2008/11/18/#bzr.txt

[00:03] <visik7> but if a replay fail due to a conflict there is no replay continue like rebase
[00:04] <jelmer> visik7, yes, you can use "bzr rebase-continue"
[00:04] <jelmer> ah, no
[00:04] <jelmer> nevermind
[00:04] <visik7> :\
[00:04] <jelmer> you can commit and run "bzr replay" again though from the point where it aborted
[00:05] <visik7> nevermind seems it's working
[00:05] <visik7> ok I think is ok
[00:06] <visik7> now I need to clean it up branching again ?
[00:06] <igc> poolie: I'd like to rename the (about to be released) diffinternals plugin to crosscheck. Is that ok with you?
[00:06] <poolie> igc, that sounds better to me
[00:07] <igc> poolie: cool. me too.
[00:08] <jelmer> visik7, yeah, branch it to a different location
[00:08] <jelmer> visik7, (replay wasn't really designed for this sort of thing)
=== mw______ is now known as mw____________
=== mw____________ is now known as mw
[00:27] <visik7> jelmer: I see anyway it worked so, thanks for the help and hope soon we will have something like git-filter-branch
[00:33] <jelmer> visik7, cool
=== kiko is now known as kiko-zzz
[00:38] <p_masho> can anyone give me a helping hand.. bit confused as used to svn and cvs(spits).......
[00:38] <visik7> good night thank you again jelmer
[00:39] <jelmer> p_masho, hi
[00:41] <p_masho> hi jelmer.. late at night here in wales.. and got this zim on my ubuntu box - http://bazaar.launchpad.net/%7Epardus-cpan/zim/pyzim/
[00:41] <p_masho> so I been hacking away... and have already modified and not creaed a branch...
[00:42] <p_masho> I'm looking for 1) - equivelent of svn copy foo.x bar.x ?
[00:43] <jelmer> p_masho, copying a branch or a file/directory ?
[00:43] <jelmer> p_masho, to copy a branch, use "bzr branch <old-branch> <new-branch>"
[00:43] <jelmer> to copy a file, simple copy it using cp and then add it
[00:44] <p_masho> Ok. advice.. am used to svn.. so be gentle.. please
[00:44] <p_masho> I go this branch bzr branch lp:~pardus-cpan/zim/pyzim = /home/zim/pyzim on my machine
[00:45] <p_masho> so I hacked that a bit and got a diff file.....
[00:45] <jelmer> p_masho, what are you trying to copy? a file?
[00:45] <p_masho> What I want to know is how I can create a new "version" on my machine copying all the mods I got
[00:46] <p_masho> and then restore the master branch to tha release from launchpad...
[00:46] <p_masho> sorrry if this sounds confusing...
[00:47] <p_masho> Suppose the question really is.. first ..
[00:47] <jelmer> sorry, phone..
[00:47] <p_masho> How do I branch off the mods I already done and restore the "main/HEAD" to main repos
[00:48] <lifeless> p_masho: to make a new branch, 'bzr branch'
[00:48] <lifeless> p_masho: e.g. 'bzr branch . ../new-branch'
[00:48] <lifeless> p_masho: if you haven't committed your changes, do so first
[00:49] <lifeless> p_masho: I don't know what you mean about the main repos, you haven't altered the mainline afaict
[00:49] <p_masho> thats the problem (as sv user) I made changes to head...
[00:50] <lifeless> you have?
[01:01]  * Sidnei is away: I'm busy
=== mw is now known as mw|out
[01:23]  * igc lunch, etc. - bbl
[01:34] <jelmer> poolie1, hi
[02:37] <poolie1> hi jelmer
[02:37] <poolie1> spiv, thanks for that note!
[02:49] <spiv> poolie1: you're welcome :)
[03:14]  * spiv -> lunch
[03:16] <poolie> spiv, still here?
[03:17] <poolie> to build on that mail, i'd like to get a page like the smart push analysis, that for each scenario says what operations we _should_ be doing
[03:20] <poolie> in other words: we have a list of what is done at the moment and we have some vectors we can apply to them but i'd like to see where the expected endpoint is too
[03:25] <Peng_> Yay. "bzr heads" + ack + sed + bash == easy upgrade from subtrees to rich-roots.
[03:25] <Peng_> "bzr upgrade' would've been even easier, of course. Whinewhine.
[03:26] <jelmer> since lifeless doesn't appear to be here...
[03:26] <jelmer> Peng_: "Patches welcome" :-)
[03:26]  * Peng_ cries
[03:27] <Peng_> Wait, I was upgrading a copy of bzr-svn, so I can blame you for putting me in this situation in the first place! :P
[03:27]  * Peng_ hides
[03:27] <beuno> Peng_, recursive-upgrade?
[03:27] <Peng_> That was actually pretty painless. :)
[03:27] <Peng_> beuno: ?
[03:27] <beuno> I know jelmer doesn't want you to know about it
[03:27] <beuno> but he wrote a plugin
[03:27] <beuno> http://people.samba.org/bzr/jelmer/bzr-recursive-upgrade/trunk
[03:27] <Peng_> Oh?
=== abentley1 is now known as abentley
[03:28] <jelmer> Peng, subtrees is more than just a version string
[03:29] <jelmer> Peng, the inventory format is also different afaik
[03:29] <Peng_> jelmer: So? I don't care. I just want to be able to convert my repos easily. I can live with it burning some extra CPU.
[03:30]  * Peng_ gets grumpy when thinking about this. :P
[03:30] <jelmer> Peng_: I wasn't sure what you were doing exactly, but just editing .bzr/repository-format isn't sufficient
[03:30] <Peng_> jelmer: Yeah, I know. I didn't do that.
[03:30] <jelmer> Ok, I'll shut up then :-)
[03:31] <Peng_> I used "bzr heads" to grab all of the heads, created a new repo and a branch, and used "bzr pull --overwrite -r revid:..." to copy all of the revisions into it.
[03:33] <Peng_> And...let's see...frozenset(bzrlib.branch.open_containing('.')[0].repository.all_revision_ids()) to verify it went well. :)
[03:34] <poolie> lifeless: can you point me to a mail about dev3 making some loggerhead pages much faster?
[03:39] <lifeless> poolie: no, there isn't one
[03:39] <lifeless> poolie: there is a branch with a merge request in LP that makes lh compatible with bb-core
[03:40] <lifeless> poolie: and its --development4 that works for lh, fwiw
[03:41] <beuno> lifeless, about that, I didn't quite understand what you meant
[03:41] <beuno> it breaks with the current bzr
[03:41] <beuno> what is it that I'm missing?
[03:41] <lifeless> beuno: works for me
[03:41]  * beuno tests again
[03:42] <beuno> lifeless, hrm, it seems to as well now for me...
[03:42] <beuno> I'll run the tests and merge
[03:43] <spiv> (back)
[03:45] <lifeless> beuno: :P
[03:45] <poolie> lifeless: but it has been measured to make some pages faster? the diff display, i'd presume?
[03:45] <beuno> lifeless, so I had a b0rked copy of bzr?
[03:45] <lifeless> poolie: unscientifically yes
[03:46] <lifeless> poolie: revision display for instance
[03:47] <beuno> lifeless, merged. Thanks for the patch
[03:48] <beuno> 101  branches owned by 24  people and 1  team, 75  commits by 12  people in the last month
[03:48] <beuno> ^  Loggerhead
[03:48]  * beuno dances
[03:48] <poolie> :)
[03:51] <lifeless> beuno: no probs
[03:51] <hydrapheetz> Hm, tempting to write a bazaar branch browser django app, but laziness D:
[03:56] <lifeless> reuse loggerhead?
[03:57] <hydrapheetz> I should try that.
[03:58]  * beuno thinks "loggerheadlib" and looks at thumper 
[04:00] <thumper> why me?
[04:00] <thumper> beuno: JFDI
[04:00] <beuno> thumper, sure sure, and world peace, and all that
[04:00] <beuno> BUT
[04:01] <beuno> there are some interesting opportunities to work on that
[04:01] <beuno> like re-writing the log FILE bit which leaks billions of litres of memory
[04:04] <beuno> maybe during UDS
[04:04] <beuno> lifeless, poolie, will you guys be at UDS?
[04:06] <beuno> ah!  you are!
[04:07]  * beuno starts ploting on how to take advantage of their presence
[04:07] <AfC> Memory being measured in terms of three dimensional volume. Neat.
[04:12] <thumper> AfC: if you squeeze memory enough, you can get liquid out
[04:14] <lifeless> silicon-juice?
=== jamesh_ is now known as jamesh
=== sdboyer|class is now known as monkeyy
=== monkeyy is now known as sdboyer
[04:59] <Peng_> Hm..what do btree indexes do for Loggerhead?
[04:59] <poolie> they should make it faster
=== abadger19991 is now known as abadger1999
[05:27] <poolie> lifeless: do you know if brisbane-core with vila's changes is now passing all tests, or not quite?
[05:28] <lifeless> poolie: its either close or all passing
[05:28] <lifeless> which isn't really helpful, sorry
[05:28] <poolie> it's approximately helpful :)
[05:29] <lifeless> he made all the intertree ones pass; I don't know about the entire test set though
=== sdboyer is now known as sdboyer|goober
[05:30] <poolie> me either
[05:47] <Peng_> No, it's not all passing.
[05:53] <Peng_> Um, there were 6 errors and 5 failures total, in test_merge, test_pack_repository and test_repository.
=== sdboyer|goober is now known as sdboyer
[06:05] <poolie> Peng_: thanks!
[06:05] <poolie> you should fix them :)
[06:42] <Peng_> What would be the use of upgrading bzr.dev to the 1.6 format?
[06:42] <Peng_> Does doing "bzr branch --stacked" require the source branch to support stacking?
=== sdboyer is now known as sdboyer|vurk
[06:54] <abentley1> Peng_: No, but default stacking policies, which Launchpad uses, require that.
[07:00] <Peng_> Oh
[07:00] <vila> hi alla
[07:00] <vila> ?? Hi all :)
[07:01] <Peng_> ...What are default stacking policies, exactly? "bzr branch" will stack by default?
[07:01] <vila> poolie: I think it's "quite close" if merged with a recent bzr.dev, but the failing ones are still a bit hard to fix (I can dig to more precise if you will)
[07:02] <vila> poolie: I think it's "quite close" if merged with a recent bzr.dev, but the failing ones are still a bit hard to fix (I can dig to be more precise if you will)
[07:02] <jml> Peng_: bzr push will.
[07:02] <jml> Peng_: tbh, I haven't tried branch.
[07:04] <Peng_> Oh.
[07:05] <Peng_> ...http://bazaar-vcs.org/bzr/ needs that?
[07:08]  * spiv finishes for the day
[07:12] <spiv> abentley1: I don't have time to give your remoterepo stacking fix a proper review atm, but at a glance it looks good
[07:12] <spiv> abentley1: if it hasn't already had a review by the time I start tomorrow I'll certainly review it then.
[07:12] <abentley1> spiv: Cool
=== abentley1 is now known as abentley
[07:16] <poolie> hi abentley!
[07:16] <abentley> poolie: hi!
[07:21] <lifeless> ciao
[07:22] <lifeless> I'm on leave now ;), I may drop by and chat when playing with pet projects, but I'll be focusing on relaxing and zoning out mainly :)
[07:32] <abentley> lifeless: have fun.
[07:54] <Peng_> lifeless: I know I'm saying this just after you left, but what are you planning to do about group_size in bzr-search?
[08:03] <lifeless> Peng_: right now? nothin
[08:04] <lifeless> abentley: tanks
[08:05] <Peng_> Heh, okay.
[08:29] <igc> night all
[08:31] <Peng_> Good night.
[08:32] <Peng_> Ehhh?
[08:33] <Peng_> I think I may have just figured out the cause of my weird Loggerhead issues: a simple typo. That's what I get for not reading the error messages well?
[08:35] <Peng_> Oh, no, the typo is new, I think.
[08:40] <lifeless> jml: right, one less TUIT; testresources now sorts properly
[08:40] <Peng_> Hm...what's the right way to get a patch into Loggerhead? Usually I just bug someone on IRC. :P
[08:41] <lifeless> push a branch to lp
[08:41] <lifeless> file a merge request in lp
[08:41] <Peng_> :)
[08:48] <mwhudson> man i wish i could just click a button to land that branch
[08:48] <Peng_> Hi. :)
[08:48] <Peng_> See? Bugging people on IRC is more efficient. :P
[08:53] <mwhudson> Peng_: merged, ta
[08:53] <mwhudson> and ooh, the merge proposal is automatically marked merged :)
[08:54] <Peng_> Thanks. :)
[08:55] <mwhudson> Peng_: do you want to join ~loggerhead-team?
[08:58] <Peng_> mwhudson: Meaning I could push to the trunk without needing to go through you?
[08:58] <Peng_> Or just to have a neat badge?
[08:58] <mwhudson> the former
[08:59] <mwhudson> (i don't think ~loggerhead-team has a badge...)
[08:59] <Peng_> No badge? That's no fun. :P
[08:59] <mwhudson> heh
[08:59] <mwhudson> i guess we'd also want you to commit to being happy to assign (c) Canonical on your work
[09:01] <Peng_> I dunno. I don't contribute enough code to be a major review burden. If I ever contributed something large, I'd want it to be reviewed anyway, and my small things aren't really high-priority.
[09:09] <lifeless> mwhudson: I wouldn't tie the things together; surely a committment to stay under the paragraph limit when pushing (until a cc assignment is done) is enough
[09:09] <lifeless> mwhudson: can always uncommit after all
[09:10] <mwhudson> lifeless: true
=== jszakmeister|awa is now known as jszakmeister
[09:17] <Peng_> Honestly, I dislike copyright assignments, but I never contribute anything large enough to be worth caring about the rights of anyway. :P
[09:17] <Peng_> (That doesn't mean I'd be unwilling to sign one.)
[09:19] <Peng_> Then I get to the part where I'm not even an adult, so I don't know if I legally *can* sign one, and then I decide to put it off and go watch TV. :\
[09:22] <lifeless> Peng_: adult is a state of mind
[09:24] <Peng_> So when I turn 18, I still won't be an adult? Damn. :(
[09:33] <mwhudson> i don't _think_ there's an age thing with copyright is there?  maybe there is
[09:35] <Peng_> I don't think so either, but copyright assignments are legal document thingies, right? so there might be one there.
[09:36] <mwhudson> hmm
=== jszakmeister is now known as jszakmeister|awa
=== andreas__ is now known as ahasenack
=== weigon__ is now known as weigon
=== doko_ is now known as doko
[12:45] <Peng_> lifeless: FYI, I let a "bzr index" of bzr.dev with a group_size of 50 run to completion. It took 54.5 minutes. :D But it peaked at only about 115 MB of RAM. :)
[13:37] <Peng_> lifeless: Also FYI, group_size = 100 took 33:00.01 and used 122 MB of RAM.
=== sdboyer|vurk is now known as sdboyer|class
[13:59] <rocky> hey... i understand what the push and pull branches that a remembered are, but what is the "submit" branch?
[13:59] <Peng_> rocky: For "bzr send".
[14:00] <rocky> so the first time i pull in merge info from a particular source branch, it gets remembered as the "submit" branch?
=== Mario_ is now known as pygi
=== mw|out is now known as mw
=== bac` is now known as bac
[15:51] <lamont> bzrlib/_dirstate_helpers_c.c:7704: warning: dereferencing type-punned pointer will break strict-aliasing rules
[15:51] <lamont> shame on bzrlib
[15:51] <lamont> (one of many, grabbed at random)
=== kiko-zzz is now known as kiko
=== ja1 is now known as jam
[16:10] <Peng_> lamont: Shame on Pyrex.
[16:11] <lamont> ok
[16:11] <lamont> only 104 occurances of type-punned in the log :-(
[16:11] <Peng_> lamont: There are generally fewer warnings with newer versions of Pyrex.
[16:12] <lamont> yeah, that's building 1.9 on hardy
[16:12] <Peng_> But I don't remember ever seeing "type-punned".
[16:12] <lamont> google will tell you what gcc really means there.
[16:12] <lamont> "ZOMG this might not do what you want in some strange corner-case"
[16:12] <lamont> or something like that
[16:14] <jdong> it's a clever feature, not a bug (tm)
[16:14] <lamont> jdong: DENIED
[16:33] <Kinnison> the warning is completely valid
[16:33] <Kinnison> just compile -fno-strict-aliasing if you don't want it
[16:56] <darkness51> Hola a todos, como les va?
[16:56] <darkness51> tengo una duda, cuales son los comandos para usar bzr en forma distribuida
[16:57] <Kinnison> darkness51: ingles?
[16:59] <Jc2k> morning thumper_laptop
[17:03] <darkness51> Kinnison: i am understand english, but i don't write fine
[17:04] <darkness51> what command use form distribute work
[17:05] <Kinnison> darkness51: 'bzr push' can be used to upload your branch to another location. E.g. bzr push sftp://my-fileserver/home/me/projects/mything
[17:08] <darkness51> what is the use of bzr pull??
[17:10] <Kinnison> The 'pull' command lets you fetch changes from a remote branch into your local one. E.g. Pretend I have a project 'proj' and I work on my laptop and my desktop. I have a server and I keep a branch on there for synchronising.
[17:10] <Kinnison> I make a change on my desktop and push it to the server with 'bzr push'
[17:10] <Kinnison> Then on my laptop I can do 'bzr pull' to get the change onto my laptop
[17:11] <Kinnison> Naturally this requires you to tell it the location of the server first time. After that it will remember for itself.
[17:13] <darkness51> ok, thank's Kinnison
[17:14] <elmo> tuna 17:13 ~/scratch % bzr branch lp:~ubuntu-core-dev/livecd-rootfs/trunk/
[17:14] <elmo> You have not informed bzr of your launchpad login. If you are attempting a
[17:14] <elmo> write operation and it fails, run "bzr launchpad-login YOUR_ID" and try again.
[17:14] <elmo> can't it tell? :-P
[17:15] <fullermd> Well, not with the brain of a tuna   :p
[17:15] <Odd_Bloke> My Launchpad login isn't the same as my local login.
[17:18] <beuno> Odd_Bloke, maybe you need to set it in .ssh/config
[17:18] <beuno> Host bazaar.launchpad.net
[17:18] <beuno> User Odd_Bloke
[17:18] <beuno> or whaeva'
[17:18] <elmo> Odd_Bloke: I meant, can't it tell if it's a write operation
[17:18] <Odd_Bloke> elmo: Ah, I see.
[17:18] <Odd_Bloke> Not easily, IIRC. :)
=== mark1 is now known as markh
[17:19] <james_w> not with the current layering
[17:30] <jam> morning vila
[17:30] <jam> morning #bzr
[17:30] <jam> markh: what are you still doing up? :)
[17:32] <jam> elmo: We can know that "bzr branch" is a readonly operation. I believe the problem is that people often do a readonly operation, and then shortly thereafter do a write operation which fails, but we have lost the info about the LP url they used to get the branch.
[17:32] <jam> The most common one is:
[17:32] <jam> "bzr co lp:foo"
[17:32] <jam> cd foo
[17:32] <jam> bzr commit -m "this will fail, because of something about HTTP"
[17:33] <vila> jam: morning jam, evening jam (jetlag making me falling asleep again :-/)
[17:34] <jam> vila: sorry to hear that. I do believe I had a harder time coming home than I did going to Sydney
[17:34] <jam> I'm not sure if it is heading East
[17:34] <jam> or if it is coming home to a baby with an ear-infection
[17:35] <jam> my guess is the latter
[17:35] <abentley> Odd_Bloke: if you're using bzr 1.9, you don't need to configure ssh.  At most, you need to launchpad-login again, but any lp: lookup will do that for you.
[17:35] <vila> baby's ear-infections are bad for dad's sleep for sure :-/
[17:36] <abentley> elmo: And no, it can't tell.  We don't "open for write" or "open for read".  We just open, and after that we're dealing with real URLs.
[17:37] <vila> jetlag seems better than last week though, but I have a lot of sleep to catch up since I almost *not* sleep from Saturday morning to Monday morning (including the full day in Singapore as an added bonus for delayed flight :)
[17:37]  * elmo trollingly proposes rich-rich-root which records the original lp: url
[17:37] <elmo> abentley/jam: fair enough :-p
[17:39] <jam> abentley: then again, that may change in the future. We are looking at having a Branch.open_or_create_locked() which would allow us to know that we will be writing to it.
[17:39] <jam> elmo: And yes, we've discussed recording the original lp: url rather than its redirection
[17:40] <jam> vila: ouch on the extra day. I at least managed ~4 hrs of sleep on the plane, and I arrived home at 9pm, which is a great time to reset your internal clock and go straight to bed :)
[17:44] <vila> jam: well, I did sleep well last night, so hopefully I will too this night and be ok tomorrow :)
[17:44] <jam> So did you manage to get out of the Airport in Singapore, or was it just 24-hours of airport-chair pain?
[17:55] <mthaddon> hi folks, I'm getting the following error from config manager - "bzrlib.errors.TagsNotSupported: Tags not supported by BzrBranch5('file:///blah/blah/blah); you may be able to use bzr upgrade." using bzr 1.6.1 - should I "bzr upgrade" that directory?
[17:55] <jam> mthaddon: It sounds like you have a remote source with a newer format than the local source.
[17:56] <jam> And the remote source also has tags included.
[17:56] <jam> (which we are failing to fetch)
[17:56] <jam> I would recommend upgrading to whatever the remote format has.
[17:56] <jam> (You can upgrade even further, but the safe bet is to just stay in step)
[17:56] <mthaddon> jam, how to I verify the remote format version - bzr info?
[17:56] <jam> I'm not sure about how CM works, but perhaps just deleting the branch will recreate it at the right format.
[17:57] <jam> bzr info URL should work, IIRC
[17:57] <mthaddon> thx
[17:57] <beuno> if it's a remote branch, just with http to get the proper format
[17:57] <beuno> bzr+ssh always says nonsense
[17:59] <Peng_> You could use nosmart+bzr+ssh to get the real info.
[18:05] <mthaddon> bzr info bzr+ssh says "Standalone branch (format: unnamed)", bzr info nosmart+bzr+ssh says "Standalone branch (format: pack-0.92)"
[18:06] <Peng_> mthaddon: Run "bzr upgrade" in your local branch then.
[18:06] <vila> jam: well, the airport-chair pain was in Brisbane where they announced 1-hour delay... 7 times :-)
[18:07] <mthaddon> Peng_, that seems to have fixed it - thx
[18:07] <jam> vila: what airline?
[18:07] <jam> and yeah, ouch
[18:07] <jam> if they told you in advance it was going to be 7 hours, you could have at least done something
[18:07] <jam> I guess you at least weren't already *on* the plane
[18:08] <vila> when we finally got to Singapore, the connecting flights were long gone so they transfered us to an hotel and re-book the same flights only 24h later
[18:08] <jam> There was something about JetBlue in the states doing that
[18:08] <vila> Quantas from Brisbane
[18:08] <jam> http://gettingtomaybe.blogspot.com/2007/02/jet-blue-delay.html
[18:08] <jam> Quote: News broke last week that passengers on Jet Blue flights were subjected to 10 hour delays inside the plane, while on the runway.
[18:08] <tetha_> ouch, thats even worse than my regular train delays
[18:08] <jam> It was a big deal, as they didn't even have food for the passengers, etc.
[18:09] <jam> Since they are budget flights
[18:09] <vila> The hotel were nice, but I was so sleep deprived already that I only half enjoyed the sight-seeing in Singapore :-)
=== mvo__ is now known as mvo
[18:10] <vila> Quantas did pretty well, giving vouchers in Brisbane for refreshments and so on, the hotel was nice too (even if I ended up in a smoking *level*, ha ha the irony)
[18:10] <tetha_> oops, now I did it. my proof for tomorrows algorithm class finally brake apart
=== mw is now known as mw|food
=== beuno_ is now known as beuno
=== fta_ is now known as fta
[19:40] <jelmer> jam, What happened to your "This week in Bazaar" posts?
[19:44] <lamalex> anyone experience with LP here
[19:44] <lamalex> nm, i'll go to #launchpad
[19:44] <lamalex> sorry
=== mw|food is now known as mw
=== thumper_laptop is now known as thumper
[20:21] <jelmer> crap, I just realized I'm halfway through reimplementing Subversion in Python..
[20:24] <LarstiQ> jelmer: ai
[20:24] <fullermd> A painful realization, to be sure.
[20:25] <fullermd> 'course, you could always reimplement half of git in python to recover  :p
[20:25] <dobey> or you could just fix bugs in bzr and make it fast
=== kiko is now known as kiko-afk
=== bac is now known as bac_afk
[20:57] <jml> lifeless: sweet
[20:58] <jml> lifeless: does it solve the TSP or what?
[20:59] <jml> bzr alias upll="pull"
[20:59]  * LarstiQ has bpvl aliased
[21:03] <thumper> LarstiQ: to what?
[21:04] <LarstiQ> thumper: bzr pull -v | less
[21:04] <mwhudson> jelmer: because writing libsvn bindings is too painful?
[21:05] <thumper> jelmer: how is the bzr-git work going?
=== Mario_ is now known as pygi
[21:36] <jelmer> thumper: Pretty good; I'm working on two things at the moment - getting more familiar with git's pack format to be able to implement that for bzr-git and factoring out code that is shared by bzr-svn and bzr-git
[21:40] <james_w> jelmer: I found git's pack format to be pretty clear when I looked, how are you finding it?
[21:44] <jelmer> james_w, Yeah, it's quite simple
[21:46] <jelmer> btw, I was thinking of renaming the python-git code you originally wrote to "Dulwich"
[21:46] <jelmer> (after the Monty Python sketch about a cocktail party there featuring a Mr and Mrs Git)
[21:46] <james_w> heh :-)
[21:50] <thumper> jelmer: have you talked with Jc2k about git and python?
[21:50] <thumper> jelmer: I think he did something...
[21:50] <thumper> jelmer: or perhaps he was working in C, I'm not entirely sure
[21:50] <jelmer> thumper, Yeah, we talked about it a bit during the summer
[21:50]  * thumper nods
[21:50] <jelmer> thumper, That was indeed in C - based on the upstream git code
[21:51] <thumper> ah ok
[21:52] <eydaimon> Unable to obtain lock chroot-14417232:///srv/bzr/ ... If you're sure that it's not being modified, use bzr break-lock chroot-14417232:///srv/bzr/.... $ bzr break-lock chroot-14417232:///srv/bzr/ ... bzr: ERROR: Unsupported protocol for url "chroot-14417232:///srv/bzr/... that's not terribly useful
[21:52] <eydaimon> how can I get it to break this lock?
[21:52] <jelmer> eydaimon, I think the error message may be a bit clearer in newer versions of bzr
[21:52] <jelmer> eydaimon, generally, you should be able to run "bzr break-lock" on the URL that was giving you problems
[21:53] <abentley> jelmer: Really?  I thought it was still leaking underlying paths in the latest versions.
[21:53] <eydaimon> jelmer: but it says unsupported protocol for url
[21:53] <eydaimon> how do I deal with that?
[21:53] <jelmer> eydaimon, You should be able to run "bzr break-lock" on the URL you were trying to pull from/push to/branch from/etc
[21:54] <eydaimon> ok
[21:54] <abentley> eydaimon: You ignore the URL it specifies.  Use the normal URL for the branch.
[21:54] <jelmer> abentley, Sorry, I must be mistaken then
[21:54] <jelmer> ISTR some patch for this going in
[21:55] <eydaimon> jelmer: thanks, that fixed it
[21:56] <jelmer> eydaimon: You're welcome. Hopefully the URL in that error message will get fixed at some point.
[21:56] <eydaimon> that's probably helpful :)
=== bac_afk is now known as bac
[22:11] <Lethalman> hello
[22:11] <Lethalman> I have commit A -> B locally, then pushed to the remote repository
[22:12] <Lethalman> I then uncommitted B and committed C
[22:12] <Lethalman> when I push it says of course that branches diverge
[22:12] <Lethalman> how can I force to push A -> C?
[22:12] <beuno_> Lethalman, push --overwrite
[22:12] <bob2> push --overwrite
[22:12] <bob2> if you're sure
[22:12] <Lethalman> ah
[22:12] <Lethalman> yes i'm sure
=== beuno_ is now known as beuno
[22:13] <bob2> or merge from then push to
[22:13] <Lethalman> thanks very much
[22:34] <rockstar> abentley, can you give me a rundown of the proper way to push a loom to launchpad?
[22:39] <abentley> rockstar: the loom must be in format 6.  Then pushing it should work, but won't stack.
[22:41] <rockstar> abentley, I apparently forgot to record before I pushed.  So I deleted the pushed branch, recorded, and pushed again.
[22:41] <rockstar> But when I pull from another system, the loom appears to only have one thread.
[22:42] <abentley> rockstar: I'm the wrong guy to answer this, because I never push looms on purpose.
[22:42] <rockstar> abentley, :)
[23:07] <jfroy|work> jelmer: do I need to set some things up to run bzr-svn's unit tests?
[23:07] <jfroy|work> I'm getting failures across the board
[23:07] <jelmer> jfroy|work, no, just run "bzr selftest --starting-with=bzrlib.plugins.svn"
[23:12] <jfroy|work> jelmer: eh, it didn't even make it to the end :p
[23:12] <jfroy|work> OSError, "too many files open"
[23:13] <jelmer> I've never run the testsuite on Mac OS X, so I'm not sure how well it works there
[23:13] <jfroy|work> Apparently it massively fails.
[23:13] <jfroy|work> Hum, any known issues with Python 2.6?
[23:14] <jelmer> oh, yeah, python2.6 still has issues
[23:14] <jelmer> sqlite3 has slightly different behaviour, for example
[23:14] <jfroy|work> right
[23:14] <jfroy|work> I'll try to run the tests on 2.5
[23:15] <jelmer> I'll be back in ~45 minutes
=== sdboyer|class is now known as sdboyer
[23:24] <markh> jam: you here and got a sec?
=== jamesh_ is now known as jamesh
[23:45] <jfroy|work> jelmer: sent you a merge bundle as the first parter of a workaround for my situation
[23:58] <thrope> is it possible to use local svn repo with bzr-svn