# /srv/irclogs.ubuntu.com/2008/02/29/#bzr.txt

spiv beuno beuno I know it used to come up frequently, this is the first report I've seen in a while. 00:00 oh, could be 00:02 time is a very slippery thing lately 00:02 how does bazaar compare to git? 00:47 seriously, i have only used git, but bazaar has a cooler name 00:47 lol 00:47 It's less salty, with more bite. 00:47 I find bzr's interface much easier to use, much more friendly 00:47 i want to name some project 'balthazaar' 00:47 it feels more high-level than git IMO 00:47 though the performance in my experience isn't as great 00:48 seydar, maybe something like this will help: http://bazaar-vcs.org/BzrVsGit 00:48 is there a sample session online? 00:48 ooh windows support. i'm not too fond of supporting windows ;-) 00:48 seydar: http://doc.bazaar-vcs.org/latest/en/mini-tutorial/index.html 00:48 seydar: that's a tutorial called "Bazaar in five minutes" 00:48 I really disliked git's way of committing things before I expected it to commit 00:49 seydar: there's other documentation at http://doc.bazaar-vcs.org/latest/ 00:49 I've recently started using git for dealing with subversion... I've found its performance to be more acceptable when paired with svn versus bzr-svn 00:49 but at the same time, I grumble about git's UI every 5 minutes 00:49 does bazaar have a witty hat, like 00:50 'git 'er done' from larry the cable guy? 00:50 I believe they do sell t-shirts like that :) 00:50 and what's up with git's refusal to acknowledge empty directories exist? :) 00:51 i saw linus torvalds in one 00:51 * beuno wonders if igc will take the bzr tshirts to the sprint 00:51 ok. so bazaar it is 00:51 I think it's a side effect of refusing to acknowledge directories exist. 00:51 lol 00:52 as I said, I've been looking at git lately for my subversion interactions with large svn repos 00:52 Actually, who DOES treat directories directly as entities?  bzr does, mtn does...  I don't think hg does... 00:52 also file identity 00:52 but only because I'm kinda forced to 00:52 balthazar is such a badass name 00:53 whats mtn? 00:53 beuno: igc has no idea re tshirts sorry 00:53 monotone I believe 00:54 seydar: monotone (venge.net) 00:54 * fullermd nods. 00:54 seydar: btw, you do acknowledge the implications of asking #bzr whether to use bzr or git, right? :D 00:54 totally 00:55 but I really do think we're giving an objective opinion, if there is ever such a thing 00:55 Which there isn't, on most things (especially UI), because they're inherently subjective. 00:56 I mean, Tom Lord probably thinks arch has an easy and intuitive UI   ;) 00:56 haha 00:56 though I find that hard to believe! 00:56 And he'd be right, for him. 00:56 And most people would probably hate my GUI setup...  but it's perfect for me. 00:57 oh, and arch had directories as entities 00:58 thanks guys 01:01 === jelmer_ is now known as jelmer igc, didn't you send an email a few months ago to send bzr tshirts for contributors? 01:04 That was poolie I think. 01:05 beuno: probably. If so, I did it on behalf of poolie 01:05 right, it was poolie. 01:05 not sure why I was convinced it was you 01:06 it was poolie 01:06 === _Verterok_ is now known as Verterok Right, hg doesn't version directories. 01:45 * Peng wanders off. 01:45 What's the name of the other bzr GUI that isn't qbzr? wildfire or something? 02:12 wildcat, IIRC 02:12 awmcclain, or bzr-gtk? 02:13 wildcat! 02:13 * igc lunch 02:26 Hrm. I'm guessing wildcat doesn't work with 1.2.0... I get a "file not locked" exception when I try to diff. 03:31 awmcclain, I think it's a very alpha sort-of-thing at the moment 03:32 And bzr-gtk is x11, right? 03:40 awmcclain: yes 03:41 awmcclain: it should work on any platform where pygtk works 03:41 Which includes win32 I believe. 03:41 I'm running os x but I'm not a huge fan of the x11 emulation. :) 03:41 (Although installing it might be a bit of a hassle) 03:41 So... it looks like i'll have to brave pyqt 03:42 awmcclain: There's the native OS X gtk now, right? 03:42 RAOF: o really? 03:43 If you don't mind less than fully tested code.  http://developer.imendio.com/projects/gtk-macosx/ is one google hit. 03:43 I recently saw a blog post with native mac gtk screenshots, which is why I have any idea about this :). 03:44 === fullermd_ is now known as fullermd is there anyone around? 04:46 carnage4ever: yes 04:50 spiv: trying to somehow get bzr to run a precommit shell script BEFORE commiting files... 05:03 is that possible? 05:03 carnage4ever: it is, although you need a python plugin to do it. 05:07 carnage4ever: https://lists.ubuntu.com/archives/bazaar/2008q1/036955.html may help 05:10 === bigdo1 is now known as bigdog New bug: #196881 in bzr "Exception redirecting merge output to a file" [Undecided,New] https://launchpad.net/bugs/196881 05:56 === [1]carnage4ever is now known as carnage4ever === [1]carnage4ever is now known as carnage4ever The protocol-v3 branch is ready for dogfooding. 08:33 cool 08:36 heya 08:37 what answer is there to something like  http://pastebin.ca/923001 and http://rafb.net/p/6xkuJL96.html ? 08:37 dholbach: looks like a possible launchpad bug 08:44 spiv: oh? how so? 08:44 dholbach: or maybe a bug in the launchpad plugin for the client 08:45 is there any information I should try to get for it? 08:46 dholbach: but IIRC "lp--NNNN:..." URLs are a server-side implementation detail. 08:46 dholbach: "bzr info -v", and the actual bzr commands being run by that script 08:46 bzr update; bzr commit -m "something" is what the script runs 08:47 let me get the version 08:47 spiv: http://rafb.net/p/LKucRO74.html 08:48 dholbach: yeah, I'm pretty sure it's a launchpad bug.  File it against launchpad-bazaar. 08:51 dholbach: as a workaround, hobbsee should be able to "bzr switch sftp://hobbsee@bazaar.launchpad.net/%7E5-a-day/5-a-day-data/main/" 08:52 spiv: thanks a lot 08:54 https://bugs.launchpad.net/launchpad-bazaar/+bug/196913 08:54 Launchpad bug 196913 in launchpad-bazaar "Cannot lock LockDir(lp--1218658708:///~5-a-day/5-a-day-data/main/.bzr/branchlock): Transport operation not possible: readonly transport" [Undecided,New] 08:54 dholbach: thanks for the bug report 08:54 de rien 08:54 dholbach: if you can get the particular bzr command that triggers the error in the bug report, that'd be good. 08:55 dholbach: or link to the add-5-a-day script 08:56 done, thanks 08:57 abentley: ping 09:15 http://bundlebuggy.aaronbentley.com/request/<1203467195.15574.136.camel@lifeless-64> 09:15 abentley: I think you had partially reviewed this 09:15 dholbach: 5-a-day stuff should be faster now; we had some supermirror issues 09:16 lifeless: rock and roll 09:19 you guys kick ass 09:19 lifeless (and others): do you think it's worth splitting up the branch into per-people-branches? (less possible locking issues)? we're at 53 committers now 09:22 dholbach: are you having locking issues? 09:22 lifeless: some people run into this every now and then (not that often yet, but I expect the project to grow) 09:23 dholbach: bzr should handle it 09:23 OK 09:23 if it becomes a problem, then sure 09:23 I'll keep you in the loop :) 09:23 nag me to do 5 a day :) 09:23 5-a-day keeps the doctor away! :) 09:23 https://wiki.ubuntu.com/5-A-Day#Log :) 09:24 jelmer: hi. I've found the -r option for replay a tad confusing. its docs say "see help revisionspec", so I assumed it'd work as in merge, so that replay -r 372..373 would replay just 373, but it replays both 372 and 373. -r 373 does what I want, but that's not how merge behaves. it does what -c 373 would do, but -c is not available in replay. 09:47 jelmer: do you see what I mean? 09:47 sounds cnfusing 09:49 dato: send a patch :] 09:49 Hmm, looks like svn-import is up to something now. 09:56 * Peng wanders off, nervous of swapping to death suddenly. 09:57 * dato wonders if --overwrite will work well when pushing to svn. 10:06 * Peng wonders if converting 60k revisions with about 400 MB of free RAM is a good idea. 10:08 jelmer: Man, svn-import is way too verbose. It's averaging almost one line of output per day! 10:09 revno: 371 10:18 revision-id:dato@net.com.org.es-20080228115922-wi3npwo1rnfmilbg 10:18 parent: dato@net.com.org.es-20080222115116-62obf3ebstr8l3lh 10:18 a space has been lost after revision-id? 10:18 jelmer: (also, any plans for a new bzr-svn in unstable? bzr can't migrate to testing without it) 10:19 dato, this weekend 10:34 ok 10:35 jelmer: do you have a sec to tell me how/if I can solve this bzr-svn problem? 10:35 dato: sure 10:37 I had a bzr branch synced with a svn branch. my replay went bad (see above), and I replayed one more revision than desired. sadly, I pushed that to svn. 10:38 so I replayed well in a copy of the bzr branch, and tried to push that (expecting to get the "branches have diverged" message), but it's just taking ages to do anything 10:39 should I try with --overwrite directly, or... maybe svn rm, + svn cp -r $HEAD-2 ? 10:40 dato: The argument to replay is a range of revisions and supporting -r272..273 to mean [272,273] makes sense to me 10:41 dato: Patches for improvements are welcome though 10:41 dato: It should just work 10:41 what should work, sorry? 10:41 dato: The push giving "Diverged Branches" 10:43 and with --overwrite? 10:43 jelmer: we use [from, to) for nearly everything else in bzrlib though 10:46 jelmer: or (from, to] perhaps I should say 10:46 jelmer: I think, that 'diff -r x..y' should show the aggreate changes that reply -r x..y will aply. 10:46 lifeless: Yeah, consistency is also nice, indeed 10:48 === cprov-out is now known as cprov === [1]carnage4ever is now known as carnage4ever === [1]carnage4ever is now known as carnage4ever === mrevell is now known as mrevell-lunch === mrevell-lunch is now known as mrevell === `6og is now known as Kamping_Kaiser is it normal that bzr-bookmarks doesn't work with bzr pull ? 14:15 I don't know, what is bzr-bookmarks? 14:16 a bzr plugin 14:18 to have bookmarks 14:18 lifeless: also the launchpad plugin doesn't work with bzr pull 14:18 bzr pull lp:~easy-radio/easy-radio/bargraph 14:18 bzr: ERROR: Not a branch: "/home/asabil/Devel/INSA/5IF/OT/easy-radio/trunk/lp:~easy-radio/easy-radio/bargraph/" 14:18 asabil: you are missing the required dependencies for sftp for bzr 14:41 lifeless: ??? why ? I use the sftp transport all the time ? 14:42 python-paramiko is installed 14:42 asabil: huh, interesting 14:44 asabil: what does 'bzr plugins' list ? 14:44 svn, multiparent, launchpad, rebase, gtk, record, bzrtools, xmloutput, bookmarks 14:45 asabil: this is very interesting. I think something is buggering up pull :) 14:48 let me disable xmloutput 14:49 no, that's not the cultpit 14:50 bookmarks sounds like something inclined to diddle with pull 14:51 lifeless: there's a bug about it reported on lp iirc, pull doesn't like lp: uris 14:52 https://bugs.launchpad.net/bzr/+bug/181945 if you havn't seen it yet 14:53 Launchpad bug 181945 in bzr "bzr pull lp:upstart fails" [High,Confirmed] 14:53 lifeless: the bug was there before I installed bookmarks 14:53 asabil: ah, interesting 14:57 TFKyle: thanks 14:57 seems to be pull not handling redirects 14:58 now it seems like lp's server is stalled or something :/ 14:59 am unable to push 15:00 asabil: checking 15:01 thanks 15:02 asabil: yes, rogue process, fixing in a couple of minutes 15:04 your lilnk may get bounced 15:04 oki 15:04 I interrupted already 15:04 \asait should be faster now 15:31 Hmmm. "bzr pack" seems to make "bzr status" *slower*. 15:36 Not much, but a bit. 15:37 Lo-lan-do: it may be forcing things out of memory? 15:38 I ran it several times in a row :-) 15:38 Both before and after the pack. 15:38 It took about 0.58s before the pack, and it now takes 0.61s. 15:39 Lo-lan-do: pack and status use dufferent data; its almost certainly just measurement noise 15:40 Lo-lan-do: specifically status won't read data from the repository 15:40 Hm. Okay :-) 15:40 bazaar.launchpad.net is refusing connections for me, does anyone else sees this too? 15:41 yes 15:41 we're working on it urgently 15:41 ok, thanks 15:42 bazaar.launchpad.net is down; it's being addressed urgently 15:42 sorry, paste error 15:42 its back now 15:48 pqm.bazaar-vcs.org's invalid https cert looks worse under ff3 15:59 poolie: yes it does :) 16:03 awilkins: just to let you know, I pushed a BazaarClient with some fixes and command syntax updated to bzr-1.2 17:16 === BasicMac is now known as BasicOSX === doko_ is now known as doko === gotgenes is now known as gotgenes|away jelmer: have you considered upgrading or at least making available a packs version of bzr-svn's branch? :D 18:02 I'm in a low-patience branching mood today 18:02 === mw is now known as mw|food night all 18:58 night poolie 18:58 have a safe journey 18:58 night all 19:03 james_w: we're in London now 19:03 ah, welcome. 19:03 what does: "899.934 not updating child fraction" in the .bzr.log file? 19:19 It's related to progress bars. The details escape me. 19:34 === mw|food is now known as mw jdong: heh 19:54 yeah, I guess it's about time to upgrade 19:54 jdong: please note that it is *not* advised to use the 0.4 branch of bzr-svn atm 19:56 jelmer: ah, ok. Also, are there any known issues between bzr-svn and subversion 1.5.x/1.6.x? 20:17 jelmer: I recall trying it earlier and getting something along the lines of a mismatched # of arguments traceback 20:17 jdong: I haven't tried 1.6.x, 1.5.x should work fine 20:17 jelmer: ok, just an FYI the ForeignBranches howto says to branch subversion trunk which is apparently "1.6.x" now 20:18 jdong: Can you be more specific ? :-) 20:18 jelmer: *grumble* let me reproduce it ;-) 20:18 return apply(_ra.svn_ra_do_update, args) 20:20 TypeError: add_nodes() takes exactly 2 arguments (4 given) 20:20 jelmer: ^^. Lemme know if that's totally wack. I did really mutilate the svn stack to get it to build in this RHEL4 env 20:21 jdong: That looks like a python-subversion issue 20:21 jelmer: this is subversion branches/1.5.x IIRC.... 20:22 jelmer: wait is bzr-svn 0.4.7 supposed to work against bzr 1.2.0? 20:23 jdong: Something funky is happening at your side 20:23 jdong: There is no symbol add_nodes in either bzr-svn or python-subversion 20:23 /mit/jdong/.local/lib/python2.5/site-packages/bzrlib/index.py: def add_nodes(self, nodes): 20:24 it seems to be in bzrlib? 20:24 any chance you can post the full backtrace somewhere? 20:24 O_O 20:25 oh 20:25 wow 20:25 lol that was confusing 20:25 so I post this traceback to Ubuntu pastebin... 20:25 and Ubuntu pastebin HAPPENED to spit back a CGI error in the form of a python traceback 20:26 heh 20:26 and all I could think is "hey that's not the one I put in" 20:26 lol lemme find another pastebin 20:26 (-: 20:26 http://paste.ubuntu-nl.org/57869/ 20:26 the more I work with svn, the more I love bzr. thanks folks for making bzr real. 20:27 WELL I guess the selftest segfaulting is a good sign that I did something idiotic compiling svn. 20:31 lifeless, ping 20:43 === gotgenes|away is now known as gotgenes I'd like to merge changes from a branch into another one but only in relation to one file - can I do that? 20:49 "bzr merge branch filename" doesn't seem to work 20:49 jelmer: any hints on the backtrace? 20:50 mdke: the only way I know of is to merge the branch, then revert all but that one file 20:50 mdke, AFAIK, cherrypicking is not supported yet 20:50 jdong: Whoops, sorry 20:50 jdong, beuno: ok, shame. Thanks 20:51 === jkakar_ is now known as jkakar cp it is then :) 20:51 spoken almost like a git developer :) 20:51 mdke, old n' trusty :p 20:51 jdong: Doesn't make an awful lot of sense 20:52 yes, cp seems to work 20:52 No, you can bzr merge /some/branch/some/file 20:52 jdong: I guess the backtrace gets messed up by the python subversion bindings 20:52 fullermd: doesn't seem to work with a remote branch on launchpad at least 20:53 (it doesn't record anything of course, just like cherrypicking revs, but it works) 20:53 ah I guess that would be bug #81758 then 20:53 Launchpad bug 81758 in bzr "'bzr help merge' should describe merging a single file" [Low,Confirmed] https://launchpad.net/bugs/81758 20:53 jelmer: well, I can tell these RHEL4 servers hate me 20:53 it may actually be a core bzr bug 20:53 but it's hard to tell without a proper traceback :-/ 20:53 jelmer: confusing why a nearly identical setup on ubuntu doesn't err out 20:54 jdong: Does "bzr selftest svn" pass? 20:54 jelmer: no, it.... segfaults.... 20:54 which is another worrisome sign 20:54 uhm, yes :-) 20:54 * jdong nukes his ~/.local and starts over 20:55 when in doubt..... wipe ito ut. 20:55 spoken in the true voice of American diplomacy. 20:55 (-: 20:55 ok, let's try just installing bzr and running its selftest first... 20:57 has anyone used bzr-lomb? 22:02 hi, I'm running bzr on Mac (Leopard OS X) and I don't know why, but when I run a commit or branch, it just hangs... I've left it running overnight and it still doesn't finish. Any ideas/pointers? 22:08 makko, maybe take a look at ~/.bzr.log? 22:11 Does any one know of a nice (aka TortousCVS) graphical interface to bazarr on OSX? 22:11 beuno: I did. For example, I just wanted to get the latest bzr.dev (as in the user manual), and it just hangs. My bzr.log says: 396.636 creating branch in file:///Users/normsu/root/bzr/bzr.dev/.bzr/ ... 22:12 makko, what version of bzr would that be? 22:13 beuno: I got bzr from macports (1.2.0) and I'm trying to fetch bzr.dev with "bzr branch -v http://bazaar-vcs.org/bzr/bzr.dev" 22:13 poolfool: have you seen QBzr 22:14 makko, can you report a bug with the contents of your .bzr.log? 22:15 beuno: OK, I just wasn't sure if this was an actual bug or something kooky with my own setup/environment. 22:16 makko, thanks. I'm sure someone else will help you debug it 22:17 beuno: OK, thanks for your kind assistance I'll try your suggestion. 22:18 makko, np. good luck :D 22:18 === asak_ is now known as asak benjaminpeterson: No ... sorry ... but I guess I was looking for something a little more tied into finder. 22:34 I am kind of new to OSX (1 month new), and I mostly use the console ... but at work I use TortoiseCVS and it gets kind of addicting. 22:35 benjaminpeterson: Quit a nick by the way ... hurts to type ... almost 22:35 poolfool: Hmm, I don't actually know of any applications that interface with the MacOS finder. Is that possible? (Are their APIs?) 22:37 benjaminpeterson: I would guess not ... Apple kind of sucks that way ... but I figured I would throw the question out any ways. 22:38 === benjaminpeterson is now known as bpeterson bpeterson: Much easier ... are you a Mac OSX guy? 22:39 poolfool: I switch between KDE and Mac 22:40 poolfool: I agree that Apple should make this possible. Some cool things could be done... 22:40 maybe I'm crazy, nerdy, and so on... but am I the only one to prefer working with a textmode VCS? 22:46 I mean, I've used Subclipse, TortoiseSVN, bzr-gtk, etc all before, and find them more cumbersome than their command line counterparts 22:46 with the sole exception of bzr visualize and similar tools 22:47 ancenstry is more clear in a GUI form 22:47 jdong: I love textmode 22:48 jdong: I typically use a simple editor with syntax highlighting and perform all my other operations with bash 22:49 I like gui for the quick clues that a file may be out of date (icon overlay in Microsoft Windows with Tortoise), but commits, diff, log generation for reports are much easier text based. 22:49 bzr missing 22:50 That said, I just seem to live via gui. That is kind of why I was looking for something that might tie into finder ... not yet another graphic program to learn. 22:50 poolfoo1: interesting... I find bzr diff, bzr log, bzr status easier by commandline 22:50 jdong: sorry, maybe I didn't read/write that just write. I agree that diff, log and status are much easier from command line\text. 22:51 A GUI is a way to get LOTS of command lines at once :) 22:52 What about web interface (bzrweb?) to revision control? I like the bonsai/ViewVC interface a lot for a while. 22:53 poolfoo1: seems like loggerhead is the preferred one nowadays 22:53 launchpad has a demo of it 22:54 poolfoo1: http://bazaar.launchpad.net/~bzr/bzr/trunk/files 22:54 I still like bzr-serve or whatever that other older one is called 22:54 because you can start it on some arbitrary port on a single command 22:54 loggerhead (the last time I played with it) was a whole suite of power tools just ready to rip an finger/arm/leg off. I like the small light weight ability of bzweb for small projects/groups or single use as a simple gui. 22:54 loggerhead would be much better for something like launchpad/sorceforge/... 22:55 poolfoo1: yeah the older one was easier for personal use 22:55 i like to know what my computer is doing, so bash gives the most control and detail 22:56 I find commandline easiest for stuff I know how to do, and GUI's easier for, frankly, BS-ing my way along when I really don't know what I'm doing. 22:59 again, personal opinion. 22:59 it is more intuitive 22:59 What is the craziest/coolest use of bazaar to date? I heard someone ( lifeless ) suggest using bzrlib for a wiki ... sounded pretty cool to me. 22:59 if you *know* what you want to do, it's much easier to express it in a CLI 22:59 poolfoo1: benjamin mako hill is doing a bzr wiki for his doctoral thesis 22:59 poolfoo1: I wrote a ruby+bzr TODO list program a few weeks ago 23:00 so I can operate disconnectedly 23:00 jdong: do you have the link? ... but what about a revision controlled filesystem ... something more complex then the Linux COW patch. 23:00 poolfoo1: I also use bzr heavily while doing Ubuntu package development or packaging. It helps me keep track of my current changes a lot better 23:00 i suppose you can bumble around in GUI with out causing horrible damage 23:00 poolfoo1: I don't have a link handy, no 23:00 bpeterson: well a GUI's beter at answering "Hmm, what can I do?" 23:01 i.e. poking around a new program 23:01 I found git extremely bewildering on day one 23:01 IMO it's approaching arch fame :) 23:01 but once I kinda knew what I was doing, it was pretty easy to use 23:01 jdong: when you use it for ubuntu you are using it as a RCS? ... how about using it somewhere that is revision control (wiki/apple time machine/... ) but you just don't realise or think of it that way. 23:01 jdong: when you don't want to look through the docs :) 23:01 jdong: is git really as fast as Linus says 23:06 New bug: #197125 in bzr "bzr hangs on Mac OS X Leopard (10.5.2)" [Undecided,New] https://launchpad.net/bugs/197125 23:11 Is there any way to make bzr ignore permission changes on files, for a checkin? 23:14 bpeterson: git is fast indeed, impressively so 23:15 bpeterson: something like "git log > /dev/null" for 500,000 revisions happens nearly instantly 23:16 i.e. complete in 0.5 seconds 23:16 bpeterson: but personally I don't think most sized projects benefit from that speed 23:16 bpeterson: git has been much slower for me because in the time it takes me to answer "What the hell is git pull . refs:master/trunk?", bzr could've done the operation 10 times :) 23:17 jdong: Have you tried git on Microsoft Windows recently? Is there a not cygwin version? 23:18 My big selling point for bazaar is Win32 at work, Linux & MacOSX at home ... no problems to date ... 23:19 poolfoo1: AFAIK all the git-on-windows implementations are either quirky or very slow 23:19 poolfoo1: bzr is definitely much much more portable 23:19 Which would be my big selling point .... 23:20 poolfoo1: for me, that and (2) simplicity (3) outstanding support for dumb protocols 23:21 those two are things that most VCSes lack IMO 23:22 jdong: you mean the fact that bzr can checkout over http/scp/ssh ? 23:23 poolfoo1: right. The fact that bzr works just fine over standard protocols like SFTP, HTTP, FTP and so on without the need for anything special on the remote end 23:24 Git can read over HTTP, but that's it 23:24 if you want to push, you can't do that over FTP. You can only do it over SSH if git were installed on the remote machine 23:24 jdong: so the purpose of a dedicated bzr protocol and server? speed ... but I am just looking to get info. 23:24 === mw is now known as mw|out poolfoo1: right, the smart server tends to be faster 23:25 poolfoo1: rather, has more potential to be faster 23:25 poolfoo1: though the recent "packs" storage format evened the playing field a lot 23:25 jdong: potential ... yea .... about that ... 23:25 poolfoo1: branching a large packs-format repo over http maxes out my connection at 8MB/s 23:25 for the most part 23:25 which is as good performance as one can hope for, right? :) 23:26 Wow ... 8MB/s ... that's a lot better then any connection I have had in a long while; at work Dual T1 with poor load balance right now and CrapCast (ComCast) at home. 23:29 poolfoo1: well I'm glad my$40,000 tuition allows MIT to buy me some pretty decent internet :) 23:30 I've bursted almost that high on Comcast. 23:30 jdong: Hahahaa, ahh, the days of school internet... 23:30 indeed :) 23:30 I'll miss my 18.0.0.0/8 subnet too 23:30 Aaaahh yes ... the last time I had decent net speed I was at Colorado State University (CSU) .... the good old days. 23:30 where everything from the printer next to me to my iPod Touch has a public IP 23:31 jdong: In a sense that's a little frightening, though. :-) 23:32 can anyone explain how to write a pre-commit hook script with Bzr 1.2? 23:32 mkanat: :) access control should be strong enough to withstand that anyway :) 23:32 damageboy: http://doc.bazaar-vcs.org/bzr.dev/en/user-reference/bzr_man.html#hooks 23:33 Is there some place for good bzr scripts? 23:35 jdong: I hope so! :-) 23:36 Can I run a bash script that changes the files I'm about to commit (potentially even creates more changes...)? 23:36 From the pre_commit hook that is...? 23:36 damageboy: I dont't believe so. Bazaar has already calculated the changes to commit. 23:44 damageboy: I'm not sure, but I'd imaging that would work, unlike on SVN 23:57 damageboy: I think I might want to try that myself, I have some reasons to do it in my present project :-) 23:58 Anyway, time for bed. 23:58

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!