[00:01] hmm [00:01] so it is [00:01] and ist isn't [00:02] ? [00:10] not all log viewers show bugs [00:10] but yeah, I can see the case [00:10] so we want a [00:10] #bug:lp:3423423432 [00:10] or something [00:10] and a hook that can edit the message and rev props late [00:16] yeah [00:29] is there some setting that needs to be present before I can receive karma on launchpad for committing code? [00:29] no [00:30] but you ahve to push the code to lp [00:30] also, karma isn't updated instantaneously [00:51] bbiaw [01:24] good evening === dOxxx1 is now known as dOxxx [01:25] when should bzrlib.trace.show_error etc be used instead of the ui_factory.show_error etc? [01:47] http://bazaar.launchpad.net/~sproaty/whyteboard/development/files how is my 'whyteboard' folder shown at version 170 when I've commited several times to files inside it since? [01:47] the dir was amde in 170 [01:48] sproaty: Bazaar tracks directories as their own objects. Presumably it's showing when the directory itself was modified, rather than its contents. [01:48] * Peng shrugs [01:49] figured [01:49] kinda a shame# [01:49] https://bugs.launchpad.net/loggerhead :D [01:51] heh, cheers [01:51] night [01:56] I wasn't joking. [03:04] if I'm writing a bzr plugin which works with launchpad branches via the lp api, should I be using the lp_api module from the launchpad plugin or just using launchpadlib direcrly? [03:05] directly* [03:35] hi. I started 'bzr upgrade' on a launchpad branch to upgrade it to 2a, but I haven't upgraded the repo yet. will it be a problem to kill the upgrade and do the repo first? [03:38] oh, heh, it finished. Never mind! [04:14] i checked out a branch from launchpad to fix some bug and pushed changes to my own branch on lp. I would like to remove that revision from launchpad, is it possible? [04:15] enli: Completely remove all traces of it? Is this just, like, you misspelled something, or did you accidentally commit the nuclear launch codes? [04:16] Peng: the code is working, its just that i want to change the commit message [04:17] enli: Ah. How bad is it? Mostly it's not worth doing. You can see all sorts of fun errors in lp:bzr's history. [04:17] not too bad, i am just playing with it.. wondering if it is possible :D [04:20] enli: It's not hard, but it's not clean either. [04:21] enli: You'd have to "bzr uncommit", "bzr commit again" and then "bzr push --overwrite". This creates a completely different revision, so people who have already pulled the old one will get confused. [04:21] enli: Like I said, it's usually not worth it. [04:22] ok, so if my latest revision was before uncomit, 198 and i do uncommit, commit, push what will be the revesion number? [04:23] enli: 198 [04:23] enli: But the ID and everything else will be different. [04:25] Peng: i dont know yet what ID and everything else is, but i will take your word and not uncommit it, thanks a lot [04:26] enli: bzr log --show-ids will display it, FWIW. [04:26] enli: It's not very interesting. It's just a more complicated, unique ID for the revision. [04:27] hmm, so if nobody has yet pulled my branch, it will be pretty safe to push --overwrite.. yes? [04:38] Peng: i am out, thanks again [05:50] having a problem when committing on my server that bzr hangs [05:50] Any help much appreciated [06:14] hi [06:15] is it possible, with the understanding that i will break trees, to remove a commit, and all subcommits, from a branch? [06:16] sure, uncommit [06:16] i'm not uncommitting that far back [06:17] meoblast001: uncommit takes a -r argument. [06:25] Peng: so i can remove 1 specific commit, while keeping everything else in tact? [06:25] if it's the most recent one [06:25] bob2: <3 [06:25] it's not [06:25] that's the problem [06:25] do you care this much |<------>|? [06:26] If this is really the nuclear launch codes, you can use rebase or fastimport or somesuch to rebuild the branch. [06:26] care about what? [06:27] about removing this commit [06:27] yes, i suppose [06:28] meoblast001: Just curious, why? [06:29] well, currently, until i get legal stuff worked out, i need to make sure i own copyright on every ounce of code in my project [06:29] there is a small amount of unnecessary code made by others [06:29] so i wanted to remove it completely [06:30] Do projects usually complete expunge the old code in such situations? [06:30] completely* [06:31] don't think so [06:31] but i'm very paranoid [06:31] already freaking out with stress [06:32] well, how will you separate it out? [06:32] e.g. if you branch from before the first potentially-dodgy commit, what do you want to do? just exclude the entire commits by other people? [06:32] well, it's only really one [06:32] i've been lucky enough to not get many contributors [06:32] which is good, considering i have copyright issues to work out [06:33] is this free software? [06:33] yes [06:38] i think i'll work this out tomorrow [06:38] i'm going to bed [06:38] good night [08:33] anyone know the syntax for drush generate-content [11:59] hey guys, is there something wrong with website? [11:59] I can't connect to it [11:59] Hi, I've done 'bzr add foo' but haven't committed yet. I want to remove foo from the list of outstanding things to commit. How? [11:59] mathrick: I concur. [12:00] ralph: bzr rm --keep --new foo [12:01] mathrick: Thanks. I understand. Bye. [12:41] vila: ping [12:53] vila: ping: unknown host vila ... hmm maybe i will try later :-) [15:08] hey all [15:08] trying to install bazaar for the first time on win xp sp3 with this tutorial [15:08] http://wiki.bazaar.canonical.com/WindowsInstall [15:09] now unfortunately i got python 2.6 running and i cant seem to find the pycurl package for 2.6 only for 2.5 [15:09] what should i do? [15:11] or is there even a crucial reason to use the python based installation instead of the stand alone one?? [15:20] hi, i'm desperate. how can i get branch base path in precommit hook? === khmarbaise_ is now known as khmarbaise [15:21] just solved it. http://curl.haxx.se/mail/curlpython-2009-11/0001.html for a python 2.6 pycurl msi file === goundy is now known as elgg === elgg is now known as Guest91366 [15:58] its me again... i ran into following error while compiling bzr-gtk. http://paste.pocoo.org/show/186339/ [15:58] what shall i do now? === Guest91366 is now known as TabMe [16:45] hi! does the rebase plugin have any active maintainer? [16:45] I tried to use it but could not get it working [16:46] using google does not bring up any tutorials or anything [16:47] ah, OK, it's maintained alright.. [16:48] http://people.samba.org/bzr/jelmer/bzr-rebase/trunk/ has had it's last update on Sat 2010-02-27 15:28:05 +0100 [16:49] oh, well.. maybe I some day have time to look in to the inner workings of the plugin to be able to use it [16:50] Wellark: yeah, it's maintained but there isn't much documentation other than the command-line help [16:50] it's quite similar to rebase in git and hg though [16:51] ok, well it's probably just a user error on my side [17:35] I saw that inkscape recommended keeping a checkout the remote branch, and doing development in a separate branch, and then merging that into the otherwise pristine branch and then pushing, instead of pushing from the branch you actually wrote the feature in. [17:35] and then i did some googling and saw this thread: http://www.mail-archive.com/maria-developers@lists.launchpad.net/msg00220.html [17:36] so i was wondering if the reverse merge thing they were talking about was on the roadmap [18:55] is there something like git's commit --amend? [19:00] bzr uncommit && bzr commit? [19:02] well, yeah, but it makes me write the commit message again [19:02] bzr uncommit && bzr qcommit :) [19:09] at least some links in 2.1 user guide are broken, they specify the old-style URLs: http://doc.bazaar.canonical.com/bzr.2.1/en/user-reference/bzr_man.html#revision-identifiers [19:10] which doesn't resolve, because this chapter is available at http://doc.bazaar.canonical.com/bzr.2.1/en/user-reference/revisionspec-help.html [19:10] the faulty link is here: http://doc.bazaar.canonical.com/bzr.2.1/en/user-guide/specifying_revisions.html === radoe_ is now known as radoe [20:04] #chromespot [20:04] oupse [20:04] sorry guys [20:04] I missed the /join [20:04] this is not advertisment -_- [21:06] https://blueprints.launchpad.net/bzr/+spec/nested-tree-support links to http://bazaar.canonical.com/NestedTreeSupport, which 404s [21:30] For my own bzr server; how hard is it to convert something like this bzr+ssh://user@code.domain.com/home/user/connect/foo to something like this kalliki:~user/connect/foo ? [21:30] I remember there being some nifty way to do something like that but I forgot how [21:30] MTecknology: what do you mean by "convert"? [21:30] and never got it to work before [21:31] mathrick: so I can do bzr branch kalliki:~user/connect/foo and it will change it to the really long one [21:32] mathrick: sorry for the poor question - long day already [21:34] MTecknology: "converting" it as such is simple, as long as you have the access to a shell or smart server and install your own hooks [21:34] but no idea how hard it'd be to make bzr actually do that [21:35] you'd need plugins both on the client side and on the server side [21:35] because client needs to grok foo:bar as a remote ssh reference, and the server needs to do the actual translation [21:36] MTecknology: perhaps it'd be easier to make a plugin which just works on the client side and translates ~foo to /home/foo by default, unless you provide a map to override that for a particular host [21:36] I'm just trying to make things simple for my users [21:37] then you definitely want to avoid relying on a particular server + client combination [21:37] client side plugins sounds doable [21:37] make it a client-only plugin, it'll be a bit more manual for machines where ~foo isn't /home/foo, but at least it'll work with any host as long as you use the proper client [21:38] MTecknology: if you need to override the branch reference syntax, take a look at lp:bzr-colo [21:38] I should be able to make something that can just be unpacked into ~/.bazaar/pulgins/, right? [21:38] it installs a colo: prefix [21:38] yes [21:39] oh, right, and beware of clashes with the built-in colon prefixes [21:39] date:, branch:, tag:, submit:, before: and last: at least [21:40] though of these I think only branch: and submit: are valid branch syntaxes [21:40] I was thinking probably just k: or kalliki: [21:40] well, and lp: of course [21:40] MTecknology: but you want to provide that for any host [21:40] otherwise it'll be just confusing for users [21:40] well, maybe not more than lp: [21:41] but still, if you attempt to support widely-used syntax (SSH), you'd better make it work like everything else that supports it [21:41] so, where should I get started building this plugin? [21:44] I'll obviously need to start understanding python a little bit :P [21:46] mathrick: do you know where the plugin is that handles lp:? or is that probably much more in depth than what I need? [21:47] MTecknology: lp: and colo: are the same deal [21:48] MTecknology: I think bzr plugins -v should tell you [21:49] the plugin you want is called launchpad [21:49] I'm seeing .pyc files, do I need to compile the python before it can be run? [21:49] eh, no [21:49] that's done automatically by python whenever you invoke a .py file [21:49] oh [21:50] running ubuntu/debian? [21:50] ya [21:50] if so, you have them precompiled [21:50] sec [21:50] looks like test_lp_directory.py is the file that I should look into [21:50] quite possibly [21:59] mathrick: wow; this doesn't look friendly :P [22:00] lp is kinda huge [22:00] yup [22:01] I don't think my mind thinks in python [22:01] but you shouldn't be looking at test_* [22:01] that's, as the name suggests, tests [22:02] hmm, the netloc thing might not really be suitable for the kind of thing ssh foo:bar syntax needs [22:03] MTecknology: I'd wait until somebody more familiar with bzr's guts comes [22:04] test_lp_directory.py has 'lp:' in it; which makes me want to look at lp_directory.py [22:05] mathrick: ok, thanks for pointing me in the right direction [22:06] MTecknology: lp_directory is where it's at, but it might not be suitable a method for registering a catch-all prefix [22:06] which is why I suggested waiting [22:06] and you're welcome [22:09] It's not really that bad because it remembers the branch push/pull locations; just be easier when we get more going on [22:10] yeah, though I could see it useful in an environment with many branches going on [22:10] if that's the case, a generic branch-reference-shortener could be useful [22:10] so you could say b:mtecknology:foo for example [22:10] MTecknology, I can probably help [22:10] or b:main [22:11] asabil: I can probably eagerly listen :D [22:12] mathrick: how would I do that? [22:12] http://pastebin.com/HHgSxp5M [22:13] MTecknology: by registering a b: URL scheme, like lp:, and then having a map from shorthands to full locations somewhere in the config file [22:13] but that only makes much sense if you really operate in a forest and need to refer to many branches often [22:14] asabil: but that's only for a fixed prefix:, no? [22:14] can you make it catch any not otherwise recognised prefix? [22:15] mathrick, well you wanted something like lp: no ? [22:15] ya [22:15] well, I wanted something like scp host:dir [22:16] This is long: bzr+ssh://user@code.domain.com/home/user/connect/foo [22:16] that's what I'm trying to change [22:16] to k:~user/connect/foo [22:16] just upgraded to 2.1, but now I get the following error when pushing to a remote repository: "AssertionError: _remember_remote_is_before((2, 1)) called, but _remember_remote_is_before((1, 6)) was called previously. [22:17] MTecknology, what I pasted answers your question [22:18] asabil: what about the username@ part? [22:18] MTecknology, you can do whatever you want with the url [22:18] you parse it the way you want [22:19] So I need to format this? -> return "bzr://code.kalliki.com/repo" + name [22:19] yep [22:19] for example [22:28] asabil: Is there any way to decopmose the whole path into little pieces? [22:28] So I could grab the first part of the path [22:29] split ? [22:29] ya [22:29] "hello:world".split(":") [22:29] for example [22:29] oh [22:30] that first part of the path will be the user name as well as the home directory [22:30] so k:foo/bar -> bzr+ssh://foo@blah.com/home/foo/bar [22:31] you remove the prefix [22:31] and then split on th "/" [22:31] then toss the first part into a variable [22:31] user, directory = uri[2:].split("/", 1) [22:32] (it is not the most beautiful python code :P) [22:32] if url.startswith('prefix:'): [22:32] I change that to my prefix? [22:32] yes [22:33] the uri[2:] it to remove the 1st 2 characters [22:33] oh [22:34] what's the else: for? if I matched that it start's with k: [22:34] the else shouldn't happen [22:36] if url.startswith('k:'): if url starts with k: name = url[2:] then remove the first two letters and store in name else: but if it doesn't start with k: [22:37] OH! [22:37] does indentation matter in python? [22:38] I had it wrong when you copied things and I read it as the return happens in the else :P [22:44] asabil: that's what goes in __init__.py ? [22:45] yes [22:46] cool - an error [22:48] ? [22:49] I tried return 'bzr+ssh://' + name.split("/", 1) + '@code.kalliki.com/home/' + name [22:50] name.split("/", 1) returns an array [22:50] whose size is maximum 2 [22:50] oh [22:50] you need to extract the element [22:55] asabil: ok, I thought it was returning the 1th item from it :P [22:55] I guessed so [22:57] hi, I'm not a bzr user but I want to check out some code from trunk/tip whatever you guys call it :) [22:58] will someone be kind enough to point me the URL to check out the 0.10 branch for https://launchpad.net/ipython [22:58] elpargo, bzr branch lp:ipython [22:59] or lp:ipython/0.10 [22:59] if you want the 0.10 series [23:00] yea I believe that is what I want as it has the bugfix but there is no 0.10.1 [23:00] asabil: what's this lp? is that with an existing checkout? [23:02] elpargo, lp is a shorthand for launchpad [23:03] asabil: so if I want to force an error to happen, now do I do that? [23:03] do the same as what is in the else: [23:04] asabil: interesting. That seems handy. [23:04] elpargo, you are welcome [23:11] hello [23:11] I'm tryig to serve a branch downloaded from launchpad with "bzr serve" and I get this error: [23:11] bzr: ERROR: Invalid http response for http://localhost:4155/.bzr/branch-format: Bad status line received [23:33] asabil: :D absolutely perfect :D [23:33] cool [23:34] asabil: wanna see? [23:34] yeah sure [23:34] http://dpaste.com/168836/ [23:35] asabil: team|user/project/branch :D [23:36] looks good [23:36] asabil: namesplit = name.split("/", 3) [23:36] sorry [23:37] asabil: thanks :D [23:39] mathrick: ^ In case you're curious what the solution was