[00:17] doc patches welcome :) [00:20] lifeless, bug 613247 may interest you. [00:20] Launchpad bug 613247 in Bazaar "TestCase instances may not be collected till selftest finishes (affected: 1, heat: 6)" [Undecided,New] https://launchpad.net/bugs/613247 [00:20] mgz: I saw, it scared me. \o/ qa. [00:21] should have found the time to work that one out fully ages ago, given how annoying it's been and easy it is to fix. [00:21] ^ah, this is a special new bug filed for the underlying testing thing, rather than just the logging module related fallout [00:25] Should bzr just work with 2.7 or only some versions? Is there a list somewhere or compat matrix of some sort? [00:26] lifeless: sorry to be dumb. I updated to https://edge.launchpad.net/~bzr-beta-ppa/+archive/ppa, but it doesn't include the bzr package, so how do I get that package to 2.2? [00:26] Meths: there are some bugs, but it's not far off working. I've got another five or so to file with bzr and/or python [00:27] mgz: Okay, thanks. [00:27] mrjazzcat: I'm not sure, I'm not really up on the bzr ppa story these days. [00:28] lifeless: makes sense. I'd rather point this customer to a package than to source, that's all. [00:29] sure; spiv or poolie may have more answers [00:31] lifeless: I see the answer. bzr beta package is here; https://edge.launchpad.net/~bzr-nightly-ppa/+archive/ppa [00:32] mrjazzcat: it's only in the nightly for now [00:32] we should get it in the beta ppa soon [00:33] poolie: hey Martin. Got it. Thanks. Good luck with it all. cheers === rubbs is now known as world === world is now known as rubbs [00:37] hello. I'm working on an external tool (a git remote-helper for interacting with bazaar) and I'm having difficulty with figuring out some things about how to use bzrlib.. === mrjazzcat is now known as mrjazzcat-afk [01:58] LeLutin_: what in particular? [01:58] hi spiv [02:07] poolie: for now, I'm trying to discover branches that can be found on a URL. I think I may have found what I needed (BzrDir.open_containing) but I'm not sure yet.. [02:08] yes, that's where you should start [02:08] poolie: oh.. also, I was wondering what made urls like "lp:something" work. I checked out the code from the bzr-fastimport plugin and it seems to automatically receive a "canonical" url [02:08] there's a directory service [02:08] i think looked up by get_transport [02:08] that ends up asking lp what actual url it should use [02:10] hmm ok. I tried giving Branch.open an "lp:.." url and it raised an exception saying that $pwd/lp:.. is not a branch [02:11] you may need get_transport/open_transport [02:11] what are you trying to build? [02:12] oh wait.. must be a plugin that's responsible for handling those then [02:12] yes, the lp plugin does them [02:12] calling bzrlib.initialize should register it [02:13] oh ok. I'll try that [02:13] hm, i'm not sure if that loads plugins by default [02:14] to answer your question: I'm trying to build a remote-helper script for git that will make it possible to add bzr branches/repositories as native git branches [02:15] poolie: is it possible to selectively load bzr plugins? [02:15] interesting [02:15] probably, i don't recall how [02:17] if you look in plugin.py something like calling _load_plugin_module should do it [02:19] poolie: oh ok.. well... thinking a bit more about it, maybe it's a bad idea. people could have other transport plugins that implement different types of urls [02:20] right [02:20] unless you specifically expect some plugins to cause trouble, i'd just load everything [02:20] most of them should only register their capabilities and not actually do anything until asked [02:24] right [02:27] hmm.. AttributeError: 'module' object has no attribute 'initialize'. ipython to the rescue [02:33] hmm. the bzrlib installed on my laptop doesn't have an initialize function [02:35] I'm at bzr 2.1.1-1 (in lucid lynx) [02:50] LeLutin_: ah in that case you have to set up some things manually [02:50] primarily set up the ui_factory and load plugins [02:56] poolie: is it possible to find the API documentation for another version than the last one? [02:57] LeLutin_: hm [02:57] for now, I have this: http://people.canonical.com/~mwh/bzrlibapi/bzrlib.html -- but it's reffering to the latest revision of bzr [02:57] you can run 'make apidoc' in a tree [03:05] oh.. have to d/l the whole 2.1 branch then [03:08] aha, there's a bzr-doc package :) should be faster to d/l than having to wait for launchpad [03:24] poolie: good, bzrlib.plugin.load_plugins() seems to have added the necessary. [03:24] I thought using BzrDir.find_branches(..) would get me a list of branches, but I don't know where to find the transport object that I need to give to it :\ [03:51] LeLutin_: transport.get_transport() [03:53] oh, good. I had found BzrDir.root_transport but it spit out an error about list_dir not being implemented for http :( [04:34] how do I bzr log commit made by a specific committer? [04:44] spiv: hi, did you get anyone to land your codebrowse-oops branch yet? [04:51] mwhudson: not yet [04:54] spiv: ok, let's see if i have the necessary bits set up then :-) [05:00] spiv: can you add a commit message? [05:02] I'm not really here today. [05:03] ok, i'll think of something [05:33] GungaDin: i'd either use grep or maybe write a log formatter that does it [05:33] hi spiv [07:11] poolie: hi. could you get someone to look at bug 613066? it's currently blocking work on a branch, although i hope that recreating the branch and copy over the files will work around it (haven't had time to test it yet) [07:11] Launchpad bug 613066 in Bazaar "IndexError when running bzr status (affected: 1, heat: 6)" [Undecided,New] https://launchpad.net/bugs/613066 [08:14] hi BjornT_ [08:24] BjornT_: can you tar up that tree and put it on eg chinstrap? [08:26] poolie: sure, just a minute [08:27] it sounds a bit like a dupe but i don't think we've hit exactly that [08:27] also if you recall what you did in this tree that would help [08:31] poolie: the tree and repository (it's a lightweight checkout) is in ~bjorn/bug-613066 on chinstrap [08:32] poolie: i'll look through the command history to see if i'll find something special. the only thing i can remember is that i used bzr-pipeline and switch pipes just before doing changes, and then getting this error when trying to run bzr diff [08:36] poolie: more specifically, before the error i did: bzr pump; bzr lp-propose; bzr switch-pipe :next [08:36] after that i modified some files, and got this error running bzr diff [10:47] hi. how does bzr handle the case when one branch adds a file 'test' and another adds a dir 'test'? [10:52] It would be a conflict, just like if two branches both added files 'test'. [12:18] i have registered a branch on LP and did an initial commit. later i've added more files to the local dir and tried to push them to LP with 'bzr diff' which was unsuccessful. any idea what am i missing here? TIA. [12:29] svaksha: i am not sure i understand. is 'bzr push' failing? [12:35] parthm: i'm only able to push the diff's for the old commit. the newly created files dont get pushed. [12:37] svaksha: are you seeing an error message on 'bzr push'? [12:37] svaksha: you need to add the new files to the repo with 'bzr add' [12:37] then commit, then push that [12:38] unless the files are added, they show as "unknown" in "bzr status", and don't show up in the diff [12:38] bzr add w/a argument adds all unknown files in the tree [12:39] abeaumont: that gives the error, ERROR: No WorkingTree exists [12:39] svaksha: the add needs to be done wherever you do the commit [12:40] * svaksha checks again [12:40] a working tree is the stuff with the actual content files, when you can make changes, run the code, commit, produce diffs, etc. [12:40] s/when/where/ [12:46] ddaa: i repeated the process --bzr add, diff,commmit and push, but it only commits to the local dir and push/add does not push new files into LP [12:46] * svaksha is missing something [12:49] svaksha: what does 'bzr push' say? [12:50] parthm: ssh: bazaar.launchpad.net: Name or service not known [12:50] bzr: ERROR: Connection closed: Unexpected end of message. Please check connectivity and permissions, and report a bug if problems persist. [12:50] ^^ latest ..earlier i didnt get this error [12:51] svaksha: looks like a transient network error. i have seen this sometimes but it works when i trying again. [12:53] parthm: i hope so. /me will check later then. thanks for the help [12:53] ddaa: thanks [12:56] parthm, it worked just now. probably was a network issue afterall :) [12:56] svaksha: :) === oubiwann is now known as oubiwann-away === oubiwann-away is now known as oubiwann === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-lunch [16:47] is it possible to set options for bzr-grep in the bazaar config file? [16:48] LeLutin: You can set an alias === beuno is now known as beuno-lunch === Ursinha-lunch is now known as Ursinha === deryck is now known as deryck[lunch] [18:02] jelmer: yes of course :) [18:05] jelmer: hmm well in fact I can't really set an alias for a command with a space in it.. alias 'bzr grep'=.. is not valid [18:05] LeLutin: Don't include the 'bzr ' bit, that is implicit. [18:07] I already have an alias for grep but bzr-grep doesn't use it [18:12] LeLutin: not when invoking "bzr grep" ? [18:15] hey there, i'm looking for a way to check what's the current revision i use [18:18] jelmer: right. the alias is defined as "alias grep='grep --colo=auto'". when I do a normal "grep something" the color argument is there but when invoking "bzr grep ..." it's not activated [18:19] Ddorda: "bzr revno" inside the workdir [18:19] LeLutin: awesome, thanks === beuno-lunch is now known as beuno [18:21] LeLutin: is this in ~/.bazaar/bazaar.conf ? [18:25] jelmer: nope. this is in my .bashrc :P I don't know how to add the necessary config in bazaar.conf === deryck[lunch] is now known as deryck [18:50] jelmer: nm, I've found a web page about how to setup bzr command aliases and it works fine. [18:50] jelmer: thanks for your help [18:51] np [19:22] FAIL: ShortReadvError: readv() read 0 bytes rather than 6046 bytes at 4096 for "d33aaa4aea5de86f6370880d52c949f5.cix" [20:18] mtaylor: shoot your file system [20:20] Right between the inodes? [21:38] lifeless: ok [21:38] lifeless: well, my local solution was to delete the shared repo, recreate it and re-branch [21:38] lifeless: but, tbh - that's still sort of fail [21:40] mtaylor: your file system only wrote some of the file to disk. It is epic fail. [21:43] lifeless: indeed. but I'm not sure I have any control over that filesystem ... so I'd sort of like the thing on top of it to re-fetch from remove if local file is ass ... [21:43] fwiw [21:43] mtaylor: it doesn't have a clue where remote is at the point it finds out that failure [21:44] mtaylor: it would be like drizzle going back to an appserver a week later and saying 'hey, can you give me that users full name again? thanks!' [21:44] hehe [21:44] mtaylor: or are you saying this happened without a crash of some sort in the middle? [21:45] lifeless: it is possible there was a disk-full situation on the machine [21:45] ah [21:45] mtaylor: and your fs didn't tell userspace. win. [21:45] mtaylor: you might like to put that in the 'perhaps we should fsync' bug [21:45] ah. that explains it - we're running xfs there [21:46] * mtaylor shoots stewart [21:46] bzr assumes, reasonably, that when it writes and closes a file, the fs considers it 'ok' [21:49] lifeless: just for my edification - in the future, is there any way to say to bzr "hey, could you just delete any corrupted shit you've got" so that on a subsequent pull operation it would just fetch unknown objects as usual? [21:49] rm -rf :( [21:49] k [21:49] it would be nice to have a fsrepair thingy [21:49] well, that's what I did this time [21:49] and then I was like "wow, I haven't branched drizzle from scratch in a VERY long time" [21:58] LeLutin: there is also a 'bzr alias' command that operates very similar to the bash alias command === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-afk [23:17] abentley: Hi! [23:18] abentley: I've had another look at implementation infrastructure for custom merge directives in bzrlib, and was wondering if you could have a quick look to see if my approach is reasonable, when you have a moment.