[00:02] KhaZ: should be fine, if you're talking about a place to push and pull from [00:04] Aye. Alright, I'll give that a shot. Thanks. [00:05] best performance would be to use bzr+ssh [00:05] if you only have LAN's though file:// is fine [00:17] Yeah. We'll see; maybe I'll switch over to installing SSH on Windows sometime; but I'll try out samba for now. [00:21] the bzr install comes with bzr+ssh support, you just need pageant [00:21] (I think) [00:28] Right; but it doesn't set up a server, does it? [00:28] i.e., if I have two fresh windows boxes, and I use the bzr installer, it won't set up an svn server such that I acn push/pull between the two, I think? [00:29] s/svn/bzr/ ? :P [00:29] no it won't, but samba implies that you have a unix machine there anyway ;) [00:29] unless you meant SMB rather than Samba. [00:29] Sorry, I should mention, "Windows Sharing", not samba. [00:29] Yeah; SMB, rather. I just always spell out samba. [00:29] My bad. [00:29] np [00:29] *spell it out as Samba, implying they're the same when they're not. [02:14] igc: so, CHKInventorypath2id is doing much more work than needed [02:14] igc: I'm fixing it. [02:14] lifeless: thank-you, thank-you, thank-you [02:15] lifeless: lots of performance hot spots will be improved by doing that I suspect [02:16] igc: its this todo: [02:16] # XXX: Todo - use proxy objects for the children rather than loading [02:16] # all when the attribute is referenced. [02:16] however, I'm not going to do the todo, for now I'm just going to fix path2id directly [02:19] igc: for reference, just double clicking on the path2id node in the profile shows it doing 300 IO's [02:19] igc: which gives a hint that its doing more work than needed :P [02:21] lifeless: in kcachegrind? That doesn't work for me? What panel? [02:28] igc: callee map [02:29] igc: the tree - double click on the path2id bzrlib.inventory:1993 line [02:30] s/line/node [02:30] actually, its not the callee map, the UI confused me [02:30] its the 'call graph' [02:31] I'm just testing the fix now [03:19] igc: rev 4544 should be good [03:19] igc: all review feedback applied; bugs fixed, path2id hopefully faster [03:21] lifeless: cool. I'm wrapping up a patch and have a medical appointment after lunch. I'll take a look after that [03:22] igc: cool. it would be great if we can get this assessed before ~3, when my day ends. [03:22] (I'm still being woken up silly-early by neighbours :( ) [03:24] lifeless: my medical appt is 2.40 so I suspect before 3 is unlikely sorry [03:24] ah well [03:24] monday then [03:33] fullermd: ping :) where should I file a bug about lt2.x ? [03:44] according to this tutorial http://doc.bazaar-vcs.org/latest/en/tutorials/using_bazaar_with_launchpad.html#publishing-your-changes [03:44] i can push my changes to launchpad via: "bzr push lp:~team-name/project-name/branch-name" [03:44] but what if my project is not part of any team? [03:45] replace team-name with your name [03:45] (lp usercode actually) [03:45] lunching [03:47] sorry, i don't understand the last two lines of what you said [03:47] lp usercode? lunching? [03:47] was that meant in answer to my question? [03:47] or were you anouncing you were going out to lunch? :) [03:47] pattern: I think so [03:48] pattern: he's going out to lunch :) [03:48] ok... just checking :) [03:48] so I would push to lp:~garyvdm/qbzr/./// [03:48] pattern: but his 2nd last line was to you :) [03:48] so I would push to lp:~garyvdm/qbzr/.... [03:48] pattern: i.e. I generally push to lp:~spiv/bzr/branch-name [03:49] got it [03:49] thank you everyone [03:50] The ~name at the start controls who Launchpad grants write access to. If it's a team, the whole team can update the branch. If it's just a person, then only that person. [03:51] (for now) [03:52] so how can people submit patches to projects that aren't part of a team? [03:52] (assuming they don't control the branch) [03:52] pattern: They can request a merge [03:52] by the person that does [03:53] They do that by pushing to a branch in their own name [03:53] If you do have more than one person that needs to push to the mainline of a project, it's easy to get a team created. [03:54] my project is very small right now [03:54] i don't think i need a team [03:55] jml: You got me to bite - tell me more... [03:55] just trying to understand how i can get contributions from others [03:55] garyvdm: i think jml is talking about package branches [03:55] where it will become a bit more subtle for the official branches [03:57] what mwhudson said. [03:57] pattern: they can email you with the contribution. There's a "bzr send" command that will even make a mail with the new bzr revisions as an attachment for you. [03:58] pattern: or you could "bzr merge" (or "bzr pull") from their branch if they've published it somewhere (e.g. on Launchpad).o [03:58] pattern: Launchpad has a merge proposal tracker as well, so you could ask contributors to use that to submit patches. [04:00] pattern: basically, the two questions are "how do you want contributors to contact you about their contributions?" (e.g. email to a particular address) and "how do you want to receive the contributions?" (e.g. a bundle in an email from bzr send, or by using a branch they've published somewhere, etc.) [04:05] that merge proposal tracker sounds interesting [04:07] where can i find out more about it? [04:12] pattern: https://help.launchpad.net/Code/Review is probably a good starting point. [04:12] pattern: https://help.launchpad.net/, https://help.launchpad.net/Code/Review [04:12] * spiv also goes to lunch :) [04:13] thank you very much [04:14] jml: I think its really really clear as it is today [04:15] jml: When you look at DACL in windows and posix they are generally a lot more complex and harder to understand [04:15] lifeless: initial benchmarking looks much better thanks [04:15] lifeless, [04:16] igc: cool [04:16] lifeless: I'll run the full benchmark over lunch (it takes 90 minutes) [04:16] jml: (just saying) [04:16] thanks to the 'speed' of branching outside a repo [04:17] lifeless, yeah. it _is_ less clear to have people with implicit write permissions, but it's something the distro community really really want, and we think the loss of clarity is worth it. [04:17] jml: for the distro official branches I totally agree [04:18] well [04:18] I totally get the need, and that we need to support that [04:18] I'm not 100% on board the particular way to address that [04:19] that said, one way to phrase it would be that distro official branches have an implicit per-branch team that we just elide from the namespace [04:19] which is...roundabout [04:20] yes. [04:20] oh, ETA? [04:21] difficult to say. [04:21] it's a one or two day patch, and my highest priority task. [04:21] so probably in two or three weeks time. [04:22] Actually, I meant you, at my place. [04:23] I'm assuming that the changes for the distro are either not applicable, or at least optional, for other branches ;) [04:23] ~6:30pm? [04:23] sure, earlier is fine if you like; later and I'll be climbing up the walls in search of food [04:25] I'll try to get there earlier, but I doubt it'll happen -- there's a lot to do before then. [04:32] lifeless, bug 397705 is targeted to 1.17. Is there still more stuff to land for 1.17? [04:32] Launchpad bug 397705 in bzr "inventory delta consistency checks are missing known problems and not universally applied." [Critical,Fix committed] https://launchpad.net/bugs/397705 [04:35] * RenatoSilva is thinking about suggesting his boss to leave CVS and start using Bazaar... [04:37] jml: there is more on that branch [04:37] jml: it doesn't have to go in 1.17; its waiting review at the moment. [04:38] jml: It probably speeds up a bunch of things (while slowing down others just a data, and becoming faster at the same time) [04:38] lifeless, I don't grok the bit in parens. [04:41] spm, how does bazaar's pqm spell the name of the 1.17 branch? [04:42] jml: do you mean: https://code.edge.launchpad.net/~bzr-pqm ? [04:42] not quite [04:42] merge http://bazaar.launchpad.net/~jml/bzr/1.17-build-updates lp:bzr/1.17 [04:42] Command failed! [04:42] All lines of log output:Sender not authorised to commit to branch lp:bzr/1.17 [04:42] that's what I get when I try to land a branch to 'lp:bzr/1.17' [04:42] ah right, one sec [04:43] lifeless: starting the full benchmark now. Still issues with the test suite btw - see the associated bug for details. [04:43] * igc lunch & medical stuff - bbl [04:43] jml: bzr+ssh://bazaar.launchpad.net/~bzr-pqm/bzr/1.17 [04:44] spm, thanks. [04:52] igc: thanks [04:53] jml: or http:// [04:56] igc: thanks [04:56] igc: I've switched context to another bug but I'll fix those up later [04:56] igc: I'm sure any other changes won't have performance implications [05:23] jam: abentley: could you please move the conversation out of bug 390563? That bug is for a totally different cause, and *still* hasn't had all its code changes landed. [05:23] Launchpad bug 390563 in bzr "absent factory exception from smart server when streaming 2a stacked branches" [Critical,In progress] https://launchpad.net/bugs/390563 [05:24] lifeless: I don't want to disrupt the conversation. [05:25] abentley: I'd be happy to copy over the last few posts to the correct bug if that would help [05:26] its a shame malone doesn't provide facilities to remix bug conversations [05:26] lifeless: Nevermind, I can drag it to the list if I have more questions for jam. [05:28] abentley: I don't want to prevent important discussions; I'm just worried that its making it harder to have the discussion about delta generation thats also on the same bug :( [05:28] lifeless: I understand. [05:31] committing a tarball of bzr to a bzr branch [05:31] no irony at all [05:32] jml: buildout? [05:32] yeah. [05:34] I'm very unconvinced [05:36] Can any one word this more elegantly: [05:36] "You are reverting all changed files. Are you sure that you don't also want to revert pending merges?" [05:37] Warning for when a user tries to revert all file, and not pending merges from qrevert. [05:37] perhaps [05:37] "You are reverting all changed files but leaving some pending merges. Do you want to [also revert the merges] [revert the changed files] [cancel the revert]" [05:40] tweaking [05:41] "You are reverting all changed paths without also reverting pending merges. Do you want to [also revert pending merges] [revert all changed paths without merges] [cancel the revert]" [05:42] "You are reverting all changed paths without also reverting pending merges. Continue?" [Yes] [No] [05:42] The ability to chose the other option is provided by the dialog: http://imagebin.ca/view/nys-K4s.html [05:43] Thanks lifeless. - That sound much better. [05:45] my pleasure [05:53] Morning amanica. Up early! [05:53] garyvdm: morning, couldn't sleep. [05:53] you too btw. wasn't gonna say something [05:54] I'm working night shift - then I hack after work and sleep during the day. [05:54] I blame lifeless, for complaining about stuff I did :) [05:54] ah ok [05:55] once your mind keeps on drafting that mail, its over === abentley1 is now known as abentley [06:07] All the Bazaar-nics should appreciate this: http://www.chapterthree.com/blog/josh_koenig/project_mercury_preconfigured_drupalvarnish_ec2_ami [06:07] It's a Drupal-based AMI that auto-updates using Bazaar [06:08] sweet [06:26] oh wow [06:26] don't you love OMG bugs [06:40] mm, sometimes [06:40] lifeless: did you want to talk about anything in particular wrt hpssvfs? [06:41] no, I was making a meta comment; that the patch is trivial enough to just land, but if discussion is sought perhaps the list works better. I don't know - still tossing ideas around [06:41] I have a rathole bug at the moment [06:41] revert -> tree transform -> iter_changes -> find_ids_across_trees [06:42] ah ok [06:42] will let you back to it [06:42] going into code myself if i'm lucky... [06:45] abentley: are you around still? Do you have an opinion on tree.has_id for missing paths? [07:08] \o/ got that bug fixed [07:08] amanica: you may be interested in https://bugs.edge.launchpad.net/bzr/+bug/129880 [07:08] Ubuntu bug 129880 in bzr "Can't bzr rm a directory that contained a modified+emigrated file" [Medium,Confirmed] [07:29] hi all [07:32] lifeless: thanks, sure, I'll append it to my evergrowing todo list [07:32] good morning vila [07:32] amanica: Hey ! [07:41] back [07:41] hi vila! [07:42] igc: hi [08:27] * spiv -> gone [08:52] amanica: ping [08:59] vila: pong [09:00] I see you wanted to work on bzr mv again, so I wanted to talk a bit about some problems I encoutered on OSX and that seem to be related [09:01] amanica: roughly, If unicode paths are used and BzrMoveFailedError and derived classes are used, things go wrong [09:02] amanica: So if you touch that code again, make sure to add tests with unicode paths, and if you can test that on OSX, that would be even better, but I don't know if you have that available :) [09:03] lifeless, leaving now. [09:03] vila: I'll make a note, thanks. I only have ubuntu [09:04] amanica: ok, thanks :) [09:04] jml: cusoon [09:25] * igc dinner [09:58] mwhudson: hi [09:58] mwhudson: do you have any remaining failing imports from git? [09:59] jelmer: i had one that failed because of submodules [10:01] jelmer: i guess i should be systematic about it at some point [10:01] not at 2100 on a friday tho [10:04] :-) [10:22] jelmer: I have a failing bzr branch with for a git repo [10:23] asabil: what git repo? [10:23] asabil: and what's the error? [10:23] git://github.com/280north/cappuccino.git [10:23] invalid revision None [10:23] couldn't get a complete traceback [10:25] bzr: ERROR: The branch git://github.com/280north/cappuccino has no revision None. [10:25] that's the exact error [10:34] asabil: I can reproduce it, please file a bug [10:38] jelmer: ok will do thanks [10:48] jelmer: any idea about why bzr-rebase would fail with this: [10:49] NoSuchRevision: KnitPackRepository('file:///home/asabil/main-1.14/.bzr/repository/') has no revision ('stian.selnes@tandberg.com-20090610132048-1mk36vuwwxv7zz0c',) [10:50] asabil: not without context [10:51] jelmer: just rebasing a branch on top of another one [10:51] nothing really specific [10:51] bzr check and bzr reconcile are running fine [11:00] jelmer: here is the traceback: http://bzr.pastebin.com/d5186bfe2 [13:16] jelmer: I notice you released bzr-svn 0.6.3 but I didn't see a release mail, want me to write and send it to bazaar-announce? [13:18] LarstiQ: please do [13:18] LarstiQ: and thanks for doing so [13:18] jelmer: ETA tonight or tomorrow noon/afternoon [13:22] LarstiQ: np, thanks again [13:22] * jelmer waves from DebCamp [13:23] jelmer: ooh, say hi to Dato [13:23] and maybe more when I figure out they're there ;) [14:16] so, bzr can't handle non-ascii content, or what? [14:16] https://bugs.launchpad.net/ubuntu/+source/bzr/+bug/400670 [14:16] Ubuntu bug 400670 in bzr "bzr log -p UnicodeDecodeError" [Undecided,New] [14:17] LarstiQ: will do, if I can figure out who dato is :-) [14:18] mgedmin: Works fine for me. [14:18] newer bzr? [14:18] lifeless, jml: hooray for testresources! [14:18] fullermd: did you try with lp:zodbbrowser? [14:18] * spiv -> zzz [14:18] Exactly the branch and rev in the bug. [14:19] which bzr version? [14:19] fullermd: so, up for some yak shaving? [14:19] spiv: thanks [14:19] * mgedmin checks if he still has a copy of bzr trunk [14:19] spiv, thanks :) [14:19] The error message says it's trying to convert it to ascii, which suggests bzr isn't liking your locale. [14:19] jelmer: Am I right thinking that bzr-git doesn't support submodules and won't support them until Nested Trees works? [14:19] abentley: yep [14:20] (though you are a few versions back, and there may have been a bug fixed... none comes to mind) [14:20] lifeless: I left my clippers in my other pants. [14:21] Too much work and errands today to have time for much... [14:22] uh oh the progress bar hints that I'll wait a long time to get an up-to-date bzr.dev ... [14:22] unless it's not a progress bar but a simple spinner that looks like it's a progress bar stuck at <5% [14:22] It'd certainly be quicker to grab a tarball of something (the 1.17rc maybe). [14:23] fullermd: (where should I record 'FreeBSD doesn't have autoconf released in the last year-or-two') [14:25] lifeless: The PR form is at . [14:26] fullermd: will it be better if you file it? [14:26] Probably not. ade@ is the main person who handles autotools-related stuff, but I haven't seem him pipe up in a few months. [14:27] http://lists.freebsd.org/pipermail/freebsd-ports/2009-May/054515.html [14:29] oh wow... "Thank you for the problem report. You should receive confirmation of your report by electronic mail within a day." === JamalFanaian is now known as JamalFanaian|afk [14:35] fullermd: ok, http://www.freebsd.org/cgi/query-pr.cgi?pr=136867 [14:35] fullermd: if you have anything to add that would be great [14:40] nice, launchpad somehow managed to create the bug twice, with different numbers, and they're already marked as duplicates [14:41] It likes some bugs so much, it makes a backup. [14:42] lol [14:48] and yes, the bug is gone in bzr.dev === JamalFanaian|afk is now known as JamalFanaian === nevans1 is now known as nevans === JamalFanaian is now known as JamalFanaian|afk === mario__ is now known as pygi === pgega is now known as CTO === JamalFanaian|afk is now known as JamalFanaian [20:16] yay [20:16] roundup now kinda support bzr [20:16] :D [20:16] pygi: \o/ [20:16] not tested, and probably doesn't work with my buggy code, but oh well :P [20:18] * hey bzr or cvs, that's the question [20:19] lol [20:19] jelmer, when will we complete what we started at UDS? [20:19] pygi: at some point :-P [20:19] aha, so we'll do it during next sprint? xD [20:19] * jelmer is at DebCamp at the moment, trying to get some Samba packages into the right shape [20:21] sup [20:22] hey ronny [20:22] jelmer: europython was a blast, but the 2 weeks since it have been very unproductive for me [21:14] ie: what i really want to know... say i have a million rows in a table with a view on it, with no WHERE conditions.... i select from the view, with a WHERE.... will it take forever, and do a select all, and then apply the where later... or does it do it all at once, and make use of the indexes, and all that [21:14] er [21:14] wrong window [22:06] i've been reading about how to use bazaar with launchpad and i have some questions.. [22:06] let's say i just created my own project on launchpad, then did "bzr init" on my own local machine, and added and committed some source to it... [22:07] then i "bzr launchpad-login me" and "bzr push lp:~me/myproject/mybranch" [22:07] then, when i make some changes on my local copy of the code and type "bzr commit", will those changes automatically be uploaded to my branch on launchpad? or will i have to do another push to get them there? [22:09] pattern: You always have to push. [22:10] i see [22:11] This way, you can commit multiple changes to a branch (for a completely new feature for example) and then push it all to a new branch for merging. [22:12] that makes sense === JamalFanaian is now known as JamalFanaian|afk [22:13] now, on another subject.. let's say someone else has copied my branch from "lp:~me/myproject/mybranch" and made some changes, then pushed it to "lp:~him/hisproject/hisbranch", and notified me that he'd like to merge the changes in my project [22:14] how would i actually go about doing that? [22:14] bzr merge lp:~him/hisproject/hisbranch [22:14] that would merge with the code on my machine? [22:15] Yep. [22:15] i see [22:15] cool [22:15] And if there are any conflicts, bzr will tell you. [22:15] a nice and elegant way to do it [22:17] thanks for your help, jpds [22:18] pattern: No problem. [22:39] pattern: you need to do a commit after the merge to record it [22:40] jpds: I find the above is worth mentioning when people first encounter merge :) [22:40] lifeless: Ah, yes. :) [22:44] thank you [22:44] that's good to know [22:57] lifeless: Funny... [22:57] lifeless: http://lists.freebsd.org/pipermail/freebsd-ports/2009-July/055841.html [22:57] So I guess lt2 is already in test. [23:02] fullermd: oh cool [23:02] fullermd: (care to test ? :P) [23:03] Hey, just 'cuz I run the development head of my OS and window manager and version control system, doesn't mean I'll try ANY old bleeding edge :p [23:04] :P [23:04] I just want to know if libcpuinfo builds [23:04] I could generate a make dist tarball for you [23:04] Well, with those hacks to get along with lt 1.5, it does. [23:04] ok cool [23:04] that should be totally innocuous then [23:05] Yah. I wouldn't be blown away by shock if the lt switch caused breakage, but I would be a little surprised. [23:07] jam: if you're still here; .. a quick eyeball on https://code.edge.launchpad.net/~lifeless/bzr/bug-367632/+merge/8928 would be great [23:12] The 1.17 branch's version is set to 1.18dev :/ [23:14] when i do a push from a branch on my machine to launchpad, will a record of every revision i've committed (and the associated log messages) go with it? [23:15] or does launchpad just get the delta between the last pushed revision and the newly pushed revision? [23:17] pattern: Those revisions are what push pushes. [23:17] pattern: bzr only sends new data [23:18] pattern: but you end up with all the revisions on launchpad [23:19] i see [23:19] thanks again === Pilky_ is now known as Pilky