[03:52] <lifeless> stub: around ?
[04:44] <stub> Morning
[04:44] <stub> lifeless: pong
[04:45] <lifeless> how do I log into the produciton launchpad now ?
[04:45] <lifeless> to to imports
[04:45] <lifeless> to do.
[04:50] <stub> I haven't the foggiest
[05:14] <lifeless> argh.
[05:21] <lifeless> stub: so, how do I change a database into unicode on the fly ?
[05:21] <stub> You don't as far as I know
[05:21] <lifeless> garh
[05:21] <stub> You dump it, drop it, create it and restore it
[09:18] <lifeless> I'm about to head home and so on.. if you see stevea/sabdfl, can you ping them for me... ?
[10:37] <stub> SteveA: Robert Collins (lifeless): I'm about to head home and so on.. if you see stevea/sabdfl, can you ping them for me... ?
[10:38] <stub> lifeless was not able to log into the production launchpad to do imports
[10:43] <ddaa> I'm about to ask some pretty stupid launchpad/imports questions on the mailing list, unless somebody can help me.
[10:43] <ddaa> Here are my issues:
[10:44] <ddaa> I want to setup a test environment for imports on my workstation.
[10:44] <SteveA> hello
[10:46] <ddaa> I am able to import info files. But launchpad do not show up in the "do not use" product where the infoImporter put them (I checked there were indeed imported SourceSource objects for this product).
[10:46] <SteveA> hmm
[10:46] <SteveA> if I'm to help, I need some more background information
[10:46] <SteveA> I have never used the "importer" parts of launchpad
[10:47] <ddaa> Also, the rocketfuel launchpad keeps asking for user/pass, and I have no idea where that info is stored.
[10:47] <SteveA> i should also point out that it is 1C outside, and the heating folks havent turned on the heating here yet
[10:47] <SteveA> so I am dressed for outside
[10:47] <SteveA> user/pass is stored in the database
[10:48] <ddaa> SteveA: afaik, nobody except lifeless and I (who wrote it back in London) ever used that shit. I actually had to unbreak it for the Product refactorings before it would run.
[10:48] <SteveA> hmm
[10:49] <SteveA> so, I can help with getting the auth stuff working
[10:49] <SteveA> you need to be using an email address / password combination that is in the database
[10:49] <ddaa> The big picture is "get a test environment for imports". Lifeless seems to thinks that's not a problem, but I never managed to get there...
[10:50] <SteveA> or sighs even
[10:50] <ddaa> Okay, lemme put the stuff up and dig into the database schema to figure out what you mean.
[11:00] <mdz> spiv: ping?
[11:04] <mdz> spiv: please follow up on #1922 when you're around
[11:12] <ddaa> SteveA: Okay, I found emails and (hashed) passwords in sampledata/current.sql. Does getting a test loging involves running john to crack the passwords? I guess there is probably a saner way...
[11:13] <ddaa> * "around."
[11:18] <Kinnison> stub: ping?
[11:19] <stub> Kinnison: pong
[11:19] <Kinnison> stub: I've just pqm-merged a new patch into schema/pending
[11:19] <stub> ddaa: look for createuser.py, somewhere in rosetta
[11:19] <Kinnison> any chance that along with any of the lucille ones still pending from the list could be applied soon?
[11:20] <stub> ta. yup - back on that in 10mins
[11:20] <Kinnison> stub: fantastic
[11:20] <ddaa> stub: processing |^H/^H-^H
[11:29] <ddaa> Okay... the createuser.py, createproject.py and createproduct.py scripts all look bitrotted: they refer to RosettaPerson instead of Person. Also I think they belong in some generic launchpad directory (maybe lib/canonical/launchpad/scripts) instead of lib/canonical/rosetta/scripts.
[11:29] <ddaa> Do you think it would make sense to fix and move those?
[11:35] <stub> Yup. I don't know if anyone is already doing that or not.
[11:36] <ddaa> I only find bitrotten tools on my path...
[11:38] <ddaa> duh... the tla experience on launchpad could massively benefit from partial inventory support...
[11:40] <ddaa> 611 source files, 5235 versions-controlled files...
[11:41] <ddaa> The difference is patchlogs and explicit ids.
[11:42] <lifeless> ddaa: re test environment - are the info files in the same database ?
[11:42] <lifeless> select count(*) from sourcesource; for instance.
[11:43] <lifeless> SteveA: yes, I know its in the database
[11:43] <lifeless> but there is this whole sync apache + launchpad
[11:43] <lifeless> I'm not sure what approach is desired, and I'm not inclined to randomly poke things in production.
[11:44] <lifeless> what I'd like is to be told 'Rob, use this password + username for editing sourcesource jobs'
[11:45] <lifeless> until then, I'm kinda up-creek-without-paddle with regard to doing imports.
[11:46] <ddaa> lifeless: yessir, it's all in launchpad_test
[11:47] <lifeless> so if you go to /doap/projects/do-not-use-info-imports/unassigned/+sources/ what do you see ?
[11:48] <ddaa> "   A system error occurred."
[11:49] <lifeless> prefix the path with cd ~/botslave 
[11:49] <lifeless> PYTHONPATH=~/buildbot/launchpad/lib:~/buildbot/launchpad/lib/canonical/sourcerer/util twistd -f buildbot.tap
[11:49] <lifeless> oh bah
[11:49] <lifeless> ++skin++Debug
[11:50] <lifeless> prefix it with that, will give you the traceback
[11:50] <lifeless> you may ned to run make enable-debugging or something first
[11:50] <SteveA> lifeless: you want a username and password to use that authenticates you with launchpad?
[11:51] <lifeless> and lets me edit all the existing sync jobs.
[11:51] <lifeless> and finally, that apache will let through.
[11:51] <ddaa> lme
[11:52] <SteveA> I think ken manheimer wrote that
[11:53] <SteveA> Can you move the apache in question to using certificates?
[11:55] <lifeless> I'm happy for that to be done, have no access to do so
[11:55] <ddaa> Interesting, it raises when trying to get the 'Debug' skin... let's try w/o the debug skin thing...
[11:56] <carlos> morning
[11:56] <ddaa> (the difference is that I'm using port 8089, I just realised how it was meant to be used)
[11:56] <SteveA> lifeless: what domain name and port do you use to access launchpad for imports?
[11:57] <lifeless> https://macquarie.warthogs.hbd.com/launchpad/
[11:57] <lifeless> ok, logging out now...
[11:59] <ddaa> The error from: doap/projects/do-not-use-info-imports/unassigned/+sources/
[11:59] <ddaa>   File "/home/david/home/devel/canonical/dists/launchpad/lib/canonical/launchpad/database/sourcesource.py", line 196, in __getitem__
[11:59] <ddaa>     return ss[0] 
[11:59] <ddaa>   File "/home/david/home/devel/canonical/dists/launchpad/lib/sqlobject/main.py", line 1238, in __getitem__
[11:59] <ddaa>     return list(self.clone(start=start, end=start+1))[0] 
[11:59] <ddaa> IndexError: list index out of range
[12:00] <SteveA> spiv: ping
[12:00] <ddaa> However I do not get into postmortem, looks like it happens in some thread...
[12:00] <SteveA> spiv owns the DOAP and FOAF stuff
[12:00] <stub> Kinnison: Did you settle on a name for the distroqueue table?
[12:00] <ddaa> +1 cute and fun
[12:01] <Kinnison> stub: I don't have a preference either way. I'm happy with DistroQueue or DistroReleaseQueue. A row in the table refers to a DistroRelease but actions are taken on the aggregation of all the rows referring to DistroReleases in the same Distribution
[12:01] <SteveA> ddaa: that error says to me that the code is assuming that there is a result row.  Yet, there are no results.
[12:02] <ddaa> SteveA: I guessed that... I somehow expect people who are actually doing what I'm trying to do would have experience with my problem... I am almost certainly doing something wrong.
[12:02] <SteveA> do you have suitable stuff in your database?
[12:03] <ddaa> For some value of suitable... There are sourcesource entries for "unassigned"...
[12:03] <sabdfl> ddaa: if you create a launchpad_test db on your dev box at home
[12:03] <ddaa> sabdfl: yes
[12:03] <sabdfl> by going to launchpad/database/schema and typing make run
[12:04] <Kinnison> sabdfl: Did you ever reach a decision on whether DistroQueue or DistroReleaseQueue was preferable?
[12:04] <sabdfl> then you'll be able to login to the test system with user foo.bar@canonical.com and passwd test
[12:04] <SteveA> lifeless: I've mailed admins and cced you, to ask for apache auth to be changed to certificate based auth for what you're doing
[12:05] <SteveA> lifeless: now, we just need to ping elmo or thom to move this up the queue of sysadmin tasks
[12:05] <ddaa> sabdfl: works, thanks.
[12:06] <sabdfl> the plan as i understood it was to use a client cert for vpn, then user/pass for launchpad auth
[12:06] <sabdfl> ddaa: i think the Debug skin is broken at the moment, it works for me on some launchpad components, sporadically
[12:07] <ddaa> the post-mortem on port 8085 is even better. When the bt is available it drops into the debugger
[12:07] <sabdfl> Kinnison: i can well imagine that a large derivative might have people taking separate decisions on different releases
[12:08] <SteveA> sabdfl: want to do a phone call this afternoon?
[12:08] <sabdfl> SteveA: yes please
[12:09] <sabdfl> folks i did a lot of moving and renaming and reordering over the weekend
[12:09] <sabdfl> malone and doap are working on my latest code
[12:09] <sabdfl> soyuz and rosetta are broken
[12:10] <ddaa> (hm.. postmortem is on 8089...)
[12:10] <Kinnison> sabdfl: That's at the web-app level I think; I'm only talking about the automated processing through the queue as NEW etc need human intervention anyway and can be separated out on a DistroRelease basis if that's what's wanted. As I said; I can see arguments either way on the naming of the table so unless you do I'll suggest to stub to run with what I posted to the list.
[12:11] <sabdfl> please go with distroreleasequeue
[12:11] <Kinnison> stub: got that?
[12:11] <sabdfl> ddaa: stevea will update the skins code today, so we have the server listening on three ports
[12:12] <stub> Already going with that. I'm learning to channel Mark (at least on sensible decisions ;) )
[12:12] <Kinnison> stub: Fantastic
[12:12] <sabdfl> one gives you normal behaviour, one gives postmortem-debugger, one gives debug skin with tracebacks
[12:12] <ddaa> Goodness.
[12:12] <sabdfl> stub: will be sure to refer to you when i'm unclear on my own opinions :-)
[12:13] <ddaa> Writing up my experience as I go. I'll update the ImportProcess page eventually.
[12:13] <sabdfl> daf in NYC yet?
[12:13] <Kinnison> I thought his flight was last night; so I'd assume he's asleep now
[12:13] <sabdfl> cool. lucky boy
[12:14] <SteveA> lifeless: does pqm run 'make check' on merges to the 'launchpad' category, and merge only if the exit code is 0 ?
[12:36] <lifeless> ddaa: btw got deadlocks in the new code
[12:37] <lifeless> SteveA: yes it does
[12:37] <lifeless> ok, packing up to go home
[12:39] <sabdfl> stub: did you use my updated, commented version of the infestation table proposal?
[12:40] <stub> From pending? Yep
[12:42] <sabdfl> ok, thanks
[12:47] <carlos> sabdfl: where should be moved the Schema and Label class? (at this moment they are inside dlalo.py)
[12:48] <sabdfl> carlos:  canonical/launchpad/database/schema.py
[12:48] <sabdfl> that can hold both of them
[12:48] <carlos> ok
[12:48] <sabdfl> as well as supporting classes like "SchemaSet" if they exist
[12:48] <carlos> thanks
[12:49] <carlos> it exists as Schemas, but I will rename it.
[12:51] <ddaa> lifeless: I'd need more info about the deadlocks. The only one I identified can not actually occur since the sync-queue size in unlimited (well, should be).
[01:45] <Kinnison> sabdfl: Would it be reasonable to add to SourcepackagePublishing a datepublished column like PackagePublishing has? Also then; is it reasonable to dual-task that column based on the state column. (To provide a time-to-delete type value for packages PendingRemoval from the release?
[01:46] <sabdfl> Kinnison: rather add two columns, datecreated and scheduleddeletiondate
[01:46] <sabdfl> weloveshortdieldnames
[01:46] <sabdfl> field, even
[01:46] <Kinnison> sabdfl: 'created' is a bit of a difficult column to fill
[01:47] <sabdfl> Kinnison: am still trying to find the crack in my reorg
[01:47] <Kinnison> sabdfl: entries are created in the table pending publication
[01:47] <sabdfl> right.... datepublished
[01:47] <Kinnison> okay
[01:47] <Kinnison> sabdfl: okay, so I'll add 'scheduleddeletiondate' to packagepublishing and the two columns to sourcepackagepublishing
[01:48] <Kinnison> sabdfl: I'm going to start a Lucille/ subsection of the wiki if that's okay? (wiki.canonical.com naturally)
[03:11] <Kinnison> can I get some sanity checking on https://wiki.canonical.com/Lucille_2fPublishingNotes if anyone has a moment?
[03:25] <Kinnison> Whose is the librarian?
[03:28] <ddaa> Kinnison: I know how that feels :-)
[03:28] <Kinnison> ddaa?
[03:29] <ddaa> Had
[03:29] <ddaa> a
[03:29] <ddaa> quick look at it, but I do not know what it is about...
[03:29] <Kinnison> Aah; I know what it's about
[03:29] <Kinnison> I was just wondering who wrote it so I could ask if they have any pending changes for it before I launch a pqm assault
[03:31] <ddaa> ?????????????
[03:31] <ddaa> duh, the sources suddenly appeared in launchpad...
[03:32] <Kinnison> problem?
[03:33] <ddaa> Previously the sourcesource imported from info files did not show up in doap...
[03:33] <ddaa> Now it does, for no apparent reason.
[03:33] <Kinnison> aah
[03:40] <ddaa> Looks like one has to start launchpad _after_ running infoImporter...
[03:41] <ddaa> Merely refreshing the page is not enough, launchpad must me restarted if it's already running.
[03:46] <ddaa> What's the trick to log in into launchad so I stop getting "zope.security.interfaces.Unauthorized" exceptions?
[03:46] <Kinnison> foo.bar@canonical.com
[03:46] <ddaa> I'm asking for the trick to get the place to enter that information...
[03:47] <Kinnison> Oh right
[03:47] <Kinnison> not a clue
[03:48] <ddaa> That stuff breaks for every single step I take... it's getting annoying.
[03:51] <ddaa> Mhh... apparently I'm logged in a Foo Bar already, that's what shows in +addsources page, although it does not show in unassigned page... maybe that part of the reason why the sources suddenly appeared too...
[03:53] <ddaa> haha... the infoImporter sets the sourcesource.owner to lifeless... that's probably what breaks... It might work better with Foo Bar...
[04:02] <BradB> sabdfl: Did you add a note somewhere on the Wiki about how to propose db changes? If not, I'll do so now, and add it under Developer Resources on https://wiki.canonical.com/Launchpad_2fDevelopmentEnvironment.
[04:03] <BradB> sabdfl: The goal is to collect everything I (for i in launchpad_developers) need to know about Doing The Right Thing while doing Launchpad development.
[04:03] <BradB> In one easy-to-find place.
[04:06] <ddaa> Making progresses... I get a different error now...
[04:10] <ddaa> spiv: ping...
[04:27] <Kinnison> hey cprov 
[04:27] <Kinnison> cprov: is gina in launchpad yet?
[04:27] <kiko> hey Kinnison
[04:28] <BradB> The Makefile to create and init the launchpad_test DB seems to be broken. When I 'sudo -u postgres make', I get: createlang: language installation failed: ERROR:  could not access file "$libdir/plpython": No such file or directory, but I don't see anything on launchpad@ that tells me to expect that breakage. Anyone else seeing that problem? (I last star-merged about an hour and a half ago.)
[04:29] <kiko> BradB, could it be you don't have plpython installed into your pgsql?
[04:30] <cprov> Kinnison: hi, yes I hope :), lib/canonical/launchpad/scripts
[04:30] <BradB> That looks like the problem, yes. :) The thing is, this was working fine before though.
[04:30] <Kinnison> cprov: Cool. I've fixed the librarian so that it works. I'd like to consider getting gina to upload files into it so we can start testing it
[04:30] <BradB> So someone seems to have added that dependency, for reasons of which I'm not entirely clear.
[04:31] <Kinnison> BradB: stub is writing complex validation functions in plpython
[04:31] <cprov> Kinnison: fine, you should run gina in zhongshan or another good machine from DC, ask elmo_ 
[04:31] <Kinnison> cprov: *nod* Elmo was going to be setting up a machine for me
[04:31] <kiko> Kinnison, he's trying to scare us off from launchpad by adding incomprehensible DB spachetti
[04:32] <Kinnison> elmo_: how long will it be before I can have a machine to build a db on?
[04:32] <Kinnison> kiko: hehe
[04:32] <BradB> It'd have been nice to know about this dep being added, and for what reason (i.e. on launchpad@.) :)
[04:32] <kiko> spachetti!
[04:33] <kiko> BradB, yeah, actually, the makefile could just test and complain about it for you, how does that sound?
[04:33] <BradB> It'd be more useful for it to do that, yeah.
[04:34] <cprov> Kinnison: btw, how to reactive rev-lib support on ARCH ? I've removed my .{arch}/=revision-library days ago :)
[04:34] <cprov> Kinnison: just recreate it will solve ?
[04:36] <kiko> it's LEILO!
[04:37] <Kinnison> cprov: I think so
[04:38] <cprov> Kinnison: tks
[04:41] <kiko> how's it moving lalo
[04:42] <lalo> not terribly excitingly :-)
[04:53] <ddaa> cprov: just register the revlib again
[04:53] <ddaa> the revlib options are stored inside the revlib.
[04:54] <ddaa> I assume you are not using any of the multiple revlib voodoo which I do not know how it works and which is documented nowhere I am aware of :-)
[04:56] <cprov> ddaa: yep, I' ve done it, thanks 
[04:57] <cprov> ddaa: but I' m still waiting the star-merge started 40 minutes ago :(
[04:57] <ddaa> Sounds like you should have a local mirror.
[04:57] <ddaa> Ha... star-merge w/o a greedy revlib hurts sometimes.
[04:58] <cprov> ddaa: exactly !!
[04:58] <ddaa> But then there should be some cachedrevs along the history...
[04:59] <ddaa> if you feel like finding abentley and helping him write regression tests for the backbuilder, I'm sure that would be appreciated.
[05:00] <cprov> ddaa: sorry, I didn' t undestand what you mean ? "him" who ?
[05:00] <ddaa> abentley
[05:01] <ddaa> The sane arch devel Canonical could not hire...
[05:01] <cprov> ddaa: aaeron ? ohh yes ... :)
[05:02] <ddaa> aaron
[05:02] <cprov> ddaa: I've never seen him here,  I think, just the patches on commit@..
[05:03] <ddaa> He's not on the staff.
[05:04] <ddaa> so he's got no business being here. You can find him on #arch.
[05:04] <cprov> ddaa: ok, I' ll look for him.
[05:15] <lalo> you guys?
[05:15] <carlos> lalo: hey
[05:16] <lalo> hello carlos :-)
[05:21] <carlos> lalo: how is going?
[05:22] <lalo> not too bad :-)
[05:22] <lalo> forgot how much I *hate* looking for job
[05:23] <kiko> you should think of that *before* quitting your last one <wink>
[05:23] <lalo> :-)
[05:23] <lalo> good advice
[05:32] <BradB> spiv: Can you sanity check my changes to https://wiki.canonical.com/SQLObjectGuide?
[05:34] <BradB> I'm still a bit worried that we keep using dbName everywhere, and wonder why some people seem to get errors when that param isn't specified (particularly with the column name is identical to the property name...of course there are some namings of FK's and such for which we need to pass dbName explicitly, but 95% of the time shouldn't need to.)
[05:34] <BradB> s,with the,when the,
[05:39] <daf> I think specifying the dbName was a matter of "explicit is better than implicit"
[05:40] <daf> I don't know of any errors accounted by not having it, though
[05:40] <BradB> sabdfl claims he had errors due to it
[05:41] <BradB> Explicit is better than implicit, but when the API is documented, at some point too much explictness becomes just another place to introduce a bug in one's code.
[06:05] <kiko> explicitness also makes the channel non-pg-13 safe
[06:10] <kiko> elmo_?
[06:14] <daf> mako and I are having some net connection trouble
[06:14] <daf> we might be offline for a while
[06:19] <SteveA> daf: are you running the tests?
[06:21] <sabdfl> Kinnison: lucille page on the wiki would be excellent
[06:22] <Kinnison> sabdfl: I've started punting notes onto the wiki already
[06:22] <sabdfl> Kinnison: spiv ownz the librarian
[06:22] <Kinnison> sabdfl: they're a little braindumpish currently
[06:22] <Kinnison> sabdfl: But comments on the Lucille/PublishingNotes page would be nice
[06:23] <sabdfl> BradB: note about db update process on wiki would be excellent
[06:24] <carlos> sabdfl: Could I move Language, SpokenIn and Country objects into database/language.py?
[06:25] <sabdfl> absolutely agree that we want the wiki to consolidate the launchpad knowledge
[06:31] <SteveA> darn... there are changes that I know I made with daf on Friday that aren't in RF.
[06:32] <sabdfl> bradb: agree that having to specify the name in both places is unnecesary, but is has been faiing if i didn't
[06:32] <sabdfl> carlos: maybe put country in its own file, language and spokenin in a language.py file
[06:33] <carlos> ok
[06:40] <cprov> BradB: I' ve got the same problem with DB make, let's ask elmo ...
[06:42] <BradB> cprov: The dep on plpython is clear, it's just that the annoyance caused by the unexpected intro of that has already cost me a good hour trying to get pgsql recompiled with plpython support.
[06:44] <cprov> BradB: :0 !!
[06:45] <BradB> I'll be updating the Wiki today to mention this dep too. :)
[06:45] <sabdfl> BradB: i have a language installation thing in the db makefile that was breaking for me, so i commented it out
[06:45] <sabdfl> i committed that change, maybe it's needed elsewhere
[06:46] <sabdfl> hmm... i see it's come back
[06:46] <BradB> yeah, stub's writing complex validation stuff in python; that's what it's needed for
[06:47] <sabdfl> it's failing to Make for me
[06:47] <sabdfl> do i need to install another language module from apt?
[06:47] <BradB> sabdfl: You need plpython support compiled into postgresql.
[06:47] <BradB> I've finally got mine going.
[06:47] <sabdfl> is it not compiled into the warty postgres?
[06:48] <BradB> dunno
[06:49] <Kinnison> sabdfl: I intend to visit my parents once Warty is released to install it on their computer. As such, I'll be working a Friday and Monday from their house in wales (they have ADSL) will that be okay? [Fri 22nd, Mon 25th, Oct] 
[06:49] <sabdfl> Kinnison: no problem
[06:49] <sabdfl> musical chairs
[06:49] <sabdfl> df wales -> nyc
[06:49] <Kinnison> sabdfl: Aye
[06:50] <sabdfl> kinnison cambridge -> wales
[06:50] <sabdfl> mako?
[06:50] <Kinnison> sabdfl: I'm being bugged by my father to give him a date when I'll convert him from Windows :-)
[06:50] <sabdfl> no date could be soon enouhg :-)
[06:51] <Kinnison> My mother hates computers but has to use them; my father is an old-school geek and just wants things to stop breaking.
[06:52] <BradB> The Makefile is also checking incorrectly for the existence of the database. This has to do with the way that wc prints its output (which is indented, and has one space on the end.)
[07:08] <BradB> sabdfl: presumably it would be useful if I fixed product release infestations (at the least, a vocab that has moved has broken the interface) and implemented source package release infestations today, correct?
[07:11] <sabdfl> BradB|lunch: i have one of them working last night
[07:49] <BradB> sabdfl: Something's broken about it now (probably because ProductRelease was moved out of site), so I'll fix it now.
[07:49] <sabdfl> ok. sorry about that. we really need those page tests!
[07:50] <dilys> Bug 2096 resolved: Schema class needed by Person class but it's not in place
[07:50] <dilys> https://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=2096
[07:51] <BradB> sabdfl: I'd do one for the infestation stuff right now, but I don't think SteveA or daf have documented it on the wiki, so that I can remind myself of what they told us Friday night. :)
[07:52] <sabdfl> yeah, my memory of friday night is "cool!" but no idea how to implement
[07:53] <BradB> I remember having to start and connect to a proxy, but I need details. ;)
[07:54] <sabdfl> i think tonight i'm going to try to get productrelease and productseries smacked
[07:54] <sabdfl> then i need to turn my attention to the RC for a few days
[07:54] <sabdfl> then you'll have me back for the weekend
[08:12] <SteveA> I'll be checking in some page tests before the end of tonight
[08:12] <SteveA> and writing some docs
[08:13] <daf> I have written some docs
[08:13] <daf> I am waiting for PQM to do its thing before I send them to the list
[08:14] <SteveA> daf: please send me the docs, not the list, at first.  I'm going to be changing a couple of things.
[08:14] <daf> hmm, ok
[08:14] <daf> I've made a couple of changes which are in the PQM queue
[08:16] <SteveA> daf: I noticed that some of the changes we made on Friday weren't in RF
[08:16] <SteveA> like the new names for the skin interfaces 
[08:16] <daf> right
[08:16] <daf> those are also in this merge
[08:17] <daf> hmm, seems like a temporary problem with mail on this end
[08:30] <SteveA> I find it strange that the .css files are under canonical/launchpad/templates
[08:43] <BradB> sabdfl: Oh, I'll add a note to the Wiki on how to structure ZCML files too.
[08:50] <BradB> sabdfl: In fixing this ProductRelease vocab lookup bug, caused by not including productrelease.zcml, I'm getting config conflicts with stuff in lib/canonical/malone/sql.zcml. Just to be sure then, *all* ZCML should live under lib/canonical/launchpad/zcml/ then?
[08:50] <SteveA> all ZCML that is to do with content and its views
[08:51] <SteveA> that is, if you have a Product sqlobject class, it is descirbed by IProduct, and configured with product.zcml
[08:51] <SteveA> but, do not move the other zcml that glues the rest of launchpad together
[08:53] <BradB> IOW, to solve this conflict, the stuff from lib/canonical/malone/sql.zcml should be merged into this productrelease.zcml.
[08:57] <SteveA> hi
[08:57] <jblack> hi. What's up? 
[08:58] <SteveA> I think I want to do an undo-replay kind of thing
[08:58] <SteveA> I have been hacking on various code in my treee
[08:58] <jblack> Ok
[08:58] <SteveA> and now I'd like to sync with RF before tidying up my code wrt latest RF and commiting it
[08:58] <jblack> easily done.
[08:58] <SteveA> that is, I'd like to commit a new sync with RF before committing my changes
[08:58] <jblack> tla undo;  do the star merge stuff, commit; tla redo
[08:58] <jblack> :)
[08:59] <SteveA> what will tla undo do to files I have not yet tla added ?
[08:59] <jblack> It should handle them fine, but double check to make sure.
[09:00] <jblack> Theres an alternative if your nervous.
[09:00] <jblack> Do another get in another dir, star-merge there and commit.
[09:00] <jblack> then, go back to your old work tree, and tla replay
[09:00] <SteveA> hmm... when I do a tree-lint, I get this:
[09:00] <SteveA> These files would be source but lack inventory ids (`tla add' or a tagline perhaps?):
[09:00] <SteveA> lib/canonical/launchpad--devel--0--patch-59.patches
[09:00] <SteveA> 
[09:01] <jblack> Ok. Then undo doesn't handle new files right.
[09:01] <jblack> Do the alternative way.
[09:01] <SteveA> I haven't undone yet
[09:01] <SteveA> I haven't done anything yet
[09:01] <jblack> Oh.
[09:01] <SteveA> that's just sitting there now
[09:01] <jblack> then tla add them.
[09:01] <daf> SteveA: the CSS files didn't use to be there
[09:01] <daf> SteveA: they used to be in lp/styles, I think
[09:02] <SteveA> jblack: oh, I know what that is
[09:02] <SteveA> jblack: ok, I'll try the undo -> star-merge -> redo
[09:02] <jblack> oK.
[09:02] <jblack> tla add the new files before you undo though.
[09:22] <carlos> dinner time, see you later
[09:25] <Kinnison> Dinner sounds like an excellent plan
[09:34] <BradB> cprov: Yeah, same as before, ya gotta recompile postgres. :)
[09:35] <BradB> ./configure --with-python ...
[09:37] <BradB> cprov: Are you also noticing that the db exists check is incorrect and will claim that DB exists even if it doesn't, and then fails later saying the db doesn't exist?
[09:39] <cprov> BradB: yep, this failure I know :) but recompile Postgresql looks very bad im my point of view :(
[09:40] <BradB> yep, it was very bad from my point of view too :/
[09:40] <cprov> BradB: is there a way to exclude the plpython support from our code ?
[09:41] <BradB> I think in that case you'd want to just comment out the relevant bits of the Makefile
[09:41] <BradB> lines 25-29
[09:43] <cprov> BradB: can you briefly explain why do we need plpython ? and what is expectable by just comment the makefile lines ?
[09:44] <BradB> From what others have told me, it's because stub is writing some complex validation stuff in the DB and wants to do so in Python.
[09:45] <BradB> This may cause breakage when the DB tries to fire off some constraint that is a python function that doesn't exist. I'm not sure, it hasn't yet happened to me.
[09:46] <cprov> BradB: I'll follow you by now ... let's see what happens ... better than do nothing :), thank you for your explanantion
[09:50] <SteveA> thanks jblack, the undo-redo thing worked just fine
[09:54] <sabdfl> can the python language be a module? I see several pg language packages in aptitude, but not a python one
[09:55] <sabdfl> seems that we should have this for warty so our dev team does not each end up recompiling postgres
[09:58] <Kinnison> sabdfl: I have a warty box and it works fine for me
[09:58] <Kinnison> sabdfl: I just had to increase my personal postgres privs
[09:58] <Kinnison> well; it creates the language okay; I guess I don't know if it's working or not
[10:00] <Kinnison> and the following is in the postgresql package: /usr/lib/postgresql/lib/plpython.so
[10:02] <kiko> wow, SteveA, nice work man!
[10:03] <kiko> Kinnison, yes, it's just a matter of giving privs to the user to create languages in the DB.
[10:03] <sabdfl> Kinnison: ah, maybe it's just a privileges thing
[10:04] <BradB> To see if you have plpython support, look in `pg_config --pkglibdir`
[10:04] <sabdfl> sudo -u postgres createlang -d launchpad_test plpythonu
[10:04] <sabdfl> worked just fine
[10:04] <sabdfl> but the make still fails because it does it as my local users
[10:04] <sabdfl> user
[10:04] <kiko> exactly
[10:04] <kiko> you need to grant to your user specific privs
[10:04] <sabdfl> maybe it should TEST if you have Python installed, and if not fail telling you to install it yourself as described above
[10:05] <kiko> sabdfl, yes, exactly my suggestion to bradb a while back
[10:05] <sabdfl> kiko: the sudo command does it as postgres superuser
[10:05] <sabdfl> has anyone mailed stub?
[10:05] <sabdfl> or modified the makefile?
[10:07] <BradB> I'll mail stub.
[10:07] <daf> hmm, looks like PQM is wedged again
[10:08] <sabdfl> thanks bradb
[10:09] <sabdfl> ok, tonight i plan on finishing of productseries and productrelease
[10:09] <sabdfl> won't have librarian integration (i.o.w. it won't actually let you upload the release)
[10:09] <sabdfl> but i think i know who might be best suited to fix that
[10:09] <sabdfl> spiv: when i'm done tonight, you become the king-of-doap, ok?
[10:11] <spiv> sabdfl: Ok.
[10:11] <daf> for some reason, I now have an image of spiv as a drugs baron
[10:14] <daf> SteveA: looks like a merge from you broke PQM
[10:14] <daf> (elmo has fixed it now)
[10:16] <Kinnison> sabdfl: alter user sabdfl with createdb createuser | sudo -u postgres psql template1
[10:16] <Kinnison> sabdfl: then it should work
[10:16] <Kinnison> (assuming I got the alter-user syntax right)
[10:18] <sabdfl> Kinnison: thanks
[10:18] <daf> Kinnison: don't you mean: echo 'alter ...
[10:19] <Kinnison> daf: yeah yeah
[10:19] <BradB> This new ZCML layout is making configuration so much simpler.
[10:19] <Kinnison> it's 21:20 and I've just had dinner. My stomach has all the blood right now
[10:19] <sabdfl> Kinnison: spot on
[10:19] <daf> where did you put the food then?
[10:20] <sabdfl> BradB: you're welcome :-)
[10:20] <Kinnison> daf: On my left nipple; clearly
[10:20] <BradB> sabdfl: heh, thanks :)
[10:20] <SteveA> daf: oh darn... forgot to mirror *again*
[10:20] <daf> SteveA: that's so not cool
[10:21] <daf> ;)
[10:21] <SteveA> daf: lifeless said he'd fix it like a week ago
[10:22] <daf> SteveA: is there any reason for you not to use a script which does the mirror before sending the merge request?
[10:22] <SteveA> oh well... I'll merge, test, commit, mirror and submit
[10:25] <BradB> SteveA: You make arch sound so easy. :P
[10:26] <SteveA> it'll be the latest dance in nightclubs
[10:26] <SteveA> do the tla merge
[10:29] <daf> it's just a commit to the left
[10:29] <daf> and then a mirror to right
[10:30] <daf> put your hands on your archive
[10:30] <sabdfl> put your hands on your hips
[10:30] <daf> and star-merge your knees in tight
[10:30] <sabdfl> kiss your code good nigh-igh-igh-ight
[10:30] <daf> mmm, arch filk
[10:39] <sabdfl> spiv: before assuming the mantle of doap-king, could i ask you to act as dealer?
[10:39] <sabdfl> need you to add support for the extra fields to add/edit forms that are not currently there
[10:39] <sabdfl> in addition, the sourceforgeproject and freshmeatproject fields
[10:40] <sabdfl> so that we can fully describe these suckers when we go live
[10:40] <sabdfl> don't worry about the sourceforge scraping api, just gather the project name
[10:41] <daf> SteveA: you called the debug skin "DebugLayer", but the other class names end in Skin
[10:41] <daf> SteveA: which is canonical?
[10:42] <sabdfl> i think a skin is made up of layers
[10:42] <SteveA> they should all be "Layer"
[10:42] <SteveA> and the module should be layers.py really
[10:42] <daf> shall I make the change?
[10:42] <SteveA> there isn't such a distinction between skins and layers
[10:42] <SteveA> yes please
[10:43] <SteveA> in fact, there is no difference at all
[10:43] <SteveA> it is just a hang-over from the way plone / cmf works, which this was modeled on originally
[10:43] <SteveA> I'll change the "newskin" attribute of suburl to be "newlayer" too
[10:43] <SteveA> but not tonight
[10:49] <daf> SteveA: shall I do that too?
[10:49] <daf> SteveA: it doesn't seem difficult
[10:50] <SteveA> if you like
[10:50] <daf> see you tomorrow
[10:50] <sabdfl> night SteveA
[11:17] <BradB> Presumably bugpackagerelease.zcml can also include the container configurations?