[00:01] <awilkins> Blech, don't you just love to hate MS sometimes.
[00:02] <awilkins> I just found a weird file called DCBC2A71-70D8-4DAN-EHR8-E0D61DEA3FDF.ini my user profile
[00:02] <awilkins> It's an archive of 128-byte compressed thingies
[00:03] <awilkins> Turns out it's where WMP8 stores it's "recent files" tripe
[00:03] <awilkins> I wasted 10 minutes figuring out what that was
[00:25] <uws> I'm a little concerned.
[00:25] <Peng> Me too.
[00:25] <uws> If indeed Bazaar is becoming a GNU Project... (see http://permalink.gmane.org/gmane.emacs.devel/91263 for evidence)
[00:26] <uws> ...will I be forced to grow my beard as a happy Bazaar user?
[00:26] <uws> Please clarify. Love, /me.
[00:26] <Peng> Just a moustache will be acceptable.
[00:28] <Peng> You could always switch to Mercurial. They only require lustrous eyebrows.
[00:30] <uws> Okay. For Gnome I'm still using SVN on a daily basis... centralized haircuts perhaps?
[04:29] <bob2> http://www.advogato.org/person/apenwarr/diary/371.html <- git luvin'
[05:17] <Parker-> awww
[05:17] <Parker-> Isn't that cute
[05:45] <jamesh> bob2: is he really saying that it is remarkable that you can store files in less space than a conventional filesystem if you don't want to modify them?
[06:11] <ferringb> jamesh: compression (rw on compressed blocks is a pita), fs overhead, etc...
[06:12] <ferringb> 'course that guy is calling git a fs, which is a bit odd in my eyes ;)
[06:14] <jamesh> ferringb: I realise that.  I am just a bit surprised that he doesn't appear to realise it
[06:15] <jamesh> of course, that might just indicate that the jobs he is working on fit the model of git's object database
[07:40] <speakman> hi ppl!
[08:35] <ubotu> New bug: #148783 in olive ""Unknown error" (DBus) when trying to commit" [Undecided,New] https://launchpad.net/bugs/148783
[09:00] <james_w> http://chistera.yi.org/~adeodato/blog/entries/2008/03/05/one_day_with_git.html
[09:00] <james_w> dato: hi. Are you around?
[09:01] <awilkins> Hmm, VCS are filesystems ; git was even written by the king of filesystem design, Torvalds
[09:04] <awilkins> It would seem that people like the pre-commit feature of git ; what would the chances of puttin it in bzr be?
[09:05] <awilkins> I have to admit, thinking about it, it would be pretty cool.
[09:06] <james_w> awilkins: what do you mean by pre-commit?
[09:07] <awilkins> This git add -p thing.... putting things into a sort of special "to be committed" revision is how I'm thinking about it
[09:07] <awilkins> Like "I'm reviewing all my diffs and this bit is ok, so I'll put it in pre-commit, where it will wait until I'm finished, but now lit no longer distracts me in diff operations because it doesn't show up"
[09:08] <awilkins> A bit like shelve in reverse, I suppose
[09:09] <dato> james_w: hi
[09:10] <james_w> hi dato. How are you?
[09:10] <james_w> awilkins: I think it should be possible. I'm not sure where it would fit, maybe a new command?
[09:11] <dato> james_w: I'm actually quite well, thanks. just suffering a bit with Uni, but I guess that's normal :-)
[09:11] <james_w> dato: yeah.
[09:11] <james_w> dato: thanks for the post, I found it very interesting.
[09:12] <awilkins> james_w: Reading the (horrible) man page, it also appears that you can just "pre-commit" hunks of the file.
[09:12] <james_w> awilkins: add --interactive?
[09:12] <dato> james_w: great
[09:13] <james_w> dato: and thanks for doing the packaging of bzr. I hope you'll keep doing it while you still use bzr, you do a great job.
[09:14] <james_w> dato: and I hope that you we can improve bzr so that you don't feel the need to leave.
[09:14] <awilkins> james_w: I'm not sure, that blog page you posted earlier says git add -p
[09:14] <awilkins> james_w: The --interactive mode is where you can commit just hunks, yes
[09:14] <dato> yeah. I would even continue to do it if nobody else is available, is doesn't need much time.
[09:14] <james_w> dato: cool. Thanks.
[09:15] <james_w> awilkins: yeah, it's quite nice.
[09:15] <dato> igc: around? I was impatient and took a stab at implementing bzr-fast-export.py. I didn't need to use any code from your fastimport plugin, though. also, I dubbed it (though still mostly unpublished) bzr-crude-export.py, as not to conflict with your prospective -export.py. would you like to take a look? I have no idea if I'm making good use of the bzrlib API or not.
[09:16] <james_w> awilkins: I think making it something you explicitly do on top would be ok. I think the "git add file; edit file; git commit" not committing the current state is something that we don't want.
[09:16] <dato> igc: it can import single branches, preserving merges, even -r 0..-1 merges. importing several related branches into a same git repository is next in my TODO.
[09:16] <james_w> awilkins: unless you have staged hunks perhaps.
[09:17] <awilkins> james_w: Yes, we don't want to let the "creeping evil" of git in by trying to adopt it's features
[09:17] <james_w> awilkins: :)
[09:17] <awilkins> THe comparison with git/vi is fair
[09:17] <awilkins> (I don't use git, it's windows non-nativeness puts me off)
[09:18] <awilkins> But people forget that more and more non-techy users are now having to use VCS systems
[09:18] <james_w> yeah, I think it might be. However I wouldn't mandate that all contributors to my project use vi.
[09:18] <james_w> that would be too much, so I'll let you use emacs, but that still excludes people.
[09:19] <awilkins> I have a hard enough time teaching non-programmers to use SVN
[09:19] <awilkins> I think git would be a bridge too far
[09:19] <igc> dato: hi. I'd be happy to take your work so far as a basis for mine
[09:20] <awilkins> dato: Could I see it too? My interest is seeing an example of fast-import format generation in a language I understand better than C.
[09:20] <igc> dato: it would be great to get things round-tripping as best we can between us
[09:20]  * awilkins may have to write mks-fast-export *shudder*
[09:21] <dato> http://chistera.yi.org/~adeodato/tmp/2008-03-05/bzr-crude-export.py is yesterday's version
[09:22] <dato> right now I was converting REVISION_LIST to a dict, and there's a bug with filenames that contain non-ascii, but that's easily fixed (haven't committed yet)
[09:22] <dato> oh, and I was hoping to diagnose some problem I was having with quotes
[09:22] <igc> dato: ok, let me know when you're happy with it and I'll bundle it inside fastimport if that's ok
[09:22] <awilkins> dato: You _know_ you should be uploading a bzr branch to that server, right? :-P
[09:23] <dato> igc: ack
[09:23] <igc> and open up the plugin to multiple committers
[09:23] <awilkins> Or maybe you could both just start a launchpad branch for it
[09:23] <dato> awilkins: well, tbh I was unsure where to put it, and whether to try git *g*
[09:24] <dato> no, seriously, I'll get this figured out soon
[09:25] <dato> igc: is two revision_trees + changes_from + get_file_text the proper way to do it?
[09:26] <dato> igc: answer whenever you can, in some hours or days ;)
[09:27]  * dato leaves for class now.
[09:27] <igc> dato: I hadn't thought about it yet sorry :-)
[09:28] <igc> I'll think about it today, ask abentley, etc.
[09:28] <dato> igc: yeah, np.
[09:28] <dato> oh, that'd be awesome, but feel no pressure.
[09:28] <dato> bbl
[10:31] <lifeless> less ~/bin/drop-caches
[10:31] <lifeless> #!/bin/sh
[10:31] <lifeless> # get written data to disk (not that its guaranteed)
[10:31] <lifeless> sync
[10:31] <lifeless> # (drop the unmodified pages)
[10:31] <lifeless> echo 1 | sudo tee /proc/sys/vm/drop_caches
[10:31] <lifeless> LarstiQ: ^
[10:32] <poolie> lifeless: that tee command is not write
[10:32] <poolie> right
[10:32] <poolie> it does not write :)
[10:32] <poolie> you need um
[10:32] <soren> lifeless: Why 1?
[10:33] <poolie> 1 for on
[10:33] <soren> No.
[10:33] <LarstiQ> 1 for pagecache
[10:33] <soren> 1 to free pagecache.
[10:34] <soren> 3 to free both pagecache, dentries and inodes (i.e. everything that can be freed).
[10:34] <soren> -NOCONTEXT, so Idon't know what you're trying to do exactly, but I find that usually 3 is what people really want.
[10:35] <LarstiQ> 3 does make it slower
[10:35] <LarstiQ> soren: what we're doing is looking at cold cache startup time
[10:35] <soren> s/NOCONTEXT/ENOCTXT/, obviously.
[10:42] <jamesh> lifeless: I uploaded a pair of branches to implement half of my proposed changes to bzr-dbus
[10:43] <jamesh> lifeless: with those changes applied, it should provide equivalent functionality
[11:25] <mrevell> beuno: Did you get the Ogg file?
[11:26] <beuno> mrevell, yeap, thanks!  Ping me when it's up on Launchpad blog and I'll post on the planet
[11:27] <mrevell> cool, will do
[11:51] <lifeless> soren: its changed since I first wrote the script :)
[11:51] <lifeless> soren: thanks for th details
[12:08] <soren> lifeless: Hm... Not that it matters much, I'm sure it's always been this way, actually.
[12:29]  * spiv likes "man proc"
[13:42] <beuno> james_w, this might be useful: https://lists.ubuntu.com/archives/ubuntu-devel-announce/2008-March/000395.html
[13:47] <james_w> beuno: thanks.
[14:15] <james_w> beuno: I forgot to bring my wiki login password. Is it ok if I add to the page at the weekend?
[14:15] <beuno> james_w, it's alright, I'll add the bits you sent and you can tweak it if you feel it's not well represented
[14:16] <beuno> just thought I'd poke you to do work for me  :p
[14:16] <james_w> :-)
[14:55] <ubotu> New bug: #199147 in bzr-svn "svn: revision identifiers don't work even in bzr-svn branches" [Undecided,New] https://launchpad.net/bugs/199147
[15:01] <lazy1> How do I view the log for a directory (including all files and subdirectories)
[15:02] <beuno> lazy1, just do:  bzr log dir/
[15:02] <james_w> lazy1: that doesn't work right at the moment I'm afraid.
[15:02] <beuno> james_w, no?  I ust fired it off and it seemed to work fine
[15:02]  * beuno re-checks
[15:03] <lazy1> It shows only the log for the directory, but not for the files inside
[15:03] <james_w> beuno: it shows the first revision that adds the dir, at least for me.
[15:03] <beuno> aaah, that's right
[15:03] <james_w> and maybe renames of the directory.
[15:03] <beuno> lazy1, listen to james_w  :D
[15:03] <ubotu> New bug: #132191 in bzr-gtk (universe) "installs unuseable desktop file to bzr commit-notify" [Medium,Fix committed] https://launchpad.net/bugs/132191
[15:03] <lazy1> I will, thanks
[15:04] <lazy1> james_w, any intentions to implement this?
[15:04] <james_w> lazy1: yes, there is. I can't remember what is blocking it at the moment.
[15:04] <lazy1> OK, thanks
[15:10] <ubotu> New bug: #199150 in bzr "AssertionError during merge of two branches with similar content but no shared base revision" [Undecided,New] https://launchpad.net/bugs/199150
[15:29]  * beuno waves to jelmer 
[15:30] <jelmer> hey beuno
[15:30] <beuno> how was your flight?
[15:30] <jelmer> I came closer than ever to missing it
[15:30] <jelmer> but other than that it was ok :-)
[15:31] <jelmer> anything interesting happening @ the sprint today?
[15:31] <beuno> jelmer, no fire alarm yet, if that's what you're hinting at
[15:32] <jelmer> I meant more in terms of topics that have been discussed :-)
[15:32] <beuno> launchpad bashing was fun
[15:32] <jelmer> ah, heh
[15:33] <beuno> and the gpg-signing bit was interesting too
[15:33] <jelmer> that's probably the only nice thing about closed software from a user pov
[15:33] <Odd_Bloke> jelmer: I was about to phone you for bzr-svn support, but spiv helped me out. :p
[15:33] <Faithful> Hi, can I use Bazaar is binary files like jpg or autocad drawings?
[15:33] <jelmer> you can complain about stuff that sucks without being told "patches are welcome" ;-)
[15:33] <jelmer> Odd_Bloke, :-)
[15:33] <jdong> Faithful: yes
[15:34] <Faithful> and it still does diffs of revisions?
[15:34] <jdong> umm... internally, yes. If you try to diff two revisions with binary changes, I don't know what it'll display
[15:34] <spiv> jelmer: I now have a svn-import of Twisted at http://people.ubuntu.com/~andrew/twisted/.  Thanks!
[15:35] <jelmer> spiv: w00t
[15:41] <awilkins> Ok, % formatting a string ; why doesn't this work "my %s string %s %s" % ["sexy", "is", "cool"]
[15:41] <elmo> % () not % []
[15:42] <awilkins> Hmm, better
[15:42] <awilkins> Still getting error, but different error :-)
[16:14] <ubotu> New bug: #199166 in bzr "Move revisionloader.py into core" [Undecided,New] https://launchpad.net/bugs/199166
[16:31] <bast> hi everyone
[16:31] <bast> i've a strange problem
[16:32] <bast> i use bzr all the time
[16:32] <bast> i manage a lot of project with it
[16:32] <bast> and suddendly i can'to anything with it
[16:32] <bast> bzr status/branch/commit/ etc
[16:32] <bast> nothing works
[16:32] <bast> my processors is used at 100%
[16:33] <bast> no output
[16:33] <bast> i don't understand :s
[16:40] <awilkins> Try the same command with --no-plugins
[16:45] <phanatic> beuno: you shouldn't have said that about the fire alarm... see what happened :P
[16:46] <bast> awilkins: nothing changes
[16:47] <bast> obliged to do a ctrl+Z then to kill
[16:48] <luks> check ~/.bzr.log
[16:50] <bast> luks: i'w watching it
[16:52] <bast> 0.158  encoding stdout as sys.stdout encoding 'UTF-8'
[16:52] <bast> 0.159  bzr arguments: [u'init']
[16:52] <bast> 0.160  looking for plugins in /Users/bast/.bazaar/plugins
[16:52] <bast> 0.160  looking for plugins in /Library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/bzrlib/plugins
[16:52] <bast> 0.160  Plugin name __init__ already loaded
[16:52] <bast> 0.160  Plugin name __init__ already loaded
[16:52] <bast> 0.214  encoding stdout as sys.stdout encoding 'UTF-8'
[16:52] <bast> 0.444  created control directory in file:///private/tmp/
[16:52] <bast> 0.452  creating repository in file:///private/tmp/.bzr/.
[16:52] <bast> 0.464  creating branch <bzrlib.branch.BzrBranchFormat6 object at 0x1c1c890> in file:///private/tmp/.bzr/
[16:52] <bast> my processor is at 100%
[16:52] <bast> and the init on an empty dir
[16:53] <bast> isn't finished
[16:53] <bast> i have a macbook on mac os x
[16:53] <luks> hmm
[16:56] <bast> i used it all the time and it begin to error with no reason
[16:56] <bast> i have python 2.5.2
[16:56] <bast> bzr 1.2
[16:57] <luks> I don't really know how to debug something like this
[16:58] <luks> `python -m trace --trace /path/to/bzr` is what I personally would do, but that's extremly verbose and only useful if you know some python
[16:59] <bast> i don't know python :s
[17:00] <luks> well, you can still try and see where it ends
[17:00] <luks> but as I said, expect lots of text in the terminal :)
[17:07] <poolie> luks, bast: what are you trying to do?
[17:07] <bast> i tried another shell
[17:07] <bast> :s
[17:07] <bast> nothing change
[17:08] <bast> to know where is the bzr command it's whereis no ?
[17:08] <phanatic> http://changelog.complete.org/posts/698-If-Version-Control-Systems-were-Airlines.html
[17:08] <luks> poolie: find out why bzr hangs
[17:09] <poolie> on unix, try pressing ctrl+backslash?
[17:10] <bast> i'm on mac os
[17:10] <bast> it's don't work
[17:11] <LarstiQ> phanatic: executive summary?
[17:14] <spiv> luks: isn't there an strace-like tool you can attach to a running process?
[17:14] <phanatic> LarstiQ: tla fork + BeOS competitor... :)
[17:14] <luks> I have no idea :)
[17:15] <poolie> it was funnier 15 years ago...
[17:17] <LarstiQ> phanatic: ah :)
[17:18] <awilkins> Gah, tinymce's new plugin style is annoying, it confuses JSEclipse
[18:12] <beuno> might be of interest: http://changelog.complete.org/posts/698-If-Version-Control-Systems-were-Airlines.html
[18:17] <LarstiQ> beuno: phanatic just mentioned that
[18:22] <beuno> LarstiQ, right, fire thing threw me off
[18:22] <LarstiQ> beuno: heh
[18:23] <mwhudson> getting thrown off the millbank tower would be unfortunate
[18:24] <LarstiQ> mwhudson: it would be a lot faster than walking all the stairs each day.
[18:25] <Toksyuryel> interesting
[18:25] <Toksyuryel> I'm not sure the author like bzr :(
[18:25] <Toksyuryel> seems like a fan of git =/
[18:26] <james_w> for a few days now at least.
[18:27]  * Toksyuryel supposes that once bzr's been rewritten in C or some other compiled language, more people will wake up to how awesome it is :)
[18:27] <dato> james_w: heh
[18:27] <james_w> Toksyuryel: that should be any day now.
[18:27] <beuno> james_w, http://launchpadlibrarian.net/12471357/buildlog_ubuntu-hardy-i386.baz_1.4.2-5.4_FAILEDTOBUILD.txt.gz
[18:27] <TFKyle> hmm, someone should see if they can get bzr translated via. rpython :D
[18:28] <Toksyuryel> I'll use it as-is though... I don't care how much slower python is, I more than make up for in the time I'm not spending fighting with it to do what I want it to do
[18:29] <Toksyuryel> james_w: cool, I didn't realize it would be so soon, I figured they'd want to deal with all the bugs and missing features first, but seeing how fast this thing moves they'll probably show up soon enough anyway
[18:37] <orospakr> Toksyuryel, god, why?
[18:39] <beuno> james_w, https://edge.launchpad.net/ubuntu/+source/bazaar/1.4.2-5.3
[18:41] <Odd_Bloke> spiv: I just got http://paste.pocoo.org/show/31944/ using the stable 0.4 tip.
[18:44] <Odd_Bloke> Hmm, I also seem to be getting it with the latest stable.
[18:45] <Toksyuryel> orospakr: ?
[18:45] <orospakr> Toksyuryel, why would anyone want to rewrite bzr in a different language?
[18:48] <Toksyuryel> orospakr: There is a perception that bzr is "slow" because python is an interpreted language. While it is true that bzr the program would run a hell of a lot faster written in a compiled language such as C, most people miss that bzr is still faster than other VCSes because even though the code runs slower you get more work done faster because you aren't fighting with the software
[18:49] <Toksyuryel> orospakr: however, it's still a worthwhile change and it'll get a lot more people interested in bzr. and according to james_w it's almost done anyway :)
[18:49] <orospakr> That is my opinion, actually.  Hence my remark asking why anyone would want to rewrite it. :P
[18:49] <orospakr> ack.
[18:49] <orospakr> But the Python code has unit tests, doesn't it?
[18:49] <orospakr> Does this C implementation have all the tests and features that the original python version does?
[18:49] <luks> Toksyuryel: the fact that bzr is "slow" is not that much relevant to python
[18:49] <orospakr> I still with the Python version, thanks.
[18:49] <orospakr> s/I still/I'll stick/
[18:50] <james_w> Toksyuryel: sorry, I was being sarcastic.
[18:50] <orospakr> Toksyuryel, it really wouldn't be a worthwhile change.
[18:50] <james_w> It's exactly the opinion of the bazaar team that the increased ease of programming more than offsets any speed costs of an interpreted language.
[18:51] <Toksyuryel> james_w: oh, good; I was hoping such a rewrite would be saved until after essential features are implimented and most bugs worked out :)
[18:51] <orospakr> In fact, it would be a big waste of time.  If they want to fix any performance issues, they should profile the software and work on any problem spots.
[18:51] <orospakr> Which they have done, in fact.
[18:51] <orospakr> I don't find bzr slow, in fact.
[18:51] <orospakr> There are some facts for you. ;)
[18:53] <Toksyuryel> I remember someone mentioned before that it was in the plans, was implied to be very far off plans and happy to see that attention is being paid more where it's needed :)
[18:53] <james_w> Toksyuryel: I don't think it is anyone's plans.
[18:53] <LarstiQ> Toksyuryel: the option has been held open, but don't expect to see that happen
[18:54] <james_w> anyway, time to eat. Bye all.
[18:54] <LarstiQ> right
[18:54] <Toksyuryel> enjoy your meal
[18:54] <orospakr> Declaring a "rewrite in a compiled language" is the worst way to optimize software. :)
[18:55] <orospakr> Mostly because you go from having tested, working software to no software at all.
[18:55] <Toksyuryel> better than writting code that needs comments the length of a docteral thesis
[18:56] <Toksyuryel> but I don't think bzr is going to go there anytime soon :) too many features to add and bugs to fix
[18:57] <fredreichbier> hello
[18:58]  * Toksyuryel personally would still use bzr if it were written in BASIC or hell, even java- it'd still be faster to use than anything else because of the way it works
[18:59] <fredreichbier> how do i change the remembered default location of `bzr push`?
[18:59] <awilkins> bzr push --remember
[19:00] <Toksyuryel> anyway, time for bed for me... getting sleepy and incoherent :) good night everyone *waves*
[19:02] <fredreichbier> hm awilkins that didn't work. i just changed the value in ~/.bazaar/locations.conf, seems to work now
[19:02] <fredreichbier> thank you anyway
[19:03] <awilkins> bzr might work on IronPython one day, that's sort-of-compiled
[19:13] <apol> is there anybody from ide integrationg around?
[19:14] <apol> *integration :P
[19:16] <apol> if there isn't anyone, could anybody point me out who should I ask?
[19:41]  * apol feels alone
[19:49] <LeoRochael> This is probably a FAQ, but, I'm using the launchpad bzr packages for ubuntu, and there doesn't seem to be a bzr-1.2 compatible version of bzr-svn
[19:49] <dato> there isn't one yet
[19:50] <LeoRochael> can I compile from source or is bzr-svn 0.4.7 not compatible with bzr 1.2?
[20:07] <LeoRochael> or maybe the bzr-svn "trunk" has bzr 1.2 support?
[20:07] <mwhudson> i must admit i generally just do "cd ~/.bazaar/plugins; bzr get lp:bzr-svn svn"
[20:09] <yacc> lp:bzr-svn ?
[20:09] <LeoRochael> yacc: I assume it means the launchpad repository for bzr-svn
[20:10] <yacc> oh ;)
[20:10] <yacc> Is that a general shortcut?
[20:10] <LeoRochael> yacc: I suppose so, for launchpad hosted projects...
[20:13] <johnny> now if only launchpad was open source..
[20:13]  * johnny wants to hack on it..
[20:13]  * johnny wants XMPP messaging..
[20:16] <blueyed> Can you have multiple branches in a single directory? so that you can switch to one of them, commit and switch back?
[20:19] <johnny> i don't understand that workflow method personally..
[20:19] <LeoRochael> blueyed: bzr shelve perhaps?
[20:19] <johnny> but it seems like you can with hselv right?
[20:29] <LeoRochael> johnny: I can understand this workflow when you have a lot of external systems that point to your directory and you don't want to keep changing them to other directorys to keep two branches in parallel
[20:29] <LeoRochael> quick example
[20:29] <LeoRochael> say you're managing your ~/public_html with bzr
[20:30] <luks> use a lightweight checkout and 'bzr switch'?
[20:30] <bob2> blueyed: you can have a repository with multiple branches, then use "bzr switch" to switch a single checkout between them
[20:32] <johnny> hmm... LeoRochael   aha.. now i get it.. i do that all the time with monotone, but i wonder if i'm missing anything
[20:35] <zepard> hi people
[20:35] <zepard> I am wondering why the http://bazaar-vcs.org/ReleaseRoadmap is showing the latest version upcoming = 0.92
[20:37] <bob2> no one has updated it since september - guess lp has replaced it
[20:39] <zepard> ok its there for historical reasons I guess but google still indexing it first
[20:40] <zepard> the first page shows the 1.2 download though
[20:41] <zepard> :(
[20:42] <blueyed> my workflow is: branch the vcs-import branch from launchpad (for a mega-feature-foo branch) and while refactoring, I want to make changes there which should go to another, new branch "trunk".. so I'd like to switch to "trunk" (in the same directory, basically the current branch without the locally modified changes; needs to be created) and commit only some files there. Then switch back.
[20:43] <blueyed> http://bazaar-vcs.org/ReleaseRoadmap needs to redirect somewhere useful.. or needs an update (for a long time already).
[20:47] <zepard> redirecting would more simple but not smooth since the updated page is in lp.
[20:49] <zepard> an update then :)
[20:50] <zepard> since there is not a huge activity on this page
[20:57] <bob2> blueyed: that can work - branch from lp to some other location, then checkout that location somehwere and work in the checkout
[21:02] <blueyed> bob2: sure, but that would require me to change the directory for testing that subbranch/-feature.. think DocumentRoot entry in Apache..
[21:03] <blueyed> ..so it's far more comfortable to check this into the same other-feature branch, which is bad.
[21:08] <LeoRochael> blueyed: what about "bzr shelve" or "bzr switch"?
[21:09] <dato> MadCoder: I see you were the last to touch unquote_c_style() in 7fb1011. could you investigate, who is at fault here: http://dpaste.com/38299/
[21:10] <dato> er, EWIN
[21:13] <bob2> blueyed: eh, no
[21:14] <bob2> blueyed: you bzr switch amongst branches stored outside the checkout, which doesn't require you to move or edit different files
[21:31] <andrei_> hey guys! I love bzr :). quick question: are you going to participate on summer of code this year?
[21:32] <andrei_> cause i'd be interested, as a student
[21:33] <andrei_> i see something on http://bazaar-vcs.org/SummerOfCode
[21:33] <andrei_> but i guess these are the projects from last year, right?
[21:38] <mnemoc> hi, what's the command equivalent to svn export http://.... in bzr?
[21:38] <luks> bzr export http://...
[21:38] <mnemoc> uh
[21:39] <mnemoc> luks: thanks :)
[21:49] <blueyed> LeoRochael, bob2: I would certainly get away using "bzr branch ... new", "bzr shelve", "bzr switch new", "bzr unshelve", "bzr ci", "bzr switch ." or something similar, but sounds dirty.. ;) isn't git able to switch to another branch more easily in the current directory/checkout?
[22:00]  * mwhudson dumps the "efficiency over NFS thread" unread
[22:00] <mwhudson> am i missing anything?
[22:15] <mnemoc> can i list the tags on a remote location?
[22:25] <Peng> Oh, I didn't know you offered bzr under other licenses.
[22:39] <dato> Peng: ref?
[22:40] <Peng> Homepage.
[22:40] <Peng> "Free. Bazaar is available under the GPL v2 or later. If you want to embed Bazaar into your products under a different license, please contact us."
[22:40] <dato> ah
[22:41] <dato> thanks
[22:48] <Odd_Bloke> apol: You want beuno and Verterok, I think.
[22:49] <apol> um, hello beuno and Verterok?
[22:52] <beuno> apol, yeah, hello
[22:53] <Verterok> apol: hi :)
[22:53] <apol> i'm a kdevelop developer and i'd like to know how prepared is bzr to work with ide's
[22:54] <apol> are there any libraries and so?
[22:54] <apol> haven't investigated yet :P
[22:54] <Odd_Bloke> apol: http://bazaar-vcs.org/SprintLondonMarch08/Brainstorms#head-a1f0e77d00a77e7f4634334e12dd5703e44fd8aa has some thoughts from the sprint that is ongoing.
[22:55] <apol> is there anything of this done?
[22:55] <apol> beuno: Verterok: ^
[22:56] <beuno> apol, sorry, yeah
[22:56] <beuno> well, we've just started putting a plan together
[22:56] <beuno> so I can't point you to anything specific
[22:57] <beuno> but I can get back to you with pointers and some general information
[22:57] <Verterok> apol: for Java (aka Eclipse) I'm using bzr-xmloutput plugin
[22:57] <apol> beuno: ok, that's a good place to start :P
[22:57] <beuno> eclipse is using xml to talk to bzr
[22:57] <beuno> so there is no reason you can't do that today
[22:58] <beuno> apol, and visual studio integration is using an embedded python interpreter
[22:58] <beuno> apol, can you bind to python?
[22:59] <apol> beuno: well, I'm working on a kross plugin, that would enable this possibility
[22:59] <apol> that's why i was asking actually
[23:00] <beuno> apol, than you can talk to bzrlib directly
[23:00] <beuno> and it's fairly straightforward
[23:00] <beuno> let me get a link for you
[23:00] <beuno> apol, http://bazaar-vcs.org/Integrating_with_Bazaar
[23:01] <apol> thanks beuno
[23:01] <Verterok> apol: I'm looking http://kross.dipe.org/, and it seems to provide python bindingss
[23:02] <LeoNerd> I find quote a lot, I use "shelve" to take out changes, to "commit" the rest, then "unshelve" the others back in, in order to do a partial commit of my changes. Is there perhaps a way to do this more directly..?
[23:02] <apol> Verterok: yes sure, that's what i meant :P
[23:03] <apol> looks like integration using python is what i was looking for
[23:03] <Verterok> apol: yeap
[23:04] <apol> Verterok: why don't you use that on eclipse? can't bind to python?
[23:05] <Verterok> apol: Eclipse is Java based, for the moment there is no decent Java-python bridge
[23:06] <apol> oh ok
[23:06] <Verterok> apol: so, I use a wrapper, with multiple implementation, the current one uses bzr executable + xmloutput
[23:06] <apol> well, I'll look into that this weekend
[23:06] <Verterok> apol: when Jython reacchs 2.5, it will be jython based :D
[23:06] <Verterok> apol: great!
[23:07] <apol> maybe I'll try to apply on that kross+bzr thing for gsoc
[23:07] <apol> Verterok: yes i was thinking of jython
[23:07] <apol> seems like sun hired some hackers to work on it
[23:07] <Verterok> apol: it's in 2.3 compatibility, bzr needs 2.4 :(
[23:08] <Odd_Bloke> apol: lifeless or poolie are probably people to talk to regarding GSoC.
[23:08] <Verterok> apol: if you decide to apply for gsoc, please contact pooli
[23:08] <Verterok> Odd_Bloke: exactly :)
[23:09] <apol> Odd_Bloke: I would try to apply on the kdevelop side i think
[23:09] <apol> but i'll contact them anyway
[23:09] <beuno> apol, yes,, we really want IDE integration
[23:09] <beuno> so it's a good time to jump in and get a lot of help
[23:09] <Odd_Bloke> LeoNerd: What would you envision a more direct approach looking?
[23:10] <beuno> make sure you keep us updated on progress
[23:10] <beuno> and don't hesitate to contact me or the list directly
[23:10] <LeoNerd> bzr interactive-commit  =>
[23:10] <LeoNerd> Then it goes to a shelve-like yes/no list of hunks
[23:10] <LeoNerd> At the end, it asks  "Commit? (y/n)" and then offers to enter a message
[23:10] <apol> beuno: I will :) thanks
[23:11] <beuno> apol, thanks
[23:11] <LeoNerd> I.e. interactive commit would be literally like shelve + commit + unshelve, but with the yes/no reversed
[23:11] <apol> np
[23:15] <beuno> apol, if you could drop me an email anyway so I have your details when we have a document prepared and send them to you: argentina@gmail.com
[23:16] <apol> beuno: are you from argentina? :P (i'm spanish/catalan)
[23:17] <beuno> apol, yeap, feel free to use spanish if it's easier for you  (so is Verterok)
[23:18] <apol> :P ok
[23:18] <apol> not really easier, but more familiar you know ;)
[23:18] <beuno> apol, I do  ;)
[23:20] <apol> sent
[23:22] <Odd_Bloke> LeoNerd: If you're interested in writing it, you could do that from a plugin.  It's probably too niche to go into the core, but I (and guys on the list) could probably give you some pointers.
[23:28] <LeoNerd> Niche?
[23:28] <LeoNerd> Heh.. I only mention it 'cause I'm having an argument with a darcs fan who says that's how darcs commit works :P
[23:28] <LeoNerd> I wondered if it's be nice to have in bzr, to wave back
[23:31] <Odd_Bloke> LeoNerd: It's probably non-trivial, and probably wouldn't get used by that many people.
[23:31] <Odd_Bloke> That said, if it's there, people will use it.
[23:32] <LeoNerd> Well, I for one would use it loads
[23:32] <LeoNerd> Almost every commit I make is shelve-assisted
[23:32] <Odd_Bloke> Me too, but I want to be able to assess my tree-state and perhaps make some modifications.
[23:32] <Odd_Bloke> So I probably wouldn't use it.
[23:33] <Odd_Bloke> However, if your workflow is different, that's cool.
[23:34] <james_w> sorry, I missed the start of this conversation. Are you discussing darcs record's hunk level commit?
[23:34] <LeoNerd> Yah
[23:34] <james_w> have you tried bzr's record plugin?
[23:34] <LeoNerd> Ah, no
[23:34] <LeoNerd> Shall look into that, thanks
[23:34] <james_w> http://bazaar-vcs.org/BzrPlugins
[23:35] <james_w> http://projects.collabora.co.uk/~asabil/bzr/bzr_record/
[23:35] <Odd_Bloke> james_w: That second link is broken. :(
[23:36] <james_w> oh dear.
[23:37] <Peng> That's not good.
[23:38] <james_w> http://people.collabora.co.uk/~asabil/bzr/bzr_record/
[23:39] <james_w> Can someone edit the wiki page please?
[23:40] <Odd_Bloke> james_w: Done.
[23:41] <james_w> thanks.
[23:43] <Odd_Bloke> james_w: It seems that 'record' doesn't actually do commits, just creates patches of the changes you've made.
[23:43] <LeoNerd> Ah.. :/
[23:45] <Odd_Bloke> LeoNerd: That said, it demonstrates how to integrate the shelf stuff into a separate command.
[23:45]  * LeoNerd nod
[23:46] <james_w> Maybe we should post to the list for some clues on how to integrate it with commit.
[23:47] <Odd_Bloke> TBH, this functionality can be replicated with "bzr shelve && bzr commit -m '<message>' && bzr unshelve --all".
[23:47] <LeoNerd> Yes.. that's exactly what I do now
[23:48] <LeoNerd> I just sometimes get confused in my head and say yes when I mean no, or vice versa
[23:48] <LeoNerd> You have to shelve away the things you _don't_ want
[23:48] <LeoNerd> Whereas interactive commit would say yes to the things you _do_ want.
[23:48] <Odd_Bloke> Ah, yeah, I see.
[23:48] <james_w> I know a lot of people love "darcs record".
[23:50] <Odd_Bloke> LeoNerd: So the record plugin seems to have its various concerns fairly well separated, so a lot of what you want to do could fairly easily be scavenged. :)
[23:56] <james_w> night guys
[23:58] <Peng> I like "hg record", but it was weird to get used to after "bzr shelve".