[00:25] <poolie> hi all
[00:40] <thomi> Is there a script somewhere I can use to generate a large bzr branch for testing things?
[00:40] <poolie> yes, there is
[00:40] <poolie> um
[00:41] <thomi> I guess I could also just branch a big project from launchpad....
[00:41] <poolie> that could be good
[00:41] <poolie> there's a plugin that makes arbitrary-sized history
[00:42] <poolie> it depends what you want to do with it
[00:42] <poolie> i can't recall the name at the moment though
[00:43] <thomi> ok, no worries.
[00:43] <thomi> I guess Bazaar itself, or launchpad would be reasonable candidates
[00:49] <poolie> yep, depending on what size you want
[00:49] <poolie> and linaro-gcc and emacs are bigger again
[07:23] <poolie> hi jam
[08:08] <jam> morning all
[08:08] <jam> hi poolie
[08:15] <poolie> hi there
[08:23] <Riddell> morning
[08:27] <poolie> hi riddell, how are you?
[08:45] <Riddell> funny, I feel hungover but I didn't drink any alcohol yesterday
[08:45] <Riddell> rather I did a 6 hour paddle and a 2 hour climb.  I think I must have got dehydrated without noticing
[08:46] <AuroraBorealis> so
[08:46] <Riddell> anyway, I'm going to see if I can finish off this KDE Bazaar integration today
[08:46] <AuroraBorealis> is anyone here familiar on how...bzr packages its installers?
[08:47] <AuroraBorealis> since its the only program i know that is written in python / pyqt
[08:47] <Riddell> AuroraBorealis: it entirely depends on the platform
[08:47] <Riddell> for ubuntu it's a normal .deb package using the normal debhelper python scripts
[08:47] <AuroraBorealis> well mostly mac / windows
[08:47] <AuroraBorealis> since linux is easy
[08:48] <AuroraBorealis> like windows has a exe that runs python
[08:48] <AuroraBorealis> and mac os x has an installer but no nice app bundle
[08:48] <Riddell> I believe there's some well known process that turns python into a .exe on windows but I don't know any more than that
[08:48] <poolie> hi AuroraBorealis, jam has worked on the windows installers at least
[08:48] <poolie> there's a tool called py2exe
[08:49] <AuroraBorealis> and im assuming that works with pyqt
[08:49] <jam> AuroraBorealis: we use py2exe to create a .exe program
[08:50] <jam> Then innosetup (IIRC) to build the setup.exe
[08:50] <AuroraBorealis> which includes python and pyqt in it?
[08:55] <gour> if the bug has low priority @LP it means one cannot expect anything within next 6 months?
[08:57] <jam> AuroraBorealis: py2exe bundles up a python script with pythonXY.dll
[08:57] <jam> and creates a library.zip containing the python code
[08:57] <jam> and pulls in needed libraries, etc.
[08:57] <AuroraBorealis> ah, thanks for the info =)
[08:58] <poolie> gour: some low priority bugs are fixed but generally only when the developer has some other reason to fix them
[08:58] <poolie> eg it's on the way to fixing something else or it personally annoys them
[08:58] <gour> poolie: i'm thinking about bug #240067
[08:59] <poolie> heh
[08:59] <gour> we're on the point to decide between bzr/hg & LP/bitbucket
[08:59] <poolie> well, apparently geoff is working on it
[08:59] <gour> just asked about it at SO
[09:00] <AuroraBorealis> wikis dont seem like a trivial thing to implement
[09:00] <AuroraBorealis> but i feel they are really needed cause they are super useful
[09:01] <gour> poolie: i cannot understand that LP team is dragging this for more than 3 years...SF,Google, Github,BB...everyone has it
[09:01] <gour> people were even blogging about moving from bzr/Lp due to it (http://www.artima.com/weblogs/viewpost.jsp?thread=242653)
[09:01] <poolie> you'll probably get more traction in #launchpad
[09:02] <AuroraBorealis> its just that there are no other websites for bzr other then launchpad
[09:02] <AuroraBorealis> there was one that was starting up but its not even functional
[09:02] <gour> i must admit i never got much response there
[09:02]  * gour has asked question about it - http://stackoverflow.com/questions/7063145/bzr-launchpad-vs-hg-mercurial
[09:02] <poolie> so the big thing that launchpad has been working on for the last year or so is speed
[09:03] <poolie> both application speed and development speed
[09:03] <poolie> i mention this because it's actually first on the list of that blog post from 2008
[09:03] <gour> poolie: it's a pity that LP & bzr are losing 'customers' because of that
[09:04] <gour> iow, hard work of bzr devs is wasted due to missing features in LP
[09:05] <gour> actually, not wasted, but you get the point...
[09:05] <AuroraBorealis> it doesn't seem THAT hard to just use markdown files like github does. of course i'm not the one working on it so yeah =P
[09:07] <poolie> gour, so, xaav was making fairly fast progress on some of these bugs
[09:07] <gour> AuroraBorealis: problem is that LP people have other priorities...it seems that they even didn't get spec...although it seems it's not rocket science considering so many hostings have it
[09:07] <poolie> perhaps he will get this one up
[09:07] <AuroraBorealis> a wiki would be very nice to have.
[09:07] <AuroraBorealis> integrated.
[09:07] <AuroraBorealis> at least the start of one :3
[09:08] <gour> AuroraBorealis: the bug is over 3yrs, old :-(
[09:10] <poolie> so, basically, you're saying
[09:10] <ccxCZ> I plan to just use ikiwiki/sphinx/whatever can generate static html for a wiki
[09:10] <poolie> launchpad ought to do this next, as an official feature, because it's the single most important thing?
[09:10] <poolie> i'd like to at least show branch content formatted as rest/markdown
[09:10] <poolie> even if it's not editable in the first cut
[09:10]  * gour nods
[09:10] <AuroraBorealis> that would be acceptable ^
[09:11] <AuroraBorealis> thats kinda what github does anyway
[09:11] <AuroraBorealis> and is easily editable through the source anyway
[09:11] <gour> it would be cool to have e.g. sphinx reST docs rendered
[09:11] <gour> ..but having nothing as 'project page' really sucks
[09:11] <ccxCZ> gour: I think there's an independent service for that already
[09:11] <AuroraBorealis> id offer to help but i doubt my skillz are that good =P
[09:12] <gour> ccxCZ: there is, but its not the same
[09:12] <ccxCZ> yeah
[09:12] <gour> there is readthedocs.org
[09:13] <ccxCZ> tbh I decided to go fully decentralized and build simple stack of project management tools anyone can install (unlike launchpad)
[09:14] <AuroraBorealis> looks cool, but i still think bazaar should have something integrated to the website itself
[09:15] <ccxCZ> roundup, buildbot and probably sphinx will be the building blocks for that
[09:28] <jam> poolie, jelmer, vila: anyone up for a teddy-bear chat about the get_parent_map stuff I'm working on? I think I've got it, but it would be nice to work through it with someone
[09:29] <poolie> hi jam
[09:29] <poolie> i might be, but not tonight
[09:30] <poolie> hm, not right now anyhow
[09:30] <jam> np
[09:30] <jam> side point, we were talking about backporting bug #609187 to 2.1 for Lucid, right?
[09:39] <jam> wow, backporting to bzr-2.1 means NEWS instead of doc/en/release-notes...
[09:44] <jelmer> jam: sure
[09:44] <poolie> jam oh if you get the chance could you read https://code.launchpad.net/~spiv/bzr/faster-stacked-tree-build/+merge/70381
[09:45] <poolie> jam, yes, ideally for 2.1/lucid, but i wouldn't pay an unlimited price if there's something that makes it hard
[09:45] <poolie> beyond the location of the news file ;)
[09:45] <poolie> blah
[09:45] <poolie> poolie: jam oh if you get the chance could you read https://code.launchpad.net/~spiv/bzr/faster-stacked-tree-build/+merge/70381
[09:45] <jam> pidgin just died... very weird
[09:45] <poolie> 19:45
[09:45] <poolie> poolie: jam, yes, ideally for 2.1/lucid, but i wouldn't pay an unlimited price if there's something that makes it hard
[09:45] <jam> poolie: I feel like faster-stacked-tree just needs someone to become its new mascot/driver
[09:46] <jam> I've sort of avoided poking at it, but if you want I can move it to the DOTO queue
[09:46] <jam> TODO
[09:46] <jam> (Do Too?)
[09:46] <poolie> distcc has a 'DOTO' protocol unit for a .o file :)
[09:47] <poolie> well, if you don't mind
[09:47] <poolie> spiv is still around a bit
[09:47] <poolie> i'd better go
[09:48] <jam> k
[09:49] <gour1> poolie: thank you for the hint @SO...wasn't aware of it
[10:00] <gour> poolie: i've added --append_revisions_only to my 'upstream' repo, added a 'feature' into upstream and added a feature to 'feat1' branch. tried to merge upstream into feature and then pushed backe feat1 to upstream and got 'bzr: ERROR: Operation denied because it would change the main history, which is not permitted by the append_revisions_only setting on branch 'upstream'...so it looks as that setting for
[10:00] <gour> 'init' prevents one to do the 'mistake' and forces one to use 'correct' (for bzr) workflow?
[11:51] <lelit> hi, is there any specific reason why the cmd_merge builtin does not honor a --verbose flag (even if "bzr merge -h" says so?) like cmd_pull for example?
[12:09] <lifeless> jam: hi; did you see the loggerhead issue when we rolled it out ? with dictionary size changed during iteration
[13:45] <jam> lifeless: sorry I missed your ping. I did not see the loggerhead issue. do you have a link? I don't think I even saw a loggerhead bug report.
[13:45] <jam> ugh, looks like in all the revamping, I lost my subscription to loggerhead bugs...
[13:46] <jam> why do I lose subscriptions I want, and still get email about every lp-oops-tool that I don't care about... :(
[13:47] <jam> lifeless: https://bugs.launchpad.net/loggerhead/+bug/826136 this looks more like a launchpad codebase issue.
[13:47] <jam> At least https://lp-oops.canonical.com/oops.py/?oopsid=2051CBA110 does
[13:47] <jam> it appears to be failing in "lp.codehosting" not in bzrlib or loggerhead code.
[13:54] <jam> jelmer, Riddell, poolie, vila: I just updated http://wiki.bazaar.canonical.com/PatchPilot for the next few weeks. I just copied the rotation for the last few weeks. Double check if anything won't actually work. I tried to look at the calendar
[13:54] <jam> but that might not be complete.
[13:54] <jelmer> jam:thanks
[13:54] <jelmer> reminds me, I should send a report for last week
[13:56] <jam> jelmer: also, we wanted to pair on the quilt stuff. I wasn't sure, but it looks like I'll be available whenever.
[13:56] <jam> I think I'd like to chat with Martin tomorrow after the standup (if he's available), and you said Wed was no good.
[13:56] <jam> But I could do later on Tues, or Thurs
[13:57] <jelmer> jam: Thursday?
[13:58] <jam> jelmer: works-for-me
[13:59] <Riddell> oh aye
[14:00] <jam> hey Riddell, how was your conference?
[14:01] <Riddell> nice to catch up with desktopy people
[14:02] <Riddell> I got some ideas and requests for bzr
[15:05] <jelmer> hah, switching between colocated branches works
[15:50] <jml> mgz: hey, had a chance to look at the assertThat bug yet?
[15:59] <mgz> hey jml.
[16:00] <mgz> I have a branch, need to make a few choices about output changes
[16:00] <mgz> I think the simplest option is adding a Python 3 style repr for Python 2,
[16:01] <mgz> for anywhere %s is currently used in matcher stringifying and describe methods
[16:02] <mgz> the regexp case is a bit funky, I wonder if we could make up r"" style strings
[16:03] <mgz> I'll put up some suggestions based on your branch shortly.
[16:03] <jml> mgz: cool. thanks.
[16:04] <jml> mgz: The main thing with regexes is to avoid the double slashes.
[16:04] <jml> mgz: not sure there's any other special handling worth doing for them.
[16:05] <jml> mgz: I wonder if we could somehow add some more automated tests guaranteeing that matchers have proper unicode behaviour.
[16:05] <mgz> I've added a few tests for Mismatch describe behaviours
[16:09] <jml> mgz: that's good. thanks.
[17:33] <DerNalia> how do I fix this: bzr: ERROR: Couldn't import bzrlib and dependencies.
[17:33] <DerNalia> I just tried checking out a repo
[17:33] <DerNalia> told me to Please check the directory containing bzrlib is on your PYTHONPATH.
[18:27] <jhtran> if i'm in trunk, how do i 'checkout' an old revision based on date?
[18:27] <jhtran> such as using git, i just do git checkout <date>
[21:00] <kkrev> how do you undo 'bzr resolve somefile' so you can redo the merge for just that file?
[21:04] <kkrev> I guess 'remerge' works
[21:04] <kkrev> even after you marked resolved
[22:10] <wilx> Hi.
[22:11] <wilx> Does anybody have experience with using Bazaar for disconnected development against Subversion?
[22:12] <wilx> It seems that once I make revisions graph non-linear by a merge, I cannot push to Subversion anymore.
[22:12] <jelmer> hi wilx
[22:12] <jelmer> wilx, you have to set "append_revisions_only = False" to push changes that change the branch mainline
[22:12] <lifeless> or use dpush
[22:13] <jelmer> lifeless, that will still try to change the mainline, it just won't push extra metadata
[22:14] <lifeless> oh
[22:14] <lifeless> I thought it rebased locally as needed
[22:16] <jelmer> it will discard the local bzr revisions and replace them with revisions fetched from svn after pushing
[22:17] <jelmer> (effectively rewriting the local revisions that were pushed to no longer include data that can't be represented in svn)
[22:18] <jelmer> 'evening Riddell
[22:19] <Riddell> hi jelmer, I'm afraid I'm not too well and spent today in bed, so if I don't make it to the meeting in the morning it means I'm not better
[22:21] <jelmer> Riddell, Sorry to hear that
[22:21] <jelmer> Riddell: I'll mention it in case when you're  not around tomorrow
[22:22] <jelmer> Riddell: get well soon
[22:25] <mwhudson> jelmer: hi
[22:26] <jelmer> mwhudson, hello
[22:26] <mwhudson> jelmer: i was wondering why we still have the ~launchpad-pqm/bzr-svn/devel branches
[22:26] <jelmer> mwhudson: survived the flight back to .nz?
[22:26] <mwhudson> jelmer: rather than referring to particular revisions of lp:bzr-svn
[22:27] <mwhudson> (& so on for bzr-git etc)
[22:27] <mwhudson> jelmer: apparently!
[22:27] <jelmer> mwhudson: Why not go one step further.. perhaps we could just use by-value nested trees for the various sourcecode branches
[22:27] <jelmer> mwhudson, lp-pqm's bzr-git still has a patch to allow deregistration
[22:28] <mwhudson> jelmer: oh right, that hack
[22:29] <jelmer> mwhudson: I guess we could get that landed upstream if we'd really have to
[22:30] <mwhudson> jelmer: well carrying changes in general is a good reason for a separate branch
[22:30] <mwhudson> i'd forgotten we were doing that
[22:30] <mwhudson> i also forget a bit _why_ we do that
[22:31] <mwhudson> jelmer: do by-value nested trees actually work yet?
[22:31] <jelmer> mwhudson: by-value nested trees have worked well for years :)
[22:32] <mwhudson> oh right, i think i'd gotten the kinds of tree messed up
[22:32] <jelmer> by-reference nested trees are not yet supported
[23:07] <dOxxx> howdy
[23:08] <dOxxx> jelmer: is there going to be a stable release of bzr-svn for bzr 2.4.0? The last official release was August last year.
[23:09] <lifeless> jelmer: ah! it was your patch - 826136 :)
[23:10] <lifeless> jam: thanks for the analysis, thats excellent.
[23:20] <hazmat> is it possible with the bzr plugin api to alter existing commands, or is it only  for the addition of new commands? i'd prefer to avoid monkey patching the existing commands, so i'm looking for a replace option
[23:22] <hazmat> hmm.. found this http://doc.bazaar.canonical.com/plugins/en/plugin-development.html#extending-an-existing-command which is effectively empty
[23:22] <hazmat> ah.. ic the decorate param looks like it does the trick
[23:35] <jelmer> dOxxx, hi
[23:36] <jelmer> dOxxx, There is another release planned