[00:30] is there a way to pipe bzr log -p output through a pager that can do colorized diff? i really really like that feature from git. can less do it? [00:30] hm, not according to 'less' man page [00:31] * mgedmin raises hand [00:31] colordiff [00:31] here's my ~/bin/bzrdiff: http://pastie.org/3157485 [00:31] colorization + paging, a la git diff [00:31] hm, nope [00:31] oh, wait, -p? don't have that... [00:32] oh, wait, bzr log? [00:32] * mgedmin should stop answering questions on IRC when his brain has shut down because of excessive tiredness [00:32] i *think* git log shows colorized diffs [00:32] well, I have a ~/bin/bzrlog too: http://pastie.org/3157495 [00:33] spc is apt-get install supercat [00:33] and ~/.spcrc/spcrc-bzr_log is ... not in my home directory [00:33] ho hum [00:34] my old hard disk is in a drawer and I'm too tired to go hunt for it [00:34] also, supercat is a horrible horrible tool [00:34] on the plus side it does what I want [00:34] after I spent a few hours crafting a syntax file for it [00:35] ah yep, 'git log -p' does similar to 'bzr log -p' except ... prettier [00:35] it's as if git was written by loving people who like command-line interfaces [00:35] more colorful, perhaps, is a neutral way to say it [00:35] achiang: you can use 'less -R' preserves coloring (if you can get the command to emit it how you want in the first place) [00:37] spiv: yeah, i guess the problem is, bzr won't do color control codes (a design decision, i suppose) [00:38] Well, it can do diffs with colour [00:38] * achiang discovers bzr cdiff [00:39] So it's not a fundamental limitation, or a design decision against colouring. [00:39] Yeah, I frequently use 'bzr cdiff | less -R' [00:40] but how to make bzr log do it, so i don't have to type in every revspec? [00:40] There's *possibly* a config option. If not, there should be :) [00:40] (I think there might be an option to set the default diff colouriser/formatter? I forget.) [00:42] IIRC plugins can register alternative diff formatters, and perhaps at worst you could have a plugin tweak the default so that log uses coloured diffs. (as a workaround for getting that feature into bzr properly) [00:46] thanks for the pointers. it's a little too daunting to figure out easily, so i give up [00:46] maybe i'll try again some day [00:47] from the bzr log man page: "GUI tools and IDEs are often better at exploring history than command line tools" [00:47] sadface [00:49] Well, it's difficult to draw clear diagrams of revision graphs in ASCII art :) [00:49] Patches welcome though! [00:49] diagrams are overrated, /methinks [00:49] incidentally, is there a bzr equivalent of gitk --all? [00:49] i.e. a diagram of merges that shows more than one branch at once? [00:50] mgedmin: 'bzr qlog BRANCH_A BRANCH_B BRANCH_C ...' [00:50] thanks, good to know! [00:50] mgedmin: you might even be able to pass just a repo to it these days, or use a magic option, but it can certainly do it. [00:51] spiv: personal opinion, i don't find the rev graphs to be useful in either bzr or git. the squashed "linear" history is the view i use the most in any scm (and yes, i do understand it's not truly linear, but it's often close enough) [00:52] spiv: but i am truly appreciative of your help. #bzr is always very friendly, thanks. :) [00:52] oh, and normally, i would use bzr qlog on my local machine, but i'm looking at something remotely [00:58] That's why X is network-transparent :> [00:59] can i get your speshul t00bs that eliminate latency too? :P [00:59] I've got a 386 you can borrow. That makes network latency a non-issue... [01:01] actually, i don't understand this line of teasing. are you truly suggesting that X forwarding from a remote openstack shell is enjoyable to use? [01:02] You never know. I once ran Netscape 4 across the Internet. Mind, it was excruciating... but then, it was NS4, so it was hard to tell the difference. [01:03] bzr qlog probably works, no? [01:04] It might, but I would expect it to be... uh... what's the laughing-hysterical-like opposite of 'optimized'? [01:11] compared to remote x? :) [01:11] Pfui. FTP is never as cool as GTP :p [01:12] Would certainly be more responsive interactively. But I wouldn't be overly surprised if some of the info lookup ended up being pathologically bad. [01:17] * mgedmin hates when vim's vcscommand.vim decides to use bzr for :VCSAnnotate in Subversion checkouts -- can you spell insaaaaaane slownesss? [01:19] Are ee em oh tee ee space .... wait, that's your joke. [01:55] fullermd: actually, I think qlog against a remote URL has had a least a little bit of optimisation [01:55] fullermd: so it's probably just rather slow, rather than truly horrific. === hloeung_ is now known as hloeung [09:53] No branch/tags file in BzrBranch7('file:///srv/foo/bar/'). This branch was probably created by bzr 0.15pre. Create an empty file to silence this message. [09:53] srsly? [09:53] Create an empty file called what? where? [09:58] elmo: you can get that if you `rm .bzr/branch/tags` [09:58] provided that's all that happened, touching it and pulling will get it happy again [09:59] but if other files have gone missing as well, you may need to branch again === AuroraBorealis is now known as aurora|sc2 === Quintasan_ is now known as Quintasan [14:10] what's the bzr command to get a rundown of files that would be changed in launchpad if I committed/pushed right now? [14:10] 'status' seems to show ignored files that have been modified. === tchan1 is now known as tchan === james_w` is now known as james_w === spi_ is now known as siip [17:55] is it possible to branch only a specified subdirectory (e.g. the src subdirectory of lp:connectorj)? [17:56] no, it isn't [17:57] thanks maxb (arg this is going to be a %#$!) [18:01] you really cant? thats sort of a useful thing to do in svn [18:34] Doh! [18:34] Note to self: bzr merge -c-133 . does _not_ mean "reverse-merge changeset 133" [18:34] * LeoNerd reverts and tries again ;) [18:35] -r133..132 \o/ === yofel_ is now known as yofel [22:50] hey, i'm getting a key denied (publickey) when trying to bzr branch from an lp: address. does bzr care that i'm running it as root? bzr version 2.3.4 fwiw [22:52] kgoetz: hi [22:52] kgoetz: bzr doesn't care you're running as root - if you're branching as root, do you have ssh keys set up, and your launchpad login set? [22:52] jelmer: hey (: [22:53] I would be tempted to check the $HOME setting [22:53] if it points to /root, most likely you don't have your usual ssh keys/global bzr config [22:53] no i don't. i assumed to branch from publicly available code i wouldn't need to do that ritual :/ [22:54] sounds like bzr thinks it knows your launchpad login, but you don't have access to the keys... [22:54] since this is a server, is there a way to work around it? i'd rather not go adding ssh keys from all my hosts into my LP account [22:54] IIRC lp: use bzr+ssh for performance reasons when they know what username to ssh to [22:55] I'm sure jelmer will correct me if I"m wrong [22:55] mgedmin: you're right [22:55] kgoetz: check "bzr config" for the launchpad_username [22:55] you can unset it if you don't want bzr+ssh:// to be used [22:56] hm, bzr launchpad-login lets me set my launchpad username, but doesn't provide a way to reset it (and there's no bzr launchpad-logout) [22:56] thanks :) [22:56] so it's bzr config --remove launchpad_username then? [22:57] yeah, that should work [22:57] you are both correct. my launchpad_username was set (for reasons unknown), but ssh keys not set up [22:58] thanks! [22:58] incidentally, is there a plan to move ~/.bzr.log into ~/bazaar? i keep missing them up when tab completing [22:59] kgoetz: not that I'm aware of [22:59] jelmer: ok, thanks [23:00] kgoetz: I'm not aware of a good reason not to, but also not aware of any plans to do so [23:01] the argument i expect would be made is if you're going to move it you should put things in XDG_* locations [23:01] if ~/.bazaar lived under ~/.config, your tab-completion trouble would be gone as well ;) [23:01] mgedmin: snap (: