[00:24] <davidstrauss> I got annoyed with the lack of up-to-date RHEL/CentOS RPMs and built some. Anything I should do with them?
[00:27] <igc> mwhudson: yes. I'm looking to do that
[00:27] <mwhudson> igc: cool :)
[00:36] <mib_65qsfnbw> I installed bazaar but it seems that bzr dosent work behind proxy with authentication...Is there any way to make bzr work?
[00:53] <bob2> does including the password in http_proxy work?
[00:58] <jbalint> how do i cancel a merge? (bzr status stil shows all pending merges0
[01:00] <mwhudson> jbalint: bzr revert
[01:00] <mwhudson> bzr revert --forget-merges if you want to keep the text changes
[01:01] <jbalint> ah, reverted already , needed that forget-merges. thanks
[01:02] <mwhudson> revert . will revert the text changes but not the merges
[01:02] <jbalint> ok
[01:03] <mwhudson> 'revert' on its own does both
[01:03] <jbalint> oh i see
[01:10] <alf> hello, I just branched bzr.dev and tried to upgrade it to --1.9-rich-root but the upgrade just hangs with 99% CPU
[01:11] <bob2> that might be a bug, but why a rich root format?
[01:12] <alf> bob2: for no specific reason, just playing around
[01:14] <alf> I have found https://bugs.launchpad.net/bugs/177874 but in that case there is an explicit error (and the bug its quite old)
[01:14] <alf> ubottu: you read my mind :)
[01:22] <davidstrauss> First stab at updated RHEL/CentOS packages: http://fourkitchens.com/blog/2009/01/21/bazaar-111-bzrtools-1110-rpms-red-hat-enterprise-linux-5-centos-5
[01:22] <davidstrauss> They seem to work.
[01:37] <jbalint> is file_id format depends on the repo/branch format?
[01:45] <thumper> jbalint: I think so
[01:45] <jbalint> would it ever change by branching?
[02:07]  * igc lunch
[02:07] <mib_65qsfnbw> bob2: i tried http proxy but its saying that bzr: ERROR: socket.error: (111, 'Connection refused')
[02:07] <mib_65qsfnbw> bob2: also http://www.mibbit.com/pb/Z1UtTw
[02:08] <mib_65qsfnbw> so i filed a bug report
[02:26] <lifeless> would anyone object to my putting back the set_plugin_path on import of bzrlib?
[02:27] <lifeless> its really annoying not having it
[02:27] <beuno> hi lifeless
[02:27] <lifeless> hi beuno
[02:27] <beuno> what would be the consecuences of doing that?
[02:27] <beuno> I know it breaks stuff not having it
[02:29] <lifeless> I don't see any consequnces
[02:29] <lifeless> bzr --no-plugins shouldn't be loading stuff anyhow, and there is an omplicit plugin path anyhow
[02:30] <beuno> hard to object to that  :)
[02:32] <jbalint> anybody have a tip on where to find svn2bzr? this page is leading nowhere  http://bazaar-vcs.org/svn2bzr
[02:33] <mib_65qsfnbw> Please help me in getting this removed: http://www.mibbit.com/pb/Z1UtTw
[02:50] <lifeless> mib_65qsfnbw: I'm not quite sure what you're asking
[02:50] <mib_65qsfnbw> I installed bazaar but it seems that bzr dosent work behind proxy with authentication...Is there any way to make bzr work?
[02:52] <lifeless> the lp: syntax doesn't work, but all our other http stuff should work fine
[02:52] <lifeless> as far as I know that is
[02:53] <mib_65qsfnbw> lifeless: i am newbie to this bzr thing can you be please more elaborative ( what is http stuff)?
[02:53] <lifeless> I'm sorry, I have a very bad connection here; I'll let someone else elaborate
[02:53] <lifeless> igc: perhaps you are around and could help mib_65qsfnbw ?
[02:53] <mib_65qsfnbw> lifeless: fine thankyou
[03:04] <jearl> mib_65qsfnbw: I believe that what lifeless is saying is that while you can't check out the code using bzr clone lp:geany
[03:04] <jearl> mib_65qsfnbw: you might be able to check it out using
[03:05] <jearl> mib_65qsfnbw: bzr clone https://code.launchpad.net/~vcs-imports/geany/trunk geany
[03:06] <jearl> mib_65qsfnbw: my own checkout hasn't finished, so I can't be sure :)
[03:12] <mib_65qsfnbw> jearl:  "bzr branch http://bazaar.launchpad.net/~vcs-imports/geany/trunk" this one wotked for me
[03:13] <mib_65qsfnbw> I got it by googling
[03:14] <mib_65qsfnbw> jearl: Ho would I know which packages can be got using the above command( bzr branch http://bazaar.launchpad.net/~vcs-imports/genay(or any othe app)/trunk )
[03:15] <igc> lifeless, mib_65qsfnbw: can't help right now sorry (heading to the hospital for the rest of the day)
[03:15] <igc> see everyone tomorrow
[03:19] <thumper> if the cbranch command does a branch then checkout, is there a sbranch (or something) that does branch and switch?
[03:20] <fullermd> mib_65qsfnbw: Well, all of them.  lp:whatever is just a shortcut that looks up the actual URL.  It's just figuring out what the URL is...
[03:22] <fullermd> You can probably figure it out by looking at the project on LP with a browser.
[03:23] <mib_65qsfnbw> fullermd: OK
[03:23] <mib_65qsfnbw> gotta go I have a lecture to attend see you in 2 hours...
[04:10] <jbalint> how do i get svn-import to create separate bzr branches for my svn branches? I just get on bzr branch with each svn branch and trunk in the bzr branch
[04:12] <lifeless> jbalint: have you checked the faq?
[04:12] <jbalint> this one? http://samba.org/~jelmer/bzr-svn/FAQ.html
[04:12] <jbalint> i try separately --scheme and bzr svn-branching-scheme --set
[04:13] <jbalint> but it seems each time i run svn-import it reverts it to none in ~/.bazaar/subversion.conf
[04:17]  * AfC wonders why bzr is so slow on his (borrowed) Ubuntu system.
[04:18] <AfC> I tried --no-plugins to see if that made a difference, but only about 0.08s
[04:18] <AfC> (just taking times from .bzr.log)
[04:19] <AfC> whereas bzr status on a 2 revision 5 file tree is taking 1.2s
[04:19] <AfC> Maybe I just didn't notice, but it sure feels slower than my other system. All things are not equal, of course, but...
[04:19] <AfC> bzr 1.11 from PPA
[04:21] <fullermd> Does it have the C stuff compiled?  Maybe that's missing...
[04:21] <lifeless> AfC: is it a bound branch ?
[04:22] <lifeless> 1.2s smells like a network round trip on this network
[04:23] <davidstrauss> The operation mysteriously permeated the air with the lingering stench of latency.
[04:23] <davidstrauss> :-)
[04:27] <AfC> fullermd: I wondered about that, but `dpkg -L` showed a bunch of .so files, so...
[04:28] <AfC> lifeless: no. Just a straight up been-working-on-this-brand-new-branch-for-only-an-hour-or-two.
[04:28] <AfC> lifeless: (on the assumption that it doesn't change, I'll show you later in case you care to poke it)
[04:29] <AfC> lifeless: (you have better things to do)
[04:30] <davidstrauss> What's the process for getting packages listed on the downloads page? I know I can technically edit it, but I'm curious if there's some procedure for vetting the sources.
[04:42] <Ycros> hey guys, I'm trying to pull in an external (svn) library into my project, and I'm wondering if branching it with bzr-svn, upgrading my repo format to 1.9-rich-root and bzr joining it in is the sane approach, or if there's a better way
[05:00] <lifeless> AfC: please do show me
[05:01] <AfC> lifeless: will do. I'm up the hill right now, but next time I cross paths with you I'll mention it.
[05:02] <AfC> Ycros: that's an interesting question; I imagine there's no right answer, so you're probably best off to try a few different ways and see what you end up with.
[05:02] <AfC> Ycros: (and then you'll be the expert! Ta-da, sucker)
[05:03] <Ycros> I'm trying things out now
[05:03] <Ycros> I think bzr join --reference might be what I'm looking gor
[05:04] <fullermd> Apparently, it's nested-tree-week on #bzr...
[05:09] <Ycros> maybe... hmm. What I really want is something like svn:externals. I'm not just what join --reference just did, it looks like the external tree now appears as a bunch of unknown changes
[07:03] <vila> hi all
[07:20] <davidstrauss> vila: hi
[09:03] <ralph> Hi, can a revision's comment be changed?  Like    rcs -mrev:msg   ?  In a corporate setting, for political reasons, this can be required.  Or, as just now, I've put in a very misleading comment.
[09:04] <ralph> The manual makes no mention of it that I can find, and I fear it's not possible;  the message being part of a revision's unique key in some way?
[09:11] <luks> it isn't
[09:11] <luks> if it's the last revision, you can uncommit and recommit it
[09:12] <luks> otherwise you can do some trickery, but you will end up with technically a different branch (at least since the changed revision)
[09:12] <ralph> Will that muck others up?  I've done a `bzr push' with the wrong comments in?   At the moment, I'm planning to do several uncommits in turn, saving the diffs as a patch, and then work forward again.
[09:14] <luks> if anybody has a mirror of the branch, they will end up with diverged branches
[09:14] <luks> the uncommitted revisions will *look* the same, but they will be different for bzr
[09:14] <ralph> Thanks luks.  Does that stop simple `bzr merge' from working?  There's just the two of us.  I push and he sends back to me occasionally.
[09:15] <luks> it doesn't, but then you will have two copies of each revision in the branch
[09:15] <ralph> k, think it's worth doing to clarify comments.
[09:15] <fullermd> Including the one with the wrong message.  Kinda defeats the purpose, especially in a nuclear scenario.
[09:15] <luks> if you are just two, you can maybe tell him to pull --overwrite
[09:16] <luks> oh, right, forgot about that :)
[09:16] <ralph> Thanks, will look into the --overwrite option.  Basically, if he's no local changes then he can just --overwrite.
[09:16] <fullermd> If you can talk other people into --overwrite'ing, it doesn't have THAT wide an impact.  Where it really gets ugly is when they've already started work on top of the old stuff.
[09:16] <luks> yep
[09:18] <luks> btw, bzr replay from bzr-rebase can save you some time doing the uncomitting
[09:18] <ralph> Thanks.
[09:36] <vila> try again guilhembi.. I want to ping you :)
[09:37] <guilhembi> vila: hello! Yesterday evening my internet connection went down, and this morning it just went up a few seconds ago.
[09:37] <guilhembi> After a second reboot of the livebox
[09:37] <vila> here you are !
[09:37] <vila> change ISP :)
[09:37] <guilhembi> vila: bah
[11:39] <glatzor> hello james_w, is it possible to use bzr bd inside of a python script?
[11:40] <james_w> hey glatzor
[11:40] <james_w> you could import it and run the various parts
[11:40] <james_w> unfortunately there is quite a lot of logic in the Command class
[11:40] <james_w> or you could just run it under subprocess
[14:13] <lampliter> are there any bzr plug-ins for netbeans?
[14:13] <LarstiQ> yes, I believe so
[14:14] <lampliter> interesting.  Any pointers?  Using Google didn't reveal anything terribly useful
[14:15] <lampliter> for what it's worth, I discovered that Emacs plus tramp is bzr aware
[14:16] <LarstiQ> hmm, don't see it on the wiki at BzrPlugins, not at launchpad.net/bazaar
[14:16] <lampliter>  and is not listed at net beans
[14:16]  * LarstiQ trwals the mailing list
[14:16] <jelmer> I don't think it was in a usable state yet though
[14:17] <LarstiQ> lampliter: people have certainly talked about it on the mailing list
[14:17]  * LarstiQ fails to find anything concrete right now
[14:17] <lampliter> okay, I'll check there
[15:11] <beuno> jam, I know I said it already, but having --short as my default changed my life completely  :)
[15:11] <jam> beuno: I'm glad you like it as much as I do
[15:11] <jam> I wish we could get away with making it the default
[15:11] <sjokkis> hi, i'm trying to create a branch of my files in trunk. am i supposed to copy the files using the filesystem, or is there some bazaar-command that eludes me here?
[15:11] <jam> without encountering the "not similar to everyone else" problems
[15:12] <jam> sjokkis: "bzr branch" ?
[15:12] <sjokkis> i tried that, but it tells me that trunk isn't a branch
[15:13] <sjokkis> "bzr: ERROR: Not a branch: "/blah/blah/project/trunk/"."
[15:14] <sjokkis> so what gives, jam?
[15:14] <beuno> sjokkis, what command are you running?   sounds like you're adding extra stuff to the path
[15:14] <sjokkis> i ran "bzr branch . ../branches/port" from inside trunk
[15:15] <sjokkis> i can do this from inside my local checkout, right?
[15:15] <beuno> sjokkis, you probably shouldn't use .
[15:15] <beuno> so if you're inside trunk,  bzr branch ../trunk ../branches/port
[15:16] <sjokkis> i get the same error doing that
[15:18] <sjokkis> i think this was originally imported from svn using the bzr2svn utility. if that matters
[15:18] <sjokkis> eh, svn2bzr
[15:21] <ollie> hey wondering fi anyone could give me any advice or thoughts on a workflow setup for web development
[15:21] <ollie> i have setup a repository on a dev box and given each of my developers vhost enabled checkouts of the trunk of my project
[15:21] <ollie> so everyone does updates and commits from their own checkouts
[15:22] <ollie> this just gives some simple separation
[15:22] <ollie> I have not looked into setting up branches yet
[15:24] <ollie> I now want to publish this trunk to a live box. I thought i could just do my live site as a checkout of this trunk
[15:24] <ollie> and then manually do a bzr update from the live box to get the latest revisions
[15:24] <ollie> i would then setup bzrignore to ignore the server specific settings files
[15:25] <ollie> wondering if this is a good approach or whether someone could suggest something more suitable
[15:27] <ollie> doh nobodies home :(
[15:30] <sjokkis> beuno, jam: neither of you guys can help me figure out how to do this?
[15:31] <jam> sjokkis: sorry, I'm a bit distracted right now.
[15:31] <jam> To start with, figure out where the actual branch is at
[15:31] <jam> cd trunk; bzr info
[15:31] <jam> It sort of sounds like things may not be what you think
[15:31] <NfNitLoop> ollie: That sounds like a reasonable workflow.
[15:31] <jam> like maybe the branch is "trunk/foo"
[15:31] <jam> or the branch is "project" and "trunk" is just a directory underneath
[15:31] <sjokkis> seems the branch is one level above trunk/
[15:32] <sjokkis> so yeah, the branch is project/
[15:32] <NfNitLoop> ollie: If you're not worried about having a working copy on your live site, you could just "bzr push" there instead of pulling from there.
[15:32] <sjokkis> that's fucked up...
[15:32] <NfNitLoop> OH, wait, by "live site" you mean you're checking out working code (PHP or something?) into a web site.
[15:32] <ollie> NfNitLoop: yeah
[15:32] <ollie> through the cms extra files e.g. images will be added
[15:32] <sjokkis> jam: what's the best way to sort this out?
[15:33] <ollie> so i think i will have to merge new changes in
[15:33] <jam> sjokkis: it sounds like you may want to do another conversion, but using "bzr-svn" instead of svn2bzr
[15:33] <jam> specifically, if you have the bzr-svn plugin installed
[15:33] <jam> you then run "bzr svn-import" to do the conversion
[15:33] <NfNitLoop> ollie: well, if all you're ever doing is 'bzr pull' from your "trunk" branch, you shouldn't need to do any merging.
[15:34] <sjokkis> well i've already done quite a bit of work since i moved over to bzr, jam
[15:34] <sjokkis> jam: so if possible i'd like to keep my current repository. or export it. something
[15:35] <ollie> NfNitLoop: ok. One thing i was thinkning was giving my developers each a branch instead of them all doing checkouts of the trunk. However i havn't much experience with merging from one branch to another
[15:35] <jam> there would certainly be ways to get your newer changes brought across
[15:35] <jam> there are other things we could do
[15:35] <jam> but I have the feeling you really don't want to have things in your current layout
[15:36] <jam> and without re-converting, it will be problematic to merge between branches, etc
[15:37] <sjokkis> well i'll be happy to create a new layout etc
[15:37] <sjokkis> i just need to keep my history etc
[15:37] <sjokkis> etc
[15:37] <james_w> if I merge 2 unrelated branches by just creating the revision with the appropriate parents then I can continue to use both branches and merge between them, is that correct?
[15:38] <james_w> it's only the UI that prevents me from doing this by forcing me to do a cherry-pick merge, which leaves the branches with distinct revision histories?
[15:38] <jam> sjokkis: so what I would recommend would be to try a new conversion, and see if that gets the layout you want
[15:38] <jam> (don't delete your existing one yet)
[15:38] <jam> and then we'll look at carrying across your more recent changes
[15:41] <sjokkis> jam: i don't have root on the machine my repo is on. can i do this on another machine and move the repo afterwards?
[15:41] <jam> sjokkis: As far as I know you can
[15:42] <jam> I'm not sure what access you need to the SVN repo, but I'm pretty sure just read access
[15:42] <jam> as you can use bzr-svn with repos that you don't have write access to
[15:43] <sjokkis> jam: actually screw this. it's a smallish project, and i'm the only one working on it. we can just copy the files manually
[15:43] <sjokkis> and forget about the history
[15:56] <ollie> should the .bzr file show up as modified under bzr st. I would of thought that bzr wouldn't include this in its version control
[15:57] <Peng_> It doesn't track the ".bzr" directory.
[15:58] <Peng_> It will track files like ".bzrignore" or ".bzr-whatever-you-want".
[15:58] <ollie> hmm
[15:59] <Peng_> You sure it's actually ".bzr", not some sort of case-manipulated variation, or some Unicode oddity?
[16:00] <ollie> well i made it using bzr ignore ./config
[16:00] <ollie> so i guess bzr would of created the file in the correct way
[16:00] <Peng_> You mean the file ".bzrignore"?
[16:00] <ollie> just suprised me that this is included in the version historuy
[16:00] <ollie> ye
[16:00] <ollie> s
[16:00] <Peng_> Oh. Yeah, it's versioned.
[16:01] <ollie> thats a problem as i need to ignore different things on different systems
[16:02] <ollie> so is that not possible
[16:05] <Peng_> ollie: You don't *have* to version it if you don't want to. 'bzr rm --keep .bzrignore && bzr ci .bzrignore -m 'oops, stop tracking .bzrignore'"
[16:05] <Peng_> s/'/"/
[16:05] <ollie> is bzr ci the same as bzr commit?
[16:05] <Peng_> Yeah.
[16:05] <Peng_> It's an alias.
[16:06] <ollie> ok
[16:09] <garyvdm> ollie: Why would you want to ignore different things on different systems?
[16:10] <ollie> its complicated :(
[16:13] <awilkins> Anything in the inventory already will continue to be tracked though
[16:14] <awilkins> Regardless of your ignore settings. And whether you ignore them or not :-)
[16:23] <jam> ollie: If you need to ignore things on different locations you can:
[16:23] <jam> a) ignore the things in both places
[16:24] <jam> (if you are ignoring foo in place X and bar in place Y, ignoring foo and bar doesn't hurt anything)
[16:24] <jam> b) Use the global ignore file in ~/.bazaar/ignore
[16:24] <jam> which you can customize based on the machine you are no
[16:24] <jam> on
[16:24] <ollie> ok brilliant thanks a lot jam
[16:24] <jam> c) If you ignore something and the "bzr add X" it will still be versioned, because versioning something supersedes ignoring it
[16:54] <jrwren> I'm a little confused on send/merge and the docs aren't helping. Can I do a send & merge entirely offline (sneaker net) if a co-worker and I have branches with a shared common ancestory?
[16:55] <LarstiQ> jrwren: yes
[16:55] <LarstiQ> jrwren: and provided you include enough information in your merge directives
[16:56] <davidstrauss> jrwren: bzr send -o filename
[16:56] <davidstrauss> jrwren: and then bzr merge filename
[16:56] <jrwren> i did bzr send -r REVNOWEBOTHHAVE.. -o filename
[16:56] <jrwren> it fails on merge
[16:56] <Peng_> What's the error message?
[16:57] <jrwren> http://pastebin.ca/1314855
[16:57] <jrwren> that was supposed to be a readable version of revision number we both have...
[16:58] <Peng_> Does he really have it? "bzr log -r revid:ben@ben-pc-20090121162140-gx1ag11qswfwp556"
[16:59] <davidstrauss> ooh, traceback
[16:59] <davidstrauss> that's not supposed to happen
[16:59] <davidstrauss> jrwren: have you confirmed the issue in bzr 1.11?
[16:59] <jrwren> i have not. I'll try it now.
[17:00] <jrwren> nope, I don't have that rev in my branch.  AFAIK, that is the rev to which I'm trying to get.
[17:04] <jrwren> i skipped moving to 1.11 because it is rc1 a few days ago. I guess I should have done it then.
[17:15] <jrwren> same error with 1.11rc1 http://pastebin.ca/1314868
[17:15] <jrwren> maybe I need to include more information in my merge directives, but I don't know what that might be.
[17:19] <jam> jrwren: my guess is that you don't have a proper mirror of the branch you are trying to merge into
[17:19] <jrwren> normally use a centralized workflow, we commit/update from the same branch in the same repo.
[17:19] <jam> "send" needs 2 branches
[17:20] <jam> (not by my choice)
[17:20] <jam> so you need 1 branch which is exactly what you started from
[17:20] <jam> and then another which is the new stuff
[17:20] <jrwren> why does send need 2 branches? can I achieve similar behavior with 1 branch ?
[17:20] <jrwren> ah...
[17:20] <jrwren> so he can branch -r 323 (our shared last match) and use that as the shared branch ?
[17:21] <jam> basically, abentley felt that people made too many mistakes when trying to use 1 branch, so opted to not allow it
[17:21] <jam> jrwren: right
[17:21] <jrwren> thank you.
[17:21] <jrwren> jam: this is the 2nd time you were of HUGE help to me.
[17:21] <jam> I'm not sure whether I hope it is the last or not :)
[17:21] <jam> I hope you don't need the help, but I'm always happy to help
[17:24] <beniwtv>  hi all... I'm trying to check out a branch that lives on one of my servers, doing a bzr branch http://url/nameofbranch/ but it gives me ERRROR: <url> not a branch. It traced down that bzr wants to access the .bzr dir that loggerhead doesn't export and gives 404. So now I would like to know how to configure loggerhead and bzr properly so I can get my branches? Is there anything I'm missing?
[17:25] <davidstrauss> beniwtv: are you trying to check out from  loggerhead page?
[17:25] <beniwtv> davidstrauss: yes *ducks?* :)
[17:26] <davidstrauss> beniwtv: loggerhead does not give bzr the data it needs to branch
[17:26] <davidstrauss> beniwtv: it's just a human-usable browser
[17:26] <LarstiQ> beniwtv: are you running it behind apache?
[17:26] <LarstiQ> beniwtv: in that case you can configure apache to redirect .bzr to the right place
[17:26] <jelmer> beniwtv, there is a open bugreport in loggerhead about supporting this
[17:27] <davidstrauss> beniwtv: you may also want to look at "bzr serve" to allow public branching
[17:27] <beniwtv> LarstiQ: No, loggerhead only. Should I use it with apache? Any tutorials on this?
[17:27] <beniwtv> davidstrauss: OK, thanks, I'll have a look at that
[17:27] <davidstrauss> beniwtv: I typically use Loggerhead for human browsing and bzr serve behind xinitd for public branching.
[17:30] <davidstrauss> beniwtv: And I run Loggerhead behind an Apache proxy
[17:30] <beniwtv> davidstrauss: I see. However, I would need two ports, for loggerhead and bzr serve. So if I want to have only one port, is this possible? (I would like that for simplicity)
[17:31] <stavros2> hello
[17:31] <stavros2> how can i add the bzr ppa key so the repo is verified?
[17:32] <LarstiQ> beniwtv: what I do, is run loggerhead on localhost:8080
[17:32] <davidstrauss> beniwtv: Yes, you would need two ports
[17:32] <davidstrauss> beniwtv: To run bzr serve and Loggerhead
[17:32] <LarstiQ> beniwtv: and then use apache with ProxyPass for public access
[17:32] <davidstrauss> LarstiQ: Yes, but how do you combine that with allowing public branching? That's why I run a separate bzr serve?
[17:33] <beniwtv> LarstiQ, davidstrauss: OK, thanks... That will get me going I think :)
[17:33] <LarstiQ> davidstrauss: well, our 'public' branching goes via sftp://
[17:33] <davidstrauss> LarstiQ: But then it's not a smart server
[17:33] <LarstiQ> davidstrauss: correct
[17:34] <LarstiQ> (or bzr+ssh)
[17:34] <davidstrauss> I think it would be difficult to secure public branching over bzr+ssh
[17:34] <LarstiQ> davidstrauss: I would like to provide the convenience of branching from loggerhead urls
[17:34] <davidstrauss> LarstiQ: agreed
[17:34] <LarstiQ> davidstrauss: secure? You mean, control access of who branches what?
[17:35]  * LarstiQ is in a different environment :)
[17:35] <davidstrauss> LarstiQ: no, in terms of preventing anyone from changing the "public" account password, preventing port forwarding, etc.
[17:35] <LarstiQ> davidstrauss: so, for that, I'd use apache rewriting for http://loggerhead/branch/.bzr/
[17:35] <davidstrauss> LarstiQ: that would work
[17:36] <LarstiQ> davidstrauss: that rings even less of a bell. My situation is a bunch of developers in two rooms fully trusting everyone.
[17:36] <stavros2> can i authenticate bzr packages?
[17:37] <davidstrauss> LarstiQ: I'm doing things like hosting repositories for Drupal development
[17:37] <LarstiQ> stavros2: Bazaar releases are signed with gpg, is that what you mean?
[17:37] <LarstiQ> davidstrauss: aah. Right.
[17:37] <davidstrauss> When I say "public" branching, I mean by Joe Internet.
[17:37] <stavros2> LarstiQ: well, apt is telling me that they can't be authenticated and won't upgrade them
[17:37] <davidstrauss> No authentication
[17:38] <LarstiQ> davidstrauss: yeah, handing out shell accounts to Joe Internet is less of an idea :)
[17:38] <LarstiQ> stavros2: ah, apt. What repository are you using, the ppa?
[17:38] <davidstrauss> http:// is also not a smart server, right?
[17:38] <LarstiQ> davidstrauss: it can be with some work
[17:38] <davidstrauss> LarstiQ: I mean right now
[17:39] <LarstiQ> davidstrauss: doc/en/user-guide/http_smart_server.txt
[17:39] <stavros2> LarstiQ: yes
[17:39] <stavros2> LarstiQ: for intrepid
[17:39] <davidstrauss> I do have Loggerhead and bzr set up so you can check out http://vcs.fourkitchens.com/drupal/6 as bzr://vcs.fourkitchens.com/drupal/6
[17:39] <davidstrauss> It's not too much of a pain
[17:46] <luke-jr> how happy is bzr-svn with back-and-forth type stuff?
[17:46] <jelmer> luke-jr, you mean pushing back into svn and then pulling from svn ?
[17:46] <jelmer> luke-jr, it's quite happy about that (-:
[17:48] <luke-jr> actually, I mean using 'bzr diff;bzr commit' and then doing 'svn up; svn commit' ;)
[17:48] <luke-jr> I noticed 'bzr diff' in a Subversion WC actually worked
[17:48] <luke-jr> even after an extra 'svn commit'
[17:48] <luke-jr> but does it update the .svn info when I 'bzr commit'?
[17:48] <jelmer> yes
[17:49] <luke-jr> nice
[17:50] <luke-jr> if I 'bzr branch' from a svn URI, make a bunch of commits locally, then 'bzr push', do my commits get to Svn repo as if they happened when I committed them locally? ;)
[17:50] <jelmer> yes
[17:50] <luke-jr> and finally, will Bad Things (TM) happen if I tell bzr-svn not to use svn props for metadata?
[17:51] <jelmer> luke-jr: You can't do that when working in a .svn checkout
[17:52] <luke-jr> yes, different scenario
[17:52] <jelmer> luke-jr, if you do a push without svn props ("bzr dpush") then if you have a second copy of your bzr branch, it won't be able to pull from svn ("Branches have diverged")
[17:52] <luke-jr> I see :/
[17:52] <jelmer> luke-jr, you can also use bzr-svn 0.5 and svn 1.5 on the server side
[17:53] <luke-jr> ?
[17:53] <jelmer> that can use svn revision properties, which don't show up in "svn diff" and commit notification emails
[17:53] <luke-jr> ah
[17:53] <luke-jr> even over HTTPS?
[17:54] <jelmer> yes, the protocol is not relevant as long as it's Subversion 1.5 on the other side.
[18:02] <ollie> whats the process for merging treeless repo branches into a trunk? i have say /web/project/trunk/blog/ and /web/project/branches/tim/blog and  /web/project/branches/lucy/blog. The devs do checkouts of their respective branches into apache enabled directories so they can see their work at http://dev.lucy.project.com.
[18:02] <ollie> However I am not really clear on how I would merge the changes within the branches
[18:02] <ollie> and whether using branches like this is really the right approach
[18:04] <jam> ollie: well, you need a checkout to do a merge
[18:04] <jam> so you could do:
[18:04] <jam> bzr co /web/project/trunk
[18:04] <jam> cd trunk
[18:04] <jam> bzr merge /web/project/branches/tim/blog
[18:04] <jam> bzr commit -m "bring in Tim's changes"
[18:04] <jam> etc
[18:05] <jam> I'm not 100% sure if you are trying to automate something, or if I'm missing what you need
[18:05] <ollie> the key thing is it needs to be done from a checkout
[18:06] <luke-jr> bzr: ERROR: Invalid http response for https://svn.openmethodscorp.com/svn/VXMLB/vxmlb/branches/CCXML/.bzr/branch-format: Unable to handle http code 401: Authorization Required
[18:06] <jam> igc: what are you doing awake?
[18:06] <jam> :)
[18:06] <luke-jr> how do I tell bzr my auth info? :/
[18:06] <ollie> jam: finding it a little tricky to create this layout when i am not 100% sure how everything in bzr works
[18:07] <LarstiQ> luke-jr: you can use ~/.bazaaar/authentication.conf
[18:07] <jam> ollie: well, I don't really know what you are trying to accomplish either. What content is going to be where, and why are you merging individual's blog content back into trunk anyway?
[18:09] <igc> jam: well I'd like to be asleep but it just isn't happening
[18:09] <luke-jr> LarstiQ: use it how?
[18:09] <ollie> i am basicially building a dev box and live box. The idea is that each developer can work on a project then i can merge them into a live version and push/pull/update that onto the live box
[18:12] <jelmer> luke-jr: Please prefix your URL with svn+
[18:13] <LarstiQ> luke-jr: http://doc.bazaar-vcs.org/bzr.dev/en/user-reference/bzr_man.html#authentication-settings
[18:15] <LarstiQ> luke-jr: specifically for svn, you could use svn against the repo, make sure it stores your credentials, and then bzr-svn will pick up on that
[18:15] <LarstiQ> luke-jr: you can give it a hint with https://user@..
[18:16] <LarstiQ> luke-jr: (or do that via authentication.conf)
[18:18] <luke-jr> LarstiQ: I already am using svn against the repo, it didn't
[18:18] <luke-jr> "The svn+ syntax is deprecated, use https://svn.openmethodscorp.com/svn/VXMLB/vxmlb/branches/CCXML instead."
[18:19] <LarstiQ> luke-jr: *blink*
[18:19] <LarstiQ> luke-jr: I meant the subversion application itself, say, "svn ls https://svn.openmethodscorp.com/svn/VXMLB/vxmlb/branches/CCXML"
[18:19] <LarstiQ> luke-jr: bzr-svn can use the ~/.subversion credentials store
[18:20] <luke-jr> LarstiQ: I already have been using subversion itself for months
[18:20] <luke-jr> on this same branch
[18:21] <LarstiQ> luke-jr: ok, then providing your username should be all you need to do.
[18:21] <LarstiQ> luke-jr: if not, something has broken down (what versions of bzr and bzr-svn are you using?)
[18:22] <luke-jr> bzr: ERROR: Permission denied: ".": OPTIONS of 'https://svn.openmethodscorp.com/svn/VXMLB/vxmlb/branches/CCXML': authorization failed (https://svn.openmethodscorp.com)
[18:22] <luke-jr> 0.10 + 0.5
[18:23] <LarstiQ> 1.10 I hope, not 0.10 :)
[18:23] <jam> ollie: sorry, grabbed some food. Anyway...
[18:23] <LarstiQ> luke-jr: how did you specify the username?
[18:23] <luke-jr> yes, 1.10
[18:23]  * LarstiQ has a look at the bzr-svn buglist
[18:23] <jam> I would say users should generally have a checkout of 'live'
[18:23] <luke-jr> LarstiQ: I put it in the URI, and removed it from the error for security
[18:23] <jam> and then another place where they work on their content
[18:23] <jam> when they want to publish they would do
[18:23] <LarstiQ> luke-jr: ok, that should have done it. Something broke then.
[18:24] <jam> cd live-checkout
[18:24] <jam> bzr up
[18:24] <jam> bzr merge ../my-work
[18:24] <jam> bzr commit -m"Publish XXX"
[18:39] <luke-jr> AHA
[18:39] <luke-jr> wrong username :x
[18:47] <deepjoy> .list
[18:51] <luke-jr> is it possible to do a copy with bzr-svn?
[18:58] <luke-jr> jelmer: LarstiQ: http://pastebin.com/m1231cf6c
[19:29] <jam> poolie2: ping
[19:31] <luke-jr> does 'bzr mv' work with bzr-svn right, or does it just add?
[19:46] <luke-jr> meh
[19:46] <luke-jr> bzr dpush at least doesn't send timestamps right ☹
[20:51] <jml> LarstiQ: I haven't done anything
[20:51] <LarstiQ> jml: k
[20:53] <beuno> igc, you rock
[21:34] <kfogel> In ReST format (on http://bazaar-vcs.org/Scenarios/Bugfix), how do I do a link to Scenarios/ReviewIncorporate?  Right now, the linking text is supposed to be "this one", and I've tried various syntaxes, most recently "[Scenarios/ReviewIncorporate this one]".
[21:35] <jam> kfogel: I thought I updated that
[21:35] <kfogel> jam: forgot that one
[21:35] <jam> it would be `this one <Scenarios/ReviewIncorporate>`_
[21:35] <kfogel> It's no problem, I should learn ReST anyway.  But the syntax references, of course, still refer to the moin syntax :-).
[21:35] <jam> `text <hyperlink>`_
[21:35] <kfogel> ah, forgot the underscore when I tried that one
[21:35] <kfogel> jam: thx
[21:36] <Kobaz> Repository branch (format: unnamed)
[21:36] <Kobaz> is that bad?
[21:37] <jam> Kobaz: not specifically. Most likely caused by using a 1.9 repo with a --pack-0.92 format branch
[21:37] <jam> we just don't have a short name for it
[21:38] <jam> bbiab
[21:38] <Kobaz> i just did a init-repo and then an init
[21:38] <Kobaz> and it created a branch with the unnamed format
[21:38] <Kobaz> it's never done that before that i know of
[22:04] <jsled> So … we branched to early.  We have a release branch and a trunk branch, and we're double-commiting everything; there are definitely commits unique to both branches, however.  It seems that once we encountered one of these commits and didn't merge it over to the release branch, we lost the ability to properly do merges going forward.
[22:04] <jsled> That is, every `bzr merge -r X..Y` just shows up as a normal ("diff|patch") commit rather than a subordinate "pending merge:".
[22:05] <LarstiQ> right
[22:05] <LarstiQ> jsled: are you in a position to do `bzr merge -r Y` instead?
[22:06] <jsled> A) What's the actual condition that triggers this difference, and B) do we need to basically do an "empty" merge for each commit we *don't* want ot pull over to have bzr keep things straight, or is there a better way?
[22:06] <jsled> LarstiQ: yeah, I'm thinking of doing a big "full" merge to keep all the meta-data straigth, then changing policy.
[22:06] <jsled> Looking to confirm my mental model of how bzr's working, &c.
[22:07] <LarstiQ> jsled: since bzr merge -r X..Y does not (would like to have that fixed) store any metadata, just brings across the diff.
[22:07] <jsled> Well, I notice so long as a do them in sequence, it does at least treat them as subordinate merges.
[22:08] <LarstiQ> that might be the case if you have X in your ancestry I guess
[22:08] <jam> Kobaz: it should only do that if you supplied an option to the "init-repo" but did not do the same for the "bzr init"
[22:08] <jsled> i.e, if trunk has rev 1,2,3 and branch was branched at '1'.  Then `merge -r 1..2` and `merge -r 2..3` both track.  As soon as a I skip one (or so it seems), then they're not.
[22:08] <LarstiQ> jsled: this operation is called cherry picking btw (one form of), and recording them is a wishlist item.
[22:08] <jam> (eg, "bzr init-repo --1.9 foo; cd foo; bzr init bar")
[22:09]  * LarstiQ nods at jsled 
[22:09] <jsled> LarstiQ: okay.  And we've used both -c and -r before:X..X, which I'm believing are equivalent.
[22:09] <jsled> (maybe -c is just shorthand, even.  immaterial, really.)
[22:09] <LarstiQ> jsled: for a range of 1, yes
[22:10] <jsled> right, yeah.
[22:10] <jam> igc, poolie1: Are you guys around?
[22:15] <lifeless> hi jam
[22:16] <jam> hi lifeless
[22:38] <enigma> How do I make a branch without a working tree? I don't see an option to the branch command.
[22:39] <enigma> So is it a two step process? "branch" followed by "remove-tree"?
[22:39] <enigma> Is there a one-command way of making a branch without a working tree?
[22:39] <delaney> Hi, I have a standalone project, inside a my lib folder there is a bunch of svn'd repositories.  I want to bzr the whole project but still have ability to do svn updates as needed for the libraries.  How do I set this up in bzr?
[22:41] <enigma> I would create a bzr project and check in everything but the lib folder.
[22:41] <enigma> Actually...check in an empty lib folder.
[22:42] <enigma> you can leave the svn directories sitting in the lib folder and bzr should ignore them just fine.
[22:42] <delaney> but i want local bzr version control on the libs as well
[22:43] <enigma> You mean on the files in the svn checkouts in the lib folder?
[22:43] <spiv> enigma: you can create the branch inside a repository made with "bzr init-repo --no-trees"
[22:44] <enigma> spiv: OK. Thanks.
[22:44] <enigma> Why not just run "svn export" from time to time and put those files into lib?
[22:44] <enigma> Then you wouldn't have all those ".svn" directories littered about.
[22:45] <enigma> And you don't have to worry about bzr-svn trying to do something with the .svn directories.
[22:51] <mwhudson> i think the new bzr st output confuses meld
[23:49] <luke-jr> any way to get 'rm' (not 'bzr rm') to not count as 'bzr rm'?
[23:51] <mwhudson> luke-jr: you mean have bzr commit complain when files go missing?