[00:34] <mtaylor> lifeless: did you have an example of using bzr-builddeb in your subunit slides for some reason?
[00:35] <mtaylor> lifeless: I remember seeing a usage of bzr bd on someone's slides, and I remember it looking like a better workflow than what I currently use with bzr bd... but I didn't write it down
[00:40] <dutchie> is there a way to have hooks propagated to all the repositories?
[00:45] <mtaylor> dutchie: if you are asking what I think you are, then no
[00:45] <mtaylor> dutchie: but it's possible that you are asking a question to which the answer is yes :)
[00:46] <dutchie> well, am I?
[03:18] <EdWyse_Mobile> What would you recommend as a public html frontend for bazaar in CentOS 5.4?
[03:22] <Peng> EdWyse_Mobile: Loggerhead?
[03:25] <EdWyse_Mobile> Thanks Peng. Loggerhead doesn't lend itself very well to centos as most of the dependencies aren't installed, plus I hate to have another service running. I'd prefer either a python cgi or php script if they exist. Any other possibilities?
[03:36] <Peng> Dunno if they're maintained, but bzrweb and bzr-webserve
[03:36] <Peng> Loggerhead is the main choice, though, and its dependencies aren't a big deal to install.
[03:50] <spiv> I think you could run loggerhead as a CGI
[03:50] <spiv> You just need to find a WSGI thing that runs as a CGI.
[03:54] <igc> hi all
[03:57] <Peng> That would be hilariously slow.
[03:57] <Peng> And inefficient. And probably chew inodes unless you disable the caches.
[04:54] <EdWyse_Mobile> Looks like redmine does what I want. It's ruby...
[05:03] <lifeless> dutchie: Not sure what you are asking
[05:03] <lifeless> mtaylor: no, I didn't. But  can give you a better bd workflow if you need ;)
[06:25] <lifeless> poolie: if you up and want company, I'm breaking fast now
[07:57] <EdWyse_Mobile> FYI, if someone wants a non-loggerhead frontend, I just installed Redmine and it's working beautifully. It's nice and snappy and it doesn't run as it's own service so it could be implemented on shared-hosting situations. It does require ruby though.
[08:00] <Peng> Oh, I forgot about Redmine. Oops.
[08:00] <Peng> You can do Loggerhead with FastCGI/etc. if you want to.
[08:00] <Peng> just fyi
[08:30]  * igc dinner
[08:45] <bialix> heya bzr
[08:46] <poolie> hi all
[08:46] <bialix> hi poolie!
[08:47]  * bialix waves at poolie :-D
[09:11] <igc> hi poolie, bialix
[09:12] <bialix> hi igc :-)
[09:18] <jam> igc: vincent's ready for you to call
[09:20] <igc> jam: it crashed because of not enough bandwidth. Maybe try again?
[09:20] <igc> trying ...
[09:20] <jam> igc: yeah, try again
[09:20] <jam> vincent's video is a bit off
[09:20] <jam> just a sec
[09:21] <jam> igc: still waiting if you can try again
[09:21] <igc> jam: it claims its starting
[09:23] <jam> igc: I don't think its ringing on this end
[09:23] <dutchie> OK, I didn't really explain what I wanted very well last night
[09:23] <dutchie> what I want is to write a hook that automagically ends up being run for everyone who has a branch
[09:24] <jam> igc: also, vila doesn't seem to see you as online
[09:24] <igc> jam: crashed again. I'll reboot and try again
[09:24] <jam> igc: k, see you in a few
[09:32] <bialix> hi jam, vila_
[09:32]  * vila_ waves !
[09:34] <jam> morning bialix
[09:35]  * bialix waves back :-D
[09:57] <igc> vila: hooray!
[11:21] <poolie> igc, can you answer https://answers.launchpad.net/bzr/+question/98910
[12:14] <jam> lifeless: news_merge_files = NEWS
[12:44] <lifeless> igc: gnight
[12:44] <igc> lifeless: night
[17:12] <bvk> hi, does bzr-cia plugin work over http_proxy ?
[17:19] <jelmer> bvk: Hi
[17:19] <jelmer> bvk: It uses the Python xmlrpclib module, I'm not sure if that uses http_proxy
[17:29] <poolie> jam, can you tell https://code.launchpad.net/~ed-marsfire/bzr/giveback/+merge/18180 how to run tests easily on windows?
[18:58] <rocky> i feel like i'm missing something stupid, how do i determine the active revision of the "working tree" i'm in?
[18:58] <rocky> revision/versino
[18:58] <rocky> *version
[19:05] <keithhub> bzr revno
[19:07] <fullermd> 17
[20:17] <phoenixz> How can I make a checkout without generating the WT?
[20:18] <phoenixz> so that basically, I end up with only the .bzr directory?
[20:18] <mzz> phoenixz: branching into a no-tree repository does what you want, I think
[20:18] <mzz> assuming you don't mean "checkout" in the way bzr uses the term
[20:18] <phoenixz> mzz: oh crap, BRANCH! :)
[20:18] <phoenixz> mzz: branch, my bad..
[20:19] <mzz> although hmm, I wonder what "bzr checkout" inside a no-tree repo does
[20:19]  * mzz checks
[20:19] <phoenixz> mzz: I'd figure, it would generate the WT, no?
[20:20] <mzz> looks like it does
[20:20] <mzz> you can just branch and then bind
[20:21] <mzz> or if you don't want the repo separate you can branch or checkout and then use bzr reconfigure to nuke the working tree, afaik
[20:21] <phoenixz> mzz: nah, the --no-tree option was what I was looking for
[20:21] <mzz> and bzr branch takes a --no-tree argument, actually. I didn't know that.
[20:21] <phoenixz> mzz: no? I thought you just told me to use the --no-tree option to branch... :P
[20:22] <mzz> phoenixz: no, I meant with init-repo
[20:22] <mzz> which is a rather roundabout way to do what you want
[20:22] <phoenixz> mzz: ah... well, branch works perfect for me! :)
[23:35] <SpamapS> so. I think I did something really dumb.. its hard to explain, but it all happened on launchpad so hopefully we can figure it out...
[23:35] <SpamapS> I merged in some changes from one branch..
[23:35] <SpamapS> then decided I didn't want them, so to undo it, I did
[23:35] <SpamapS> bzr revert -r $prev_revision
[23:35] <SpamapS> then bzr commit
[23:36] <SpamapS> now.. when I go to merge those changes from before.. they won't merge, I presume because they've already been merged in...
[23:36] <SpamapS> so.. can I somehow tell merge to ignore those two cancelling revision?
[23:37] <SpamapS> s
[23:37] <NfNitLoop> bzr revert --forget-merges
[23:37] <NfNitLoop> Oooh.
[23:37] <NfNitLoop> you did a commit without doing --forget-merges.
[23:37] <NfNitLoop> You've basically "blocked" that merge.
[23:38] <SpamapS> yeah thats what I'm afraid of
[23:38] <NfNitLoop> What you can do...
[23:38] <NfNitLoop> is apply those changes again as a new commit.
[23:38] <NfNitLoop> bzr diff -r x..y | patch -p0 (I think?)
[23:38] <NfNitLoop> bzr commit
[23:39] <NfNitLoop> though...
[23:39] <SpamapS> so the changes I want are in lp:~coryb/gearmand/time-order ... and I want to apply them to their parent branch (lp:gearmand) .. what are my x/y values then?
[23:39] <NfNitLoop> if this branch is local (i.e.: not public) and that commit wasn't long ago, you may just want to fix history.
[23:40] <SpamapS> no it gets worse.. because I did this in my branch, which was merged to trunk.. so now trunk blocks the merge too.. :-P
[23:40] <NfNitLoop> doh.  Yeah, fun!  :p
[23:40] <SpamapS> I'm fine w/ creating a patch to do it..
[23:40] <SpamapS> just don't know what the x/y should be
[23:40] <NfNitLoop> make a branch of lp:gearmand.
[23:41] <NfNitLoop> x = revision before your changes.  y=last revision with your changes.  assuming they're all sequential.
[23:41] <Kilroo> I'm confused as all get out by this conversation.
[23:42] <SpamapS> I just got a suggestion to use  diff -c # for the changes I want...
[23:42] <SpamapS> Kilroo: me too :(
[23:43] <Kilroo> Compared to this, my trying to figure out what a stacked bound branch would do and whether a lightweight checkout of a bound --no-trees branch is what I really want at work seems...well, ok, the stacked bound branch still seems pretty odd.
[23:43] <maxb> Rather than going via a patch, I would have hoped that bzr merge -r x..y would do the job
[23:44] <NfNitLoop> SpamapS: that only works if all of your changes are in one commit.
[23:44] <SpamapS> hmmm lets see..
[23:44] <NfNitLoop> -c N is short for -r (n-1)..n
[23:45] <NfNitLoop> Oh, hmm.
[23:45] <NfNitLoop> merge can supposedly take an -r ...
[23:45]  * maxb looks at those branches in bzr qlog and gets very confused
[23:45] <NfNitLoop> but you can't cherrypick so I'm not sure what it does.  I guess it just does the equivalent of diff|patch
[23:46] <maxb> SpamapS: What are the commit messages of what you actually want to merge here?
[23:48] <maxb> oh... I think I see
[23:48] <SpamapS> maxb: I want the entire contents of lp:~coryb/gearmand/time-order  merged into lp:gearmand ... sans rev103 and rev104 from lp:~clint-fewbar/gearmand/tokyo
[23:49] <SpamapS> so really, what I want are the changes that lp:~coryb/gearmand/time-order did, up until the last merge point w/ trunk ...
[23:49] <SpamapS> *that* seems like it would be easy to get
[23:49] <NfNitLoop> "the entire contents" is a bit vague for a merge.  contents after revision N is more helpful.
[23:50] <NfNitLoop> or revisons X..Y
[23:50] <NfNitLoop> but if you've done some merging into that branch while you were developing you may have to do your merge in a few chunks now.
[23:50] <NfNitLoop> merge in x1..y1, then x2..y2
[23:50] <maxb> I'm thinking something like:  bzr merge -r -4..-1 ../corybtimeorder
[23:50] <NfNitLoop> around all of the holes that you don't want.
[23:50] <maxb> let me test this
[23:51] <SpamapS> I just want to diff that branch with trunk up to revision 316 ..
[23:51]  * SpamapS is trying furiously to un-cross his eyes
[23:51] <Kilroo> What if you branch from before the shenanigans happened and then merge in revision ranges around the edges of the flargle?
[23:52] <Kilroo> Or is that what you're talking about already?
[23:52] <SpamapS> oh that sounds like one approach.. bzr branch revno:316:lp:gearmand ?
[23:53] <SpamapS> then merge into that... and use the diff as a patch?
[23:53] <Kilroo> Well, I don't think that was actually what I meant, but it sounds promising.
[23:54] <SpamapS> bzr branch -r before:317 lp:gearmand gearmand-trunk-b4tc
[23:54] <Kilroo> I was thinking more along the lines of branch from before the screwup, then merge in the changes you actually wanted in the right order, skipping the stuff that you wish you hadn't done.
[23:56] <Kilroo> And make that the new main branch. But it occurs to me that I have absolutely no clue how that would affect the trunk that already merged what you did before.
[23:56] <Kilroo> I'm still very new to all this stuff.
[23:57] <maxb> SpamapS: I think you want: bzr branch lp:gearmand; bzr branch lp:~coryb/gearmand/time-order; cd gearmand; bzr merge -c -3 ../time-order; (resolve conflicts, commit); bzr merge -c -1 ../time-order; (resolve conflicts, commit)
[23:59] <SpamapS> maxb: will try that..