[00:04] <mwhudson_> mm
[00:04] <mwhudson_> you know, the scanner and the puller being different processes is pretty crackladen
[00:30] <mwhudson_> er
[00:30] <mwhudson> is make run_all broken?
[00:31]  * mwhudson grrs
[01:06]  * mwhudson afk for a couple of hours
[05:57] <mwhudson> wow that took longer than i expected
[05:57] <mwhudson> jml: i hope you haven't had too lonely an afternoon
[05:57] <jml> mwhudson, no, not really.
[05:59] <jml> mwhudson, actually, now would be a great time for me to review your branches.
[06:00] <mwhudson> jml: i was planning on having dinner and then working into the evening a bit
[06:00] <mwhudson> jml: would you like to talk about my branches now, or after you've looked a bit?
[06:00] <jml> mwhudson, ok. I've got an appointment in an hour.
[06:00] <jml> let me think.
[06:09] <jml> mwhudson, sorry. that took too long.
[06:09] <mwhudson> jml: :)
[06:09] <jml> mwhudson, let's talk now, if that's ok.
[06:09] <mwhudson> jml: ok
[06:09] <mwhudson> jml: let me get my headset
[06:10] <mwhudson> jml: skype doesn't seem to think you're online
[06:10] <mwhudson> ah, it does now
[06:11] <mwhudson> jml: yay linux audio?
[06:49] <jml> g'night all.
[07:44] <noodles775> BjornT_: around? Just wondering if you're looking at the memcached vs. windmill layer buildbot issue?
[08:08] <poolie> hi noodles775
[08:08] <noodles775> hi poolie
[08:09] <noodles775> Looking forward to the European chill?
[08:10] <poolie> -6C in strasbourg yesterday
[08:10] <poolie> it's going to be a bit of a shock
[09:12] <mwhudson> ImportError: No module named memcache
[09:12] <mwhudson> what did i do wrong?
[09:13] <mrevell> Morning!
[09:14] <mwhudson> oh, make
[09:14] <mwhudson> probably
[10:20] <jtv> hi mrevell!  how's the head?
[10:21] <jtv> al-maisan: hi!  I filed bug 499350, but it may already be included in your work on bug 484819... just making sure it's not forgotten.
[10:21] <mup> Bug #499350: BuildQueue still tied to BuildPackageJob <Launchpad itself:New> <https://launchpad.net/bugs/499350>
[10:21] <mup> Bug #484819: The job dispatch time estimation needs to be generalized to cope with the various build farm types <Soyuz:In Progress by al-maisan> <https://launchpad.net/bugs/484819>
[10:21] <mrevell> Hi there jtv, still sore, believe it or not, but not too bad.
[10:50] <jtv> al-maisan: if it's stuff you're not covering, maybe it's a nice task for me—or maybe it should go into the sprint...  Chat?
[10:51] <al-maisan> jtv: sorry .. status meeting + got distracted on #lp .. let me look now.
[10:51] <jtv> :)
[10:54] <al-maisan> jtv: I nailed down the first bullet point in the forthcoming job dispatch time estimation branch (http://pastebin.ubuntu.com/344638/), noodles775 is looking at the 2nd: build farm candidate job selection
[10:55] <noodles775> al-maisan: am I? I'm not currently and it's not on my immediate plans (my next venture on that work will be bug 491330)
[10:55] <mup> Bug #491330: Update Builder history UI for generalised builders <Soyuz:Triaged> <https://launchpad.net/bugs/491330>
[10:55] <noodles775> ie. IBuilder.getBuildRecords()
[10:56] <al-maisan> noodles775: oh, I must have misunderstood then, I thought you were looking into build candidate selection, no?
[10:56] <jtv> al-maisan: I'll remove the one about specific_job from the bug
[10:56] <al-maisan> jtv: thanks
[10:57] <noodles775> al-maisan: no, I wrote an email about an idea for simplifying it, but that's it.
[10:57] <noodles775> it might be that I do work on it later, not sure, but my immediate work will be getBuildRecords()
[10:57] <al-maisan> jtv: anyway, the build candidate selection is somewhat similar to the dispatch time estimation
[10:58] <jtv> al-maisan: it's complicated, I suppose, by the fact that it returns packagebuildjobs as well as buildqueues
[11:01] <al-maisan> jtv: the relevance of the getForBuilds() method is someewhat reduced by the fact that it's used only once so far (in the soyuz browser layer)
[11:02] <jtv> al-maisan: that's a relief—it means it won't get in our way.  It'll just ignore our jobs.
[11:03] <al-maisan> jtv: re. difficulty: I am recycling from a recent email on the subject
[11:03] <al-maisan> In the end this is somewhat of a balancing act -- you want to implement
[11:03] <al-maisan> general build farm mechanisms in a general fashion while allowing for
[11:03] <al-maisan> some degree of build farm job type specific customisation.
[11:06] <jtv> al-maisan: it doesn't look _too_ hard to generalize all of these with a single approach...  "Add your BuildFarmJob type to this list of classes"
[11:09] <al-maisan> jtv: the devil is in the detail .. please have a look at the email I just forwarded to you
[11:10] <jtv> al-maisan: thanks
[11:10] <jtv> I will
[11:49] <jtv> al-maisan: btw Job Of Interest is now TLA #23498 in the GTF
[11:49] <mup> Bug #23498: FTBFS/mipsel: Please stop building java bindings for mipsel <subversion (Ubuntu):Invalid> <subversion (Debian):Fix Released> <https://launchpad.net/bugs/23498>
[11:50] <jtv> no mup, that's not a bug
[11:50] <al-maisan> jtv: what is "GTF"?
[11:50] <jtv> al-maisan: http://xs4all.nl/~jtv/gtf/
[11:51] <al-maisan> jtv: Ah! Very nice :)
[11:51] <jtv> al-maisan: with a contribution by DEK.  :)
[11:51]  * al-maisan looks up DEK
[11:52] <jtv> But you, too, can now use the GCP logo to show that you're a contributor.  :-)
[11:52] <al-maisan> :)
[13:07] <Aim_> Hello
[13:10] <Aim_> anyone ever tried ditching apache?
[17:44] <maxb> Aim_: I have heard it said that the production Launchpad actually uses Pound
[17:46] <elmo> haha, what
[17:46] <elmo> not for 3 or 4 years now
[17:46] <elmo> LP is a mix of apache, squid and haproxy
[17:46] <elmo> LP.net that is
[18:00] <mrevell> night all
[20:01] <mwhudson> good morning
[20:02] <ajmitch> hi
[22:10] <jml> mwhudson, hi
[22:14] <poolie> hi jml
[22:15] <jml> poolie, hello
[22:38] <mwhudson> jml: good day
[22:39] <jml> mwhudson, good day to you too.
[22:56] <mwhudson> hm, bzr 2.1b4 got released
[22:56] <mwhudson> i guess i could upgrade us to that