katie db for auckland on auckland?02:33
elmono, it's on jackass - auckland is the public mirror machine, jackass is the private master archive server02:33
debonzidaf, realy? I've sent a fix for sourcepackage pages some minutes ago.02:34
sabdfldo you think you could send a dump of the db to kiko and me?02:34
kikoah, killer02:34
dafdebonzi: is the merge complete yet?02:34
spivdaf: Yep.02:34
dafyeah, I see it02:34
debonzidaf, I think so02:34
debonzidaf, nice :)02:34
elmosabdfl: it's 2.5Mb compressed - shall I make it available via http and send you the URL ?02:35
dafdebonzi: is it working?02:35
sabdflthanks elmo02:35
dafhmm, no it's broken again02:35
debonzidaf, I've got black page02:35
kikoelmo, url is good.02:36
daflast page accessed before it broke was /soyuz/distros/ubuntu/src/warty02:36
dafsame error as before02:36
SteveAdaf: I see what the problem is.02:36
cprovdaf: It was me :) something is broken on SQLObject02:36
SteveAAnything more for soyuz?02:37
dafcprov: oops :)02:37
SteveALet's move on.02:38
SteveAdaf: the immediate goals for rosetta, 02:38
cprovspiv: can we talk on #soyuz ?02:38
SteveAand how are things progressing towards it02:38
dafimmediate goal: fix all bugs which prevent the alpha release02:38
sabdfldaf: can you give me a status on rosetta first please?02:38
dafyou can see these here:02:38
spivcprov: Ok.02:39
dafsome of those have been fixed, but are waiting on merges to rocketfuel02:39
SteveAdaf: the wiki page that describes what "alpha release" means?02:39
dafSteveA: good idea02:40
dafhttps://www.warthogs.hbd.com/RosettaAlpha describes our goals for the alpha release02:40
elmokiko/sabdfl: sent02:40
kikothanks james02:41
dafonce we have our alpha-critical bugs fixed, we'll have real projects imported into a test server which we'll make available to our testers02:41
SteveAwe'll make the test server available as rosetta.ubuntulinux.org02:41
dafthen, we'll fix any bugs we or our testers find as well as adding functionality for our beta release02:42
SteveAdaf: the other day, we were discussing the "import POT/PO -> alter through the web -> export POT/PO" cycle02:42
dafSteveA: yep02:42
SteveAwhere does that stand now?02:42
dafit has less bugs in it than it did the other day02:42
carlosdaf: Could we add the beta release date to the wiki?02:42
SteveAwhat can I do with it?02:42
dafyou can do all of those things02:43
SteveAI can take a POT file and PO file for ProjectX, and use a script to import it (if I have access to the DB)02:43
SteveAI can take a POT file and a PO file for ProjectX, and upload it to a web form02:43
SteveA(good so far?)02:44
dafyou can import via a script02:44
dafweb upload is not done yet02:44
SteveAno web form yet?02:44
SteveAanything holding that up?02:44
dafthat's bug #191902:44
dafI might need some hand-holding for implementing it, but other than that, there's nothing stopping it02:45
SteveAWhen can we do it?02:45
dafI can do it today02:45
SteveAYou need some help with the HTML form / Zope 3 aspect of it?02:46
dafI'll start working on it and ask you for help when I get stuck02:46
SteveAThat is very straightforward.  Let's do it at 1500 UTC.02:46
SteveAwe'll start with the zope3 aspect, get that out of the way, and write a wiki page on how it works02:47
sabdflso typical process might be:02:47
sabdfl - > upload a POT02:47
sabdfl - > upload a PO02:47
sabdfl - > add some translations02:47
sabdfl - > fetch a better PO containing translations02:47
dafof course, you need to have the project / product / template set up first02:48
sabdfl - > upload a new POT02:48
sabdfl - > add some translations for the new strings02:48
sabdfl -> fetch a better PO02:48
carlossabdfl: yes02:48
SteveAI want to check the "ready or not" status of the rest of the process.02:48
sabdfland finally:02:48
sabdfl - > upload a new PO that has additional translations02:48
sabdfl - > add some more translations through the web02:49
sabdfl - > fetch an even *better* PO02:49
carlosdaf: we need a policy about who will be able to upload .po/.pot files from the web...02:50
dafand, of course, each step might be carried out by one person or different people02:50
sabdflbut initially this can all work on trust02:50
sabdflin a small community02:51
dafcarlos: yes, this will be a permissions issue -- do you want to file a non-alpha-critical bug for this?02:51
SteveAwe can say that any person who can log into the system can upload files02:51
SteveAthat is, any $Person02:51
dafsure, initially trust will do fine02:51
carlosdaf: sure02:51
SteveAdaf: which of the "->" points from sabdfl are not yet working?02:52
dafSteveA: uploading through the web02:53
SteveAbut, it works with a script02:53
SteveAso, just the web front-end on that02:53
dafSteveA: also, I'm not sure about "- > upload a new PO that has additional translations" -- we need to test combinations of TTW/import translations02:53
daf"not sure about" as in "I'm not 100% confident in it yet"02:54
SteveAdo you have a set of good test-cases written down?02:54
SteveAas in, a list that covers the different kinds of situation02:54
dafno -- that would be a good thing to have02:54
carlosdaf: new bug report, I will fill it02:55
SteveAplease file a bug to do that02:55
sabdflare there any checks to make sure that the PO file being uploaded matches the POT, and the language is correct?02:55
SteveAthanks carlos02:55
sabdflis the language specified inside the PO file, or only in the PO filename?02:55
carlossabdfl: only in the PO Filename02:55
carlossabdfl: no, we don't have such tests02:55
dafsabdfl: we don't check against the POT, but we do mark messages in the PO file which are not in the POT as obsolete on import02:56
sabdflis there any "key" that links a PO to a POT?02:56
carlossabdfl: perhaps we could try to do something with the language mailing list, but I don't think we have time to do it for Alpha release and it will not be robust enough02:56
SteveAthe po/pot system wasn't really designed with rosetta in mind :)02:56
dafso true :)02:57
sabdflok, the nightmare scenario being someone uploads a PO file to the wrong POT :-)02:57
dafwhen that happens:02:57
SteveAwould that look odd, because of a lot of obselete messageS?02:57
carlosSteveA: it should not be a problem02:58
daf - either there is no overlap between the intended POT and the uploaded PO, in which case you just get a lot of obsolete messages in the PO file02:58
lulucan we do any sort of validation on it before allowing it through?02:58
daf -  or, there is overlap, and you get bogus translations02:58
carlosthe big problem is when they upload a PO file to the wrong language02:58
dafcarlos: yeah02:58
daflulu: we could probably have some weak heuristics about "does this PO file look like it belongs to the template/language specified"02:59
SteveAthere are tools out there that attempt to identify the language of some text02:59
sabdflthat sort of thing will blow our translation memory away02:59
sabdflbut, let's just make the web pages simple, and trust people not to screw it up02:59
SteveAwhat do you mean by "translation memory" ?03:00
dafSteveA: a "translation memory" is a database of existing translations03:00
dafSteveA: usually coupled with a mechanism for reusing them03:00
sabdfllater on, we will still think that the word "Achtung" is "Attention" in Spanish if someone uploaded a PO file for German to the Spanish language place03:00
luludaf:mmm - perhaps a weak heuristic may be better than nothing. not critical for alpha but later may be good.03:01
sabdflwe would ave seen the German translation, and recorded it as spanish03:01
dafRosetta is basically a translation application with a translation memory03:01
dafwe're not utilising that memory yet, but we will do in the future03:01
dafhence the Achtung problem03:01
sabdfland since we flatten the history of sightings (we dont store a record of every sighting each time we review a PO file) we have no easy way to tell which sightings came from the mistaken upload03:01
SteveAok.  file a bug on the achtung problem, so that we have a place to record thoughts about dealing with it.03:02
dafwill do03:02
carlossabdfl: if we have the po file we could develop a script that removes the wrong strings03:02
SteveAsounds like a restore from backup03:02
dafcarlos: an "anti-import"? :)03:02
carlosdaf: yes, it will not restore the previous status03:03
carlosdaf: but will remove the broken translations03:03
carlosand I think that will be enough03:03
SteveAthis may be an area where we can use the zodb.  it has the capability to store historical data.03:04
SteveAanyhow, let's focus on rosetta, and progress towards our immediate goals03:04
sabdflwe do store the firstseen date, so we should also be able to eliminate a bunch of sightings on that basis03:05
sabdflanyhow, enough of the achtung problem03:05
SteveAany other problems/sticking points with rosetta?03:05
sabdfldid you separate out potmsgset from pomsgset?03:05
dafsabdfl: not yet03:05
dafsabdfl: none that I can think of03:06
sabdflwhat;s the rationale?03:06
sabdflok, no rationale :-)03:06
dafthe rationale is manyfold:03:06
carloswell, we need to know the process to import projects/products03:06
dafat the moment, we have one table representing two different things03:06
carloswhen, who and what03:06
=== carlos waits until the rationale ends
dafthis makes implementing those two different things in the SQL code more difficult03:07
sabdflok, I'm amenable03:07
SteveAthe "if statements all over the code" problem03:07
dafSteveA: exactly03:07
sabdflgo for it03:07
dafsabdfl: great03:08
dafthe "none that I can think of" was meant to be directed at Steve -- bloody tab-completion :)03:08
carlosthen, the only thing we need to know is the data that should be imported into launchpad for alpha release03:09
carlos(the only remaining thing)03:09
SteveA"the packages in ubuntu that are the most important to get translated"03:10
carlosSteveA: that's too abstract :-)03:10
carloswe need the list03:10
carlosa way to get the code/po/pot files03:10
SteveAwho knows what packages are in need of translation?03:10
luluis it on ubuntu wiki? under Warty?03:10
SteveAin the future, we'd just consult Rosetta, and look at the graphs :-)03:11
carlosand how many times/day/week/month we should update our information (.po and .pot files)03:11
dafhttps://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=1942 -- accidental imports can cause faulty memory03:11
SteveAwho can we ask on the warty team, who would know about which packages need most help with translations?03:13
SteveAcolin watson was doing l10n of the installer03:13
dafwe can only test things which have already been imported03:13
luluSteveA: we need all the packages don't we? 03:13
SteveAwe can ask for things to be imported, if it is important to get help with their translations03:14
dafso I suggest we choose from a list of what's been imported when we're ready to do it, as imports are still going on03:14
carlosSteveA: we don't need the list of packages with translations, with a list of packages is enough, we can try to discover the ones with translations to import into rosetta03:14
dafI don't think we should dwell on this too much now03:14
SteveAlulu: we need to prioritize.03:14
sabdfllet's start with smaller applications, where we have a good relationship with upstream03:14
luluSteveA: yes I agree - ok - but it would be a good test to provide the correct test data we need to cover any Warty problems we may have...03:16
SteveAwe'll get there03:16
lulusabdfl: who will chose the apps?03:16
SteveAhow about a bug that says "come up with the first 10 apps to import" ?03:17
dafSteveA: sounds good03:17
SteveAthose 10 can be listed in the bug comments03:17
sabdfllulu: daf03:17
sabdflwarty team can also make nominations of course03:17
sabdfllike warty installer :-)03:18
SteveAAre we done talking about rosetta for now?03:19
sabdflok let's move on03:20
SteveAWhat is the current immediate goal for malone?03:21
SteveAI was talking with mdz about what's needed to start using malone for ubuntu, sometime between warty and hoary03:22
SteveAhe suggested that it would be a good idea for us to start using malone for launchpad bug tracking, 03:22
SteveAso that we'll know it is somewhat solid in real use, before starting to use it with ubuntu03:23
sabdflfair enough03:23
carlosSteveA: makes sense03:23
SteveAone feature needed for using it for launchpad is also needed for using it with ubuntu: assigning someone to fix a bug in a source package.03:23
sabdflwe can make launchpad a project, with products rosetta, soyuz and malone, and file bugs in it03:23
SteveAthe ubuntu team use this all the time in their bugzilla.  it is currently absent from the RSD.03:23
sabdflthe difference being assignments to product vs assignments to sourcepackage, as in the distro view03:24
SteveAI've asked matt to get someone from his team to make a quick review of the RSD, and say whether anything is missing from it that they would need to use it with ubuntu03:24
sabdflbut quite possibly getting it right for ourselves makes the distroview easier03:24
SteveAfor using it with launchpad, we may want to have a "this is a secret / private" project flag.03:25
SteveAI'm not sure about that, though.03:25
SteveAif we use it for launchpad, would we want it to be publicly visible?03:25
SteveAthere are some advantages to working in the fishbowl :-)03:26
SteveAbut, there are commercial aspects too, perhaps03:26
sabdfllet's work on the same instance that the rosetta alpha will use03:26
sabdflthen when we have it working, we can copy the data over to emperor and not lose any bugs03:26
sabdfland by then we'll have some ability to indicate private projects03:26
SteveAok.  Malone needs a few extra things before we can use it for launchpad, even on a semi-private server.03:27
SteveAI've noticed the rosetta team making good use of bugzilla's dependency tracking features03:28
SteveAwe'll need mail-out of bugs03:28
SteveAstuart can start working on these when he gets back.  any news about andrew v. ?03:29
SteveAor, maybe we can cover that later on the phone03:29
SteveASo, we have the following milestones for Malone:03:31
SteveA    - implement features needed for use with launchpad03:31
SteveA    - dogfood with launchpad on rosetta server03:31
SteveA    - implement "private project" features03:31
SteveA    - implement features for ubuntu03:31
SteveA    - export data to emperor, and start using with ubuntu03:31
sabdfli still need to catch up with him03:31
sabdflis that a wrap?03:32
SteveAAnything else we need to talk about for malone?03:32
SteveAgoing, going...03:32
SteveAThat's it.  Thanks for having the meeting, everyone.03:33
sabdflthanks guys03:33
SteveAWe'll have another (probably shorter) one next week.03:33
sabdfllet's meet weekly from here onwards, steve please coordinate03:33
SteveAI'll mail some notes on this meeting to the list later today, with any Actions not covered by bugs that have been filed in the course of the meeting.03:34
kikosounds good03:34
sabdflSteveA, daf, whats the status on having a single Project object for Rosetta and Soyuz and Malone...?03:34
luluthanks :o)03:34
luluSteveA:  Steve - please can u chat to James on McQuarie setup if not done already. Arch stuff.03:34
SteveAI have that in my tree.  I need to sync with current changes, and get it checked in, and tested.03:34
SteveAlulu: yes, will do03:35
lulucheers mate :o)03:35
sabdfllulu: we close to having a plone setup i can see?03:35
sabdfllulu: can i call you to discuss now?03:35
lulusabdfl: - we've had problems but Upfront are finalising them today.03:36
lulusabdfl: elmo and Upfront are co-ordinating as we speak03:36
carloslunch time, see you later!03:47
kikolater guys03:47
cprovlater 03:50
kikodaf, can you kick rosetta.wh for me?04:13
kikoit's dead again.04:13
dafkiko: there's a bug04:16
spivdaf: Obviously ;)04:16
daf...which won't be fixed by restarting the server04:16
dafwhich actually works 99% of the time04:17
spivAny idea what it is?04:17
dafSteveA has an idea04:17
luluSteveA: can we give all team leaders an account on Rosetta so they don't have to rely/disturb/cajole Daf every time. Each team needs to have access if possible.04:17
kikodaf, if restarting the server won't fix things, what will?04:20
dafkiko: fixing the bug that breaks the server04:20
dafSteveA: any idea on the timescale?04:20
kikodaf, but restarting the server would at least allow me to browse it :)04:34
dafkiko: ok, but don't say I didn't warn you if it breaks again :)04:35
kikoI'm not holding it against you, just fix the #@!#@!#! thing <wink>04:35
SteveAdaf: I know what the *immediate* problem is.04:36
dafkiko: ok, done04:36
SteveAbut we have a greater problem.  apache on that server is doing basic auth to protect the server from inquisitive eyes.04:36
dafSteveA: you know the symptom, but not the cause?04:36
SteveAThe server now does basic auth itself.04:36
dafah, oops04:36
SteveAWe need to stop apache doing basic auth, I think.04:36
dafwhat about the inquisitive eyes?04:37
dafsecurity through obscurity?04:37
SteveAdo we have much to hide?04:38
SteveAat this point, probably not04:38
SteveAthere's nothing important in the database behind that rosetta04:38
SteveAthere's no TTW programming stuff in Zope04:38
dafSteveA: I'm going to have lunch -- we're meeting in 23 minutes, right?04:39
SteveAwould you like a tad longer for lunch?04:39
dafa tad longer, yes04:39
SteveAwe could meet in 50 minutes instead04:40
dafsounds good04:40
daf15:30 UTC?04:40
SteveAok.  that's 1830 my time, 1630 yours 1530 universal04:40
dafsorted, see you later04:40
SteveAlulu: I'm happy for team leaders to have accounts on Rosetta.  I'll mail admins.04:41
luluthanks Stevo :o)04:42
SteveAkiko, debonzi, cprov, spiv: who out of you needs an account on rosetta?04:43
dafSteveA: I suggest the following:04:44
dafwe have a launchpad user/group04:45
dafwhich has a running screen session04:45
dafanybody who needs access is in the launchpad group04:45
dafand has access to the screen session04:45
dafso you can easily attach to the session and examine the server output04:46
dafmost server fixes I'm doing consist of the following04:46
dafssh -t rosetta screen -x04:46
dafControl-]  d04:46
dafor variations on that theme04:47
SteveAdaf: go and eat! :-)04:49
dafeating is blocking on food cooking :-)04:50
=== lalo [~lalo@] has joined #launchpad
lululalo: hi - I'm sending you an email with what's been discussed today - so you're in the loop.05:07
lululalo: done05:09
lululalo: no problem - some important stuff discussed.05:10
=== kiko is now known as kiko-fud
laloaugh. stupid, stupid.05:29
lalo/me bangs head against wall05:29
carloslalo: ?05:31
laloI applied timezones in the wrong direction again :-/ and registered in my brain that we had a meeting in the middle of the afternoon, when it was instead in the early morning05:31
SteveAdaf: it is time!05:32
dafSteveA: it is05:34
SteveAlet's start by making a page that you can upload something to05:34
SteveAlet's just do this as a view on /rosetta/05:35
SteveAonce we have that so that it is accepting uploads, you can remove that page, and apply the same thing to PO/POT files05:35
SteveAshall we do this with screen again?05:37
SteveAthen I can see what you're coding etc.05:37
dafI think I've fixed the routing problems05:37
SteveAwhat host and port shall I use?05:37
dafyou should be able to "ssh -t stevea@earthsea.homdns.org screen -x daf/"05:37
SteveA $ telnet 2205:39
SteveArouter still stopping ssh ?05:40
dafhmm, it was working earlier05:40
dafoh, hmm05:41
dafthat should be
SteveAI'm in05:42
=== daf fixes that DNS record
dafokay, I'm just writing the template in the other screen window05:43
SteveAhow do I get there?05:44
dafescape key + 005:45
dafescape key twice also switches between windows05:46
SteveAnot really05:46
dafI can't remember if the escape key is shared or not05:46
dafthe default is Control-a, but I prefer Control-] 05:47
SteveAnothing happening05:47
SteveAcan you boot me out?05:47
SteveAI'll try connecting again, with an explicit escape set in screen perhaps05:47
SteveAoh, back in jabber05:55
luluSteveA: is jabber logged?06:02
SteveAnot publicly06:02
luluare there discussions on Ubuntu etc that should be logged for our records?06:04
daflulu: we're only using Jabber for one-to-one conversations, as far as I know06:04
SteveAI'm pair-programming with daf just now, so I'll have to talk about this later.06:05
luludaf: cool  just checking so we don't lose valuable input from our great team.06:05
=== lalo done with the meeting log
daflulu: sure :-)06:06
luludaf: good thanks.06:07
lalodaf, SteveA: I got no feedback on #1935. I'm tempted to simply submit the merge, as it's blocking some other stuff.06:07
carlosSteveA: I need to execute a unix command (msgfmt) to export .mo files and I need to use stdin and stdout to send the pofile (text data) to that command and get back the .mo file (binary data) to serve it from launchpad. I'm thinking on using the popen2 module. Is that ok?06:08
SteveAcarlos: you mean the new one, with python 2.4 ?06:09
SteveAor, the popen2 call in python 2.3 (and earlier) ?06:09
lalodaf: and, would you like me to work on something in particular or can I just pick one at random?06:09
carlosthe later one06:09
SteveAlalo: can you summarize what the fix is?06:10
carlosSteveA: I don't think I'm using python 2.4 at this moment06:10
carlosFrom the python manual: 6.8 popen2 -- Subprocesses with accessible I/O streams06:10
daflalo: I think a sensible policy is that any bugs which are NEW (as opposed to ASSIGNED) are up for grabs06:10
laloSteveA: don't use lazyUpdate when running zopeless06:10
daflalo: and hence if you're working on something, make it assigned to you06:11
SteveAcarlos: I suggest using popen306:11
SteveAso you get the error stream06:11
lalodaf: yes :-) what I meant is, to ask you whether you have a specific one in mind you'd like me to poke06:11
SteveAlalo: sounds good. 06:11
daflalo: sorry, I have a habit of answering questions people didn't ask06:11
daflalo: I don't have anything specific in mind06:11
carlosSteveA: ok, If I get an error I send that stream as the message, right?06:12
SteveAerror output should go to that stream06:12
carlosI mean, to launchpad exception06:12
daf#1938 and #1939 are interesting06:12
SteveAdaf and I need to focus on our pair programming now06:12
=== lalo gets at it
=== daf assigned #1919 to himself
=== lalo wonders why the heck you can't reassign and accept in one operation
lululalo: that would be good to keep in mind for malone - could you add it to the wiki - wish list for Malone?06:20
lululalo: thanks so we don't forget it.06:21
lululalo: thanks hon :o) 4 days to Rosetta Alpha....06:30
lalodateLastActive is not updated when you translate from the web?06:32
daffile a bug :)06:33
lalothat's not a bug, it's a question :-) I'm not sure it's supposed to06:34
lalodue to the whole 42 conversations about active/current/inPofile/etc06:35
daflalo: care to write a document describing the various sates of message sets, sightings, etc?06:36
lalonot sure I have all information I need :-) but I'll try06:36
lalodaf: the timestamp is updated, in case you wonder06:45
dafdid I add a comment next to that saying something like "XXX: Should we update this always?"06:47
dafword for the day: prosopography06:47
dafmust have been something else then :)06:47
dafhmm, they shuoldn't be getting touched unless the translation has actually changed06:49
lalowhich we have no easy way of knowing06:50
laloshort of comparing the strings, I suppose06:50
dafwhich is what we do :)06:50
lalodo we really? I don't see the code to do this06:51
lalothat's where I'm looking06:51
laloyet, I would be a bit worried if we did, too - string comparison is inefficient by nature06:51
dafit's only O(n), and n isn't very big06:52
carlosdaf: we have "datefirstseen" and "datelastactive"06:52
dafI'd be far more concerned about the database access needed to get the old translations for comparing with06:52
laloah, found it06:53
carlosI think the datelastactive is updated every time we import a .po file with that translation06:53
=== kiko-fud is now known as kiko
carloswe should do the same everytime someone submits the translation without any change06:53
daf"is" or "should be"?06:53
carloshmm "should be" :-)06:53
dafwhat about if somebody loads a page with 5 translations, updates one, and then submits?06:54
=== debonzi [~debonzi@] has joined #launchpad
dafthe 4 translations they didn't even look at get updated06:54
dafwhich doesn't seem right to me06:54
dafat the moment, I think the timestamp is only updated when the translation is changed06:54
laloyes, that was exactly my point06:54
debonzikiko, 06:55
carlosdaf: but then, the .po ones will have always preference from the website ones06:55
lalothe fact that a translation is submitted doesn't mean the user actually even looked at it :-)06:55
carlosdaf: I was sharing that idea until 10 minutes ago :-)06:55
carloslalo: or perhaps he agree that it's the correct translation06:55
dafcarlos: they're not "website ones" because they're only in the form because they came from the PO file06:55
carlosdaf: or from previous website editions06:56
dafyes, exactly06:56
carloswell, I suppose it's not a problem until we implement a way to suggest translations from the UI06:57
lalocarlos: can't reproduce 193906:57
=== lalo ponders the fine art of gaim plug-in development
carloslalo: if you translate the evolution's pot07:00
dafcarlos: I've noticed something else07:00
carloslalo: and then export it... you should see the problem07:00
lalocarlos: I find it rather puzzling that your msgstr[0]  and msgstr[1]  are almost identical (wrt whitespace), but yet [0]  was modified and [1]  wasn't07:00
carloslalo: go to the page where that string is and add a new translation (but don't touch that msgset)07:00
=== sabdfl [~mark@wblv-228-55.telkomadsl.co.za] has joined #launchpad
dafcarlos: there are spaces missing from the Spanish translations on the input fields07:00
lalosome bitrot seems to be involved :-P07:00
dafcarlos: do you see it?07:01
carloslalo: I think it's a website bug more than a po handling one. I think the problem comes with the addtion of the return char graph07:01
carlosdaf: I know there are missing spaces07:02
carlosdaf: because I translated them and I didn't saw them in the english text07:02
laloyet, how come [1]  is not touched?07:02
carlosso I did not added them07:02
dafso, not a Rosetta bug? :)07:02
carloslalo: I don't know, the code should tell you it :-)07:02
carlosdaf: yes, it's a bug, a UI bug07:03
carlosdaf: If I don't see the spaces, I cannot add them to my translations 07:03
dafoh, I fixed that one07:03
dafI was worried that the translations in the database had spaces, but that they were not being included in the text fields properly07:04
dafdid you see my fix for the spaces?07:04
carlosdaf: one moment...07:05
carlosdaf: yes, I see it07:05
carlosdaf: it's ok, the only problem I have now is that I see them too small (included the "return" char)07:06
carlosfirst, I thought that was ... instead of a special graph07:06
dafdoes making your font bigger help?07:06
carloshmm, I need to do it too big to see it correctly07:07
dafany suggestions?07:08
=== sabdfl [~mark@wblv-228-55.telkomadsl.co.za] has joined #launchpad
carlosdaf: I did a user test with my sister07:09
dafgreat, how did it go?07:09
carlosdaf: Ctr++ two times until she was able to see it07:10
carlosshe thought it was ..07:10
dafI'm not sure if there's a better way to show spaces07:10
carlosIs there any way to increase the size?07:11
dafof the font size in general, or of that particular glyph?07:11
dafwe could increate the size for messageIDs only07:12
dafthat would be a simple CSS fix07:12
carlosfor that particular glyph07:13
carlosmy sister says that the "enter" glyph is easy to detect without any change07:13
dafwe could probably do <span class="openBoxes"></span>07:14
carlosin fact I did not notice it but the space glyph is the one used with the movile phones to represent the space, so it's the correct char :-P07:14
daf.openBoxes { font-size: 120% }07:14
dafheh, I hadn't thought of that07:15
=== cprov [~cprov@] has joined #launchpad
dafwhich font are you using, by the way?07:15
carlosdaf: my sister told me that, I didn't saw it :-P07:15
dafah, ok07:15
carlosdaf: sans-serif07:16
carlosthe default one from epiphany07:16
lalough. can we sort the languages list in the preferences form by name rather than code?07:16
daflalo: I don't see why not07:16
=== lalo registers himself as translator for lojban
carloslalo: sure07:17
lalodaf: I'll file a low-pri bug07:17
daflalo: :D07:17
daflalo: do you speak lojban?07:17
carloslalo: the languages with the country code don't work yet, I just remember a bug report I need to fill :-)07:17
dafwhere has limi been of late?07:19
lalodaf: speak, no, but I can read it (and translate into it) with some effort07:20
carlosdaf: I don't understand your question about limi...07:20
lalodude, I can't for the life of me find Spanish in this widget :-P perhaps normal priority07:20
daflalo: cool!07:20
carloslalo: hmm, seems like the #1939 is fixed,07:20
dafcarlos: sorry, "of late" is another way of saying "recently" in English07:21
carloslalo: is at the end of the list07:21
carlosdaf: hes out for holidays07:21
dafoh, right, I see07:21
lalocarlos: yay for our magic bug-fixing fairies07:21
dafwhen is he back?07:21
carloslalo: :-)07:21
carlosdaf: on Monday07:21
dafah, that's good07:21
=== daf has bugs for limi
carlosdaf: fill them with bugzilla, he has an account :-)07:22
luludaf: limi should be back on Monday giving 150%. We are hoping he will be in london, but this is unconfirmed as yet.07:23
luluif there is anything you guys can handle on your own, please do so, as he's got lots of work to do.07:24
daflulu: it's nothing high-priority, just some small fixes that hopefully won't take up much of his time07:24
=== lalo ponders making a fresh bowl of mate
luludaf: ok - make sure all the bugs are in bugzilla and he can finish high priority first, then get to those.07:25
daflulu: you were really anxious to get hold of me this morning, weren't you? :)07:25
luludaf: yup - didn't realise you went to bed at 7am.07:25
daflulu: was the logo thing a big deal?07:26
luluit's easier to have pepes online when you expect them to be!07:26
luludaf: Jane wanted to know urgently before speaking to the design agency, but no worries.07:26
dafah, I see07:26
luludaf: are these your regular hours now?07:27
luludaf: not 7am, but working more in the evenings rather than mornings?07:27
dafI was just wondering that07:28
dafI think the meeting at 12:00UTC today may have changed things :)07:28
carlosdaf, lalo: which one is the easier way to get the lenght of a file object? with seek + tell (like C)? or is there any other way to do it?07:28
dafcarlos: is this a file on disk?07:29
luludaf: perhaps if you could be online more in the mornings that would be great - leading up to the 15th - is that possible?07:29
carlosdaf: no07:29
=== lalo goes check
carlosdaf: it's a stdout returned as a file object07:29
laloyou can't seek on stdout  o.O07:30
lalounless it's a really strange stdout :-P07:30
daflulu: I can try07:30
daflulu: usually these things wrap around07:30
carlosso I need to read all the stdout and then get the lenght, right?07:30
daflulu: i.e. I start having a 27-hour cycle or so07:31
carlosdaf: that's not sane07:31
luludaf: thanks hon - that will be great :o) if you can manipulate your cycle back to normal! 24hrs!07:31
lalocarlos: yes, seek+tell is what you have. That's odd.07:31
carlosunless the days start having more than 24 hours...07:31
laloor if you can't seek, len(foo.read())07:32
dafcarlos: it's not that strange07:32
dafcarlos: my days just go out of sync with everybody else's07:32
carlosdaf: I said sane, not strange :-)07:32
lalomy "normal" days have 30h07:32
dafI think Keybuk has similar experiences07:32
laloexperiences made with cosmonauts and submarine crew point to the best possible cycle for humans being more than 24 - IIRC, more close to 30 - which may or may not prove that we're not originally from this planet :-P07:33
carlosdaf: mines were also that way, but only when I'm alone, when there is someone else in my house I cannot do that or I will not be able to sleep more than 3-4 hours because the noise07:34
lalocarlos: there is no really good reason why files can't have a method that returns the size as well as it's known... if you want to submit a bug for Python, it will probably be fixed before 2.5, with luck even 2.407:34
carloslalo: in this case, you are right, it's a stdout so it's normal you cannot get the size until you read all the data07:35
dafcarlos: ah, I am quite a deep sleeper, so I suppose it happens to me more easily07:35
carloswe are exporting now .mo files :-)07:38
lalointeresting :-)07:38
carlosand the md5sum is the same if I do it by hand07:39
carlosso I suppose it works :-P07:39
lalodoes it have the right filename? ;-)07:39
carloslalo: yes07:41
carlosthe export.po url does not exists anymore07:41
carlosdaf: how could I get the lp:url attribute from a template?07:42
carlosI want to add the links to the po and mo download urls from the pofile page07:42
dafI'm not sure I understand the question07:42
carloswe have at pages.zcml an attribute lp:url for every page07:43
carlosthat tells us the URL 07:43
carlosI'm not sure but I think I saw a template that get that information07:44
carlosor It was a dream?07:44
daffrom the PO file page, it would be <a tal:attributes="href string:${self/name}/po">Export as PO file</a>07:44
dafor soemthing like that07:44
dafthe lp:url attributes are only for documentation purposes07:44
carlosok, it was a dream :-P07:45
dafthe only thing that extracts them is the "make url-table" command, I think07:45
=== lalo ponders whether we should have a resolution in bugzilla that looks like "the magic fairies fixed it"
lalo(j/k, before someone adds it)07:48
lalo#1938 is interesting... even if the comment is "", the export should still be ignoring it, IIRC07:50
laloah, the comment is probably "\n" or something like that07:51
dafthe comment should not be touched at all07:51
=== lalo nods
carlosBug# 1913 has still 1 unresolved dependency. Show Dependency Tree.07:55
carlosdaf: is the stuart's fix merged into rocketfuel?07:55
carlosI cannot close my bug because the dependency X-)07:56
daflikewise :-(07:57
lulucheers guys - I'm off - thanks for today :o)07:58
daflifeless should be up in another 5 or 6 hours, I think07:58
daflulu: see you07:58
carloslulu: bye bye07:58
kikochao lulu07:58
luluadios amigos07:58
=== lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has left #launchpad []
carlosI'm going to leave for a while and continue working later08:15
carlosdaf: msgfmt is not installed into rosetta server right?08:34
carlosin that case, what should we show here?: https://rosetta.warthogs.hbd.com/++skin++Debug/rosetta/projects/gnome/evolution/evolution-2.0/es/mo08:36
carlosI should fix my code but we will detect a problem and we should show anything to the user08:37
lalocarlos: can you reproduce 1938? I suspect the fairies got it too, because I can't find any code in browser.py that touches commenttext08:37
carloslalo: let me check08:37
carloslalo: no, the problem is still there08:39
carlosso it's not fixed by python's god :-P08:39
=== lalo publishes Action Comics #1 featuring Superman, in an attempt to reproduce 1938, and hopes that someone else invades Austria
daflalo: something to do with the way message sets are created in PO files by the TTW translation code?08:46
lalodaf: that was my guess, but the bug is about translations that already existed in the db08:47
daf"OpenOffice.org: Is it an application or a web site? Who knows!" :D08:47
lalowell, aha08:47
lalomessage sets are, indeed, created with '' rather than None08:48
dafis that it?08:48
lalowhich is just reasonable, seeing as commentText is notNull08:48
dafok, that can't be the bug then08:48
lalostill rather odd :-P08:49
lalowell, if it's notNull then it must be fixed at the adapter level. And this part of the fix I have already.08:49
dafI think it would probably best be debugged by tracing through the export process08:49
lalonot necessary, I already found the bug in the adapter08:50
laloI was investigating further because I suspected the bug could be twofold - but that doesn't seem to be the case, so I can commit08:51
!lilo:*! More rambling in #fossie .... "a project to fiddle around with the idea of a small, modular, message-passing-event-driven communications framework..." today's subject matter would be table storage for plug-in modules and how reloads should be done.... if you have a fair amount of C and have fiddled with plug-in architectures, linked lists, memory allocation, etc., you might be of help09:08
!lilo:*! Please feel free to stop by.09:09
=== lalo makes a break, bbl
laloI'll make a task for myself of fixing the ftests when I'm back09:17
laloit's rather cumbersome to not be able to run them and be 100% sure you didn't break anything important09:18
=== kiko is now known as kiko-afk
=== npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #launchpad
=== lalo [~lalo@200-203-032-008.paemt7005.dsl.brasiltelecom.net.br] has joined #launchpad

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