[07:07] <poolie> ok good night
[07:07] <nigelb> g'nite poolie :)
[07:08] <nigelb> drat, its already pub time for you :P
[07:20] <jam> morning all
[07:59] <mgz> morning all
[10:00] <mgz> jelmer_: your recipe-status page wants a new column :)
[10:00] <jelmer_> mgz: it should automatically add one once there is a build for precise
[10:00]  * jelmer_ reruns the script
[10:04] <mgz> cool. while I'm bugging you jelmer_, are there user instructions for getting started with colocated branches in core anywhere?
[10:05] <jelmer_> mgz: not really. do you mean bzr-colo or the colocated branch support that's in the core?
[10:07] <mgz> I mean in core, as I think with the devel format support landing it should be somewhat possible to use now?
[10:08] <jelmer_> mgz: yep
[10:09] <jelmer_> basically, use --development-colo to create a directory and then use the ,branch=name syntax to specify a specific branch
[10:10] <mgz> thanks, I will play around.
[10:24] <cjwatson> hmm.  maybe I should just prepare this dpkg upload outside UDD; pretty awkward to work on getting the import fixed while the importer is stopped, I guess
[10:24] <cjwatson> oh, here, lp:debian/dpkg is up to date now
[10:27] <cjwatson> woo.  massive pile of conflicts, but unblocked.
[10:37] <mgz> do we know what the signal is to restart the importer? james_w said last night we should wait for the branch import which takes ~18 hours
[10:44] <CaMason_> Hi guys. How can I 'merge' / squash 2 commits into one?
[10:46] <mgz> CaMason_: if they're ones you just did and haven't pushed anywhere, just use uncommit twice then commit
[10:47] <CaMason_> thing is, these 2 commits are behind another commit
[10:48] <CaMason_> I want to squash 318..319 together, and I'm at 320
[10:48] <LeoNerd> You cna't without "rewriting history"
[10:48] <LeoNerd> Rewriting history is bad and evil and upsets anyone who's checked out / branched / etc... since the rewrite.
[10:48] <mgz> one option, there, again provided you've not pushed anywhere public,
[10:49] <CaMason_> nobody has - this is a local branch
[10:49] <LeoNerd> But if you want, you could e.g. checkout -r318; merge -r320; replay the rest.
[10:49] <mgz> is uncommit, shelve, uncommit, uncommit, commit, unshelve, commit
[10:49] <mgz> or what LeoNerd said is less accident prone
[10:49] <CaMason_> mgz, ah yes, that would work!
[10:49] <LeoNerd> Oh, a shelve stack. Another possible
[10:50] <mgz> I do tend to get myself into trouble on occasion when I'm juggling too many shelves and uncommits in one branch
[10:51] <mgz> making a new branch from the oldest point and cherrypicking revisions across is less destructive
[10:52] <CaMason_> I tried a cherrypick with `bzr merge -r 318..319` but it carried across the revisions
[10:54] <mgz> `bzr revert --forget-merges` will always let you lose the explicit merge info
[10:54] <CaMason_> oops wrong button
[10:54] <CaMason_> ok thanks. The shelve-stack worked. I'll remember the other suggestions though
[10:55] <mgz> cool!
[10:57] <CaMason_> also, I was testing these merges in a seperate branch. In git, I'd now do a ff-merge. Would `bzr pull` be about the same?
[10:59] <mgz> you may want --overwrite
[10:59] <mgz> but otherwise yes.
[10:59] <CaMason_> good stuff. Thanks
[10:59] <CaMason_> btw, would a shared repository make it quicker to create new branches?
[11:01] <spiv> CaMason_: yes
[11:02] <spiv> (a shared repo by itself doesn't avoid the cost of building the working tree every time, but it does save copying the history)
[11:10] <quicksilver> I was about to ask you if there was a command to do something like 'bzr ls'
[11:10] <quicksilver> it turns out there is
[11:10] <quicksilver> it's called 'bzr ls'
[11:10] <quicksilver> TIAS++
[11:22] <jelmer_> :)
[11:34] <james_w> mgz, it's on the release team's checklist for opening a new series, so we'll get a ping
[11:34] <james_w> mgz, you could ask a l-osa whether the branch-distro is finished yet, and start it if it is
[11:41] <mgz> james_w: great, thanks
[13:43] <mgz> so, apparently the branch-distro script failed, which blocks restarting the importer.
[13:44] <mgz> there's a log on carob with more information, but I'm not sure how to get there from chinstrap to look.
[13:45] <mgz> as am not allowed agent forwarding, and trying to follow the instructions on the wiki for the alternative setup just fork-bombs myself with ssh processes
[13:45] <mgz> so clearly someone more competent will have to handle this.
[13:46] <mgz> that might mean Monday.
[14:05] <fullermd> I doubt it.  Nobody's competent on monday...
[14:09] <mgz> :)
[16:19] <fullermd> jelmer_: Is that .THIS file supposed to be there?
[16:19] <jelmer_> fullermd: which one?
[16:19] <fullermd> added: bzrlib/tests/per_repository/test_fileid_involved.py.THIS test_file_involved.py
[16:19] <fullermd> +-20051215205901-728a172d1014daaa
[16:20] <jelmer_> fullermd: where is that?
[16:20] <fullermd> Is that commit you just put through PQM.
[16:20] <fullermd> s/Is/In/
[16:20]  * fullermd khan tipe.
[16:21] <jelmer_> fullermd: that's indeed not supposed to be there, thanks.
[16:22] <fullermd> PQM mailing again ++.  Makes it so much easier for me to pick people to harass in passing   8-}
[20:13] <cr3> hi folks, bzr seems to ignore buildout-templates/bin/lint.in because I have "bin" in my .bzrignore file. aren't bzrignore patterns supposed to be absolute?
[20:14] <cr3> maybe I'm supposed to use ./bin instead of just bin
[20:15] <wgz> for the bin in the root of your project only, yes
[20:15] <cr3> wgz: I now see this is actually well documented, my mistake :)
[20:16] <cr3> wgz: I guess I'll prepend './' to every pattern that's just in the root, just to be extra explicit
[21:37] <Noldorin_> hi jelmer_
[21:37] <jelmer_> hi Noldorin_
[21:38] <Noldorin_> how's it going?
[21:43] <jelmer_> Noldorin_: alright, how are you?
[21:44] <Noldorin_> jelmer_, not too bad. just very busy i'm afraid, with my own job(s) now
[21:44] <Noldorin_> jelmer_, so no time to really look into the bzr-git issue, sorry. unless i get really lucky over the next week or so, you will probably beat me to it ;-)
[22:14] <wrv> hi
[22:34] <jelmer_> hi wrv
[22:50] <elmo> does anyone know any simple-ish plugins using pre-commit hooks off the top of their heads?
[22:54] <jelmer_> elmo: I think james_w had a plugin that tried to run a testsuite in pre_commit, let me see if I can find it..
[22:54] <elmo> jelmer_: actually don't worry, I'm being an idiot - a hook, won't do what I want
[22:54] <james_w> https://launchpad.net/bzr-testrunner
[22:54] <elmo> jelmer_: thanks though