[08:03] <mgz> morning
[08:05] <jelmer> moooin
[08:13] <gour> morning all
[08:35] <gour> any fish shell user here beign aware there is bzr completion for fish?
[08:36] <gour> i used fish in the past, but there was none...then we switched to zsh (as well as to hg, fossil), but now we would like to have it for bzr
[08:37] <mgz> could you adapt the bash-completion builtin plugin? and just alter the output script it generates.
[08:39] <gour> i'll try...
[09:56] <matsubara> vila, hi :-)
[09:56] <vila> :)
[09:57] <matsubara> vila, so, should I add something to bazaar.conf to make it agree with authentication.conf?
[09:57] <vila> So, the lp plugin checks that the launchpad_username is the same than the one selected in authentication.conf
[09:57] <vila> and the launchpad_username can (so far) only be defined in bazaar.conf
[09:58] <vila> matsubara: are you issuing the bzr commands or do you rely on the jenkins plugin for that /
[09:58] <vila> ?
[09:59] <matsubara> the idea is that jenkins will run the command for me, but AFAICT it's just running bzr branch lp:<foo-branch>
[09:59] <matsubara> and I'm issuing the commands with jenkins user to test things
[10:00] <vila> matsubara: yes. I was thinking about adding -Olaunchpad_username to override bazaar.conf when needed, but if you can't provide that, that's a no go
[10:02] <matsubara> vila, I can provide that to jenkins, yes
[10:02] <matsubara> vila, but I'm getting an bzr: ERROR: no such option: -O
[10:02] <jelmer> matsubara: older versions of bzr don't support -O unfortunately
[10:02] <vila> ghaa, upgrade bzr ?
[10:02] <matsubara> is it bzr branch -Omazabot lp:maza-server?
[10:03] <vila> bzr branch -Olaunchpad_username=mazabot lp:maza-server
[10:03] <matsubara> I have Bazaar (bzr) 2.4.1 installed there
[10:03] <vila> yup, too old
[10:03] <mgz> any reason not to straight up run bzr as a seperate user?
[10:03] <vila> jenkins is the user I think and it needs the related access rights to write on disk
[10:04] <matsubara> mgz, it's going to be run by the jenkins user
[10:04] <matsubara> you're right vila :-)
[10:04] <mgz> well, set BZR_HOME to a different dir with a whole different conf then?
[10:04] <vila> 2.4.1 is weird though, 2.4.2 being out, what are you using to get only 2.4.1 (just curious, won't help here)
[10:05] <vila> yeah, heavy workaround but would work and seems to be the only way to address the issue in the short term
[10:05] <matsubara> vila, don't know. it was there when I got access to the machine
[10:07] <matsubara> and I don't manage the machine, so I'd rather not upgrade the bzr client until I can talk to retoaded who manages the box
[10:07] <matsubara> I'll try the BZR_HOME trick then
[10:16] <vila> matsubara: alternatively, since the issue is in the lp plugin, you may be able to avoid the issue by using 'bzr+ssh://' urls instead of 'lp:' ones
[10:17] <matsubara> vila, thanks for the tip. mgz's suggestion kinda worked. Now I'm having a public key denied problem and I think bzr is send the wrong key for authentcation
[10:23] <vila> you used BZR_HOMR not HOME right ?
[10:26] <matsubara> vila, yep
[10:26] <matsubara> https://pastebin.canonical.com/67878/
[10:26] <matsubara> vila, the output ^
[10:27] <matsubara> is there a way to also define which ssh key to use?
[10:27] <vila> matsubara: ssh -vv bazaar.launchpad.net
[10:28] <mgz> also `BZR_HOME=maza-bazaar-conf/ bzr launchpad-login`
[10:29] <vila> .ssh/config is still obeyed, not sure how to make ssh try several keys
[10:29] <matsubara> vila, output of ssh -vv: https://pastebin.canonical.com/67879/, it's trying to use the default id_rsa, while I want it to use the maza_lp key
[10:29] <vila> yeah
[10:29] <matsubara> mgz, I ran that command already
[10:29] <matsubara> and it returns mazabot as the user, which is the correct one
[10:30] <matsubara> vila, .ssh/config: https://pastebin.canonical.com/67880/
[10:30] <vila> ssh-add maybe
[10:31] <matsubara> I just changed s/.launchpad/bazaar.launchpad.net/ in .ssh/config and it seems to be doing something. waiting on the output
[10:32] <matsubara> aha! branched!
[10:32] <vila> yeah, the leading '.' is an authentication.conf trick
[10:32] <vila> *.launchpad.net should work in .shh/config though
[10:33] <matsubara> but then bzr client will default to that key for everything else as well, won't it?
[10:33] <vila> matsubara: and from ssh_config(1), you can have multiple IdentifiyFile directives so all keys are tried
[10:33] <matsubara> hmm
[10:34] <vila> yeah, you need an .ssh/config thta works for all your users
[10:34] <vila> well, 'need' is a bit strong, but you get the idea
[10:36] <vila> and you probably should remove the 'User' directive from .ssh/config and let bzr provide it
[10:36] <vila> (that is, provide the user when attempting to connect and rely on .ssh/config to try the various keys until the right one succeed)
[10:44] <matsubara> darn, apparently jenkins bzr plugin won't allow me to define a custom BZR_HOME
[10:45] <mgz> jenkins doesn't let you just set an var for the environment of running an external command?
[10:45] <matsubara> I'm trying to find out
[10:47] <ccxCZ> can I make bzr shell not connect to the parent when using it in bound branch?
[10:49] <ccxCZ> ah well I guess I can workaround it with cd /
[10:58] <matsubara> mgz, vila: I'm going for lunch and will figure out how to teach jenkins to use custom env variables. thanks a bunch for your help!
[11:34] <ricardo_> Hola, intento iniciar bazaar para version control system y no me permite configurar etckeeper.conf ¿Que puedo hacer?
[11:54] <jml> does bzr have a helper for printing tales?
[12:02] <hazmat> does bzr support any form of revno expansion in a file?
[12:06] <mgz> hazmat: see the bzr-keywords plugin
[12:06] <hazmat> mgz, thanks
[12:06] <mgz> generally you want to do something else in your build process though
[12:07] <mgz> jml: I don't know what a printed tale would look like I'm afraid
[12:07] <jml> sorry, I meant 'tables'
[12:07] <fullermd> What, you've never read Chaucer?
[12:09] <mgz> hm, no, 'fraid not in that case. python-novaclient uses something called prettytable but I don't think it's all that pretty.
[12:09] <jml> mgz: thanks.
[12:10] <mgz> I don't think I've ever read Chaucer pretty printed...
[12:10] <mgz> just from photocopies
[12:12] <mgz> was involved in a performance in the original language in a pub once though
[12:18] <jml> hmm.
[12:18] <jml> istr bzr had an easy way to get a dump for kcachegrind
[12:19] <mgz> `bzr --lsprof-file something.callgrind command`
[12:19] <jml> thanks.
[12:23] <jml> yay
[12:29] <jml> for those interested, I have been using your help to generate a high score table for lp:launchpad, where the score is lines of code added: http://paste.ubuntu.com/1037100/ is my latest cut.
[12:32] <mgz> what are the prizes you're giving away? :)
[12:35] <jml> mgz: dignity & honor.
[12:36] <mgz> wallyworld certainly deserves something
[12:36] <mgz> wooden spoon?
[12:36] <jml> mgz: perhaps a leather-bound print out of all the lines he's added?
[12:37] <jml> code is at lp:bzr-damage btw. Would appreciate any tips on how to do things better.
[13:07] <leo2007> how to sync branches from upstream?
[13:09] <leo2007> My current repo (emacs upstream) has the following branches http://bpaste.net/show/31321
[13:10] <leo2007> which has not been updated for a while. upstream has created new branches. I want to bring my local branches up-to-date.
[13:49] <jml> jelmer: are you responsible for bzr-stats?
[15:48] <nilg`> hi, I have a question regarding private branch and collaborative work
[15:48] <nilg`> suppose you have the trunk A, I branch A, do some commits to get B
[15:49] <nilg`> during this time A got some commits too, and becomes A'
[15:50] <nilg`> B is on a private branch, and the guys how owns that branch wants me to push his changes to the trunk (now A')
[15:50] <nilg`> so I branch B, locally, rebase it from A' (bzr rebase TRUNK)
[15:51] <nilg`> i push the stuff to the trunk, now A''
[15:51] <nilg`> the guy who owns B, in between has committed some changes, so now it become B'
[15:52] <nilg`> and he also wants to update in branch, so he rebase B's according to A''
[15:52] <nilg`> but now the commits of B' that I've pushed on the trunk are duplicated on B''
[15:52] <nilg`> how to avoid that problem?
[15:52] <nilg`> thx
[15:55] <vila> don't rebase
[15:56] <vila> merge instead
[15:57] <vila> by rebasing you're roughly saying: I don't care about what happened in branch B. When you later try to "merge" again, you do care. So if you want to track B, merge it.
[15:57] <LeoNerd> Rebasing is a form of rewriting history. As with any rewrite, you need to be careful with it
[18:20] <jelmer> jml: whoops, missed that earlier
[18:20] <jelmer> jml: fsvo responsible
[18:20] <jelmer> jml: the code was written by jam and myself
[20:05] <eridu> is there a way to have a bazaar branch that maps to a git "branch" within a git "repo"? I'd like to use bazaar in a typical multi-directory shared-repo layout, but push to github.
[20:05] <eridu> (it wasn't my choice to use github)
[20:08] <ccxCZ> eridu: bzr-git plugin has git-import command for this
[20:08] <eridu> ccxCZ: can I use the branches to push back to github?
[20:08] <ccxCZ> (you need fairly recent bzr-git for git branches to work)
[20:08] <ccxCZ> yup, or any git repo
[20:09] <eridu> woot! thank you!
[20:09] <eridu> (I was really worried I'd have to use git)
[20:23] <Zoohouse> When I try to run bzr lp-propose-merge lp:gwibber --message=Submitting my edits of the docsctrings to the plugin files. If there's no problems with the edits and it's approved, I'll continue editing the docstrings within the GTK files. Nature of edits: converted current docstrings to Python's PEP 257 styles.  I get the following error: bzr: ERROR: extra argument to command lp-propose-merge: my
[20:27] <Zoohouse> scratch the question. I needed quotes around the comment.
[21:48] <evillyEvil> Is it possible to propose a branch for a merge without having to go to Launchpad?
[21:59] <jelmer> evillyEvil: "bzr lp-propose"
[22:04] <evillyEvil> jelmer: So do I still need to push the branch to launcpad first?
[22:06] <evillyEvil> jelmer: Actually, I also got this error bzr: ERROR: No module named lazr.restfulclient.resource
[22:06] <evillyEvil> You may need to install this Python library separately.
[23:04] <jelmer> evillyEvil: yes, you'll have to install lazr-restfulclient
[23:04] <jelmer> evillyEvil: you do indeed need to push the branch first (I think)