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