[00:31] lifeless, any reason in particular for the hearts today? [00:36] bzr+ssh bzr-search support [00:36] ah :) [02:23] hi poolie [02:24] hi james [02:33] class OutOfTea(errors.BzrError): [02:33] :) [02:38] Being out of tea sounds more like a solution than a problem to me... :p [02:40] page KombuchaKip :) [02:40] it's an example error in jelmer's test code [02:44] * KombuchaKip lol [03:18] hey poo [03:18] poolie* [03:21] :-) [05:38] hi. [05:39] Is it possible to install bzr 2.5 on python 2.5 ? [05:39] (from source) [05:42] hi sobersabre, no, i think bzr 2.5 needs python2.6 [05:43] but 2.4.x will be ok [07:39] bus time [07:46] hi wgz [07:48] poolie: thanks. I did manage to install 2.4 on py2.5 [07:49] good [09:06] morning! [09:09] hey there folks [09:11] hey [09:29] hi mgz, vila [09:37] hi mrevell [09:37] Hey there poolie and everyone else. [09:49] really must finish option stuff this morning [09:56] ok [09:56] mgz i should go, but could you look at https://answers.launchpad.net/bzr/+question/179852 [09:56] and maybe some others [09:56] i have loggerhead just about baked serving tgzs [09:56] i thought of you :) [09:57] poolie: will do [09:58] ah, I'd never run an ssh server on a windows box [09:58] i would like to just hook up 'bzr serve --ssh' [09:58] how hard can it be? :/ [09:59] not very I imagine, on nix at least :) [09:59] :) [09:59] well, using paramiko [09:59] trying to take care of some of the hookup issues people have [09:59] on windows [10:02] I agree we need a better server answer, I do bzr+http with apache doing the access control [10:02] and that's not trivial to set up correctly. [10:03] jelmer: you landed a proposal I wanted to comment on too fast ;_; [10:04] but I was stuck thinking of a better suggestion, so maybe it's not really that bad [10:08] ok, night europe [10:08] night poolie! === MacSlow is now known as MacSlow|lunch === yofel_ is now known as yofel === Odd_Blok1 is now known as Odd_Bloke === MacSlow|lunch is now known as MacSlow [14:27] ah, you're working on the config test failures vila? [14:27] I was just looking at that [14:27] I have a fix [14:27] It's so silly... [14:27] mgz: By the way, hi ! ;) [14:27] the tests look dodgy, they use a relative url like "/home/user/project", which naturally doesn't resolve anywhere sensible on windows [14:28] it doesn't have to [14:28] it's the caller responsibility to provide an appropriate location and the code base already ensures that [14:29] but the location stack resolves it using abspath [14:29] for unclear reasons I felt the need to add a location = urlutils.normalize_url(location) which broke these tests [14:29] which looks at the real filesystem [14:29] right [14:29] that's the issue. [14:29] removing that code doesn't break any test and make the windows ones pass [14:30] I probably had another issue while adding that and addressed this issue in the end [14:30] okay, that's fine then. [14:31] I don't think I need to say that LocationStack should be clear about if it wants a url or a file path, and isn't currently :) [14:31] I wanted to talk with you [14:31] but thought I should reload the context [14:32] okay, fresh page. [14:32] and I knew these lines were in scope, so, I tried removing this suspicious call first... [14:32] and my sub-conscience won :) [14:33] hmm [14:33] yeah, this could be clarified... [14:33] s/could/should/ [14:33] so, I have a design twinge that path functions should be clear about whether they look at the local filesystem or not [14:34] you were basically caught out here by 'normalize_url' looking safe and cuddly, but actually calls abspath which depends on the filesystem and cwd [14:35] I think there is a FIXME in the location config code saying that we accept synonyms for local paths as file: urls which is an issue in itself [14:35] yeah, I vaguely remember adding that because I suspected some relative paths bogus handling [14:38] mgz: https://code.launchpad.net/~vila/bzr/892992-locations-on-windows/+merge/83299 [14:39] yeah, in addition to looking safe it silently made the section matching fail, hence the test failures [14:39] this hints that I should have tests more specifically targeting the section matching [14:40] in other news, for unknown reasons the osx slave suddenly decided that it was 'Unable to locate a Java Runtime to invoke.' [14:40] huh ? [14:41] rubber stamped. [14:41] blame java! [14:41] dong a software update brought a java *update* and solved the issue but the root cause for this error message is puzzling [14:42] was the precious version disabled ? How ? [14:42] nobody is tweaking this osx slave (except me ;) [14:42] macs are mysterious objects of wonderment and quizzling [14:42] mgz: thanks ! [14:43] best not approached, or you may come to regret it.. [14:43] hehe [14:44] mgz: by the way, regarding your option help tests, try bzr-diffstat, istm one option escapes your check [14:45] the plugin overrides cmd_diff, that may explain the why, but I didn't dig enough to propose a how [14:45] the global --file option at least is unchecked, which doesn't hurt much [14:46] and plugins (including builtin ones) aren't either, but would be nice to provide an easy way for them to subclass the test an opt in. [14:46] how would a plugin enumerate all the commands it registered? [14:47] some plugins already uses a list *to* register [14:47] :) [15:00] slangasek: ping, see bug #884270 [15:00] Launchpad bug 884270 in Ubuntu Distributed Development "bzr should do smarter merging of .po files" [High,In progress] https://launchpad.net/bugs/884270 [15:12] vila: pqm gave failed your branch? [15:13] mgz: no, it landed it, what made you believe that ? [15:13] * vila forces a run on babune windows slave [15:17] ...launchpad being slow, sorry [15:18] no worries, just curious [15:18] confused by looking during the window between the PQM page showing no jobs, but the review page still having the branch as approved [15:18] oh yeah, I know that one [15:19] sometimes the pqm page shows the end of the previous submission but never for long [15:58] mgz: down to 7 failures (with 3 being spurious not ?) [16:00] those are all the windows/sftp rename race [16:01] good, down to one root cause then [16:01] which is now bug 891602 [16:01] Launchpad bug 891602 in Bazaar "PermissionDenied test failures on windows with sftp" [High,Confirmed] https://launchpad.net/bugs/891602 === Quintasan_ is now known as Quintasan [19:07] hmm, more feature flags work this evening [19:33] ...are you trying to create more of a backlog for the poor patch pilot when he wakes up in the morning? [19:38] I'm just worried he'll run out of things to review. [20:50] ah, now I'm back home I can run my auto-ignore launchpad tag spam imap script [20:50] which is nice as they've now removed all the tags that accidentally got added earlier... [22:08] an interesting data point [22:09] $ bzr branch lp:nux/1.0 nux-1.0 [22:09] Branched 497 revision(s). [22:09] HPSS calls: 30 (0 vfs) [22:09] $ bzr co lp:nux/1.0 nux-1.0-checkout [22:09] HPSS calls: 48 (15 vfs) [22:09] checkout really that much more expensive in round trips? [22:27] thumper: hi [22:27] hi jelmer [22:47] hi all [22:48] thumper: hmm, thatÅÅ really odd [22:48] it should basically just be "bzr branch" + "bzr bind" [22:50] hi jelmer [23:28] Are there any plans to hook the Nautilus rename command into the bzr rename functionality when using the plugin for Nautilus on a repository? [23:29] KombuchaKip, I don't think so, there isn't a lot of activity around bzr-gtk [23:29] jelmer may know [23:29] i doubt anyone is specifically planning to do it [23:29] it seems like a sensible thing to do though [23:29] iirc hg dropped their nautilus support [23:30] beuno: Oh really? I didn't know that. Is the plugin dead? I created this bug report about it. https://bugs.launchpad.net/bzr-gtk/+bug/797438 [23:30] Ubuntu bug 797438 in Bazaar GTK+ Frontends "nautilus-bzr plugin does not hook rename" [Medium,Triaged] [23:30] do you want to try a patch for it? [23:30] it's not dead [23:30] right now i'm trying to get tarball downloads from loggerhead hooked in to launchpad [23:30] poolie: Sure I'd try a patch for it. [23:30] almost there [23:31] poolie: I'll be around, just buzz me whenever you're ready. I am using bzr from the PPA. [23:33] hi beuno, KombuchaKip [23:33] KombuchaKip: IÇ've done some work recently to port nautilus-bzr to the new Nautilus plugin API [23:33] o/ jelmer [23:33] o/ poolie [23:33] hey there [23:34] jelmer: Ah, so that's why its sort of idle right now. I forgot that everything has to be moved to Gnome 3. [23:35] * KombuchaKip enjoys channels with relations to Ubuntu much better than others because the people are nice. [23:35] KombuchaKip: it would indeed be nice to hook into the Nautilus rename command, but unfortunately the Nautilus extensions don provide that sort of hook yet [23:36] jelmer: I guess there's always inotify… ;) [23:37] spiv: heh [23:37] jelmer: I figured that was probably why it wasn't done, because it seemed intuitively straight forward since other commands are hooked. So I suppose we have to wait until Nautilus provides an API for that. Maybe you could add a comment for that at the end of this here? https://bugs.launchpad.net/bzr-gtk/+bug/797438 [23:37] Ubuntu bug 797438 in Bazaar GTK+ Frontends "nautilus-bzr plugin does not hook rename" [Medium,Triaged] [23:37] hey spiv [23:37] jelmer: I'd do it myself, but I have no experience with the API. [23:37] i did some lp fixes after gdd, don't know if you saw [23:37] 'less awful search results' basically [23:37] KombuchaKip: thereÅ's already a comment from me there to that extend :) [23:38] jelmer: Ah, I understand it now. I just saw that. [23:38] poolie: I did! [23:38] jelmer: I wonder if upstream knows. [23:38] poolie: (if you're talking about the meta description stuff) [23:39] yeah, schema.org tags, that kind of thing [23:44] poolie: How do you plan to patch it if apparently nautilus-python doesn't have a way to track file renames?