[00:00] 2.4.2 [00:01] morning === ig1 is now known as igc [00:09] lifeless: bingo ! I can reproduce it with python-2.4.4 [00:09] cool [00:09] not accpepting interrupts though :) [00:47] spiv: ping === i386_ is now known as i386 [00:55] lifeless: hi, just about to dive into your review [00:56] spiv: thanks [01:05] I have an updated deprecations patch [01:06] and I missed a couple of trivial uses of vf.join() [01:12] http://blog.red-bean.com/sussman/?p=90 [01:12] spiv: if you want to talk about the branch I'm happy to do a clal [01:21] lifeless: thanks for the detailed response to my filtering brain dump [01:21] igc: eh, that was just a off the cuff :) [01:22] np - I'm reviewing jam's patch right now and I'll digest your wisdom after that :-) [01:22] igc: I haven't had time to give a proper analysis, but its a complex problem to do really well which is why I was so keen for us to layer it with the minimum [01:24] lifeless: Agreed. I'll move the conversation to the mailing list so others can chime in as well. [01:48] igc, i think it may be ok to not worry right now about how to reapply all your filters to the tree [01:48] perhaps it does need to be fixed before we really announce it [01:49] poolie: ok. I'll send an email anyway to give people time to think about it how it ought to work [01:49] mm [01:49] i should probably not think about it today, have too much more urgent stuff [01:50] poolie: np - I can't implement a solution till next Tues at the earliest anyway [02:03] poolie: igc: I think we need *a* method to reset a tree to the current filtered contents [02:04] if 'bzr revert' did nothing, it would be most frustrating for people :) [02:14] i agree [02:15] lifeless: but things like update that normally do a 3way merge will be more difficult - they don't have the output from the old filter to merge against afaics [02:15] lifeless: re your patch [02:16] removing deprecated code [02:16] is there a current equivalent to MultipleRevisionSources? [02:16] like if we wanted to fix the problem of missing doing a fetch, is that easy with the current apis? [02:17] poolie: 'update' should be fine as long as you reset the tree to match the filters before you update [02:18] poolie: because it will always be merging the canonical forms [02:18] poolie: MultipleRevisionSources is already unused [02:18] and deprecated [02:19] graph.ParentsProvider has cleaner functionality [02:24] lifeless, poolie: email sent to the list now on this topic [02:25] I need to focus on reviewing jam's patch right now so I'll digest everyone's input next week [02:32] lifeless: yes i saw it was deprecated [02:38] spiv: updated patch sent; trivial changes [02:38] lifeless: thanks [02:40] spiv: how is it looking ? [02:40] Pretty good. [02:41] No fundamental "wtf" issues :) [02:41] So far, anyway ;) === poolie changed the topic of #bzr to: Bazaar version control system | === poolie changed the topic of #bzr to: Bazaar version control system | http://bazaar-vcs.org/ | bzr-1.4 out, May 2 [02:44] poolie: New bazaar version now? [02:44] :D [02:52] poolie: hooray! [02:52] poolie: I'm about to head off to your place... should we meet for lunch? [02:54] spiv, sure, call me when you get to the station [02:54] hersonls: yes [02:55] spiv, i may not be able to work on hpss right away but come on anyhow [02:55] poolie: Good :D [02:56] poolie: ok [02:57] spiv: review is done ? [02:59] lifeless: not yet, will continue on the train [02:59] ok [03:26] lifeless: spent any time in Dunedin? [03:27] abentley: born and bred [03:28] I'll be there on Sunday. Anything I should see and/or do? [03:28] the city gardens are beautiful [03:28] they have some of the rare and endangered bird species there [03:28] mmm winter, won'y suggest sandfly bay [03:29] abentley: when i was there i went out on the peninsula south of town [03:29] the albatross colony might be open though, thats well worth a visit; and if you like beaches head out to the spit [03:29] Sandfly bay? That sounds kinda ominous. [03:29] they have penguin and albatross nests [03:29] and giant sheep [03:29] poolie: thats NE of town :) [03:29] maybe it moved :P [03:30] Cool. Thanks for the suggestions. [03:30] there are a number of nice lookouts, the one back of the quarry has a particularly good view of the city [03:30] heading up mt cargill can be nice too [03:31] its a uni town, I'm totally positive you'll be able to find an open mic as wlel [03:31] Hehe. [03:31] really, I think its something like 10% population is floating students [03:32] the whole place ebbs and flows to the university holiday schedule [03:32] it's like the last week of term i think [03:32] (at least it is at massey) [03:34] Hmm. Yeah, I guess the only real university town I've been to is New Haven, Connecticut. Toronto and Montreal have universities, but they're also big cities. [03:34] oh,if the mei wah fish and chip shop is open, or the golden sun beside it, try some. One if unforgettably terrible, and one unforgettably good (or was 12 years ago). [03:35] I won't tell you which is which :) [03:36] poolie: congrats on the 1.4ness [03:37] thanks, sorry for the delay [03:38] now for pqm toil [03:38] fortunately it is getting more automated each time - i just sent a patch deleting more steps from teh RM guide [03:38] lifeless: I'm a bit surprised common_ancestor is deprecated. I believe I'm still using it in graph-ancestry. [03:39] abentley: deprecated in 1.4 [03:40] Ah, I'm actually using node_distances and select_farthest. [03:41] :) [03:45] abentley: have a good trip btw [03:45] poolie: Thanks. [04:29] abentley: if by any chance you're still here: what was the case with lp push or pull that was slow for you? [04:30] poolie: bzr branch [04:30] poolie: over hpss [04:31] poolie: I think his mail didn't discriminate enough between using lp, and being lp :) [04:31] spiv and I are aware of this and have both done some analysis [04:31] into an empty repo? [04:31] poolie: no repo [04:31] poolie: call ? [04:31] ok [04:38] poolie: It was branch [04:39] poolie: I'm in the middle of repeating it for you. [04:39] But the whole process takes > 1 hour [04:55] * igc lunch and pick up kids -bbl [05:02] lifeless: I'm a bit lost on how to change ReconcilePacker to update sha1s. [05:10] abentley: indeed [05:10] abentley: it needs to override the revision bulk copy and have it occur after the inventory copy; during the inventory copy you'll need to determine what the correct sha1s are [05:11] a subclass for the case where this is needed may be appropriate [05:12] In order to update the sha1s, I need to know the correct serializer, so I need the repository. Which seems like a layering violation. [05:13] well the repository chooses the reconciler etc [05:14] so it can pass in the serializer (though isn't the sha1 on disk always the correct one now ?) [05:14] oh! it needs the revision serialiser [05:14] I definitely think a subclass is appropriate here [05:27] Though perhaps if I need the repo anyway, I should be using repo methods. [05:44] lifeless: Is there any easy way of reading/writing fulltexts from a PackCollection? [05:44] yes [05:44] one sec [05:45] you instantiate a KnitVersionedFile on the pack and a GraphIndex (for revisions|signatures|inventories|knits) [05:46] I can do that on a single pack, not the PackCollection? [05:46] to read the old revisions I would do use the existing repositories iter_revisions or whatever [05:47] you can, but there is no guarantee that you can read a given uncompressed text from a single pack [05:47] basis texts may well be in different packs [05:47] For reading revisions I would use the repositories methods [05:48] for writing them serialising and adding is probably cleanest (see the output_knit variable around line 1062) [05:48] Not important for this case, since revisions are not (supposed to be) delta'd. [05:48] the existing reconcile code does 'repo.weave_store.get_weave(file_id, transaction) to read texts [05:48] that is by far the simplest way [05:49] Okay, thanks. [05:57] lifeless: review heading your way [06:01] thanks [06:01] Hmm, or at least it will be once I sort out my mail delivery... [06:07] :P [06:14] morning you hackers. [06:14] where, where? [06:14] there, there. [06:14] sup robert? [06:14] spiv: short story, is it tweak? bounce? approve? [06:15] lifeless: tweak [06:15] Morning? It's 2300 here! :) [06:15] jrydberg: VersionedFiles is taking shape [06:15] rockstar_: wait 46 minutes [06:15] starting ppa rebuild of 1.4 [06:15] lifeless: but if you want to bounce up and down, I won't stop you ;) [06:15] lifeless: Finally. [06:15] jrydberg: well, there was a bunch of prep work to do :) [06:15] jrydberg: how are you, haven't chatted in -ages- [06:16] lifeless: hehe. [06:16] lifeless: i'm fine, i'm fine. just trying to use computers are little as possible in my spare time. noticed that it helps my motivation at work. [06:19] jrydberg: heh, me too :) [06:19] or at least i am trying to be careful of that effect [06:19] i've turned to photography instead. [06:19] equally geeky, less computers. [06:19] well thats 8 hours poolie [06:19] poolie: so I'm calling it till Monday [06:20] jrydberg: this is my solution, to be self strict about work | hobby time [06:20] lifeless: you should have that mail now [06:20] abentley: thanks for that log [06:20] spiv: No problem. [06:21] well, I'll reply first :) [06:21] then call the day [06:23] lifeless: can you read this one mail i'm going to send first? :) [06:23] about PPAs [06:43] I think beta is overkill [06:43] just use snapshot for betas [06:43] IMO [06:44] * lifeless waves === quicksil1er is now known as quicksilver [08:53] night all - have a good weekend [08:53] back Tuesday for me === c1|freaky_ is now known as c1|freaky === quicksil1er is now known as quicksilve === quicksilve is now known as quicksilver [11:14] ok, I narrowed down my pqm hanging scenario, reproduced on gutsy with: [11:14] python2.4 bzr selftest bzrlib.tests.test_strace.TestStrace.test_strace_callable_is_called bzrlib.tests.test_strace.TestStrace.test_strace_callable_result bzrlib.tests.blackbox.test_serve.TestBzrServe.test_bzr_connect_to_bzr_ssh [11:15] any advice greatly appreciated while the diagnosis continues ;-) [11:16] * fullermd has none whatosoever. [11:18] fullermd: thanks for the warm moral support :-P [11:19] * fullermd is helpful like that :] [11:20] strace does sound like a good candidate for twisting up a machine, though. [11:24] the puzzling bit is that I submitted a patch that first glance has strictly nothing to do with strace :) Neither at second glance FWIW [11:25] and to make it funnier I can reproduce it on a single machine only while, still at first glance, the two are gutsy installs :) [11:25] With strace installed? [11:26] I know the tests skip without it (I don't have it installed) [11:26] fullermd: excellent idea ! But both have the same version installed [11:27] both says 'tests passed' (i.e. really late, nearly before exiting) but one hang after that [11:28] Well, maybe strace isn't standard-installed, but one has it chosen as additional. [11:29] I could believe that strace enables something that runs at cleanup time, so that could cause a late hang. [11:29] (of course, I'm in pure speculation mode here, so I could well be making it up as I go along too :) [11:34] I don't remember if I installed them myself, but that was through synaptic anyway, so both installs should be correct, but looking at the code I suspect something related to 'import subprocess' since these tests use that, there may be some weird interaction [11:35] I think my patch just revealed a latent bug [11:39] test_bzr_connect_to_bzr_ssh has several 'XXX' anyway [11:39] spiv: are you still around by chance ? [11:56] Great, you broke him. [11:59] fullermd: damn :) [12:02] That's talent. You exposed a latent bug in bzr _and_ spiv at the same time ;) [12:33] fullermd: lol, I will meditate about that during lunch :) [15:07] I've a shared repo, with a branch 'trunk'. I've done `bzr branch trunk local`. Now, I'd like to rename 'local' to something else. Is it just "mv local something-else", or does bzr need to get involved, and if so, by what command? [15:09] jsled: "mv local something-else" is fine [15:10] jsled: bzr would need to get involved for something like "mv local /somewhere/else/entirely" [15:10] as long as you stay within the shared repo you should be ok. [15:10] thanks. :) [15:10] (assuming you don't move something inside another branch, or a repo in repo or similar) [15:42] hi, is there someone using the bzr-hookless-email? for some reason it seems to hang and leak memory like crazy. [15:42] File "/usr/lib/python2.5/site-packages/bzrlib/repository.py", line 1520, in deserialise_inventory [15:43] probably that's the hanging function [16:38] dato: ping [16:38] dato: your upload of bzr 1.3.1-1 doesn't appear to be in the bzr branch on alioth [16:53] Hi all. Has anyone tried reviewboard with Bazaar? It seems to have gotten some of Bazaar support lately. [16:53] * chadmiller doesn't know how deep it goes. [16:53] chadmiller: I think statik was the person who did most of the work on that [16:53] statik, ^ [16:56] From the log, "Patch from Henrik Hedberg to add basic support for the Bazaar SCM." [16:56] That ain't statik. [16:57] hmm, must be another branch than https://code.launchpad.net/~statik/reviewboard/bazaarsupport then that got merged.. [17:01] A related question: That patch has some questionable choices in it. Instead of importing bzrlib, it runs "bzr" with arguments and catches stdout. Is that as silly as I think it is? Do you consider the internal bzr APIs stable? [17:01] Or, rather, /is there/ a public API for bzr? [17:02] ah, ouch [17:02] there is a public API for bzr in Python (in which reviewboard is also written) [17:03] that does indeed seem like a very questionable choice [17:05] * lamont struggles to remember how to import a cvs tree into bzr [17:09] and finds it [17:20] I've got a local branch that is a checkout of a branch hosted in Launchpad... I've just done a bzr upgrade to upgrade the format. What's the proper way to get my remote master branch upgraded as well? It still appears to be of format BzrBranch5 [17:21] surfous, bzr upgrade sftp://user@bazaar.lauchpad... [17:21] upgrading it through sftp on LP [17:21] but, be warned, it takes a while [17:25] beuno: Thanks for the info - I fear it's sending me on a tail-chase, though... [17:25] What prompted me to do this was trying to tag my branch. I got the following error: [17:25] bzr: ERROR: Tags not supported by BzrBranch5('bzr+ssh://surf@bazaar.launchpad.net/%7Epyax/pyax/release/'); you may be able to use bzr upgrade --dirstate-tags. [17:26] I did a bzr upgrade which brought my format up to pack-0.92 [17:26] local format, that is [17:27] trying to upgrade the remote branch tells me: bzr: ERROR: The branch format Bazaar-NG meta directory, format 1 is already at the most recent format. [17:28] So one operation tells me I'm out of date, and the other tells me I'm up to date, when ultimately, I'm just confused. [17:29] surfous, try upgrading over sftp rather than bzr+ssh [17:29] surfous, what does "bzr info -v" on the remote branch tell you? [17:30] jelmer: Cool, that seems to be doing more than what I had before... I'll see where that lands me. [17:32] beuno: surf@fischer:~/src/pyax/release$ bzr info -v bzr+ssh://surf@bazaar.launchpad.net/%7Epyax/pyax/trunk [17:32] Standalone branch (format: unnamed) [17:32] Location: [17:32] branch root: bzr+ssh://surf@bazaar.launchpad.net/%7Epyax/pyax/trunk/ [17:32] Format: [17:32] control: bzr remote bzrdir [17:32] branch: Remote BZR Branch [17:32] repository: bzr remote repository [17:32] Branch history: [17:32] 51 revisions [17:32] 4 committers [17:32] 371 days old [17:32] first revision: Thu 2007-04-26 10:50:37 -0700 [17:32] latest revision: Thu 2008-05-01 14:17:23 -0700 [17:32] Repository: [17:32] 63 revisions [17:32] 324 KiB [17:33] beuno: I know I referenced my release branch before, any this is on my trunk, but the problem was the same on each [17:36] surfous, right, for some reason, that's very uninformative, but upgrading through sftp is the way to go [17:37] beuno: And it does indeed take a while as you'd warned, but I've no problem waiting. [17:37] Thanks a lot for your help, beuno and jelmer! [17:37] surfous, :) === `6og is now known as Kamping_Kaiser [19:11] abentley, hi [19:38] given a bzr 1.3.1 tree, bzr log says that the current revision is '47'. is there any trivial way to get that out of bzr short of resorting to perl/awk/sed/grep foolery? [19:39] lamont, bzr revno? [19:39] I knew there had to be something. thanks [19:39] :) [19:40] "version" and "revision" didn't give me what I wanted, so I decided that asking would be quicker. thanks again [19:41] nothing better then taking advantage of random knowledge on IRC! === emgent_ is now known as emgent [20:15] *ignore me* #214825 === emgent_ is now known as emgent [23:26] Hi !. Is is possible to ignore all ".svn" dirs regardless how deep they are ? [23:28] does "bzr ignore .svn" not work? [23:31] let me see ... [23:32] james_w: Works, thank you very much [23:32] no problem [23:33] i have two branches ... trunk and branch1, people commit to trunk and also tag commits [23:33] how could i diff these two and only limit on tagged commits [23:34] I'm not sure that's possible [23:34] ok, thanks :) [23:34] you want to diff two branches but only show the changes that were introduced in one by certain commits? [23:35] trying to implement a code review [23:35] when people close a bug, they commit with a tag containing the bug id [23:36] laterz [23:36] night