[02:48] <poolie> SteveA: ping?
[03:33] <thumper> hey poolie, it's a little early for SteveA
[03:34] <poolie> yeah i thought so
[03:35] <thumper> both ddaa and SteveA are exactly 12 hours out of sync with me
[03:35] <thumper> makes time calculations real easy
[03:35] <thumper> but communication sometimes a little more difficult
[03:36] <poolie> yel
[04:08] <stub> thumper: I'm not sure how to proceed if we need to store the Bazaar DAGs in the relational db. Given the rough numbers Robert gave me in Singapore, it seemed that it would be impractical to explode the graphs into a table that would allow us to list all children/ancestors for a given node in a single query. And benchmarks using stored procedures to query the existing tree demonstrated that approach was too slow.
[04:13] <thumper> stub: for sure we can't work solely on an adjacency model
[04:13] <thumper> that just doesn't work for 1000's of levels of recursion
[04:13] <thumper> I've had some thoughts on a hybrid model, and using iteration instead of recursion
[04:13] <thumper> can you point me at any plpythonu docs?
[04:14] <thumper> that might be easier than plpgsql
[04:14] <thumper> which is what I have always used before
[04:14] <stub> file:///usr/share/doc/postgresql-doc-8.1/html/plpython.html
[04:14] <thumper> stub cheers
[04:15] <jamesh> so, is there a list of the things we want to do with the data?
[04:15] <jamesh> that would give a good idea of what needs to be in the DB
[04:15] <thumper> jamesh: no explicit list that I know of
[04:15] <stub> I'm aware of: Given a node in a DAG, list a) All child nodes, b) All parent nodes, c) All descendants, d) All ancestors
[04:16] <thumper> I'm wondering how much we can do through a web interface to the branches and avoid having it all in the db
[04:16] <thumper> stub: but why?
[04:16] <thumper> it is one of the reasons that I haven't invested too much time into it yet
[04:16] <thumper> that and I haven't had much time yet
[04:17] <thumper> that way we can use bzrlib to do the heavy lifting
[04:17] <jamesh> the main one I could think of was "which branches have fully merged this one?", which is roughly calculating all nodes reachable by parent -> child edges
[04:17] <stub> I'm not sure of the actual use to the end user. Just that my understanding was that if we could support that then we could support the envisaged UI reports from the relational db (not that I know what they are)
[04:17] <jamesh> and maybe the inverse (which other branches does this branch contain)
[04:18] <thumper> jamesh: that fits into what I have heard of the "auto merge detection" concept
[04:18] <jamesh> are there other features we want?
[04:18] <thumper> jamesh: does that degenerate down to is the tip of branch A in the ancestry for branch B
[04:18] <thumper> if so, then A merged into B
[04:18] <jamesh> thumper: yeah
[04:19] <thumper> bzrlib is blinding at that sort of thing
[04:19] <stub> But not given a set of 50 or 100 branches
[04:19] <thumper> perhaps we look at an xmlrpc on the supermirror accessable through the lp ui
[04:20] <stub> It works two branches at a time
[04:20] <thumper> stub: true
[04:20] <thumper> but is that what we want?
[04:20] <stub> Thats between you and Mark I think :)
[04:20] <thumper> surely we are more interested in the "have I been merged into my landing target"
[04:21] <stub> I guess specing out what we want the pages to display do is the first step. Then we can work out if it is possible and how to do it.
[04:21] <thumper> stub: that sounds like a plan to me
[04:21] <stub> go forth and mock (up)
[04:21] <thumper> my suggestion would be to not bother storing everything in the db if it isn't needed
[04:22] <thumper> especially if the complexity has already been addressed elsewhere - like bzrlib
[04:26] <jamesh> thumper: I sent an email last night about maybe changing the terms used in dbschema vocabularies, as a way to shorten the URLs in your branch filtering branch
[04:26] <thumper> jamesh: yep read it (ddaa was keen on this too)
[04:26] <jamesh> does it sound okay to you?
[04:26] <thumper> I looked at it when I was first approaching the issue
[04:26] <thumper> it has to do with how we create the vocabularies from the dbschema objects
[04:27] <thumper> I didn't want to do too much with it at that stage due to the changes in the pipes that SteveA is looking at
[04:27] <thumper> and I'll probably be implementing anyway
[04:27] <jamesh> sure.
[04:27] <thumper> so I'll make sure that it works
[04:27] <jamesh> I think it is independent of Steve's proposed changes
[04:27] <thumper> and I fully agree about using value instead of text to avoid i18n issues
[04:27] <jamesh> but is probably worth doing afterwards
[04:27] <thumper> well it would have meant hacking some vocab factories up
[04:28] <thumper> and I want to make that easier anyway
[04:28] <thumper> also widget changes in the pipes too
[04:28] <jamesh> like auto-register vocabularies from the DBSchema metaclass or something?
[04:29] <thumper> I think the current plan is to have the enumeration class implement IVocabulary
[04:29] <jamesh> ah.
[04:29] <jamesh> that sounds good
[04:29] <thumper> yep
[04:29] <thumper> less registering
[04:29] <thumper> (I think)
[04:29] <thumper> at least we wouldn't need the factory file
[04:30] <jamesh> as I said in the mail, it isn't just the potential i18n issues
[04:30] <jamesh> we've changed a bug importance name in the past, and that broke bookmarks for a number of Ubuntu developers
[04:31] <jamesh> and some queries that were linked off the Ubuntu wiki, iirc
[04:32] <thumper> all it would take is a different vocab factory (I think)
[04:32] <stub> Ohh... PostgreSQL 8.2 might give us facilities to do this using arrays and the new operators
[04:32] <thumper> right now
[04:32] <stub> (DAGs)
[04:32] <thumper> stub: effective limits on arrays?
[04:33] <stub> No idea. I'll wait until I have an install to investigate further.
[04:33] <thumper> stub: are we on 8.1?
[04:33] <stub> Yup
[04:33] <stub> 8.2 released just a short while ago
[04:33] <thumper> 8.3 has "with recurisive" planned
[04:33] <thumper> :)
[04:33] <thumper> FQ2007
[04:33] <thumper> maybe
[04:34] <stub> If we want to wait 6 months...
[04:34] <stub> fq?
[04:34] <stub> Hmm...
[04:34] <thumper> first quarter
[04:34] <jamesh> we probably want to wait for an 8.2 point release before considering it though, right?
[04:34] <stub> recursive stuff would be ideal if it is efficient and not just the same stuff I've already benchmarked using stored procedures.
[04:35] <stub> jamesh: They have been pretty good in the last few releases. I'd be game to give 8.2.0 a go if it passes our testing (although by that time, there will probably be a 8.2.1 anyway)
[04:36] <thumper> stub: I believe that the recursive stuff was *much* faster than stored procs
[04:37] <thumper> as it is an inherent knowledge at the db level of tree structures
[04:37] <stub> enterprisedb supports this right now...
[04:37] <stub> (using oracle syntax)
[04:37] <thumper> eterprisedb?
[04:37] <jamesh> thumper: a postgres fork that aims to make porting from Oracle easier
[04:37] <stub> commercial release of PostgreSQL
[04:37] <stub> I'm seeing July 2007 quoted for 8.3
[04:38] <thumper> ok
[08:21] <spiv> jamesh: your latest blog entry seems to have a typo: a paragraph ending in an incomplete sentence "if you do development from multiple machines if you".
[08:26] <jamesh> spiv: thanks.  Updated
[08:35] <thumper> spiv: are you still working
[08:35] <spiv> thumper: yep
[08:36] <thumper> I'd like to organise a voice call
[08:36] <thumper> skype?
[08:36] <spiv> Sure.  I'll quickly find my headset...
[08:38] <spiv> I'm on skype... "spivvo" is the name iirc.
[08:38] <thumper> spiv: ok, I'll try now
[08:41] <thumper> spiv: nope, seems like that isn't there
[09:07] <carlos> morning
[10:41] <SteveA> poolie: hi
[10:44] <stub> Launchpad will be going down in 15 minutes for a code update. Estimated downtime is 10 mins.
[10:53] <jsgotangco> ah
[11:09] <stub> That should all be done now
[11:49] <mhb> carlos: I read your mail on ubuntu-translators ... thanks for the fix.
[11:49] <carlos> mhb: sorry for the extra delay...
[11:50] <mhb> carlos: no problem. will the google indexing of translations also work soon? (or has it been deferred?)
[11:51] <carlos> mhb: it's implemented, poke kiko on it, he needs to review it
[11:51] <carlos> but anyway... I don't think that will happen before January
[11:51] <carlos> I'm not sure we will do a code update on production just before leaving for holidays
[11:51] <carlos> the whole company...
[12:01] <matthewrevell> morning jenda
[12:02] <jenda> hey there matthewrevell :)
[12:02] <jenda> How's life treating you?
[12:02] <jenda> And how's Canonical treating you? ;)
[12:02] <jenda> And - where does ubuntulog store it's logs for this channel?
[12:07] <jenda> found'em
[12:11] <matthewrevell> jenda: Not bad on both fronts, thanks :)
[12:12] <jenda> matthewrevell: Just in case you could make it, and were interested, there is a Marketing Team meeting on saturday, midnight UTC, after a looong time.
[12:12] <jenda> I think you even attended the last one :)
[12:12] <matthewrevell> jenda: Wow. Cool. Unlikely I'll be able to make it, but I'll try. Thanks for letting me knoiw
[12:12] <jenda> But don't worry about it if you can't make it.
[12:12] <jenda> np
[12:36] <Ubugtu> New bug: #75874 in soyuz "please don't require Standards-Version" [Undecided,Unconfirmed]  http://launchpad.net/bugs/75874
[12:50] <cprov> SteveA: hi, can you find someone to help me with a vhosts related error on dogfood ? https://devpad.canonical.com/~andrew/paste/fileJQQHqQ.html
[12:52] <SteveA> cprov: I can help you with that.
[12:53] <cprov> SteveA: nice, is there something missing in dogfood config ?
[12:55] <SteveA> cprov: where can I see the current dogfood config?
[12:56] <cprov> SteveA: uhm, in your tree, config/dogfood/launchpad.conf (I'm using RF-head there)
[12:57] <SteveA> ok
[12:57] <SteveA> I'll pull down the latest and take a look
[12:58] <cprov> SteveA: thanks
[01:10] <SteveA> cprov: ok, so the dogfood config doesn't have a vhosts section in the launchpad section
[01:10] <SteveA> and if dogfood runs a web server, then it should do
[01:10] <SteveA> you can copy the one from staging, say, and then customize it
[01:10] <SteveA> ping me if you need help knowing what to put where
[01:11] <cprov> SteveA: sorry, power outage over here ..
[01:11] <cprov> SteveA: will do, thanks o lot for identifying this.
[01:37] <cprov> carlos: hi, any idea about how long patch-67-19-0.sql will take to be applied in dogfood ? (1 month old lp_prod copy)
[01:40] <carlos> let me check...
[01:40] <carlos> cprov: no idea, kiko did that change so I didn't do any timing on it
[01:42] <cprov> carlos: ok, I have to wait, I guess
[01:43] <matthewrevell> So, how's Fix-it Friday going?
[01:45] <salgado> fixing a lot of small annoying issues on team membership notification
[01:48] <matthewrevell> salgado: Ah, that sounds good. What sort of thing?
[01:49] <salgado> matthewrevell, bugs: [WWW]  66787, [WWW]  73079, [WWW]  47227, [WWW]  47476, [WWW]  49637, [WWW]  55156.
[01:49] <salgado> ah, crap. copy and pasting from the wiki doesn't always work
[01:49] <matthewrevell> :) Thanks, I'll look them up
[01:50] <salgado> matthewrevell, the links are at https://launchpad.canonical.com/PendingReviews
[01:50] <matthewrevell> aha, superb
[01:51] <salgado> it's mostly improving the messages in the notifications, avoiding duplicate notifications in some special cases and adding some new notifications
[01:51] <salgado> I also took the opportunity to clean up/refactor some code, since it's been a long time since I last touched it
[01:53] <matthewrevell> salgado: Cheers. I'm going to type up a small report. Nothing too grand, as we're starting off small.
[01:54] <salgado> that's great! :)
[02:27] <kiko> morning
[02:27] <kiko> how is everybody
[02:28] <aa_> excellent, thank-you
[02:28] <kiko> good to hear
[03:15] <cprov> kiko: ping
[03:25] <kiko> cprov, pong
[03:26] <cprov> kiko: hi, can you do a quick review for me, the bug is kind of urgent, it's blocking distro-team
[03:27] <cprov> kiko: fix for bug #75874
[03:27] <Ubugtu> Malone bug 75874 in soyuz "please don't require Standards-Version" [Critical,In progress]  http://launchpad.net/bugs/75874
[03:27] <cprov> kiko: https://devpad.canonical.com/~jamesh/pending-reviews/cprov/launchpad/soyuz-fixes/full-diff
[03:59] <aa_> sorry to bother, can we get commit emails from the supermirror, or was I right in seeing that this specification has not been fulfilled?
[04:03] <LarstiQ> you can have commit mails sent locally
[04:03] <LarstiQ> it would really be 'push mails' for the sm
[04:04] <aa_> sorry yes, push emails
[04:04] <aa_> by locally, you mean in my local branch?
[04:04] <kiko> aa_, tim's working on that; should be ready for the next rollout
[04:05] <aa_> tim peters?
[04:07] <LarstiQ> aa_: hah, we wish :)
[04:07] <ddaa> aa_: Tim Penhey
[04:07] <ddaa> his nickname is thumper
[04:08] <LarstiQ> aa_: by locally, I mean having the email plugin installed, and having a post_commit=bzrlib.plugins.email.post_commit section in ~/.bazaar/locations.conf or bazaar.conf
[04:08] <ddaa> Great to have him around, so he can actually implement all those cool things with bzr integration that we wanted to do forever but never had the time for.
[04:08] <LarstiQ> aa_: it will then send out a mail for every commit you do in the covered branch.
[04:08] <LarstiQ> ddaa: I can pile all such requests on him?
[04:09] <ddaa> LarstiQ: we've got a bug tracker for such requests :)
[04:09] <ddaa> actually the reason this email stuff is being done now is that there is an explicit need for it from Ubuntu
[04:09] <aa_> LarstiQ: ok, that is a perfect solution for now, thanks
[04:09] <ddaa> there's still so much more to do than we can possibly do...
[04:11] <ddaa> I'm just voicing out my unrelenting pleasure at seeing how having Tim in appears to unstuck so many good things.
[04:11] <LarstiQ> ddaa: good to hear
[04:12] <ddaa> don't think too hard, we already have a full agenda for the next six months :)
[05:22] <matthewrevell> Is there a way to view Launchpad users in order of karma?
[05:28] <jenda> matthewrevell: if you find it tell me :)
[05:29] <jenda> I've always been looking for that :)
[05:29] <matthewrevell> jenda: Ah, interesting :) If it's not there, I wonder how many people would find it useful.
[05:30] <BjornT> matthewrevell: i don't think you can see all the users overall, but for each product/distribution you can see the top contributors.
[05:31] <matthewrevell> BjornT: Thanks. I'll give it a go for Ubuntu, then.
[05:35] <BjornT> matthewrevell: btw, if you need something for your Fix-It Friday report, you could add bug 75783. it fixes a timeout problem when listing open bugs statistics for packages a person is a bug contact for.
[05:35] <Ubugtu> Malone bug 75783 in malone "+packagebugs page times out due privacy related clause in the queries." [High,In progress]  http://launchpad.net/bugs/75783
[05:36] <matthewrevell> BjornT: cool, thanks
[06:45] <mhb> I wonder - would it be possible for a translation team of a non-LP project (alsa, for instance) to somehow start a project in Launchpad and manage the translations via Rosetta?
[06:47] <kiko> mhb, what do you mean by "non-LP project"? launchpad is open for any open source project to use
[06:49] <mhb> kiko: I know
[06:51] <mhb> kiko: what I mean is if only a part of a project (one certain translation team) is able to start a LP project and if those people have the possibility of uploading .po templates by themselves
[06:53] <kiko> mhb, certainly -- you don't need official endorsement to use launchpad to translate any project.
[06:54] <mhb> kiko: I have guessed that. Is it easy to upload new .po templates to a project in LP?
[06:54] <kiko> yeah, pretty easy.
[06:54] <kiko> danilos or carlos can assist you if you have trouble.
[06:55] <danilos> mhb: sure, the only thing is that we'll have to manually approve the first POT upload (since there are some settings we have to do), and from then on, it'll all go automatically
[06:59] <mhb> danilos: so after the first upload the templates are fetched from the project the day they are released?
[07:00] <danilos> mhb: no, you'd still have to upload them manually
[07:02] <danilos> mhb: there are some plans for adding VCS integration (using bazaar branches), but we're not there yet
[07:02] <mhb> danilos: thanks
[07:03] <matthewrevell> Other than BjornT and salgado, does anyone have any Fix-it Friday news?
[07:03] <danilos> mhb: no problem; if you run into any problem while trying this out, feel free to ask
[07:05] <mhb> danilos: maybe I have one more question, but a trivial one. Are the ubuntu translator teams able to get translation rights on other projects?
[07:05] <danilos> mhb: sure, it's up to your product to select "translation group" which will be used there
[07:06] <danilos> mhb: you can even keep it completely open, though that may not be very good for the quality of your translations
[07:06] <mhb> danilos: I believe that. Thanks for the answers.
[07:07] <danilos> mhb: np :)
[07:41] <j^> hi, is there any way to browse/view the source of a bzr branche hosted by launchpad in the browser?
[07:52] <kiko> not currently, j^
[08:02] <j^> kiko: are there any plans to add that?
[08:06] <carlos> j^: yes
[08:12] <SteveA> j^: it'll happen soon, and certainly by the end of January.  I secretly expect it to appear much sooner than that.
[08:23] <mhb> one more rosetta-related question
[08:25] <mhb> update manager has msgid strings like " msgid "Every %s days" " which don't need extra plural forms in English but they do in other languages. This string has no msgid_plural defined. If I define it myself in the .po file and then upload it, would that change anything?
[08:30] <carlos> mhb: no
[08:30] <carlos> mhb: you need to 'fix' the application
[08:30] <carlos> Rosetta will store your translation, but you will not see it unless the application uploads a fixed .pot file
[08:31] <carlos> s/uploads/generates/
[08:34] <mhb> carlos: How can I find out who's responsible for that?
[08:35] <carlos> mhb: mvo is the author
[08:35] <mhb> carlos: is he allowed to change .pot files for Edgy so we can fix it there?
[08:36] <carlos> mhb: well, the fix is not just 'fix' .pot files
[08:36] <carlos> I would do it myself
[08:36] <carlos> he needs to change the application
[08:36] <carlos> to be able to use plural forms for that message
[08:36] <carlos> or your translation will be ignored
[08:37] <mhb> aww
[08:37] <carlos> once he fixes it and releases an update for Ubuntu (I don't think that will happen in Edgy, but you can ask for it just in case is possible) Rosetta will get the new .pot file automatically
[08:39] <mhb> that's what I really dislike. Bugs that l10n team can't fix but he will be criticised for them.
[08:41] <mhb> carlos: thanks for the info. I'll go poke mvo and discuss it with him.
[08:44] <carlos> mhb: well, that's a bug in the application, so it's quite easy to state that... The problem is that is not a critical or security problem, that's why I don't think the fix would land in Edgy...
[08:45] <mhb> carlos: Don't worry, I see your point. I just don't like such nasty bugs.
[09:00] <carlos> kiko: UI changes that you asked for TranslationReview are implemented, (included the javascript) and needs a reviewer
[09:01] <ctrlsoft> thanks for making lp usable without images! works like a charm over slow connections
[11:30] <Ubugtu> New bug: #75942 in launchpad "Karma page does not build correct name when it ends with "s"" [Undecided,Unconfirmed]  http://launchpad.net/bugs/75942
[12:15] <Ubugtu> New bug: #75950 in malone "Debian tasks have lost all content" [Undecided,Unconfirmed]  http://launchpad.net/bugs/75950