[00:09] <poolie> hi all
[00:10] <poolie> can someone help me understand why https://code.launchpad.net/~mbp/launchpad/612171-diff-generation-spam/+merge/79244  is out of date?
[00:10] <poolie> because i pulled from trunk a bit earlier than lp did?
[00:10] <poolie> s/did/scanned it
[00:10] <poolie> no, nm
[00:10] <poolie> because i set the wrong target
[02:31] <mwhudson> err how many packages can you copy using +copy-packages at once without it timing out?
[02:31] <wgrant> mwhudson: Somewhat between 0 and infinity.
[02:31] <wgrant> Probably on the lower end.
[02:31] <wgrant> Occasionally 0, for stuff like linux.
[02:31] <mwhudson> wgrant: yeah, i'm definitely failing at substantially less than infinity
[02:32] <wgrant> I've done 10 packages for small stuff.
[02:32] <wgrant> It's better than it used to be, but I'm the only one who has tried to optimise it and I'm on a feature squad now :(
[02:32] <wgrant> It's my main priority once I get back onto maintenance, though.
[02:32] <mwhudson> what's the measure of 'size' here?
[02:32] <wgrant> number of sources + number of binaries, mainly.
[02:32] <mwhudson> ah ok
[02:32] <wgrant> + number of matching sources in the target archive
[02:33] <wgrant> If any of them exist already in the target archive, it gets very expensive.
[02:33] <mwhudson> i'm copying from lucid to oneiric in a ppa
[02:33] <mwhudson> so yeah, they all exist in the target archive
[02:33] <wgrant> Yeah, that invokes extra consistency checks.
[02:33] <mwhudson> \o/
[02:36] <mwhudson> er
[02:37] <mwhudson> why do i get things like
[02:37] <mwhudson> python-versiontools-mercurial 1.8-0ubuntu0~lava1 in oneiric amd64
[02:37] <mwhudson> python-versiontools-mercurial 1.8-0ubuntu0~lava1 in oneiric armel
[02:37] <mwhudson> python-versiontools-mercurial 1.8-0ubuntu0~lava1 in oneiric i386
[02:37] <mwhudson> python-versiontools-mercurial 1.8-0ubuntu0~lava1 in oneiric powerpc
[02:37] <mwhudson> when i copy an arch: all package?
[02:37] <wgrant> It's architecture-independent.
[02:37] <wgrant> So it's published on all architectures.
[02:37] <mwhudson> oh right, publications
[02:37] <wgrant> It lists the copied publications.
[02:37] <mwhudson> makes sense
[02:37] <wgrant> Not releases.
[02:37] <wgrant> Right.
[02:37] <wgrant> A little confusing!
[02:38] <mwhudson> this is fairly boring
[03:09] <mwhudson> wgrant: btw, the number between 0 and infinity appears to be "2"
[03:10] <wgrant> mwhudson: Heh.
[03:15] <StevenK> mwhudson: For the particular archives you picked -- it can vary wildly
[03:15] <mwhudson> ok
[03:15] <StevenK> (Sadly)
[03:15] <mwhudson> i managed to copy the 25 or so packages i cared about eventually
[04:31] <Darxus> 6 hours estimated till build, boo :P
[04:32] <Darxus> Better than the initial estimate of 10 hours.
[04:32] <wgrant> Darxus: Most of the builders are gone for a few days due to oneiric.
[04:33] <wgrant> I've shuffled some builders around to try to even things out a bit.
[04:33] <Darxus> wgrant: I figured it was something like that, thanks.  Really no urgency, my stuff can wait.
[04:34] <Darxus> I think you meant precise, not oneric.
[04:34] <StevenK> No, we mean oneiric
[04:34] <wgrant> precise isn't doing much yet.
[04:34] <StevenK> The builders have been borrowed to add "bulk" to releases and cdimage .ubuntu.com
[04:34] <wgrant> oneiric is using lots of bandwidth and servers :)
[04:35] <Darxus> Ohhh, okay, thanks.  I thought you were rebuilding everything for precise.  It's cool that you have the ability to re-allocate build machines for image serving.
[04:35] <StevenK> Not yet
[04:35] <StevenK> Precise is still having the toolchain uploaded
[04:35] <wgrant> Darxus: https://launchpad.net/builders
[04:35] <wgrant> Note that the official distribution builders are separate, and presently idle.
[04:36] <Darxus> Neat, thanks.
[04:37] <Darxus> Heh, PPA builders have 554 jobs queued.
[04:37] <wgrant> I shuffled some amd64 builds onto i386 earlier to try to even the queues out.
[04:37] <Darxus> Why so many more jobs for i386 than amd64?  Many packages not building 64 bit yet?
[04:37] <Darxus> Ahh.
[04:37] <wgrant> Darxus: architecture-independent packages build on i386.
[04:37] <Darxus> Ah.
[04:38] <Darxus> You should automate that re-assignment :)
[04:38] <wgrant> Indeed. The main thing stopping me from doing that is, somewhat ironically, the build time estimation.
[04:39] <Darxus> What, working on improving the algorithm?
[04:42] <Darxus> The launchpad build system is great, I really appreciate it.  SpamAssassin upstream is a major pain to work with.  So I have my automated daily builds set up, and once a month I'm copying the daily builds to a -test PPA, and when a month of testing goes well, copying the builds from -test to -monthly.  Today is my first time copying from -test to -monthly.  So hopefully I'll find out if anybody else is interested.
[04:43] <Darxus> SpamAssassin's Project Management Comittee claims we don't have the man power to do monthly releases, and I can show it takes no man power.
[04:48] <poolie> Darxus: you know there's a command-line copier in hydrazine or lptools?
[04:48] <poolie> i forget which
[04:53] <Darxus> poolie: I tried figuring out how to automate it via the API, got frustrated, and really, the web interface is fine for my purposes.  I only want it to happen when I've verified the month of testing went well anyway.  Thanks.
[04:54] <Darxus> Hadn't heard of hydrazine though, I'll look into it.
[04:56] <poolie> Darxus: launchpad.net/hydrazine, and its' called 'lp-promote-ppa'
[04:56] <Darxus> Thanks.
[07:59] <czajkowski> mrevll ping
[08:22] <czajkowski> mrevell: ping even
[08:22] <mrevell> Hey czajkowski
[09:03] <davidcalle> Hi
[09:07] <davidcalle> I have an issue with loggerhead, all the text is highlighted in red in non code files, making them hard to read, what can I do to avoid that?
[09:07] <lifeless> it will be pygments (mis)detecting the language
[09:09] <davidcalle> lifeless, I get it, but it doesn't do it in README files, for example. Is this documented somewhere?
[09:10] <hrw> any news on https://answers.launchpad.net/launchpad/+question/173422 maybe?
[09:10] <hrw> 'bzr: out of memory' stuff
[09:11] <lifeless> davidcalle: can you point me at a misdetected file ?
[09:11] <davidcalle> http://bazaar.launchpad.net/~davidc3/onehundredscopes/rottentomatoes/view/head:/SPECS
[09:12] <davidcalle> lifeless ^
[09:13] <lifeless> I'm not sure whats up ;)
[09:13] <lifeless> I think we use the automagic api
[09:14] <lifeless> which suggests your content is being detected as a file type that its not (and then highlighted as bad syntax or something)
[09:15] <davidcalle> I will try to change the "something : something" syntax.
[09:24] <davidcalle> lifeless, fixed with a new syntax. Thanks :)
[09:29] <lifeless> davidcalle: np
[12:18] <lamont> jelmer_: about?
[12:19] <jelmer_> hi lamont
[12:19] <lamont> python-debian 0.1.21 - I assume that's not in ubuntu only because there hasn't been an opportunity to upload it?
[12:22] <jelmer_> lamont: yeah, I'm going to get it merged for precise
[12:23] <lamont> ta
[12:36] <lamont> jelmer_: and I assume I need python-debian available before bzr  builds, and that available before bzr-builder builds?
[12:36] <jelmer_> lamont: bzr doesn't need the newer python-debian, bzr-builder needs the newer bzr and python-debian
[12:38] <lamont> your python-debian lucid backport needs to be format 1.0, for future reference. (fixed)
[12:39] <soren> I'm getting these again: http://paste.ubuntu.com/707965/
[12:40] <jelmer_> lamont: I thought lucid supported 3.0?
[12:40] <lamont> jelmer_: it's an archive issue, not a builder issue
[12:41] <jelmer_> lamont: Ah, ok - thanks
[12:41] <soren> The upload gets accepted, though.
[12:45] <jelmer_> soren: I get them occasionally too, haven't yet figured out what causes them. Using sftp seems to work around it though.
[12:46] <lamont> jelmer_: let me guess... all of your lucid backports have that issue, don't they?
[12:46] <lamont> have I mentioned just how much I totally hate 3.0?
[12:46] <soren> jelmer_: Lovely.
[12:47] <lamont> W: bzr source: quilt-series-but-no-build-dep
[12:47] <lamont> I wonder if that's going to kill this build...
[12:47] <soren> jelmer_: I think last time I mentioned it, lifeless seemed to have an idea what was going on.
[12:49] <jelmer_> lamont: Sorry, yes - they all use 3.0. Are you ok changing them back to 1.0 or should I have a look?
[12:49] <lamont> jelmer_: I'm just smashing source/format to 1.0 and uploading... if that fails, then I'll be back to pester  you... feel free to start down that path before I bug you if you're bored
[12:58] <jelmer_> lamont: I'll have a look then, I suspect the quilt patches will need some changes
[13:44] <czajkowski> gmb: it's called drinking and talking to canonical and pimping you out
[13:46] <gmb> czajkowski: I'm certain it's inappropriate for me to buy you a t-shirt that says "gmb's pimp" on it, but I can't help but think it would be accurate.
[13:46]  * gmb remembers this channel is logged, and that this is the wrong channel for this conversation :)
[13:46] <czajkowski> lol
[15:01] <lamont> jelmer_: interestingly, it compiled just fine
[15:01] <lamont> bzr that is
[15:02] <jelmer_> lamont: was this with quilt present?
[15:02] <lamont> jelmer_: logistically speaking, I'll work on having all the bits built by "sometime monday", for testing and initial rollout starting "sometime tue/wed"
[15:02] <lamont> no
[15:02] <lamont> OTOH, when I unpacked the 3.0 source, quilt applied the patches
[15:03] <sagaci> what's the downtime like for staging.lp.net... are we talking a few minutes or is it a longer-term thing
[15:03] <jelmer_> lamont: ah, that would explain it
[15:03] <lamont> yeah
[15:04] <jelmer_> lamont: thanks for the update; I'll be around next week except for Friday if you need me to look into anything
[15:04] <lamont> colol
[15:04] <lamont> cool, even