BradBSounds like the rosetta instance that's currently being used fits into "test", so that leaves us needing to get the dev LP up and running, and then on the next test deployment, include soyuz and malone along with the next ver of rosetta12:02
=== kiko-fud is now known as kiko
dafBradB: the Rosetta alpha server is not suitable for testing other things12:05
dafBradB: it is on its own code branch, is not regularly updated, and has modifications in place so that Rosetta is the only application available12:06
BradBthe test apps shouldn't get regularly updated12:07
sabdflwhat;s the diff between the dev and the test lp?12:07
BradBdev is the one that gets updated every day with the latest and greatest changes (or breakages)12:08
sabdflbradb: i don't think people work well in too many branches12:08
BradBtest is the one that users test, and should actually expect to work relatively well, and change relatively infrequently.12:08
sabdfli'd like to be moving to production as frequently as is sane12:08
sabdflweekly, or fortnightly12:08
sabdflso there's not really much difference between dev and test12:09
sabdfland i don't think it's worth having it12:09
sabdflnot at this stage12:09
sabdflwhen lp is huge with tens of thousands of users, yes12:09
sabdflbut right now we want it to evolve FAST12:09
sabdflso prefer to keep it lighter weight12:09
BradBok, so first step, we want testing LP that is running all three apps, right? what happens to the rosetta alpha?12:10
SteveAdev is the one that gets updated every N minutes, not every day12:10
carlosI think we should update soon the rosetta alpha version with some bug fixes12:11
SteveAthe main point of the dev one is so that lu and others can see changes that developers make hour-by-hour12:11
carlosdaf: did you merged the patchsets to fix the language_country language selection?12:12
SteveArather than waiting a day to see some contentious change take effect12:12
sabdflSteveA: you can see the hour-by-hour status on your own dev box12:13
SteveAbut can lu?12:13
sabdflbut we should hjust blow that db away every half hour too12:13
SteveAoh, you mean the dev server on mawson12:13
SteveAI thought you were proposing to do away with that12:13
sabdflthe dev box is a daily build, with data we care about enough to backup regularly12:13
sabdfli'm tired, so i may not be explaining this clearly12:14
sabdflheres what i would suggest we have:12:14
sabdfl(a) crack of the hour (rosetta.wh.h.c equivalent) that nobody uses except to poke people at a brand new feature, blows away its own code abd db every half hour12:15
sabdfl(b) dev / dogfood box, updated daily to latest db and code schema, with specific merges by the team leads to fix critical issues that come up12:15
sabdfl(d) production server12:15
sabdflstub will focus mostly on (b)12:16
sabdflthen (c) and (d) during a push to production12:16
sabdflwe should push to production weekly or fortnightly12:16
sabdfltime based releases12:16
sabdfluse tests to ensure the whole system is always deployable12:16
sabdflhow's that sound12:17
dafsabdfl: I suggest you write up your proposal to the mailing list12:17
daffor the benefit of those not present12:17
SteveAsounds good to me.  I imagine staging will be hit by automated tests, not people so much.12:17
sabdfl     SteveA agreed12:18
sabdflstaging is a place to bring in a dump of the production db, run the db update scripts, run the tests, and see if it all looked to work. then repeat on production server12:18
sabdflit's a defined process, not a place for arbitrary work12:19
sabdflBradB: any success on mawson?12:19
BradBsabdfl: i stopped deploying, based on the current discussion (which started somewhat earlier)12:20
sabdfldoes this proposal sound sane to you?12:20
carlosit makes sense12:21
sabdflwe bring up malone. we get kinnison ad cprov to bring up soyuz.12:21
BradBthree seems like enough, but if you really want to maintain a forth, i won't object.12:21
sabdfland when daf is ready they migrate the rosetta alpha to the dogfood platform12:21
BradBfourth even12:21
sabdflr.w.h.c is on autopilot12:22
sabdflall our work is focused on the dev12:22
sabdflstaging is a pritine environment for stub to blow away, test, blow away again12:22
sabdflproduction is production12:22
sabdflthis is less fragmented than having a rosetta "test" and a malone "test" and a soyus "test"12:22
BradBisn't this more complicated than what i suggested of just dev, test, and live?12:23
BradB(one dev LP where anything goes, one test LP that should be fairly sane, and one live LP where lives hang in the balance.)12:24
sabdflit's the same, just addin the pristine staging server we always planned on12:24
sabdflso dogfood = dev12:24
sabdflsame proposal12:24
sabdflkeep going :-)12:25
=== carlos is tired
carlosdaf: please, send me an email with a "go" to the merge if you are able to test it today so I can do the merge tomorrow morning.12:26
carlosgood night12:26
=== BradB starts build-config again, on patch-659.
BradBI can have that up and running in however much CPU time it takes, less the appropriate Apache configuration.12:27
sabdflelmo will sort out apache config in a tick12:27
sabdflbasically, no reason not for it to be identical to the current rosetta.wh.h.c12:28
sabdflclient cert for vpn effect12:28
dafwe should probably get rid of the rosetta.w.h.c domain12:28
dafas an alias for mawson12:28
sabdfldaf: right, just run it on mawson, on a different port?12:28
sabdflit's purely to show off brand spanking new features12:29
dafrosetta.w.h.c is the same machine as mawson.ul.o12:29
sabdfldiff port12:29
sabdfldiff db12:29
dafyou're thinking of rosetta.sf.o12:29
daf(I suspect)12:29
sabdflthat's the "alpha"12:30
dafhttps://mawson.ubuntulinux.org is exactly the same as https://rosetta.warthogs.hbd.com12:30
sabdflrosetta.wh.h.c is the "update every half hour to crack of the minute db and code"12:30
sabdflNOW it is12:30
sabdflbut bradb is bringing up the dev / dogfood server  on mawson too12:30
sabdflor not?12:31
BradBi am12:31
BradBwaiting on the build-config12:31
dafsabdfl: are we going to use rosetta.w.h.c for something difference?12:31
sabdflso... r.w.h.c is just for kicks, its just a "public" snapshot of your and my dev box12:31
sabdfldaf: please read what i wrote12:32
sabdflthe dev server will only be updated once a day12:32
dafI'm saying I think we should stop using that domain12:32
dafto avoid confusion12:32
sabdfl*relatively* stable12:32
sabdfldaf: agreed12:32
dafok, then we were agreed all along :)12:32
sabdfllet's call them the bleeder, the feeder, the staging and the breeder :-)12:32
=== sabdfl ducks and weaves
sabdflthe bleeder should take none of our time12:33
sabdflit just happens in the background12:34
sabdflthe feeder is where we really test12:34
sabdflit's running a relatively full db12:34
sabdfland it's hosting guests working on rosetta beta code12:37
sabdfland it's hosting our own launchpad bug tracking instance of malone12:37
sabdflstaging is just for stub and lifeless12:37
sabdflproduction is production12:37
sabdflok, bleeder, feeder and breeder are terrible names :-)12:37
sabdflbut bleeder, dailydev and production might be better12:37
=== debonzi_ [~debonzi@200-206-134-238.async.com.br] has joined #launchpad
=== debonzi_ is now known as debonzi
=== BradB goes to get food while build-config continues chugging
=== BradB is now known as BradB|food
=== BradB|food is now known as BradB
=== debonzi [~debonzi@200-206-134-238.async.com.br] has joined #launchpad
=== debonzi [~debonzi@200-206-134-238.async.com.br] has joined #launchpad
BradBdaf: How bad will it be if I blow away rosetta's DB on mawson?03:15
BradBdaf or elmo_: ping03:27
BradBI need someone to stop rosetta so that I can create malone_alpha03:28
stubBradB: Ignore what I said in email about the header rewriting - just changing the destination address in the envelope (what is happening now) is fine. That said, I'd already implemented the (optional) full header rewriting before I realised it so I'll check it in anyway.03:30
stubShould I document this sort of thing on the Wiki or in a README.email ?03:33
stubMmm.... Malone is starting to look like a real system :-)03:35
BradBwiki or README, heads or tails, hm03:35
BradBI'd say a README03:35
stubMmm... I prefer that too... keep it in the same RCS.03:37
BradBas for malone: once edit notification works (which i almost was able to wrap my brain around earlier), the 80%'ll be done.03:37
BradBI guess that email thing is pretty cool. Where do bounces go though?03:39
BradBe.g. if the envelop says From: test@bbnet.ca, and the From: header says steve@canonical.com, I hope it bounces to bbnet.ca :)03:40
stubAt the moment there is an Errors-To: header being set to 'nobody@example.com' along with a TODO.04:05
stubBradB: I opened a bugzilla bug about configuration information via ZCML, as it should be configurable rather than hard coded.04:06
stubFeel free to change the default if your mail server doesn't mind everyone spamming it :-)04:11
stubWe will need to do bounce processing in the future to detect expired email addresses, but I think it is ok to just check them in /dev/null for now (?)04:21
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Header rewriting and email documentation (patch-670)04:23
dafBradB: "rosetta's DB"?04:42
BradBon mawson04:54
BradBstub: Probably, only reason I was asking about bouncing was perhaps slight over paranoia of real emails getting spammed when tests with expired recipient addresses get run.04:55
stubYup. I think it is very valid paranoia - the email system needs to get this stuff right.04:57
stubErrors-To: should cope, or the header rewriting if we want to be doubly sure.04:58
dafBradB: you mean the "launchpad_test" DB on mawson?05:43
dafBradB: trash that as you like05:44
BradBis that what you're running on? oh.05:45
BradBpresumably i can't trash it as i like though05:45
BradBnope, that definitely wouldn't be a good idea ;)05:46
BradBbut i need rosetta stopped momentarily, so that i can create malone_alpha05:46
BradBdaf: can you stop rosetta on mawson momentarily while i do that?05:46
dafoh, ok05:47
dafwhen I say you can trash the launchpad_test DB, I mean it05:47
dafit doesn't contain any critical information05:47
BradBother than the schema :)05:47
dafthe schema cna be recreated also05:47
BradBthere's no guarantee that it'll be the same as what i'm using05:47
dafthere's a separate DB for the alpha, which actually should not be changed05:48
BradBoh, ok, well anyway, i'm creating a separate db anyway05:48
dafit sucks how Postgres won't let you create new databases while other databases are being used05:48
BradByeah, i don't understand why05:49
dafok, stopped05:49
dafdo I need to stop the Alpha also?05:49
BradBnope, db recreated, you should be able to restart the thing you just stopped05:50
dafthe Rosetta alpha was using the DB05:51
dafthe test LP was using the DB05:51
dafbut I only needed to stop one of them?05:51
BradBeh, how do i start an LP app on this machine?05:54
BradBif I run it as the launchpad user, i get a perms error on a log file05:54
BradBif i run it as bradb, i can't access the db05:55
dafwhich log file?05:56
BradB /home/bradb/launchpad-malone-dogfood/launchpad.log05:57
dafif you want to run it as launchpad, you'll probably need to make a copy of the checkout as launchpad05:57
BradBhow are you doing it with rosetta?05:58
dafe.g. sudo -u launchpad "cp -a ~bradb/launchpad-malone-dogfood /home/launchpad; cd /home/launchpad/launchpad-malone-dogfood; make run"05:58
dafactually, the existing instances are still running as me05:58
BradBthe launchpad user doesn't have a home dir :)05:58
dafI should change it, but I always have more urgent stuff to do05:59
dafoh, I think the launchpad user has permissions in /srv or someplace like that05:59
BradBhow did you auth to pgsql then?05:59
BradBif you're running rosetta as daf05:59
dafI think the DB has permissions for either me or LP to access it05:59
dafboth databases05:59
dafbecause you created the new DB as bradb, the launchpad user can't access it06:00
BradBi created it as launchpad06:00
dafoh, hmm06:00
=== daf shrugs
BradBdaf: can you access https://mawson.ubuntu.com:9020/malone?06:18
dafit won't work06:21
dafbecause it is firewalled06:21
BradBah, so it was more than just apache config :)06:22
dafApache acts as a proxy06:22
dafApache listens to ports which are allowed through the firewall, and proxies ports which aren't06:23
BradBproxy on block ports? i don't follow...06:24
dafApache listens on port 44306:26
dafand forwards requests it recieves based on the virtual host06:26
BradBi'll get james or thom to set me one up tomorrow then06:27
dafbest to send the email now06:29
=== BradB sends, sleeps
=== BradB is now known as BradB|zzz
=== stub [~stub@dsl-] has joined #launchpad
sabdflSteveA: around?11:45
=== Kinnison [~dsilvers@81-178-238-108.dsl.pipex.com] has joined #launchpad
KinnisonI guess I should be here too11:55
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Tidying lucille ready for renaming views etc (patch-671)11:57
Kinnisonthanks babe.11:57
Kinnisonsabdfl: Right; I'm doing the view rename/column rename we discussed on wednesday11:58
sabdflKinnison: cool, thank you11:58
Kinnisonsabdfl: The views I have to rename are... BinaryPackageFilesToPublish SourcePackageFilesToPublish PublishedBinaryPackageOverrides and PublishedSourcePackageOverrides11:59
Kinnisonsabdfl: any suggestions on better names?11:59
sabdflOverrides needs rethinking, and better to do it with grep and %s/x/y/ sooner rather than later11:59
sabdflIt was the field names that I thought needed to be the same as the field names they map to12:00
sabdflbinarypackagename instead of bpname12:00
Kinnisonyeah; the column names I'm clear on12:00
sabdfllet's use a given name as consistently as possible12:00
KinnisonThose views list packages pending publishing12:02
KinnisonI.E. sourcepackagepublishing and packagepublishing records which have status == PendingPublishing12:02
SteveAsabdfl: yep12:04
sabdflhi SteveA12:05
sabdflSteveA: blush, forgotten what i was going to ask12:06
stubAnyone mind if I rename the databases, so launchpad_dev will be what is used when you do 'make run' and all the tests use launchpad_ftest? I'm fixing tests atm, and some of them are broken because they talk to the launchpad_test database.12:10
Kinnisonsabdfl: How about PendingBinaryPackagePublishings and PendingSourcePackagePublishings ?12:10
stubI know Daf is for it (from email about a week ago...)12:10
sabdflKinnison: sounds perfect12:10
sabdflstub: yes please!12:11
Kinnisonsabdfl: and for the other two... PublishedSourcePackageLocations and PublishedBinaryPackageLocations ?12:11
sabdflthe old overrides... is it really just locations?12:12
sabdflisn't it also quite a bit of "we want this, we don't want that"?12:12
KinnisonIt's not really locations at all12:12
Kinnisonit's component/section/priority12:12
KinnisonThe 'want this, don't want that' choice is the filelist which is the next bit on my list of things to do12:12
KinnisonSo PublishedBinaryPackageArchiveLocations ?12:13
sabdflloooong but true12:13
KinnisonThen there'll be Published{Source,Binary}PackageFileList12:13
sabdflhold on12:13
sabdflis there any other view of "published packages"?12:13
sabdflif not, then the rest is spurious12:14
sabdflif this is our primary view of the packages that are published, then just call it "PublishedSourcePackage"12:14
=== Kinnison ponders
sabdfland the fact that it includes the archive lcoations is nice12:14
Kinnisonthe primary way to see what's published is to look at the sourcepackagepublishing and packagepublishing tables12:14
sabdflwe do need to rename the Overrides table12:15
Kinnisonthese views are collating information for lucille to generate her output12:15
sabdflhmm.... so what tables do they pull data from?12:15
KinnisonE.g. 'SourcePackageFilesToPublish' draws from... sourcepackagepublishing sourcepackagerelease sourcepackagereleasefile libraryfilealias distrorelease sourcepackage sourcepackagename and component12:16
sabdflPublishedSourcePackage then12:17
sabdfland PublishedBinaryPackage12:17
sabdflsimpler better faster gotta love it12:17
KinnisonThese views are listing *files* not packages12:17
Kinnison(assuming the ...ToPublish views are what you're on about now)12:18
sabdfl... i can live with it12:18
=== elmo_ [~james@] has joined #launchpad
=== Kinnison thinks we're talking at cross-purposes
Kinnisongive me a sec to regroup12:18
=== Kinnison -> thinking room
sabdflKinnison: you're right, sorry12:18
sabdflPublishedSourcePackage would be for the alread-published files12:19
sabdflwhich tables does that draw from?12:19
sabdflsame list?12:19
=== Kinnison returns
KinnisonPublishedSourcePackage draws from: SourcePackagePublishing, DistroRelease, SourcePackageRelease, SourcePackage, SourcePackageName, Component and Section12:23
KinnisonColumns it provides are:12:25
Kinnisonid, distroreleasename, sourcepackagename, componentname, sectionname and distribution12:26
Kinnisonwell; theoretically it provides sourcepackagename.name12:26
KinnisonWhat should I call that column?12:26
Kinnisonsourcepackagenamename ?12:32
=== Kinnison pokes sabdfl
sabdflKinnison: pong12:51
Kinnisonsabdfl: ^^^^^^^^^^12:51
sabdfljust sourcepackagename12:51
Kinnisoneven though that has connotations of being a foreign key onto the sourcepackagename table?12:51
sabdfla bit of a fiddle, that one12:52
sabdflsourcepackagenamereally ;-)12:52
sabdflso PublishedSourcePackage i'm happy with12:52
KinnisonI don't mind any of: sourcepackagename sourcepackagenamename or "sourcepackagename.name"12:52
Kinnisonalthough sqlobject might barf at the latter12:53
sabdflwhat about the name of the other views?12:53
sabdflKinnison: works for me12:54
KinnisonI'll do that; and then I'll do PublishedSourcePackageFile and PublishedBinaryPackageFile for the new views I need to do too12:54
sabdfland rename Override?12:55
Kinnisonlaunchpad_dsilvers=> \d override12:55
KinnisonDid not find any relation named "override".12:55
KinnisonWe're renaming PublishedBinaryPackageOverrides to just PublishedBinaryPackage I thought we agreed that up there somewhere?12:55
sabdflKinnison: yes, that's the view, right?12:58
Kinnisonthere are no 'override' tables12:58
sabdflbut there is an underlying Override table somewhere12:58
KinnisonNupNup :-)12:58
KinnisonThe overrides *are* the publishing records :-)12:58
sabdflwhere do you store the info about how the distro wants to publish a given package?12:58
sabdfl"when the NEXT package foo comes in, publish it in component/section/priority"12:58
sabdflthat's what I understood an override to be12:59
Kinnisonwhen foo 1.0-2 comes in, you copy the publishing records from 1.0-1 pointing them at the new package12:59
KinnisonOr at least; that's how I was going to do it12:59
KinnisonSince that's the obvious way; then if you don't have enough context to do that; the package is clearly NEW and should be queued in the NEW queue instead01:00
=== Kinnison -> lunch
sabdflKinnison: i think we need an explicit policy table. when a new package comes in you look at the policy table to decide what to do with it02:11
sabdflwe can start with a straight line-per-package approach like the debian overrides system02:11
sabdflbut later it needs to be "accept packages that are in this bundle" (gnome / kde / etc)02:11
sabdfland "reject packages that are in that bundle (perl / dev / etc)02:11
sabdflso that derivatives get to shape their distro at a higher level than package-by-package.02:12
sabdflSteveA: kamion just reminded me what i wanted to ask02:23
sabdflwiki / zwiki decision is now urgent02:23
sabdflwhat's your view on the best course of action?02:23
SteveAWhat decision exactly do we need to make?  Whether to use ZWiki, integrated with the UL plone site, or to use moinmoin not so integrated?02:26
SteveAIt seems that there are no tables used on wiki.ubuntu.com02:36
=== BradB|zzz is now known as BradB
SteveAI did wget -m http://wiki.ubuntu.com/ and grepped the results case-insensitively for "<table".  There were no tables as part of the content.02:38
SteveASo, we can use ZWiki, and move content across from wiki.ubuntu.com02:39
BradBelmo_: ping02:42
elmo_BradB: 02:44
elmo_SteveA: ??02:44
elmo_stevea: check out hardware support ..02:44
BradBelmo_: Two things:02:45
BradB1: Will you have a chance to do the Apache config for Malone soonish?02:46
BradB2. Do you care if I create another postgres superuser on mawson? (bradb)02:46
elmo_2: yes, run stuff as launchpad please02:47
BradBThere's a problem with that.02:47
elmo_services should not be tied to one user's account02:47
SteveAelmo_: ah -- wget didn't get second level pages...02:47
BradBThe problem is that the zope instance needs to write logfiles. the launchpad user can't write logfiles in ~bradb.02:48
BradBUnless I'm meant to have another dir under /srv02:48
elmo_yeah, create something under /srv02:51
elmo_hmm, there's meant to be a /srv/launchpad or something - if there isn't just let me know02:52
elmo_have you guys agreed on the url stuff then?02:52
BradByep, i put it in an email to you guys (i.e. the admins) last night02:52
stubSteveA: How recent is lib/canonical/functional.py ? I think it has superceeded lib/canonical/tests/functional.py but I'm not sure.02:53
elmo_yeah, I know, just wanted to make sure it's final02:53
BradBelmo_: afaic, yes :)02:53
elmo_oh, sorry I should have read that mail02:54
elmo_malone.w.h.c doesn't work - there's like 24 hour+ turn around time on that DNS02:54
elmo_if you pick *.ul.o or *.u.c or *.c.c etc., I can do it.. otherwise..02:54
=== carlos [~carlos@69.Red-80-33-181.pooles.rima-tde.net] has joined #launchpad
BradByeah, like i say, i'm not picky. m.u.c would be fine.02:55
sabdflSteveA: ok, and the moin support is working, and installed in our copy of zwiki on www.ul.o?02:55
sabdflelmo_, SteveA: the zwiki supports native html for tables02:56
BradBelmo_: can you create a dir /srv/malone? /srv/rosetta.* doesn't seem like an appropriate place for it.02:56
sabdflif necessary we can munge them over as html tables directly02:56
sabdflBradB: morning!02:56
BradBhi :)02:56
elmo_sabdfl: that's super user unfriendly for editing02:56
elmo_sabdfl: okay for malone.u.c02:57
sabdflbradb: re /malone is this for the dev server? it will host ALL the lp componets though02:57
SteveAstub: lib/canonical/functional.py is the newest and best02:57
sabdflelmo_: it's much better for searhcing02:57
sabdfli think we are going to put together a dedicated Plone class for the hardware db02:58
BradBsabdfl: so launchpad.ubuntu.com?02:58
sabdflBradB: perfect!02:58
BradBelmo_: Can you make that dir /srv/launchpad instead please? :)02:59
elmo_tramps.   this is why I asked if you guys were done "discussing" :P03:00
BradBi had got the impression that i was just putting up a malone df, but i guess not :)03:01
SteveAsabdfl: the latest released ZWiki is 0.35.0.  This is not installed in our plone sites.  It has support for moinmoin pages without tables.  The preview mode was added one week later, and has been improved since.03:01
SteveAThere will be a new wiki release on 1 November containing the preview mode, and any further work done on moinmoin markup.03:01
SteveAIf we want to get the preview mode before then, we'll need to check zwiki out of its DARCS repository.03:01
sabdflSteveA: how extensive are the tables in hte hardware section03:02
elmo_what client cert do we want protecting this?  the testers one?03:02
elmo_or a 'warthogs' one?03:02
sabdflelmo_ accept either, allow testers into /rosetta but for everything else require warthogs03:02
SteveAsabdfl: there are 11 hardware pages.  most have a table on them.03:02
BradBgood thing I checked if there was already a database called "launchpad_alpha" :)03:03
=== BradB uses one called launchpad_dogfood instead
SteveAI could convert the tables to restructured text format in under 30 minutes in vim03:03
sabdflSteveA: is the ReST table format efficient? moin is terrible03:04
SteveAit is an improvement over moin, but still a bit awkward.  Here's an example:03:04
SteveAI'm looking at the "simple table" format03:05
BradBelmo_: Can you stop rosetta too please so that it doesn't hang when I try to create the DB?03:05
elmo_"stop rosetta" ?03:06
elmo_you mean UP it?03:06
BradBkill it03:07
BradBdaf's running it on mawson03:07
SteveAwhy would rosetta hang?03:07
SteveAit uses a different db, doesn't it?03:07
BradBrosetta doesn't hang03:07
elmo_postgres' fucked - I'm going to try UP-ing rosetta to see if it helps03:07
SteveAwhy would your stuff hang on account of rosetta running?03:07
BradBme trying to create launchpad_dogfood hangs, because "another process is accessing the database"03:07
elmo_err, UP-ing mawson03:07
elmo_this is why we changed the machine name isn't it :)03:08
SteveAso postgres doesn't like you doing stuff to the databases when a process is using a different database?03:08
SteveAthat sucks03:08
SteveAmakes me want to run several different postgreses03:08
BradBit seems to me that there must just be a problem with the pgsql config on mawson03:09
elmo_no, you tramps, it's a bug03:09
BradBmaybe it's fixed in 8.0b3?03:09
SteveAsabdfl: I think HTML tables are the way to go, unless people generally use a decent editor to edit them.03:10
elmo_we're not using beta's in production03:10
elmo_and changing major versions between production and devel is a reciepe for madness03:10
BradByeah, wasn't suggesting that, but optimistically looking towards the 8.x series :)03:10
sabdflelmo_ are we runing 7.4.5?03:10
elmo_yeah, 7.4.5-303:11
elmo_let me UP the box first03:11
elmo_mawson going down..03:11
=== BradB logged out
sabdfl(13:18:07) Kamion: kind of impressed that the wiki handles Korean totally seamlessly03:16
sabdfl(13:21:17) lamont: thom: you're welcome03:16
sabdfl(13:22:09) sabdfl: Kamion: i'm keen to move to the Zwiki which is a little more restrictive but *should* do the same03:16
sabdflwill zwiki handle i18n?03:16
sabdflSteveA: the ReST table format seems very sane03:17
sabdflif we can get kupu working nicely for the wiki then html might be fine03:17
sabdflbut let's not rush into that03:17
sabdflSteveA: think we can big-bang the content across in a morning?03:17
=== debonzi [~debonzi@] has joined #launchpad
sabdfli was thinking to move it over a page at a time, and lock the moin-wiki page once it had moved03:18
SteveAThere's a discussion on #zwiki about hebrew.  I just asked about the details.03:18
SteveAthe problem with the ReST table format is that if you add a row with a column that is longer than all those that currently exist,03:19
SteveAin the simple format, you need to go through and make every row have a longer column03:19
SteveAIn the full table format, you can use wrapped text03:20
SteveAso, although it is trickier to set up, it is easier to maintain03:20
elmo_right, mawson's back up03:21
elmo_please try again, and if it still hangs, yell03:21
SteveAlimi is very negative about kupu -- he says it works only in a very few browsers.03:21
SteveAits ancestor, epoz, works in most browsers.03:21
SteveAbut, I think epoz is not integrated into zwiki, whereas kupu is03:22
BradBsabdfl: Do we want to restart whatever Rosetta daf had running on mawson, or is this new LP dogfood thing enough?03:23
SteveAthe rosetta alpha testers will be a bit upset, I guess03:24
SteveAdaf should be around any minute03:24
SteveAI have a meeting with him and carlos in 30 minutes03:24
SteveAmarius says:03:28
SteveA<SteveA> wiki question mgedmin:  how easy is it to put non-ascii characters into a ReST zwiki page?03:28
SteveA<mgedmin> easy03:28
SteveA<mgedmin> well, depends03:28
SteveA<mgedmin> if it is Zope 2.7, you have to edit zope.conf03:28
SteveA<mgedmin> and specify rest-input-encoding utf-803:28
SteveA<mgedmin> and also rest-output-encoding utf-803:28
BradBelmo_: Bad Gateway when trying to access launchpad.ubuntu.com; is it being forwarded correctly to 9020?03:29
SteveAhere is a ReST page on zwiki.org: http://zwiki.org/SandBox   There are some odd characters in it.  Someone who uses non-ascii characters can try adding a comment.03:30
elmo_no, I hadn't finished03:30
elmo_try again now03:30
sabdflSteveA: agree on kupu. let's use ReST, Moin, HTML for now03:30
BradBelmo_: looks good, thanks, just doing some quick sanity checks now.03:30
stubI gave up on kupu - epoz got overengineered and became urgh...03:31
stuberm.... kupu got overengineered. epoz is cool.03:31
sabdflbradb: the old server should still continue... it's the half-hourly snapshot on a db that gets blown away thing03:32
sabdflit should totally run on autopilot03:32
sabdflon separate ports03:32
sabdflwith client cert restricted to warthogs only03:32
sabdflSteveA: korean appears to be working fine there03:33
sabdflSteveA: ok, when can we big-bang the wiki across?03:33
sabdfland could you keep an eye on zwiki, and make sure we run the latest release?03:34
SteveABrad needs to tweak zope.conf, and install the latest ZWiki (perhaps latest from DARCS, so we get the improvements to preview mode)03:34
sabdflyes we definitely need preview mode03:35
SteveAThen we should check that we can enter all manner of characters into a test page on ul.org03:36
SteveAWe probably want to set up zwiki so that people cannot use DTML pages or other pages with server-side scripting in them.03:37
SteveAI'm not sure how to do that when zwiki is mixed in with plone03:38
SteveAit would be nice if we can convert all the moin wiki markup to ReST.03:38
sabdflSteveA: i'd rather leave moin available as an option03:39
SteveAthat way, there will be just one wiki markup in use.03:39
sabdflthe guys are screaming along making lots of changes to the wiki03:39
sabdfland i don't want to impose a change unless we have to 03:39
sabdflhappy to say "sorry, tables aren't supported yet in the new wiki"03:39
sabdflbut keep the disruption to a minimum03:39
SteveAso, pages with tables will use ReST?03:39
sabdflor html03:40
sabdflwhatever the authoer prefers03:40
SteveAthe rest remain in moin03:40
sabdflcan we lock a page in moin?03:40
SteveAI don't know03:41
SteveAthere's always "soft locking": "THIS PAGE HAS BEEN MOVED TO http://ubuntulinux.org/wiki/whatever"03:42
=== SteveA uses darcs to get the latest zwiki
SteveAit doesn't look as scary as tla03:46
SteveAit doesn't say very much.  just "copying patches..." then lots of ........03:47
BradBany objections to me blowing away launchpad_test on mawson?03:55
BradB(well, wreaking havoc on it, anyway)03:55
dafit's trashable03:56
SteveABradB: https://chinstrap.warthogs.hbd.com/~stevea/zwiki-2004-10-22.tar.gz03:57
SteveAthat's ZWiki from today03:57
SteveAI haven't run it.  I just thought I'd save you the trouble of installing darcs on a mac :-)03:57
BradBdeploying lp is as painful as a very painful thing04:00
sabdflBradB: can we redecorate pages in our plone site that are below /wiki/ ?04:00
sabdflget rid of the right-hand bars?04:00
sabdflBradB: what's painful? Is it retrieving it from arch?04:01
BradBi keep getting:04:01
BradB    for row in self.table.select():04:01
BradB  File "/srv/launchpad.ubuntu.com/launchpad-dogfood/lib/sqlobject/main.py", line 1198, in __iter__04:01
BradB    return conn.iterSelect(self)04:01
BradB  File "/srv/launchpad.ubuntu.com/launchpad-dogfood/lib/sqlobject/dbconnection.py", line 507, in iterSelect04:01
BradB    select, keepConnection=True))04:01
BradBTypeError: iteration over non-sequence04:01
BradBwhen trying to access basically any screen that hits the db04:01
BradB(i've even tried running directly from launchpad_test, just in case trying to run off launchpad_dogfood was config'd incorrectly)04:02
sabdflBradB: i've been seeing that on rosetta.wh.hbd.com for a while04:03
sabdflI wonder if postgres on that box isn't *&^*&^'d04:03
BradBhttp://paste.husk.org/1840 is the config I built from04:03
BradBmaybe i'm not using the right version of something04:04
BradBit seems to me that the problem is that the app is failing to connect to the db, but that's a bit of a guess, because the error message is computer science04:04
sabdflBradB: can you do a psql launchpad_test and see that it looks ok?04:05
BradBit does04:05
sabdflBradB: wrong sqlobject version04:09
BradBi thought that might be the case04:09
sabdflyou want: ./launchpad-malone-dogfood/sourcecode/sqlobject        rocketfuel@canonical.com/sqlobject--test--0.604:09
BradBi thought rocketfuel@canonical.com/sqlobject--test--0.5.1 was the misnamed 0.6 we had though04:09
sabdflalso, i don't think we should call it launchpad-malone-dogfood04:10
BradBit isn't04:10
sabdfljust call it launchpad-dev04:10
BradBthat was just on a local build-config04:10
sabdflor launchpad-dogfood04:10
sabdflBradB: my config has 0.604:10
BradBit is launchpad-dogfood already :)04:10
sabdfltry the updated sqlobject04:11
sabdflthat would also explain why rosetta.w.h.c was being b0rked04:11
BradBdaf used the same version?04:11
dafno, I think the development server I was running has 0.604:13
sabdfldaf: t was getting exactly the same error for me04:13
sabdfliterating over a non-sequence crap04:13
BradBthe upgrade fixed it04:14
BradBthanks for pointing out the prob04:15
sabdfldaf, please fix the other one04:15
SteveAfyi, rosetta progress meeting starting on #canonical-meeting04:15
daf~/launchpad-devel/launchpad/sourcecode/sqlobject$ tla tree-version04:15
BradBemail notification works too, but it's still always coming to me, which may be a good thing until the next version is deployed with stub's email changes04:15
BradBhm, i guess a clean way of starting this as a daemon would be nice :)04:19
BradBbin/launchpadctl start04:20
BradBWe'll also need to be able to change the DB name in exactly one place, if we're to have any hope of running this thing on four different databases (which I'm assuming will all be named differently.)04:23
SteveAshould be a config option in launchpad.conf04:23
SteveAwe'll need to do a similar thing for the zodb to use04:23
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Rename databases. Implement new test harnesses. Port tests to new harnesses where appropriate. Fix all unit and functional tests. 'make test' runs without errors. (patch-672)04:24
stubBradB: At the moment it should be possible to stick .zcml files in the override-includes directory that, well, overrides the default settings. This might be a way to keep the dogfood config in an easy-to-upgrade fashion04:31
stubSteveA: You want we should have test_on_merge.py do a full test run now?04:33
SteveAI think we should use a launchpad.conf.in file, with the default settings, and put all settings that could be changed on a deployment into launchpad.conf.04:33
SteveAthere's a bit of work required to put settings we need into launchpad.conf, though04:34
SteveAstub: do all tests pass right now?  If so, great, let's do it.04:34
=== stub gestures up the screen to patc-672
SteveAcool.  let's go for it04:36
SteveABradB: how does plone handle internationalizing content?04:36
SteveAsabdfl: have any plans been discussed about internationalizing the ul.org website?04:36
BradBSteveA: yeah, but i don't have an experience i18n'ing content04:38
BradBSteveA, sabdfl: Should I read my scrollback right now about zwiki et al. or focus on getting edit notification working in malone (and deploying stub's new email machinery?)04:40
BradBSteveA: LinguaPlone is the standard product there, last I checked.04:40
BradBlimi develop{ed,s} it04:41
SteveABradB: I'll send a mail about zwiki04:41
SteveAI'd like you to upgrade zwiki and zope.conf on ubuntulinux.org before monday UK time04:42
SteveAso we can get to work moving pages across on monday04:42
BradBBradB & # brekkie run04:43
=== Kinnison -> tea and cakes
SteveAmalone question: how are the scripts to pull open bugs out of bugzilla and into malone looking?04:58
SteveAwe'll need to use those to do dogfooding of malone04:58
stubhmm.. where did all those tests with hardcoded launchpad_test come from? They wern't there a minute ago, I swear!05:02
sabdflSteveA: i'd love to i18n the web site but will only go with what plone standard provides05:10
sabdflSteveA, stub: i don't know about auto-pulling open bugs from bugzilla05:10
SteveAlet's ask limi's advice on what approach to look at next week05:10
sabdflwe currently have code that will update the status of a remote bugzilla bug when it changes05:11
sabdflbut not to automatically find new bugs over there and create them in maone05:11
sabdfli suggest we just transition to malone by turning off new bug creation in bugzilla05:11
SteveAI'm thinking more of pulling open launchpad bugs out of the launchpad bugzilla05:12
sabdflpeople can wrap up old bugs in bugzilla and create new ones in maone05:12
SteveAand putting them into malone05:12
sabdflfor critical bugs we can create them in malone and put remove watches on them in the bugzilla05:12
sabdfl(14:22:02) limi: * patching for this revision (alexander.limi@canonical.com--2004/launchpad--devel2--0--patch-20)05:13
sabdfl(14:22:05) limi: unable to open file "./.arch-inventory" (Permission denied)05:13
sabdfl(14:22:07) limi: PANIC: I/O error05:13
sabdfl(14:22:18) limi: :] 05:13
sabdflanybody seen this before?05:13
SteveAI've seen panics from tla, but not one about .arch-inventory05:14
=== debonzi is now known as debonzi|lunch
BradBls -l ./.arch-inventory05:24
sabdfl(16:33:03) limi: $ ls -al ./.arch-inventory 05:39
sabdfl(16:33:04) limi: -rw-r--r--  1 limi  limi  98 19 Oct 11:44 ./.arch-inventory05:39
dilysNew bug 2131 for Launchpad/Launchpad: Launchpad should have a graceful way to startup in daemon mode05:42
dilysNew bug 2132 for Launchpad/Rosetta: internationalize Rosetta05:46
dilysNew bug 2133 for Launchpad/Launchpad: Launchpad database must be configurable in exactly one place05:48
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: View renaming for lucille's publishing stuff (patch-673)05:48
BradBsabdfl: hmph, I wonder what ./.arch-invetory it's referring to then.05:48
sabdfldilys: you rock05:50
=== debonzi|lunch is now known as debonzi
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Merged my BIG change to split pomsgsets and potmsgsets into their own tables (Closese: #2083) (patch-674)05:53
dilysNew bug 2134 for Launchpad/Rosetta: Create and migrate the user languages preferences to a table PersonLanguage06:00
debonziBradB, ping06:05
dilysBug 2037 resolved: main search form should search products as well as projects06:06
BradBdebonzi: pong06:08
debonzihi BradB I saw there is a bug in SourcePackage that is MultipleJoin... Is there a easy way to get from the attribute bugs only the ones with BugSeverity.CRITICAL for example?06:10
BradBnope, not that i'm aware of06:11
debonziuhmmm... so I think I need a new query and a SourcePackageBugAssignment.select06:12
BradBthat or a list comp06:13
debonziBradB, what you mean with list comp? some python code to get them?06:13
BradB[c for c in sp.bugs if c.severity == CRITICAL] 06:14
BradBs/c/b/, but yeah06:14
debonziright.. it is nice because whould avoid and circular import since SourcePackageBugAssignment already imports SourcePackage, but don't you think it is too expensive?06:15
BradBdepends on what you're using it for06:16
BradBare you just doing a search page for all critical bugs for one SP? if so, then the solution is almost written in the question. :)06:17
debonziright but I don't know precise how big can be a SourcePackageBugAssignment for a given SourcePackage of course.. do you have and ideia06:17
debonziyes.. thats what I whould like to do06:17
BradBThen just do a simple select on SourcePackageBugAssignment. :)06:18
BradBthat way you needn't add any special methods to anything06:19
dilysBug 2083 resolved: Do the split of .pot and .po records from POMsgSet to two different tables06:23
debonziuhm? why do you say I don't need? let me explain I little better what I want: I whould like to show some bug informations in the sourcepackage page. so, on the template I have context/sourcepackage/bugs and I can use it to should for example the total number of bugs for that sourcepackage. But I also want to show the number of critical bugs06:23
debonziso, I beliave I need something like context/sourcepackage/bugscritical06:24
debonzifor that I think I do need to write a new method inside SourcePackge SQLObject.06:24
debonziI am wrong somewhere?06:24
sabdfldebonzi: there's a good idea in there somewhere06:25
sabdfli would rather do it like this06:25
sabdflwe can write a function SourcePackage.bugs(severity_min=None, status=dbschema.BugStatus.OPEN)06:26
sabdfland expose that partly through SourcePackageView06:26
sabdflthe important thing to me is that it be a general, well written function06:27
sabdflthat will report on bugs associated with the source package06:27
sabdfland cann be constrained in interesting ways, like by status and severity and priority level06:28
BradBdebonzi: I had the impression (as per above) that you were just doing a search page and trying to show critical bugs for an SP, in which case self.critical_bugs = SourcePackageBugAssignment.select(SourcePackageBugAssignment.q.sourcepackageID == 1, ...) would have been the simplest way to do it. Sounds like you're doing something slightly more complex though.06:30
=== debonzi thinks about
=== kiko [~kiko@200-206-134-238.async.com.br] has joined #launchpad
debonziBradB, sabdfl Im thinking about it. If I have more questions Ill ask again.. thanks :)06:35
=== cprov [~cprov@] has joined #launchpad
sabdflbradb: is it possible to turn off the right hand portlets for  pages under /wiki/ on our plone site?06:50
BradBI'll take a quick look now to see if what I think might work actually does.06:52
BradBsabdfl: yep, it worked06:54
BradBi'm betting the old one's still cached at the real URL though06:55
BradBoh no, seems to work properly event at the real URL for me06:55
BradBsabdfl: does that look ok then?06:55
BradBSteveA: (#zope3-dev) is ignoring the class directive for an editform a bug then?06:58
sabdflBradB: shiny!07:02
sabdflexcellent thanks07:02
sabdfli'm going to get some of our wiki extraordinaires to play with the zwiki there07:02
dilysMerge to rocketfuel@canonical.com/launchpad--devel--0: Make merge check use full test suite (patch-675)07:03
sabdflyowser. chug chug. test suite just grew07:11
sabdflah, dilys07:11
=== BradB is now known as BradB|lunch
=== kiko is now known as kiko-fud
!lilo:*! Of possible interest, a recently created channel: #accounting, a support channel for small not-for-profit entities.... stop by if you have skills and want to answer questions, or if you have questions and need answers :)08:24
debonzisabdfl, I was thinking in the way you sugested for the implementation I need and looking at the already exists but I still can't see a good way to solve my "problem". When you say SourcePackage.bugs(...) what do you mean with it? A classmethod inside SourcePackage?08:59
debonziSourcePackage already have and attribute called bugs that since I have a SourcePackage.SelectResult all the rows in it have a list (bug attribute) with all its bugs.09:03
debonzis/(bug attribute)/(bugs attribute)09:03
sabdflgiven an instance of SourcePackage you want to get a list of relevant bugs assigned to that source pacakge.09:04
sabdflyou could have different functions SourcePackage.criticalBugs and Sourcepackage.closedBugs and Sourcepackage.openBugs etc09:04
sabdflor just have a single function to which you can pass some good methods09:04
sabdfllike SourcePackage.getBugs(minSeverity=HIGH, status=OPEN)09:05
debonziRight... that is ok.. 09:05
debonzibut I whould like to use the bugs attribute that I already have to get this things.. but seems that there is no way09:06
sabdfltry to write a new method.09:07
sabdflultimately it will use Bug.select(querystring)09:07
sabdfluse the arguments to the method to build a useful querystring09:07
=== sabdfl hs to leave shortly
debonziI whould like to avoid one sql query for each "kind" of bug since I already have a list with all of them09:07
sabdflwell, you could also have one function which gets all the bugs once09:08
sabdflthen other functions which run over that and get subsets09:08
sabdflbased on your criteria09:08
sabdflanyhow, got to go, have a good weekend!09:08
debonzisabdfl, ok thanks :)09:08
=== BradB|lunch is now known as BradB
=== sabdfl is now known as sabdfl|out
BradBstub: have you thought about pushing your CC change upstream?09:40
BradBi've got a foo. foo is an sqlobject, which has one attrib, x. foo.x == 1. how do i capture foo's state into an object such that when i go foo.x = 2, i can look at old_foo.x and see that it was 1?10:16
BradBoh, i thought of something10:16
=== BradB tries

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