[00:16] hi spiv [00:17] and riddell [00:59] okay so i'm still getting 'checksum mismatch' when doing a new branch from svn into a standalone branch === wallyworld is now known as 92AADA0EO [01:10] i think its clashing with your bzr-svn cache [01:10] which might be in ~/.bazaar/ [01:10] you could try savinga copy and moving it away === medberry is now known as med_out [02:04] I'm trying to extract a working tree from a branch for which I have only the .bzr directory. (Background is that I did a bzr branch, but that failed just at the end when it could not create a symlink on windows. Looks like the repo is intact though.) I cd'd to the branch directory where .bzr was, and did "bzr checkout". That got Error 183, file exists, on the .bzr directory itself. [02:04] hm [02:05] I'd like to redo the checkout under cygwin, which can create symlinks, but the above error is stopping that. [02:05] try 'bzr checkout broken other' - ie into a different directory [02:05] Is there something wrong with my syntax for "bzr checkout"? [02:05] Ah, ok. [02:05] And then move the files back in? [02:07] Almost worked... That didn't get the "file exists" error, but...it still couldn't make the symlink. [02:08] I thought that was the usual workaround for symlinks on Windows. [02:08] oh, it failed under cygwin? [02:08] Right [02:08] :/ [02:08] Odd. [02:09] I've asked the person who put the symlink in to take it back out, but it might be a while... [02:09] :( [02:09] cygwin can create symlinks -- just checked. [02:10] Hmmm, I wonder if the fix that added the "unable to create symlink" error message is checking for attempted symlink creation *before* trying it... [02:11] The workaround of using cygwin was mentioned before that fix. https://bugs.launchpad.net/bzr/+bug/81689 [02:11] Ubuntu bug 81689 in Bazaar "Branches with symlinks can't be checked out on Windows" [High,Confirmed] [03:07] after bzr branch foo, how do I back up a rev? (need to figure out which of the last 5 - 10 commits broke it) [03:11] CarlFK, if this is a throwaway branch and you've got the real code saved elsewhere, you can back out each change in turn by doing "bzr uncommit" then "bzr revert". [03:12] ptressel: it is, thanksl [03:12] Welcome! [03:12] CarlFK: you can use "bzr revert -r REV" to revert just the files to the state at a particular rev [03:12] CarlFK: and then plain "bzr revert" to restore the files back to the current revision [03:13] So that doesn't affect anything but the working tree? [03:14] Safer... ;-) [03:14] Right. [03:16] So one could go backwards down the revisions with "bzr revert -r -1", "bzr revert -r -2", etc. === tchan1 is now known as tchan [05:37] . [05:38] does anyone have an example of a plugin that extends or overrides a builtin command? [05:39] the docs for 'extending an existing command' are 'TO BE DOCUMENTED.' [05:40] loom overrides status [05:41] lifeless: ta [05:41] loom is also one of the older plugins; its stule may be a little... idiosyncratic [05:44] is there a concept of repository specific plugins? [05:44] i.e. any user of the (shared) repository runs the plugins [05:44] I can see why there wouldn't be ... ;-) [05:44] no, security considerations [05:44] there is the ability for a plugin to do nothing on repositories it doesn't know about [05:44] bah humbug [05:44] e.g. in the repo config you provide the config for the plugin [05:44] and you install the plugin globally for all users. [05:45] oh, that'd work [05:45] the plugin then checks for a config and if none found on a repo does nothing. [05:46] lifeless: can I get bzr to tell me the default plugin paths or shall I UTSL? [05:47] gmm, I thought it was in bzr --version [05:47] but no [05:47] ah [05:48] check in ~/.bzr.log [05:48] lifeless: aha, ta [05:49] basically your plugin becomes a submodule of bzr [05:49] see the setup.py for loom etc === ptressel is now known as ptressel|bbl [07:04] Hi, could someone please explain how "bzr split" works? [07:04] I need to create a branch out of a directory in an existing tree. [07:05] Whatever changes I commit to files inside that directory should be stored only in that directory's own repo - is this how it works? [08:50] morning all [08:50] Hum, is a way to do a "lightweight" checkout from a lp repo? Downloadin lp:openerp-addons is around 500Mo wheras I ust want to "get the latest version", not especially all the commits... [08:58] Merwin: "bzr branch --stacked", except there's bugs that make that much slower (and transfer more bytes) than it needs toi [08:58] * spiv goes to pick up his kid [09:10] hola [09:24] jelmer: hello, any hint about this one: http://pastebin.com/4Wp3nr3H [09:33] English experts, what means 'not wild' in 'Not wild this is added for all tests regardless of the flags.' ? [09:34] vila: I think it's short for "I'm not wild about X" (I'm not enthusiastic about it) [09:35] jelmer: makes sense, thanks [09:35] gour: It means bzr-git found an error trying to create an object from scratch again - the resulting object didn't match the sha it expected [09:36] gour: bzr-git is now a fair bit more pedantic about it, older versions got it wrong in some situations === ptressel|bbl is now known as ptressel [10:12] vila: FWIW, I agree with jelmer about "not wild" [10:13] spiv: great, yeah, made sense, never encountered that usage before and I was wondering if it was a typo... [10:14] I thank mgz (and you guys ;) for teaching me more sophisticated English ;) [10:18] Nah, if it were a typo, you'd have intuitively understood it ;p [10:47] how many hours does it take to do a branch of emacs? [10:54] Riddell, jam probably has some experience with that.. [10:54] it used to take quite a while, I'm not sure if that's changed with 2.4 [11:03] jelmer: so, there is no workaround to update the repo? [11:07] gour: no, you'll have to create a fresh clone [11:12] fullermd: hehe [11:13] jelmer: ok. i did it [11:18] Riddell: from gnu or from lp? [11:18] AIUI it is quite a bit faster from lp [11:45] hi [11:45] hi nooga [11:45] vila: subunit-sum could go in subunit ;) [11:45] i've created new project on my laptop and i'd like my friend to co-work with me via local network [11:46] how do i allow him to access my branch and make changes? [11:47] nooga: if you give him ssh access, he can use bzr to make a checkout if you both have write access to the bzr branch [11:47] lifeless: thanks, that's the middle-term plan, I don't want to wait for it to be released to be able to use it (and there may be some tweaks required to land in subunit which may break the way I intend to use it) [11:48] so i need to make a user account for him in my system and let him log in via ssh [11:49] nooga: just sftp access or the ability to run the bzr command should be sufficient [11:49] lifeless: one concern raised by jam (for example) is to use a single 'counter' dict detail (including all counters) in the subunit stream instead of one detail per counter, what's your feedback on that ? [11:50] nooga: After that, he should be able to run something like "bzr co sftp://yourlaptop/path/to/bzr/dir" or "bzr co bzr+ssh://yourlaptop/path/to/bzr/dir" [14:25] jam: cicp-add approved, can you land it ? I'd like to include it in 2.4b4 [14:25] jam: oh, forgot to mention in th review: line too long :-} [14:26] vila: copied from a line that was too long, but I can try to fix [14:26] hehe === med_out is now known as medberry [14:47] vila: re-wrapped, and sent to pqm, you just have to wait for the backlog (and hope there aren't NEWS conflicts) [14:47] in the test suite now, I guess [14:47] jam: yeah, so no news conflicts [15:03] hey [15:03] I have done it again [15:04] merged trunk into my working copy with changes [15:04] is there an easy way to fix it or do I have to manually edit each file? [15:04] I know I could shelve each hunk, but that assumes they are not mixed changes === zyga is now known as zyga-food [15:38] vila: landed [15:39] jam: yup, saw it, I'm having an issue with an added test for selftest-config-stats I can't reproduce locally though :-/ [15:39] sorry to hear that, bbiab [15:39] jelmer: do you have other landings planned ? [15:42] ha crap, I'm pretty sure this is an outdated subunit/testtools combo on pqm :-/ [15:43] vila: nope, nothing important in the queue for bzr.dev [15:44] jelmer: ok, I'll cut 2.4.b4 as soon as I sort this mess out, better if pqm is free for that :0 [15:51] git rules [16:32] basic python question, I'm running a for loop over a method which returns names, what's a good way to count the number of times each name is returned? [16:38] Riddell: without more context, I'd go with a dict of names -> number of occurrences [16:39] vila: but if I do mydict[name] += 1 then it fails on the first time because mydict[name] doesn't exist yet [16:40] yeah, that sucks :) except KeyError: mydict[name] = 1 [16:41] .setdefault() [16:41] mydict.setdefault(name, 0) [16:41] mydict[name] += 1 [16:41] or use defaultdict if it's in the oldest Python you support [16:42] ha right, I knew it somewhere in the back of my head but it refused to pop up ;) [16:42] that's what I'd do at least [16:44] groovy, thanks james_w, vila === joey` is now known as joey [17:10] can someone please clarify the differnce between a repository, branch, and working tree? are these meanings universal between VCSs? [17:13] They are roughly universal in core concepts, but they often have rather different UI [17:13] A repository is a place in which history is stored on disk. [17:13] A branch is a representation of an ongoing line of development [17:14] A working tree is an area on disk where new revisions are prepared before committing [17:14] Bazaar repositories are *really* only about storage [17:15] Mercurial and Git repositories also behave as a container for a set of branches - whereas Bazaar's container for a set of branches is "Your filesystem" [18:09] Hi, does anyone know if it is possible to call the visual diff viewer in bazaar explorer from command line? [18:10] rather than the standard command line output you'd get just by calling bzr diff [18:21] does anyone know if it is possible to call the visual diff viewer in bazaar explorer from command line? [18:27] if i init-repository with --no-trees then the branches will not have working trees, but i'm not sure what this means? [18:30] pretty quiet in here so far buttons, I'm taking a look at your question though [18:32] well, basically my question is just "what is a working tree" i've read the docs and help but i'm not understanding [18:32] just a conceptual question [18:32] amdahlj, I think you're looking for "bzr qdiff" [18:32] thanks jelmer [18:32] will try it out [18:32] Yes, you are right jelmer, thank you very much [18:33] Buttons840, a working tree is an area with the actual files in it so you can prepare a commit - I think somebody else explained it well earlier: [18:33] A repository is a place in which history is stored on disk. [18:33] A branch is a representation of an ongoing line of development [18:33] A working tree is an area on disk where new revisions are prepared before committing [18:33] (from maxb) [18:34] if you want to be able to edit files, you need a working tree [18:34] jelmer: so how is --no-trees even possible; that's like saying "don't allow changes to repository" or "don't allow changes to the files on the disk" at least, that's my view [18:35] Buttons840: You can still push revisions into or pull revisions out of a branch without a working tree [18:36] Buttons840: --no-trees is useful if you have a server where your history is stored, but where you don't do direct editing [18:37] so what information is stored in a working tree? i though no information was stored until i commit? [18:37] thought* [18:38] Buttons840: no history exists until you commit [18:38] Buttons840, the files in the working tree that have been added with "bzr add" are in the working tree [18:38] Buttons840, say you want to store a file named "README" in a bzr branch [18:39] Buttons840, only in a working tree would that file actually exist in the file system so that you can open it with your editor [18:41] ok, that makes more sense [18:41] thanks [18:45] are there branch naming conventions i should conern myself with? [18:47] Buttons840: not really [18:53] how come running "bzr log -r-3..-1 lp:myproject" downloads half a meg, slowly, from launchpad? [18:54] MarkAtwood: it's not a very well optimized code path, [18:54] does it pull down the entire changelog over the network, and then just display the 3 most recent entries? [18:54] it's also a fair bit better in newer versions of bzr, though still not as quick and efficient as we'd like [18:55] MarkAtwood: no, it's more efficient than that but it falls back to reading some of the actual files in the repository rather than streaming just those three entries [18:56] Bazaar (bzr) 2.2.1 [18:56] Python interpreter: /usr/bin/python 2.7.0 [18:56] Python standard library: /usr/lib64/python2.7 [18:56] Platform: Linux-2.6.35.9-64.fc14.x86_64-x86_64-with-fedora-14-Laughlin [18:56] ok [18:56] The initial wait of a few seconds for every connection to Launchpad is a bug in Launchpad that jam has been working on [18:57] the speed will improve further as we implement more calls in the smart server directly [18:59] ok [19:00] this is a pattern i use a lot, because i coordinate the merges from suggested merges into lp:drizzle/build then lp:drizzle/staging and then finally lp:drizzle [19:00] and so im always "peeking" at the top of the log for each of those, to keep track of where I'm at [19:01] lp:drizzle/build and lp:drizzle/staging are picked up and ground up by our Jenkins array [20:24] Hi everybody. I'm using bzr 2.4b2 & bzr-svn 1.1.0dev. I'm trying to set up a custom layout for bzr-svn. The setup is almost like "trunk", where branches are found under branches/ but the actual trunk is in trunk/product. Adding "branches = branches/*;trunk/product" to my subversion.conf leads to an error when I try 'bzr svn-layout URL': bzr: ERROR: does not support co-located [20:26] I tried wiping out my ~/.bazaar/svn-cache to no avail [20:28] briandealwis, hi [20:28] hi jelmer [20:28] briandealwis, can you post the full backtrace? [20:28] sure, hold on [20:29] jelmer: http://pastebin.com/wu9RGZNw === yofel_ is now known as yofel [20:44] briandealwis, is it just "bzr svn-layout" or do other commands give this as well? [20:46] Hmm, good question: I hadn't tried. But trying to branch the trunk seems to work. [20:47] jelmer: huh, and 'svn-layout URL/trunk/product' works [20:47] as does 'svn-layout URL/branches/<>' [20:48] but 'svn-layout URL/branch' or 'URL/trunk' fails [20:49] briandealwis: Looks like it's just invalid handling of a particular exception at the moment [20:49] briandealwis: can you file a bug about this? [20:49] will do [20:49] thanks jelmer [20:49] thank you :) [20:54] there is really no bzr record/staging? [20:55] I have to shelve everything and then unshelve and commit? [20:58] santagada: hi [20:58] santagada: yes [20:59] jelmer: is there any plans to implement something like the hg record extension? [20:59] I just want to filter my commits [21:00] santagada: What does hg record do? [21:01] jelmer: I think it is mostly the reverse of what shelve do, the same as git interactive commit [21:03] jelmer: it asks if you want to commit each piece of a file/the whole file [21:04] jelmer: very simple and makes me wonder why bzr has only the inverse (shelve) [21:04] well it is useful, if you want to test each patch before commiting [21:05] is there an easy way to exclude a particular file from qdiff? [21:06] something like -x for commit [21:07] santagada: I don't think there is anything like that (yet) === zyga-food is now known as zyga [22:43] hi all [22:45] poolie: hey ! [22:52] 'evening poolie, vila [22:54] jelmer: _o/ [23:11] vila, thanks for updating all the bugs [23:12] closing the closed ones, i mean [23:12] not just today [23:12] poolie: hehe, -w option on check-newsbugs is a treat :) [23:14] is it? [23:14] is that for --write? [23:14] tnhat's cool [23:14] --webbrowser [23:15] when they are all opened in tabs, it's really easy to update them [23:15] i think we talked at the sprint about systematic release naming [23:15] i think we should [23:15] i noticed openstack are doing this too [23:15] really ? I may have missed that one [23:15] you mean, giving names to bzr releases ? [23:18] yes [23:18] at the moment the naming's a bit arbitrary [23:18] hmm [23:19] i think we talked about matching the letter of the ubuntu release they synchronize with [23:19] you're ~2h late for 2.4b4, or are you thinking about the series instead ? [23:19] ha, right series then [23:20] well, I'd like that too, if only to make references simpler [23:20] the series [23:21] and/or the final release [23:21] i don't think betas need names [23:22] yup, nor point releases [23:23] so we need a scheme [23:23] perhaps famous markets in the world [23:23] haaa, worth searching a bit [23:24] or islands [23:24] needs something with reasonable coverage of every letter [23:24] yup, well, at least the remaining ones ;) [23:35] what island begins the O? [23:35] Oahu? [23:36] or Old Spitalfields, as a market (according to wp), though the 'Old' is cheating a bit [23:36] Oronsay [23:36] I'm not convinced the world has that many famous markets [23:36] there's a lot of Oronsays apparenttly [23:36] yeah, that is a bit of a problem [23:37] using a scottish one would be nice [23:37] I agree :) [23:38] Orkney, Outer Hebridies [23:39] Oigh-Sgeir, could start a trend of hard to pronounce ones :) [23:41] Riddell, :) [23:47] ok, 2.4b4 frozen, uploaded, already fixed ... and babune on its way to turning blue again (including OSX !) leaving only windows as the ugly duckling ;) [23:48] that was quite a day (including my elder daughter starting her final exam and generating a lot of stress across the whole family ;) [23:48] so, have fun guys and see you tomorrow, well, in a few hours