| === wedgwood is now known as wedgwood_away | ||
| felipec | when a branch has diverged from the remote, is there a way to get the revision objects programatically? | 04:22 |
|---|---|---|
| felipec | like in git, I don't need to do 'git pull', I can do 'git merge', and I see the remote stuff in 'origin/master' (for example) | 04:26 |
| felipec | er, I can do 'git fetch' | 04:26 |
| bob2 | sure, bzr missing | 05:13 |
| bob2 | or doing the same with git and a repo/clone | 05:13 |
| felipec | bob2: yeah, but I want to do what bzr missing does but programatically | 05:14 |
| felipec | bob2: from what I can tell, the repot revisions are not copied locally, so I would need to query the remote branch to fetch the information | 05:15 |
| bob2 | yes of course | 05:15 |
| bob2 | if you want it to be like git, use a repository | 05:15 |
| felipec | bob2: I'm actually working on git-remote-bzr (the official bazaar support from git) | 05:15 |
| felipec | bob2: and I wasn't really aware of bzr repositories | 05:16 |
| felipec | only exploring them right now | 05:16 |
| bob2 | well | 05:16 |
| bob2 | you'll definitely want to read the bzr tutorial before trying to write an interoperable tool for it :) | 05:16 |
| felipec | bob2: too late, it's already written, and already released | 05:16 |
| felipec | what is the oldest version of bazaar that supports repositories/ | 05:17 |
| bob2 | thousands of years ago | 05:18 |
| felipec | hmm, branch.repository.find_branches() seems to be rather slow | 05:26 |
| felipec | is there some remote repository I can test listing the branches? | 05:38 |
| bob2 | thousands on launchpad.net | 05:39 |
| felipec | bob2: for example | 05:39 |
| felipec | preferably a small one | 05:39 |
| felipec | all the projects I see on launchpad.net have the repository inside the branch | 05:49 |
| bob2 | ? | 05:51 |
| bob2 | I really would read the bzr tutorial | 05:52 |
| bob2 | launchpad just exposes the branches to you, how it's stored does not matter to you as a remote user | 05:52 |
| bob2 | and this is unrelated to the above | 05:52 |
| bob2 | (if you don't want to query the remote for info every time, you want to clone it into a local repository and ask that) | 05:54 |
| felipec | bob2: I already have | 05:55 |
| felipec | and I cansee the repository: RemoteRepository(bzr+ssh://bazaar.launchpad.net/~felipec/%2Bjunk/test/.bzr/) | 05:56 |
| bob2 | yes | 05:56 |
| felipec | also, I created a local repository, and I get nothing while calling get_branches() | 05:57 |
| bob2 | that's not related to your goal of not having to query remote branches | 05:57 |
| felipec | bob2: that's not my only goal | 05:57 |
| bob2 | what is your goal | 05:57 |
| felipec | bob2: another goal I have is to be able to clone all the branches in a repository | 05:57 |
| felipec | bob2: I have many goals | 05:57 |
| bob2 | suspect it'll be hard to get useful help here without a broader explanation then | 05:59 |
| felipec | bob2: I want to list all the branches in a repository | 06:00 |
| felipec | what more explanation do you need? | 06:00 |
| bob2 | good luck =) | 06:00 |
| felipec | even 'bzr branches' doesn't list them | 06:01 |
| felipec | bob2: yeah, thanks for *nothing* | 06:01 |
| bob2 | :/ | 06:01 |
| felipec | that's what you get for trying to support a dead SCM | 06:02 |
| felipec | a bunch of users that have no idea what they are doing, but somehow want bzr to magically work on git, one or two developers unwilling to help, unmaintained spaghetti code, and few online resources | 06:20 |
| fullermd | In some recent version branches started needing some arg (I guess -R) to do the search. | 06:28 |
| felipec | fullermd: yeah, I eventually found that out | 06:38 |
| felipec | I can list the branches now, but with a ControlDir, not with a BzrDir | 06:38 |
| felipec | well, I can, but it's cumbersome | 06:38 |
| felipec | now, I'm trying to find the best way to clone all the branches | 06:39 |
| felipec | I guess I should 'sprout' the repository, and then 'sprout' all the branches | 06:41 |
| felipec | there's no such thing as main, or default branch in a repo, right? | 07:21 |
| vila | felipec: I don't want to sound rude or anything, but bob2 is right, if you don't start using the right words when asking questions about bazaar, we're all losing time | 07:23 |
| fullermd | Meaningless statement, yes. Really, repos don't know anything about branches under them anyway. | 07:23 |
| vila | felipec: there is no such things as a list of branches for a given repo | 07:23 |
| vila | felipec: a branch on the other hand can refer to one or several repos and all of the revisions in the history of a branch are stored in these repos | 07:23 |
| felipec | fullermd: Repository::find_branches() begs to differ | 07:24 |
| vila | felipec: I don't think you fully understand what find_branches does then | 07:24 |
| fullermd | Technically, that doesn't find branches using the repository, only under it. | 07:24 |
| vila | felipec: it walks the file system findin... yeah, what fullermd said | 07:25 |
| fullermd | All the find_branch's/find_bzrdir's evaluate roughly to `find -type d -print | xargs hi-there-are-you-a-branch` | 07:25 |
| felipec | fullermd: I did not ask to "find branches using the repository", did I? | 07:25 |
| fullermd | "in" implies (though doesn't require, true) it. | 07:26 |
| fullermd | In meaningful use cases, the distinction probably doesn't matter. | 07:26 |
| fullermd | Teally, in meaningful use cases, the whole 'repo' question in the first place probably doesn't matter. You don't care about the repo one way or another, you just want to clone some set of branches. | 07:27 |
| fullermd | And Really even, too. | 07:27 |
| felipec | 'in' is not a superset of using, so it doesn't matter | 07:27 |
| vila | ha, nit-picking about word definitions are we ? | 07:28 |
| felipec | vila: wasn't you the one that just complained about the words being used? | 07:29 |
| felipec | and btw, I didn't use any words wrongly, repositoryes do have branches in them | 07:29 |
| felipec | anwyay, moving on | 07:29 |
| vila | felipec: indeed, I told you to use the right words in the right context and you didn't care | 07:29 |
| felipec | bzr branches -> * (default) | 07:29 |
| felipec | I wonder what that (default) means, and why it's named * | 07:30 |
| vila | felipec: no, repo do not contain branches | 07:30 |
| vila | felipec: that's the git definition, not the bazaar one | 07:30 |
| felipec | bzr branches -R . -> and I wonder why that branch is called '.' | 07:30 |
| fullermd | That's a side effect of 'branches' by default doing something that isn't supported. | 07:30 |
| felipec | vila: I can 'bzr branch' it | 07:30 |
| fullermd | It's not _called_ '.', because branches don't have names. Just locations, and '.' is the location. | 07:31 |
| felipec | vila: and I can 'bzr branch' the subdirectories in it | 07:31 |
| felipec | fullermd: and what about '* (default)' ? | 07:32 |
| fullermd | That would be the something-that-isn't-supported. | 07:32 |
| vila | a branch can have a repo, so you branch it but you don't branch the repo | 07:32 |
| felipec | fullermd: BzrDir::sprout() is doing that | 07:33 |
| felipec | vila: it doesn't matter... the repo does have branches in it | 07:34 |
| vila | amazing.... | 07:34 |
| vila | I've seen people being less than friendly when asking for help but here... what can I say, I've tried but I think my time will be best used somewhere else | 07:35 |
| felipec | vila: if the repository does _not_ have branches in it, why does Repository::find_branches() returns branches? | 07:35 |
| vila | felipec: read the log both fullermd and I have already answered, repos do not contain branches, their location on the file system though is (in most usual cases) above the branches that are using it, so find_branches() walk the file system to find them | 07:39 |
| felipec | vila: yeah, you are "helping", keep telling you that | 07:39 |
| felipec | vila: all you are doing is trying to correct me, when the code shows you are wrong | 07:39 |
| felipec | repos do contain branches | 07:40 |
| felipec | as repo.find_branches() shows | 07:40 |
| felipec | the question was: there's no such thing as main, or default branch in a repo, right? | 07:40 |
| felipec | the answer was: no | 07:40 |
| vila | indeed | 07:41 |
| felipec | vila: and yet, you never answered the question | 07:41 |
| vila | felipec: if you know the code better than me why are you even listening to my answers ? | 07:41 |
| felipec | vila: which answers? | 07:41 |
| vila | felipec: everybody answered your question but you're not listening | 07:41 |
| felipec | all I heard was you trying to correct what I said | 07:41 |
| vila | when a question is incorrect, the only way to answer is to first correct the question so it can be answered | 07:42 |
| felipec | vila: but you didn't answer | 07:42 |
| felipec | and it was correct | 07:42 |
| vila | you're coming with a git background where a repo contains branche definitions, that's not the case in bzr | 07:42 |
| vila | felipec: so you care more about your question being correct than about a correct answer ? Your choice | 07:43 |
| felipec | vila: I did not ask if a repo contains branch definitions, nor did I say it | 07:43 |
| felipec | vila: it was you the one that cared about the correctness of the question, not me | 07:43 |
| felipec | all I cared for was the answer | 07:43 |
| felipec | which I didn't get | 07:43 |
| * vila stops feeding the troll | 07:44 | |
| felipec | vila: you are the troll, I am actually writing a useful bzr tool | 07:44 |
| felipec | by definition, a troll purposedly disrupts the communication in a project, so progress is inhibitted, but I'm the one doing something productive here | 08:05 |
| felipec | I take it you are not programing on bzr.bzr atm | 08:05 |
| === jamesh_ is now known as jamesh | ||
| felipec | all right, I've added support for bzr repos in remote-bzr | 10:29 |
| felipec | https://github.com/felipec/git/commit/5fa8fd57241c16f9af5ba2857ff999bf0931dd97 | 10:29 |
| felipec | I whish there was an easier way to detect if a controldir is a branch or a repo though | 10:30 |
| felipec | how does a branch finds it's repositories? checking the directories above? | 10:45 |
| smgz | felipec: yup | 10:49 |
| === mmrazik is now known as mmrazik|afk | ||
| felipec | smgz: I did fetch a branch repository into another repo, and then sprouted the branch in a directory under the other repo | 11:08 |
| felipec | it seems there are some objects in that second repo | 11:09 |
| smgz | it only ascends to the first repo found | 11:10 |
| felipec | smgz: I thought vila said above that a branch can refer to multiple repos | 11:11 |
| smgz | it can, but not via that mechanism | 11:12 |
| felipec | smgz: so under what mechanism then | 11:12 |
| smgz | it's called stacking, and don't do it. | 11:13 |
| felipec | smgz: I might need to if I don't find a way for all the branches to use the same repo | 11:14 |
| smgz | just put the repo in the parent directory (create with `bzr init-repo`) then branch into it | 11:15 |
| felipec | smgz: I am not using the command line | 11:17 |
| felipec | smgz: it's not finding it: bzrlib.errors.NoRepositoryPresent: No repository present: "file:///home/felipec/tmp/test/.git/bzr/alt/clone/master/" | 11:40 |
| smgz | what fun, you get to debug the logic when there's a .git directory involved. try --no-plugins | 11:41 |
| felipec | smgz: it's not the command line | 11:42 |
| smgz | then you get to debug your code directly. | 11:43 |
| felipec | smgz: it's not my code that is triggering that error | 11:44 |
| felipec | smgz: same happens from the command line | 11:50 |
| felipec | bzr: ERROR: No repository present: "file:///home/felipec/tmp/test/.git/bzr/alt/clone/master/" | 11:50 |
| felipec | ln -s $PWD/../.bzr/repository .bzr | 11:57 |
| felipec | and it works | 11:57 |
| === mmrazik|afk is now known as mmrazik | ||
| felipec | ah, create_repository(shared=True) | 12:16 |
| Jeeves_ | Hi all | 13:35 |
| Jeeves_ | Q: Is there any documentation available on how to use the bzrlib python modules? | 13:36 |
| smgz | Jeeves_: you mostly want to refer to the docs under developers, and the docstring/api documentation, for instance at <http://people.canonical.com/~mwh/bzrlibapi/bzrlib.html> (not sure what version that's for...) | 13:37 |
| smgz | (you can generate your own though for local browsing if you like) | 13:38 |
| Jeeves_ | smgz: I'll have a look. Because http://doc.bazaar.canonical.com/en/ seems just about how to use the program, right? | 13:38 |
| smgz | see the "Developer docs" link on that page | 13:39 |
| smgz | *linked from, <http://doc.bazaar.canonical.com/bzr.dev/developers/> | 13:40 |
| Jeeves_ | That seems to be about working *on* bzrlib | 13:41 |
| smgz | some parts | 13:41 |
| smgz | you want to read overview.html at least | 13:43 |
| Jeeves_ | ok, thanks | 13:50 |
| yacc|2 | Any idea what the branch url for bzr-svn is? | 13:52 |
| === smgz_ is now known as smgz | ||
| === wedgwood_away is now known as wedgwood | ||
| yacc_ | Any idea how to make bzr know that some url is a svn repo? Seems the auto-detect fails badly here: http://www.fpaste.org/5g1w/ | 15:20 |
| qengho | james_w: Hi. I'm interested in picking up the libv8 packaging in Ubuntu. It is Debian's, so far, with in-branch sources. Is it smart or possble to convert that to a debian/ -only branch and not break things? | 16:13 |
| james_w | qengho: you can't convert the official branch | 16:16 |
| james_w | but you could have your own version if you wanted | 16:16 |
| qengho | james_w: hrm, thanks. | 16:17 |
| === deryck is now known as deryck[lunch] | ||
| === mmrazik is now known as mmrazik|afk | ||
| === deryck[lunch] is now known as deryck | ||
| rossouwap | Hello, noob here: I'm trying to figure out why a bzr update is going extremely slowly. 3-10 kB/s. Any ideas? | 18:11 |
| qengho | james_w: Er, should the quilt ".pc" directory be versioned? That seems wonky. | 18:15 |
| qengho | In a UDD branch, I mean. | 18:15 |
| qengho | rossouwap: Hrm, what branch format? ("bzr info") What transport? Is it CPU bound? Disk bound? | 18:17 |
| rossouwap | qengho: format 2a, not sure what you meanby CPU bound or disk bound | 18:19 |
| james_w | it is a bit wonky indeed, but it's currently intended behaviour | 18:20 |
| qengho | rossouwap: it could be "normal". I'm not sure. If you're branching something big, then initialize a "shared repository" in the parent dir so that branches of the same family after this one are very cheap. | 18:26 |
| qengho | rossouwap: ("bzr init-repo .." or some place upward in the dir structure.) | 18:27 |
| rossouwap | qengho: thanks. I'll have to read more before I do anything I think (I'm still learning bzr). Was just alarmed to see an update work so slowly. Project is hosted on launchpad, and initial clone put +-800 mb onto disk. I'm sure I'm using ssh as well, as it asks for my key passphrase. | 18:31 |
| qengho | rossouwap: After your branch finishes,"mv new-dir to-be-removed && bzr init-repo . && bzr branch to-be-removed new-dir && rm -r to-be-removed;" That will fill a shared repository by branching locally and discarding a copy, so you don't see that cost again on this computer. The only disadvantage is that you can't move that branch outside of this directory. | 18:35 |
| qengho | james_w: Either there's a bug, a bad import, or I'm dumb. http://pastebin.ubuntu.com/5596373/ | 19:08 |
| james_w | qengho: that looks like a bug to me | 19:39 |
| qengho | james_w: Reported. https://bugs.launchpad.net/bzr-builddeb/+bug/1171999 | 19:56 |
| ubot5 | Launchpad bug 1171999 in bzr-builddeb "On import-upstream, unapplying checked-in .pc/ quilt patch-state makes unversioned conflicts" [Undecided,New] | 19:56 |
| qengho | Longest bug summary ever. | 19:56 |
| james_w | thanks qengho | 20:06 |
| === wedgwood is now known as wedgwood_away | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!