2008/10/25/#bzr.txt

=== RAOF_ is now known as RAOF
[05:05] <alester> is there anything more verbose about running my own server than http://doc.bazaar-vcs.org/bzr.dev/en/user-guide/index.html#running-a-smart-server ?
[05:08] <Peng_> alester: It's not very complicated, so probably not.
[05:08] <Peng_> alester: The docs for an HTTP smart server are elsewhere in that page.
[05:08] <alester> it's basically "you have a dir and everything goes thru bzr+ssh"
[05:10] <alester> I've already got a /bzr/drizzle/trunk set up
[05:10] <alester> I'll just need to init-repo at /bzr/newproject ?
[05:21] <alester> I guess what I don't get is what init-repo is actually doin'
[05:25] <alester> And, the diff between bzr branch and bzr checkout.
[05:26] <alester> That's the one that's always stumped me.  Can you explain, Peng_?
[05:31] <Peng_> Err.
[05:31]  * Peng_ vaguely gestures to the docs.
[05:40] <alester> Yeah, I have indeed read them.
[05:40] <alester> oh ewll.
[05:43] <Peng_> Sorry, but I'm bad at explaining things.
[05:44] <alester> One is bound, the other isn't
[05:44] <alester> but that doesn't help me. :-)
[05:44] <alester> oooh, now it does.
[05:44] <pickscrape> alester: if a branch is bound it means when you commit to it, the commit also goes to the branch it is bound to
[05:44] <alester> and you have to be online to do so?
[05:44] <pickscrape> And update gets revisions from the upstream branch.
[05:45] <pickscrape> If the upstream branch is online, yes.
[05:45] <pickscrape> Though you can unbind and make it into a normal branch.
[05:45] <alester> so I can't use a checkout repo in airplane mode
[05:45] <alester> which is my key need here on this one.
[05:45] <pickscrape> Yes, just unbind it
[05:46] <alester> ok
[05:46] <alester> so in a checkout, you can only commit to one place
[05:46] <pickscrape> Say you have upstream branch A, and you do bzr checkout A B to create checkout B.
[05:47] <pickscrape> This is the same as doing bzr branch A B, except that the new branch is bound to A
[05:47] <pickscrape> This means that when you commit in B, the commit also goes to A.
[05:47] <pickscrape> You could get the same effect by doing bzr branch A B, and then doing bzr bind A from within B
[05:48] <alester> Mine is going to be central, just me.
[05:48] <alester> but what you're saying is switching between bind and unbind is the diff between a checkout and a branch
[05:48] <alester> and I can switch between them as necessary
[05:48] <pickscrape> Checkouts basically give you a centralised workflow, svn style. With the added flexibility that you can unbind it at will.
[05:48] <pickscrape> Yes
[05:49] <alester> oK, thank you
[05:49] <pickscrape> Best thing for you to do really is to experiment checking out a local branch. Doesn't have to be over a network.
[05:51] <alester> I've been using bzr for a while on launchpad, but have not set up my own projects utnil now
[05:52] <pickscrape> Just use bzr init A anywhere, create a file in it, commit, then bzr  checkout A B and start experimenting etc.
[06:04] <alester> I need to get bzr-svn not looking in my home directory.
=== eMBee_ is now known as eMBee
=== Spaz is now known as Kittens
=== markus_ is now known as thekorn
[10:45] <Skellus> It might sound not entirely legible (as for all these repository kind of things I am total newbie) but how can a repository be unpacked?
[10:50] <Peng_> What's your goal and what do you mean?
[10:51] <Skellus> I mean
[10:51] <Skellus> that there is a OGG Vorbis "codec" for Actionscript
[10:51] <Skellus> and it can be
[10:51] <Skellus> supposedly
[10:51] <Skellus> downlaoded from
[10:51] <Skellus> http://people.xiph.org/~arek/bzr/fogg.dev
[10:51] <Skellus> which is Bazaar repository
[10:52] <Skellus> the guy says I have to use a command set like this
[10:52] <Skellus> bzr branch http://people.xiph.org/~arek/bzr/fogg.dev fogg
[10:52] <Skellus> but inside there iss no source files
[10:52] <Skellus> some odd files which seem to be logs
[10:52] <Skellus> and one ~450 kb file which has extension .PACK
[10:52] <Skellus> which I suppose has to be the source
[10:52] <Skellus> but I have less than no idea what to do with that thing
[10:54] <Skellus> But due to my limited knowledge about it
[10:54] <Skellus> the problem might be as well lying somewhere else
[10:57] <Peng_> Skellus: "bzr co"?
[10:58] <Peng_> Skellus: That is, cd to fogg and run "bzr co"/
[10:58] <Peng_> Err, without the slash. :P
[10:58] <Skellus> You mean
[10:58] <Peng_> Wait, what are we talking about?
[10:58] <Skellus> I should go into
[10:58] <Skellus> cmd
[10:58] <Skellus> I mean
[10:58] <Skellus> command line
[10:58] <Peng_> Skellus: Have you run "bzr branch http://people.xiph.org/~arek/bzr/fogg.dev"?
[10:58] <Skellus> nope
[10:59] <Skellus> yes
[10:59] <Peng_> Oh.
[10:59] <Skellus> yes
[10:59] <Peng_> What?
[10:59] <Skellus> (got confused here for a second)
[10:59] <Skellus> Yes, I run
[10:59] <Skellus> bzr branch http://people.xiph.org/~arek/bzr/fogg.dev fogg
[11:00] <Skellus> this to be precise
[11:00] <Peng_> OK.
[11:00] <Peng_> What's in "fogg"?
[11:00] <Skellus> hidden .bzr directory
[11:01] <Skellus> and nothing else
[11:16] <Peng_> Skellus: Locally? After you branched it?
[11:16] <Peng_> Skellus: If so, run "bzr co" inside it.
[11:16] <Peng_> But...I don't get it. That shouldn't happen.
[11:16] <Peng_> Working trees are created by default.
[11:17] <Skellus> so you mean in command line I should get to that directory and run "Bzr co", right?
[11:17] <Skellus> I did it
[11:17] <Skellus> and it told me
[11:17] <Skellus> File exists
[11:17] <Skellus> can't open file
[11:17] <Skellus> not open
[11:17] <Skellus> create*
[11:22] <Skellus> Ah, nevermind
[11:22] <Skellus> I've used
[11:22] <Skellus> bzr co "" bla
[11:22] <Skellus> and now it tells me
[11:22] <Skellus> Unable to create symlink 'test.swf' on this platform
[11:23] <Skellus> but I guess that's not the cause of Bazaar this time
[11:25] <Skellus> Peng_: But thanks either way :)
[14:49] <bhy> Hi, why bazaar always tend to consume all my memory?
[14:50] <bhy> I just did a commit with removal of serveral files, but got a MemoryError
[15:25] <stbuehler> can please someone tell me how to merge a branch in rich-root format back to a launchpad repo?
[15:26] <fullermd> Well, there's nothing special about being on launchpad.  But you can't merge from rich-root to non-rich-root.
[15:27] <stbuehler> and let me guess: i cannot update the launchpad branch to rich-root?
[15:27] <fullermd> As far as I know, you can.
[15:28] <stbuehler> i tried: bzr upgrade --rich-root lp:~lighttpd/lighttpd/sandbox
[15:28] <fullermd> Well, you almost certainly want --rich-root-pack rather than --rich-root.  But that's peripheral.
[15:28] <stbuehler> tried that one too...
[15:29] <fullermd> At one time you had to do upgrade over sftp for LP; I THOUGHT that was taken care of some time ago, but I don't know.
[15:29] <stbuehler> "bzr: ERROR: The branch format Bazaar-NG meta directory, format 1 is already at the most recent format."
[15:30] <LarstiQ> what does bzr info say it is?
[15:30] <stbuehler> i see nothing helpful in the output... "Standalone branch (format: unnamed)"
[15:31] <LarstiQ> unnamed?
[15:31] <stbuehler> launchpad says it is "Packs containing knits without subtree support"
[15:31]  * LarstiQ blinks
[15:31] <fullermd> Using lp: will do that since it walks over the SS which doesn't reveal format info.
[15:31] <fullermd> It's pack-0.92.
[15:31] <LarstiQ> stbuehler: it is pack-0.92
[15:31] <stbuehler> so what is right way to access lp branches, if lp:... is not?
[15:32] <fullermd> lp: is.  But going over the SS has certain side effects, and one of them is that you can't directly see the remote format.
[15:32] <fullermd> Format:
[15:32] <fullermd>        control: bzr remote bzrdir
[15:32] <fullermd>         branch: Remote BZR Branch
[15:32] <fullermd>     repository: bzr remote repository
[15:32] <fullermd> Switching to sftp, it shows the real formats.
[15:32] <LarstiQ> stbuehler: for upgrading, you could try sftp://bazaar.launchpad.net/~lighttpd/lighttpd/sandbox instead
[15:37] <mwhudson> (also nosmart+bzr+ssh...)
[15:38] <fullermd> Also WHAT?
[15:38] <fullermd> That's the most ridiculous sounding thing I've read in at least 2 hours.
[15:38]  * mwhudson bows
[15:41] <stbuehler> thx for your help, it looks like it at least does something now. but i don't think i will use bzr or launchpad for new projects - such problems are not really acceptable.
[15:50] <Peng_> Ouch.
[15:53] <fullermd> Our mother dresses us funny too.
[15:56] <LarstiQ> stbuehler: one of the remote transports launchpad offers doesn't work for upgrading is an unacceptable problem for using bzr?
[16:02] <stbuehler> unacceptable is: a) that there is an incompability (and no warning for it) b) i didn't find anything helpful in manuals+google
[16:12] <LarstiQ> feh
[16:14] <Peng_> What does nosmart+bzr+ssh do exactly?
[16:14] <Peng_> Oh, SFTP.
[16:14] <Peng_> That's awesome.
[16:14]  * fullermd boggles.
[16:14] <fullermd> It broke my brain enough when I thought it was a _joke_...
[16:15] <Peng_> :)
[16:16] <fullermd> Anyway, I have a hard time feeling too disgusted with a guy pissed about tripping over an incompatibility from the default format that we've been discussing round and round doing away with for a year and a half now.
[16:17] <Peng_> upgrade doesn't work over lp:?
[16:18] <Peng_> fullermd: Well, he doesn't know that.
[16:18] <fullermd> *I* do.
[16:19] <beuno> there's the plan to be able to upgrade branches via launchpad UI
[16:19] <beuno> so, that doesn't completely solve anything
[16:20] <fullermd> There's the lots of plans to do the lots of things that aren't.
[16:20] <beuno> but it's a start
[16:20] <fullermd> For the human-friendly VCS with the easy UI, we've got a lot of half-wired panels hanging out with sharp edges on the UI.
[16:21] <beuno> yeah, I agree
[16:45] <Peng_> jam: FWIW, with the attribute lookup thing on the mailing list, I get similar results on Linux.
[16:46] <Peng_> (Only it's all twice as slow for me. What kind of computer do you have?!)
[16:46]  * Peng_ gets depressed
[16:47] <fullermd> Oh, you haven't seen depressed 'till you've used a PPro to pull updates to bzr.dev in weaves.
[16:47] <Peng_> Heh.
[16:47] <Peng_> (My other computer is twice as slow again!)
[16:47] <fullermd> On a good day, there weren't any changes, so it only took a couple minutes.
[16:47] <fullermd> If there were any...  well, half hour would be the absolute minimum.
[16:57] <abentley> Peng_: nosmart+bzr+ssh is similar to sftp, but will work on systems that don't support sftp.
[17:42] <Peng_> abentley: ...How does that work? It does VFS operations over bzr+ssh?
[17:43] <abentley> Peng_: Yes.  bzr+ssh has a VFS layer it can use when there's no corresponding "smart" operation.
[17:43] <Peng_> Yeah, I know. Interesting.
[17:44] <Peng_> Was supporting nosmart+bzr+ssh intentional?
[17:44] <abentley> Peng_: I have no reason to think it wasn't intentional.
[17:45] <Peng_> What's the point? Debugging RPC/smart method/whatever-they're-called issues?
[17:46] <abentley> Peng_: Not having to switch protocols when a smart operation doesn't do what you want.
[17:46] <epsy> hi, how do I disable a plugin?
[17:47] <Peng_> abentley: What do you mean?
[17:47] <Peng_> (Are we talking about the VFS in general or nosmart+bzr+ssh specifically?)
[17:47] <abentley> nosmart+bzr+ssh.
[17:48] <Peng_> When would a smart operation not do what you want?
[17:48]  * Peng_ is confused.
[17:48] <abentley> Peng_: bzr info doesn't report useful format information when bzr+ssh is used.
[17:49] <abentley> Peng_: I have occasionally worked around other bugs using nosmart.
[17:50] <abentley> epsy: move it into a different directory.
[17:50] <Peng_> Would sftp be faster than nosmart+bzr+ssh? I mean, OpenSSH is C...
[17:51] <epsy> hm, then i got a problem..: http://pastebin.com/m7a70f89e
[17:53] <Peng_> You can disable all plugins temporarily by running with --no-plugins.
[17:53] <abentley> Peng_: C or not C is pretty irrelevant if you're using the internet.  I would expect bzr+ssh to have a high startup cost, and sftp to be less efficient overall.
[17:55] <Peng_> bzr's VFS is more efficient than sftp?
[17:55] <Peng_> Neat. :)
[17:56] <abentley> Peng: I believe so.  There isn't a 1:1 mapping of VFS operations to SFTP operations.  So some VFS operations require more than one SFTP operation.
=== abentley1 is now known as abentley
[19:10] <beuno> Peng_, ping
[19:11] <beuno> tell me about https://code.edge.launchpad.net/~mnordhoff/loggerhead/valid-html when you havee time  :0
[19:11] <beuno> :)
[19:12] <beuno> seems like something I'd want to merge
[19:35] <Peng_> beuno: I ran a couple pages through the HTML validator and fixed some of the simplest things. There's no reason not to merge it, but it's nothing major.
[19:35] <beuno> Peng_, cool, I'll merge
[19:37] <Peng_> :)
[19:39] <beuno> I didn't want you to think I was code-stalking you
[19:39] <beuno> just happened to see it
=== sdboyer is now known as sdboyer|dcc
[22:33] <strk> how do I produce a diff which includes file renames and new files and all ?
[22:33] <strk> if I can at all...
[22:33] <strk> I received a "merge-diff" once, even contained commit logs
[22:44] <Verterok> strk: bzr send -o <output-file.patch>
[22:44] <Verterok> strk: also, check bzr help send for more options
[22:48] <strk> Verterok: thanks!
[22:48] <Verterok> np
=== abentley1 is now known as abentley