[00:25] <michaelh1> Hi there.  Can I add meta data to a commit?  I want to tag a commit as 'will not be sent upstream'
[00:31] <james_w> michaelh1, there are revision properties, but I don't think that core provides command line options to set them
[00:32] <michaelh1> OK.  I can embed meta data in the comment just fine.
[02:27] <lamont> jelmer: two things are needed for hardy to fully disappear from the buildds: 1) an LTS with xen support (for the ppa host, especially - not actually a hard hard requirement), and 2) distro telling us that they're happy with using something newer than hardy to build all of the supported releases - dapper's EOL helps with that discussion, fwiw
[06:58] <vila> hello bzr world
[07:04] <jam> morning all
[08:06] <mgz> morning
[08:09] <vila> mgz: _o/
[08:15] <mgz> morning jelmer. do you know off hand where the python-debian upstream repo is?
[08:15] <jelmer> mgz: "bzr info apt:python-debian" :-)
[08:16] <jelmer> or, if you have bzr-git installed: bzr branch apt:python-debian
[08:16] <jelmer> git://git.debian.org/git/pkg-python-debian/python-debian.git
[08:16] <mgz> ...oo, new command to me
[08:16]  * mgz tries it
[08:17] <jelmer> you need bzr-builddeb
[08:17] <mgz> deprecation warnings, and it works
[08:17] <mgz> I am impressed.
[08:28] <jelmer> hmm, what deprecation warnings?
[08:28] <jelmer> (are you on oneiric?)
[08:28] <mgz> I'll put up a branch, it's a trivial spelling of fallback thing
[08:29] <mgz> two in the style of:
[08:29] <mgz> bzrlib/plugins/builddeb/directory.py:47: DeprecationWarning: Attribute 'Lookup'
[08:29] <mgz> of the 'apt_pkg.SourceRecords' object is deprecated, use 'lookup' instead. lookup = getattr(sources, 'lookup', getattr(sources, 'Lookup', None))
[09:27] <mgz> vila: babune windows seems especially unhappy of late
[09:28] <vila> mgz: :-/
[09:29] <vila> someone needs to bring it back to life :-/
[09:29] <mgz> I seem to remember we had a can't-delete-sh-file problem before and fixed it, but I don't remember how
[09:29] <vila> the infamous 'unable to delete <some tmp file>' was spurious but seem to turn into permanent
[09:29] <vila> we never fixed it, it's a jenkings issue which kind of flip-flop over time
[09:30] <vila> there is a 'hudson.remoting.RequestAbortedException: java.io.StreamCorruptedException: invalid type code: 68' in the backtrace doesn't look good
[09:31] <vila> may be related to the leaks too (the leaks in (or releated to) the subunit stream
[09:31] <mgz> yeah, seems like something is leaking through a stream, if that's even possible
[09:31] <mgz> ...vila wins again
[09:32] <vila> ?
[09:32] <mgz> you said the same thing, faster :)
[09:32] <vila> ha :)
[09:32] <vila> worth noticing then, 'cause you're the fastest most of the time ;)
[09:39] <vila> lol
[09:39] <vila> just found a cryptic YARXU commit message....
[09:39] <vila> .. wondered a bit
[09:40] <vila> ... triggered my acronym decipherer
[09:40] <vila> ... first guess: Yet Another Random Xml Update
[09:40] <vila> win ! :)
[09:41] <vila> I really wish we add a way to get better diffs (and commits) when xml files are involved. Most of the time one or two attributes in a "dict" are modified but the key/values are randomly shuffled. The end result is unreadable
[09:42] <jelmer> vila: it would be really nice to have a mechanism for diffs similar to the merge hooks we have
[09:43] <vila> indeed, they are closely related...
[09:43] <jelmer> also, we currently already have some form of custom diff displayers
[09:43] <jelmer> since we don't diff binary files by default
[09:46] <vila> and doesn't qbzr display image diffs too ? (or rather both images)
[10:03] <jelmer> vila: I wasn't aware of that. If it does, that's pretty cool
[10:04] <jelmer> I know github has some fancy stuff for displaying image diffs
[10:06] <mgz> okay, this looks tractable
[10:06]  * mgz is back on bzr-builddeb
[10:07] <mgz> or rather, the mess that is python-debian r95
[10:07] <jelmer> where everything magically became unicode?
[10:08] <mgz> yeah, but without ever checking the inputs
[10:21] <mgz> ugh, setup.py.in
[10:21]  * mgz just copies it
[10:22] <jelmer> mgz: debian/rules has the magic to generate the setup.py IIRC
[10:23] <jml> https://code.launchpad.net/~jml/udd/less-in-top-level/+merge/80037 up for review
[10:23] <mgz> wants setuptools for no good reason too.
[10:23] <jelmer> hey jml
[10:23] <jml> jelmer: hi
[10:23] <mgz> and given that it's just for a version number... which I don't even seem to be able to get at from the installed package, screw it
[10:25] <mgz> what is it with people using fancy conf stuff on python packages then not actually having __version__ anywhere
[10:25]  * mgz eyes subunit
[10:28] <lifeless> mgz: patches appreciated
[10:29] <mgz> I'd have to learn how all that worked to do that lifeless, which is exactly what I'm avoiding :)
[11:52] <vila> mgz: windows slave still failing but at least with results
[11:53] <mgz> woho.
[11:53] <mgz> I've got bugs on the failures.
[11:54] <vila> yakafokon :)
[12:16] <vila> immortal chroots are pita
[12:16] <mgz> you dig them up but they keep on growing back?
[12:17] <vila> they refuse to die
[12:18] <vila> ..can survive reboots and stay around for months
[12:18] <mgz> okay, enough random poking of plugins, it's lunchtime
[12:18] <vila> luckily it doesn't happen often (which of course makes it even harder to understand the cause)
[12:19] <vila> bon appetit
[12:27] <ccxCZ> ghost in z shell? :-)
[12:27] <AuroraBorealis> speaking of those
[12:27] <AuroraBorealis> the archive bit on some of my files refuses to clear :<
[16:00] <mgz> hm, that was a failed attempt at not getting release notes conflict
[16:31] <jml> mgz: thanks for the review
[16:31] <jml> mgz: would you please merge the branch for me?
[16:53] <mgz> jml: I don't have access either, will poke vila
[17:04] <p2000> hello
[17:04] <p2000> bzr wizards
[17:04] <p2000> how do you translate "git --format-patch" to bzr ?
[17:05] <jml> mgz: thanks.
[17:07] <mgz> p2000: `bzr help bundle` is the diff + metadata thing in bzr
[17:07] <mgz> s/is/for/
[17:08] <p2000> it would b great if i could get the metadata in plaintext as well ... any option ?
[17:08] <p2000> (havent found)
[17:09] <mgz> format-patch is a bit cuter like that, but I never worked out how to send more than one revision with it
[17:09] <mgz> I think there's a send email to file way of doing it
[17:11] <p2000> just now i realize format-patch is probably not what i'm looking for ... intension misleaded me ... i saw the specific feature in gitweb only so far ... there i was able to get a page thatt shows me all commit messages along with the patches in a flat text file ... that's what i'd like to get from bzr but don't know how to get it from git actually ...
[17:11] <p2000> s/intension/intention/
[17:11] <p2000> s/wrong/right/
[17:12] <mgz> s/True/False/ ? :P
[17:13] <mgz> p2000: like `bzr log -p` maybe?
[17:15] <p2000> yes ... looks realy good !
[17:18] <p2000> that was it! ... thank you very much mgz !!!
[17:18] <p2000> :-) same as in git i guess ...
[17:19] <p2000> the 2 dvcs's are on par feature & technology -wise at the moment ... or what do the bzr-wizards say ?
[17:20] <wgz> there's still enough minor things for people to argue about :)
[17:20] <p2000> i mean git vs bzr ... bzr seems to have been quickly developping ... sure
[17:21] <p2000> but there are no more minor differences i guess ... expect design of course
[17:22] <p2000> anyway, thanks !! hanwe
[18:45] <james_w> wgz, how does anyone add you to a team in Launchpad?
[18:46] <james_w> ah, I can avoid the javascript version and do it
[20:18] <briandealwis> jelmer, I just did a dpush to an eclipse repository and seem to have deleted most of the branches other than the branch I was pushing to and to master
[20:18] <briandealwis> (with bzr-git)
[20:40] <wgz> james_w: it's not just that I like making life hard for people...
[20:41] <wgz> thanks for the conscription!
[22:24] <jelmer> briandealwis: what versions of bzr-git/dulwich ?
[22:24] <briandealwis> bzr-git 0.6.2 and dulwich 0.8.0
[22:25] <briandealwis> It unfortunately happened after the support staff had left for home for the weekend.
[22:25] <jelmer> ouch, sorry :-(
[22:25] <jelmer> I've hit this before with tags and subsequently fixed it. I guess it's not in any releases yet
[22:25] <briandealwis> I'm pretty sure I've dpushed to this repository previously though, without this problem!
[22:27] <briandealwis> It would be nice to have a --dry-run flag :)
[22:28] <jelmer> briandealwis: that wouldn't really help with a bug
[22:29] <briandealwis> I'm kinda joking.  I had wondered why the push was taking so long.
[22:30] <briandealwis> And we can't just recover by providing another packed-refs: unfortunately the push seems to have triggered a gc on the other side.
[22:31] <briandealwis> I had meant to ask you what happens with dpush and tags, and if there's a way to not push tags
[22:44] <jelmer> bradm: there is no way to not push tags