[01:00] <nysin> I've tried importing an svn repository using svn-import of the bzr-svn plugin via bzr 1.3 using both the none and branch schemes, and it doesn't seem to preserve history across what were svn cp's
[01:01] <floam> is there anything fairly easy one can do to automatically get the revision number into a controlled source file on commits?
[01:02] <jelmer> nysin: that's correct, bzr does not support copy tracking
[01:02] <floam> what I did now is I have version = 120 in one of my files that I try to remember to manually increment each time -- but it's a losing battle
[01:02] <nysin> Oh. I took "Follow branch copies. Revision history is not truncated when a branch was copied in Subversion." from the BzrForeignBranches page to mean something else?
[01:03] <jelmer> nysin: that's copies of branches, not copies of directories or files
[01:03] <nysin> Ah, alright.
[01:04] <nysin> Which means that it doesn't matter which branching scheme svn-import uses with regards to intra-repository history tracking?
[01:05] <nysin> (The reason I tried to use scheme=none was to get that history)
[01:05] <jelmer> nysin: yes
[01:07] <nysin> Okay, thanks.
[01:10] <jelmer> nysin: Or perhaps, another way to look at it is
[01:10] <jelmer> nysin: only copies of the root of a branch are tracked
[01:11] <jelmer> nysin: so if you copied /branches/foo to /trunk and you use scheme=trunk, bzr will keep the history of that branch
[01:11] <jelmer> nysin: if you use scheme=trunk and you copied /trunk/foo to /trunk/bar, that will not be tracked
[01:11] <nysin> Hm. That's more or less what I did. The former.
[01:11] <nysin> But I guess not via a branch.
[01:12] <nysin> It started as /svn/trunk, then a big change happened, so I split off that version to /svn/branches/foo and branches/bar
[01:12] <nysin> And then finally branches/bar back to /trunk when it stabilized
[01:12] <nysin> But it was all via svn copy
[01:13] <jelmer> it should show all history for that then
[01:13] <nysin> How might I determine why it isn't, then?
[01:14] <nysin> Or, maybe it is an I'm missing it somehow
[01:14] <jelmer> is this a public repository?
[01:14] <nysin> The svn is, yeah
[01:14] <nysin> URL in a moment. bzr is all local so far
[01:14] <nysin> http://colon.ath.cx:8008/svn/ should work
[01:15]  * jelmer tries with bzr-svn 0.4.9
[01:16] <nysin>  /trunk/foo is branches/wtl
[01:16] <nysin>  /trunk/bar is branches/smartwin
[01:16] <nysin> er /branches/... for the first two in those
[01:17] <nysin> Oh, yeah, and that's the bzr-svn version I was using too, from Debian unstable.
[01:19] <nysin> ScriptManager.{cpp,h} as well as the scripts/ directory are files that should have retained a history amongst the longest...
[01:21] <jelmer> nysin: branches/wtl and branches/smartwin are unrelated in that svn repository
[01:22] <nysin> yes
[01:22] <nysin> wtl should have no offspring to speak of
[01:22] <jelmer> nysin: where exactly were you expecting it keep history but it isn't?
[01:23] <nysin> Alright, for example the scripts directory
[01:23] <nysin> which was copied around unchanged from one part to another
[01:24] <jelmer> from wtl into smartwin you mean?
[01:24] <nysin> As well as trunk/ from smartwin/, which is just a big scp copy
[01:24] <nysin> yes
[01:24] <nysin> er
[01:24] <nysin> svn copy
[01:24] <jelmer> nysin: that's a directory copy though
[01:24] <nysin> Which you said wasn't covered, yes
[01:24] <jelmer> nysin: not a branch copy
[01:24] <nysin> I guess I wasn't clear on that distinction whilst using svn - it worked about as well either way
[01:25] <jelmer> the copy from /branches/smartwin to /trunk appears to be picked up correctly here with history retained
[01:25] <nysin> hm.
[01:25] <jelmer> nysin: svn doesn't have a notion of branches
[01:25] <jelmer> which is one of its strongest advantages and one of its weaknesses :-)
[01:25] <nysin> Yeah, and its documentation discusses copies as 'light branches' of sorts
[01:25] <nysin> I really like that
[01:26] <nysin> Verifying locally though
[01:26] <nysin> the history...
[01:27] <nysin> Yeah I don't see it
[01:27] <nysin> "bzr log formatting.lua" has only one entry
[01:27] <nysin> in trunk.
[01:29] <jelmer> nysin: that's correct, because it was introduced by a directory copy (r820)
[01:30] <nysin> Yeah, which you said before etc. I had accepted that and then I got confused when you said you did in fact see some cross-branch history preserved
[01:30] <jelmer> nysin: right, with copies such as the one in r818 history is preserved
[01:32] <nysin> Alright, but when I actually go and check that in the wtl directory...
[01:33] <nysin> branches/wtl$ bzr log --short changelog.txt
[01:33] <nysin>   818 cologic   2008-01-18
[01:33] <nysin>       archive WTL 0.699d version
[01:33] <nysin> And that's it
[01:33] <nysin> It did have a history before that..
[01:33] <jelmer> nysin: I mean r818 in svn
[01:34] <nysin> Yeah my current bzr is no-branches so they match revision-wise
[01:35] <nysin> svn818 is the trunk -> wtl branch in both repositories specifically.
[01:35] <nysin> or bzr818
[01:36] <nysin> But regardless, I'm trying to see what history you're saying it's preserving? I might be missing something there
[01:36] <jelmer> That's only true for scheme=trunk
[01:36] <jelmer> Which causes /trunk and /branches/wtl to both be considered branches
[01:36] <nysin> hm? scheme=trunk produced local revision #s
[01:37] <jelmer> revision numbers are per-branch in bzr
[01:38] <nysin> Sure, so they weren't going to match svn. Which is fine and all, but just means that I don't see what correspondence you say works with and only with scheme=branch?
[01:39] <jelmer> history preservation only works when copying branches
[01:40] <jelmer> scheme=trunk causes "/trunk" and /branches/* to be considered branches
[01:40] <jelmer> scheme=none considers "" to be considered a branch
[01:40] <nysin> ah
[01:40] <nysin> Now it makes sense
[01:40] <jelmer> s/considered//
[01:40] <nysin> I've been looking at the no-branch version (I archived both so I'll go look at the branching version now...)
[01:45] <nysin> ahh
[01:45] <nysin> Much better.
[01:50] <jelmer> It would be nice if bzr could support file copy tracking at some point, even if just for log and annotate
[01:51] <nysin> When I was Googling I found some mailing list messages all the way back from the 0.10 era agreeing with that
[01:53] <nysin> And, especially, since svn at least does treat copies as 'branches' of sorts, then it seems pretty useful.
[02:08] <nysin> Oh, there's even a spec on it ( https://launchpad.net/bzr/+spec/filecopies ) which gets at the usage scenarios
[02:35] <tro> i swear i got bazaar to include the diff in the commit note (in the ignored section) somehow, but i can't remember how i did it :/
[02:35] <tro> is it a config option?
[02:36] <beuno> tro, just do bzr commit
[02:36] <beuno> without -m
[02:36] <beuno> ah, no, wait
[02:36] <tro> all i get is the modified files list
[02:36] <beuno> yes, that's right
[02:37] <tro> bzr ci --show-diff
[02:37] <tro> aha
[02:52] <Stavros> why is bzr always telling me about conflicting tags? i just moved the tag, i want it to be moved on the remote branch as well
[02:52] <Stavros> sorry, i left by mistake
[02:53] <spiv> Stavros: push --overwrite
[02:53] <Stavros> o
[02:53] <Stavros> hmm, thanks
[02:53] <Stavros> shouldn't that be done normally, though?
[03:00] <spiv> Stavros: bzr won't overwrite a tag by default, because it doesn't know which one is the one you want to use.
[03:01] <Stavros> oh
[03:01] <Stavros> so it doesn't save the change of the tag as well, huh..
[04:10] <ubotu> New bug: #212645 in bzr-svn "cannot import name CachingParentsProvider" [Undecided,New] https://launchpad.net/bugs/212645
[04:40] <ubotu> New bug: #212649 in trac-bzr "Viewing a changeset in trac 0.11 produces traceback" [Undecided,New] https://launchpad.net/bugs/212649
[06:32] <cammoblammo> What are the differences between bzr branch, clone and get?
[06:33] <mwhudson> there are none
[06:33] <spiv> cammoblammo: $ bzr help branch | grep Aliases
[06:33] <spiv> Aliases:  get, clone
[06:33] <spiv> cammoblammo: they're all the same command
[06:35] <cammoblammo> Ah. I thought I saw that somewhere, but I couldn't find it!
[06:35] <cammoblammo> Thanks
[06:47] <cammoblammo> I've got an existing working branch that I want to turn into a shared repository without the tree. What's the canonical way to do it (or can I just copy the branch and remove all the non .bzr directroies?)
[06:49] <spiv> cammoblammo: branch it into a shared repository (initialise the repo with "bzr init-repo --no-trees PATH")
[06:50] <cammoblammo> That's what I thought. I wasn't sure if that worked for already existing repositories.
[06:57] <cammoblammo> Okay. I successfully tried "bzr init-repo --no-trees --rich-root-pack SharedRepo"
[06:57] <cammoblammo> However, "bzr branch OldRepo SharedRepo" complains that the repo exists already. But of course it does...
[06:58] <cammoblammo> Hang on, shouldn't I be using the import command?
[07:04] <cammoblammo> Don't worry folks, I'm having a brain fart. I've figured it out, and yes I am embarrassed.
[07:04] <cammoblammo> ;-)
[07:16] <exothermc_> does bzr preserve perms and ownership?
[07:17] <bob2> no
[07:17] <bob2> alas
[07:17] <exothermc_> bah looking for a good solution to manage my server config files such as /etc
[07:17] <exothermc_> svn does it but only with a wrapper
[07:17] <bob2> etckeeper
[07:18] <bob2> (which layers on top of bzr)
[07:18] <exothermc_> ahh ok
[07:18] <exothermc_> does it hook into rpm as well?
[07:19] <bob2> dunno
[07:23] <exothermc_> one last question do newly created files and dirs get automatically added to the repository?  also are deletes removed?
[07:28] <cammoblammo> Can an existing branch be turned into a --no-trees cwntral repo?
[07:30] <bob2> exothermc_: bzr records deletions automatically, but not additions
[07:31] <bob2> not sure what etckeeper does on top of that
[07:31] <exothermc_> ok
[07:31] <bob2> cammoblammo: not, afaict, through the ui
[07:31] <bob2> cammoblammo: bzr reconfigure --no-trees would get rid of the working copy, though
[07:32] <bob2> cammoblammo: is the branch just too big for you to create the repository and then branch into it?
[07:34] <cammoblammo> I'm trying to reconfigure my set up so I can put my repo on a server at home rather than the free piece of rubbish my ISP gives me (no ssh, etc.)
[07:35] <bob2> ok
[07:36] <bob2> so, is "bzr init-repo --no-trees --rich-root-pack repo ; bzr branch /your/branch repo/branch" what you wany?
[07:36] <cammoblammo> Something like that. I did try that sort of thing, but I had problems. I'll try it again and see if it an operator error ;-)
[07:39] <cammoblammo> Hm, same thing. I end up with a treeless version of my branch in a directory in the repo.
[07:40] <cammoblammo> So it looks something like /home/cammoblammo/repo/branch/.bzr
[07:40] <cammoblammo> With a .bzr in the /repo directory as well.
[07:40] <spiv> cammoblammo: isn't that what you're after?
[07:42] <cammoblammo> Not quite. I was after /home/cammoblammo/repo/.bzr without anything else in there. That's how my current repo at my ISP looks.
[07:43] <bob2> standalone repos always have .bzr in the root and in the branch...
[07:44] <bob2> what you have at your isp looks like it's just a standalone branch with no working tree...
[07:44] <cammoblammo> That's what I'm trying to get here. Isn't that how it's supposed to look?
[07:45] <bob2> not if you want a shared repository
[07:45] <bob2> what was your original goal (i.e. not "make it look like the existing one")?
[07:48] <cammoblammo> I need to be able to access my repo from work. I used to run a svn server here at home, but when I changed to bzr I moved the repo to my ISP to save space on my home machine (not to mention give me more reliable access when I'm on the road). The ISP only lets me use ftp, though, which is a big performance hit. I want to move it back to try using bzr+ssh.
[07:48] <bob2> ok
[07:51] <bob2> sounds like a shared repository would work well for you.  'bzr init-repo --no-trees repo', then branch things into it.  the branches will be in subdirs under repo (and have their own, small .bzr dirs), but all the revision data gets store in repo/.bzr, so it's not duplicated.
[07:52] <cammoblammo> Well, if that's what I have to do ;-) It will look a bit silly though --- I'm only running one branch. Of course, I don't have to look at it, do I?
[07:53] <bob2> ah, ok, now I understand
[07:53] <cammoblammo> As another thought, will it work if I just copy the directory at my ISP? It's probably dirty, but...
[07:53] <bob2> don't make a repo then, just create a branch
[07:54] <bob2> that will work fine
[07:56] <cammoblammo> Out of interest, is there any reason I shouldn't just copy the repo over?
[07:58] <bob2> btw, "repo"  is generally used to mean something else (what init-repo produces) in relation to bzr, hence my confusion earlier
[07:58] <bob2> sure, you can copy it over
[07:59] <cammoblammo> Oh well then, problem solved! I've never seen that sort of approach mentioned in the docs, so I figured it must be discouraged for some reason. Then again, that's how I got it to the ISP in the first place.
[08:00] <bob2> "bzr branch /home/cammoblammo/repo/ ftp://blahblah" should just work, as well
[08:03] <cammoblammo> Hmm, let's see...
[08:04]  * cammoblammo is waiting very patiently...!
[08:05] <cammoblammo> Well, something's coming down the tubes...
[08:21] <cammoblammo> bob2: It worked, but it wasn't treeless, even though I initialised the repo that way. Oh well.
[08:23] <bob2> oops - bzr reconfigure --branch ftp://...
[08:40] <cammoblammo> bob2: Yeah, I picked up on that.
[08:40] <cammoblammo> Hang on, no I didn't.
[08:41] <cammoblammo> Let me see now...
[08:50] <cammoblammo> Well, close! I took a copy of a standard working branch and ran the reconfigure command over it and it worked.
[08:51] <cammoblammo> Now for another problem :-( !
[08:53] <bob2> ok
[08:53] <cammoblammo> I can check the shiny new repo out locally, but via bzr+ssh I get a problem: ERROR: Repository KnitPackRepository(local details) is not compatible with Repository RemoteRepository(Repo details)
[08:54] <bob2> what does "bzr info" say when run in the same location that is giving you that error?
[08:54] <cammoblammo> Do I have to change the repository format?
[08:56] <bob2> ?
[08:56] <cammoblammo> Sorry, missed your message. One sec...
[08:57] <cammoblammo> Bound branch format: pack-0.92
[08:58] <cammoblammo> The repo (if I can use that word) is rich-root-pack.
[08:58] <bob2> that's the entire output?
[08:58] <bob2> the related branch part isn't important
[08:58] <cammoblammo> Yes, apart from the Location and binding info.
[08:59] <bob2> whici is what I wanted to se...
[08:59] <cammoblammo> Oops. One tic...
[09:00] <cammoblammo> branch root: .
[09:00] <bob2> the error could be because you're inside a repo made with init-repo, or it could be because you're co'ing from rich-root-pack
[09:01] <cammoblammo> Hmm. I'm trying a clean checkout, so I'm not inside anything relevant (and I did check that).
[09:01] <cammoblammo> Is there a problem co'ing from a rich-root-pack?
[09:01] <bob2> ah, ok.  it's because co has a bug and uses the bzr default format for the checkout, instead of the remote branch format
[09:02] <cammoblammo> Yay me! I didn't do something wrong (for once). Is there a work around, or do I have to somehow reconfigure the repo?
[09:03] <bob2> you can either: "bzr branch ftp:/// foo ; bzr reconfigure --checkout foo" or "bzr init-repo --rich-root-pack repo ; bzr branch ftp:// repo/blah" wwwwwwwwww
[09:03] <bob2> note that rich-root-pack is generally only useful if the branch came from svn (or you're z bzr developer)
[09:04] <cammoblammo> Really? Someone here told me it was now the default and I should upgrade.
[09:04] <cammoblammo> Grr.
[09:05] <cammoblammo> Anyway, let's give it a go...
[09:05] <bob2> default's still pack-0.2 afaik
[09:06] <cammoblammo> Don't believe everything you see on IRC!
[09:06] <luks> you can generalize that to The Internet :)
[09:07] <cammoblammo> ;-)
[09:07] <cammoblammo> Actually, you can generalise that to anything involving, well, anything.
[09:11] <cammoblammo> Guess what guys? IT WORKED!!!!! I now have a treeless copy of my project which is functioning as a shared repository for at least two other machines, and I've lost no history along the way, but I have gained bzr+ssh functionality. YIPPEE!!!
[09:12] <cammoblammo> (Now I can get back to work)
[09:12] <cammoblammo> Dammit.
[09:12] <bob2> yay
[09:12] <cammoblammo> Oh yeah, thanks bob2 (and everyone else for putting up with the noise).
[09:13] <bob2> no worries
[09:13] <cammoblammo> See you on #emacs!
[09:13] <bob2> ciao
[11:12] <matid> Hello there. I was wondering, has any of you managed to set up svn integration in Bazaar on OS X?
[11:12] <matid> I'm having issues with python bindings.
[11:14] <matid> I compiled svn 1.6 from source and installed python bindings with 'make install-swig-py' but bzr keeps telling me "Installed Subversion version does not have updated Python bindings."
[11:15] <matid> And whenever I try to run 'from svn import fs, repos, core' in python it says:
[11:15] <matid> http://pastie.caboo.se/176037
[11:18] <matid> Any ideas?
[11:21] <DaNGeRs> alows
[11:21] <DaNGeRs> hello
[11:22] <DaNGeRs> lol[
[11:57] <jelmer> matid: did you follow the instructions on the wiki?
[11:58] <matid> jelmer: Yeah, I did.
[11:58] <matid> You're referring to http://bazaar-vcs.org/BzrForeignBranches/Subversion right?
[11:59] <matid> They refer to svn 1.5 even though trunk is 1.6 now, so atm I'm trying to build it with the 1.5.x branch.
[11:59] <matid> We'll see if it works.
[11:59] <jelmer> yep
[12:12] <matid> jelmer: No luck.
[12:13] <matid> Still the same issue, both with svn 1.6 and 1.5
[12:14] <jelmer> you probably have another svn install that's conflicting with the one you just installed
[12:17] <matid> jelmer: Yeah, I've got one that is installed with Mac OS X by default, but it shouldn't conflict with the new one...
[12:20] <jelmer> matid: I don't know then, sorry.
[12:21] <jelmer> it's getting the wrong svn library for some reason though
[12:27] <matid> OK, thanks for your help anyway.
[12:27] <matid> I'll try to work it out on my own.
[12:33] <pbor> if I have a shared-repository on a remote server (sftp), can I push a new branch on that server?
[12:34] <iKs> pbor: as far as I know yeah, there's no problem in doing that
[12:35] <iKs> bzr push sftp://yourserver/your/directory/
[12:35]  * pbor tries
[12:35] <bob2> yeah
[12:36] <pbor> the shared repo docs only mention "branch", that's why I ask
[12:40] <pbor> worked great
[12:40] <pbor> awesome
[14:41] <matthewlmcclure> I had a branch in a shared repository.  I did: $ bzr reconfigure --lightweight-checkout; then $ bzr reconfigure --branch .  Now the branch is missing revisions.  Any way to get them back?
[14:50] <Odd_Bloke> matthewlmcclure: What does `bzr info` report the repository location (of the branch) to be?
[14:51] <matthewlmcclure> $ bzr info
[14:51] <matthewlmcclure> Repository tree (format: pack-0.92)
[14:51] <matthewlmcclure> Location:
[14:51] <matthewlmcclure>   shared repository: /c/home/mlm/sandbox/bzr/repo
[14:51] <matthewlmcclure>   repository branch: .
[14:57] <Odd_Bloke> That's odd.
[14:57] <Odd_Bloke> I don't really know what else to try...
[14:58] <james_w> matthewlmcclure: installing the heads plugin and using that may help you resurrect the missing revisions.
[14:59] <matthewlmcclure> james_w: presumably because they are in my shared repo still?
[15:00] <james_w> matthewlmcclure: yeah, they won't have been deleted.
[15:01] <Bronger> Can anybody explain http://dpaste.com/43416/ to me?  I try to update an SVN repo.
[15:02] <Bronger> Sorry, forget it, my fault.
[15:05] <matthewlmcclure> james_w: thanks... is this a bug/missing feature in 'reconfigure'.  shouldn't it protect the user from this?
[15:09] <james_w> matthewlmcclure: it sounds like it may be a bug.
[15:09] <james_w> it would be interesting to know what about your setup caused it
[15:09] <james_w> it may be the fact that you were in a shared repo, or it may have been something else.
[15:10] <matthewlmcclure> i'll see if i can reproduce
[15:12] <abentley> matthewlmcclure: When you say they're "missing", what do you mean?  Log doesn't include them?
[15:12] <matthewlmcclure> right, log doesn't include them
[15:15] <abentley> Strange that it would do that.  Is there anything significant about the revision where it stops?
[15:18] <matthewlmcclure> the broken branch had a parent in the shared repo, and the last rev in the broken branch is the parent's last rev
[15:19] <abentley> I have trouble imagining how a bug like that could work.  If you can reproduce it, that would be great.
[15:20] <abentley> Oh, okay.
[15:21] <abentley> let me see if it could be the reconfiguration to lightweight checkout.
[15:21] <matthewlmcclure> that would be my guess as to when things went wrong
[15:23] <abentley> Yeah, that's it.
[15:24] <abentley> It's not really sensible to convert a branch with unique revisions into a lightweight checkout, because whatever it's a checkout of won't have those revisions.
[15:24] <matthewlmcclure> so it's official that i know just enough to be dangerous now
[15:24] <abentley> I'm sorry about this bug.  I'll add a check
[15:25] <abentley> I can help you recover the revisions.
[15:25] <matthewlmcclure> i don't actually need them in this case, but thanks for offering
[15:25] <abentley> Okay.
[15:27] <matthewlmcclure> hm... actually, i am curious if there's a way to recover a branch from a repo; say, if i had deleted the branch's directory
[15:28] <abentley> You can create a new branch with the same tip.
[15:28] <abentley> First you have to find the revision-id of the branch tip.
[15:28] <abentley> The heads plugin is useful for this.
[15:29] <matthewlmcclure> got it.  thanks
[15:29] <abentley> Then you branch from any branch in the repository, and specify -r revid:
[15:36] <ubotu> New bug: #212853 in bzr "AssertionError: 17 != 18 when trying to comuit to bzr.debian.org" [Undecided,New] https://launchpad.net/bugs/212853
[16:56] <offby1> The docs show some examples that use abbreviations for URLs, like "bzr branch m:/cool-trunk
[16:56] <offby1> where is that abbreviation format described?
[16:59] <james_w> offby1: there's an "m:/" described?
[16:59] <james_w> there's "lp:/" provided by the launchpad plugin that is shipped with bzr.
[17:01] <offby1> there's both!
[17:02] <asabil> there is bz-bookmarks
[17:02] <offby1> .../doc/en/user-guide/branching_a_project.txt, r3336, line 60
[17:03] <james_w> offby1: that's a windows drive, "m:/"
[17:05] <dato> right, but check bzr-bookmars as asabil said
[17:05] <dato> er, bzr-bookmarks
[17:06] <offby1> oho
[17:07] <offby1> I also created my local repository by doing "bzr branch lp:bzr"
[17:31] <ubotu> New bug: #212908 in bzr "fetch all from a repo with identical contents fails with pack repos" [High,Triaged] https://launchpad.net/bugs/212908
[17:45] <ubotu> New bug: #212917 in bzr ""make check" got an internal error" [Undecided,New] https://launchpad.net/bugs/212917
[20:23] <ubotu> New bug: #212979 in bzr "trying to do checkout of some django code" [Undecided,New] https://launchpad.net/bugs/212979
[20:25] <nick125> good afternoon.
[20:27] <nick125> I'm looking for a SCM to integrate into an IDE project that I'm writing, and I've been looking at bazaar as a possible option. The idea is to use a SCM to internally version saves, as well as allowing a user to use an "external" SCM. What I'm looking for is a SCM that will allow me to use a non-standard directory to store the SCM's files, so that I can have two instances of the SCM versioning the same source tree. Can bazaar do that?
[20:30] <nick125> I like bazaar since it's written in Python, which will make it easier to integrate into my IDE (since I'm writing it in Python).
[21:35] <ubotu> New bug: #213032 in bzr "allow to change author (name <e-mail>) and description in past commits" [Undecided,New] https://launchpad.net/bugs/213032
[22:03] <Cerulean> I'm probably missing something obvious, but I want to branch a repository onto my server
[22:04] <Cerulean> I can't access the repo from my server (firewall etc)
[22:04] <Cerulean> so something like bzr push that pushes the files. What can I do?
[22:25] <TFKyle> nick125: mm, writing an IDE now?
[22:27] <nick125> TFKyle: yep
[22:29] <nick125> TFKyle: What are you upto these days?
[22:30] <TFKyle> nick125: looked at crypto/digital signing stuff a while ago, a few days ago I got a fingerprint scanner and made a small thing using pylibusb to access it and get fingerprints from it :)
[22:31] <nick125> Sweet, how much did the scanner set you back/
[22:31] <TFKyle> nick125: ~$50 CAD
[22:31] <nick125> That's not bad.
[22:33] <TFKyle> the scans seem a bit low quality, dunno what sort of quality the other scanners have though (can create 128x128 images, with a few pixels being blacked out for some reason (think the register values I'm using aren't the most perfect combination though, stole them mostly from the kernel driver (that doesn't compile with newer kernels obviously))
[22:35] <TFKyle> tried just changing the one required for it to send data, but it seems to produce more heat (or more electricity, not sure), should have a look at what the windows driver does too
[22:37] <nick125> TFKyle: I wonder what kind of program you'd use for reverse engineering the windows driver..
[22:37] <TFKyle> nick125: usbsnoop/snoopy is supposedly pretty good for USB devices (logs the traffic to the USB device)
[22:40] <nick125> TFKyle: Aah, fun. I've never tried to reverse enginner anything before.
[22:41] <nick125> (Well, least over USB)
[23:09] <lifeless> poolie: up?
[23:10] <lifeless> nick125: so, we don't have code in bzrlib to maintain two completely separate trees like you want to; knowing why might help :P
[23:12] <nick125> lifeless: Would it be possible to simply use a different directory for .bzr?
[23:15] <nick125> lifeless: Also, the reason I want to have two trees is so that my "internal" save bzr tree will not conflict if the user uses bazaar for an external SCM.
[23:21] <asabil> nick125: why not simply support bzr alone ? :D
[23:22] <asabil> you can still write repository handlers for other scms ?
[23:23] <nick125> asabil: Well, the idea is that I'd integrate really closely to a SCM to keep an internal save revision tree, that way, if you delete a file or something, you can restore it easily. The only problem with using bazaar for the internal SCM is that it might conflict if the user wants to use bazaar for their "external" SCM.
[23:26] <beuno> how would I get the list of uncommited changes to a working tree?  I'm getting a bit lost in the code
[23:31] <beuno> nevermind, got it :)
[23:32] <lifeless> nick125: well, you could submit a patch:P.
[23:34] <nick125> Is the .bzr path hardcoded into bazaar?
[23:50] <lifeless> no
[23:50] <lifeless> and yes
[23:50] <lifeless> see, we support svn natively :P
[23:51] <lifeless> so there are factory functions to drive a lot of things; you could for instance subclass the stock format and create a .idesave control directory
[23:51] <lifeless> which would be a bzr one internally but just change the .bzr -> .idesave
[23:51] <nick125> lifeless: Hmm, how hard would that be to implement?
[23:52] <lifeless> I think the setup.py of the plugin would be longer than the code
[23:52] <nick125> ouch :p
[23:52] <lifeless> (or you could import bzrlib and register after import, no bzrlib plugin needed then, in which case the code will still be small :P)
[23:53] <nick125> I wonder how hard it is to interface with bazaar in Python
[23:53]  * nick125 looks at bzrlib
[23:53] <lifeless> as a user of ides from time to time, I would prefer to see ide save stuff to be done as a branch off my actual vcs; but thats me
[23:54] <lifeless> pydoctor generates good docs, there is developer docs in doc/en/developer/
[23:54] <lifeless> or something like that :)
[23:54] <lifeless> bbiab
[23:54] <nick125> Hmm...using a branch could be useful.