[00:57] Good morning. [00:59] allo allo === Adys_ is now known as Adys === poolie changed the topic of #bzr to: Bazaar version control | try https://answers.launchpad.net/bzr for more help | http://irclogs.ubuntu.com/ | Patch pilot: jam | bzr 2.1.1 is out | bzr 2.1.2 is having binaries built for it [02:33] hi spiv [02:34] hi poolie [03:24] anyone know, can you do keyword lambdas in python 2.4? [03:26] by which I mean [03:26] lambda foo=bar:quux [03:28] lifeless: I've never seen that before [03:28] thumper: :) [03:28] thumper: I'd never had cause to do it before today. [03:28] lifeless: what are you trying to do? [03:28] it Just Worked in python 2.6 [03:29] thumper: theres an interface which is (now) called with keyword parameters, that a test stubbed out with a lambda. [03:30] ah [03:31] lifeless: seems to work on 2.4 http://pastebin.com/k4X1x8pE [03:31] parthm: thanks! [03:32] You've been able to do that with lambda since approximately forever. [03:32] It was particularly useful back when Python didn't have nested scopes. [03:32] cool [03:33] hi there spiv [03:35] Hey poolie [03:36] hi spiv [03:36] do you have any particular thoughts on bug 582157? [03:36] Launchpad bug 582157 in Bazaar "pull from launchpad is slow (affected: 1, heat: 8)" [High,In progress] https://launchpad.net/bugs/582157 [03:36] spiv: http://pastebin.com/fqiwpZ7S [03:39] poolie: I like your remarks about the two parts of the problem [03:40] :) [03:40] i was hoping you'd say i was obviously wrong and there was an easy fix :) [03:41] everyone.move_to(ireland) [03:42] Heh. [03:42] spiv: I'm not entirely sure why the is None check is needed; pqm failed two calls to stat() with it though [03:42] bah [03:42] 'bzr st' [03:43] or the IoM, 10% tax and low ping time :) [03:44] spiv: or possibly I'm misreading the trace; trace shadowing a top level import is a bit annoying. [03:44] Hmm, maybe we should fix that [03:46] poolie: so how close are we to finding that SSH is the limiting factor? [03:47] good question [03:47] we know that https from chinstrap is not much faster [03:47] it would be nice to try plain tcp with no encryption [03:47] And SSH with no encryption, perhaps. [03:48] if it is something at the encryption level it might be either encryption cpu overhead or added latency or windowing effect [03:55] mm that could be good [03:55] spiv, what are you up to next? [03:58] Firing off a merge request, then lunch ;) Then https://bugs.edge.launchpad.net/bzr/+bug/551525, then maybe https://bugs.edge.launchpad.net/bzr/+bug/522637 [03:58] Launchpad bug 551525 in Bazaar "reconfigure --unstacked doesn't quite work for lp branches (affected: 1, heat: 6)" [High,Confirmed] [03:58] spiv: no comment on my pastebin ? [03:58] that would be nice [03:59] bug oh especially the latter, but both would be great [04:00] lifeless: the test change seems fine, there's an incomplete sentence in the docstring change, and I lack enough context to judge if that's sane and obviously correct, or nonsense ;) [04:01] i'm going to try Maverick on my desktop to give feedback on some kernel bugs i hit on the new mb [04:01] Hmm, SSH with no encryption seems hard with modern versions of SSH. Oh well. [04:01] i may try ssh without encryption [04:01] istm the next thing is plain tcp [04:01] this might require sysadmin help [04:02] spiv: yes [04:02] You could try SSH with a supposedly faster cipher at least, like blowfish. [04:02] spiv: the doctstring is winging its way up, I'll iterate on it sometime this week. [04:02] spiv: its only missing a single bracket. [04:03] That's ok, the original sentence was arguably missing a comma ;) [04:04] I'd like --lsprof-file [04:04] to tell the smart server to lsprof the request (and lock out other threads yada yada) [04:05] and attach the callgrind format data as another part of the response. [04:06] ! :) === oubiwann is now known as oubiwann-away [07:12] Strangest thing just happened. Doing bzr push bzr+ssh:/// to a server where we don't have RSA keys and so have to type the password on console, [07:12] suddenly a GTK dialog popped up asking for the target user@host's password. [07:12] WTF? === AfC1 is now known as AfC [07:13] Was that Bazaar trying to be helpful, or something else $known trying to be helpful, or a "problem"? [07:13] something else probably [07:13] new ubuntu version ? [07:13] lifeless: Lucid [07:13] both sides [07:15] It was reminiscent of the Subversion credential caching behaviour, but I wasn't expecting Bazaar to do that at all, and especially not for an SSH login password. [07:16] And not I can't duplicate it. If I can't figure it out, I may have to judge that the box got owned [07:16] there is a gtk UI for ssh keychains [07:16] its part of the gnome stack [07:17] Hm [07:17] It was just really anonymous looking. Didn't look that authoritative. [07:17] lifeless: (anyway, thanks) [07:18] look in your process list [07:18] for agent [07:41] hi all [07:51] hey vila! [08:35] hi there vila [08:36] hey poolie [08:39] vila, did i tell you i replaced my broken pc with a 6-core amd? [08:39] it makes selftest --parallel pretty nice [08:39] it gets easily disk bound though [08:39] wow, welcome to the club :-) [08:41] poolie: python -c 'from bzrlib import osutils ; print osutils.local_concurrency()' ? [08:41] 6 [08:41] 6 or 12 ? [08:41] ok [08:41] probably 6, no ht [08:42] AIUI amd have rather different ideas in that space [08:42] so running the full test suite is down to what ? 6mins 8 mins ? [09:06] something like that [09:10] hmm [09:10] tomorrow [09:10] remind me to land my 'delete 1/3 of the registered repo formats' patch. [09:10] :) [09:10] its intermingled with other stuff [09:10] vila, is texinfo blocked, or still on your plate, or ..? [09:11] but I can pull it out and do it with a prod, and I think it will be quite noticable [09:11] still on my plate [09:14] hi jkakar [09:15] Hiya lifeless! [10:06] memo to self, put http://www.webperformancematters.com/journal/2007/7/24/latency-bandwidth-and-response-times.html in the bzr perf bug === euph0ria_ is now known as euph0ria === oubiwann-away is now known as oubiwann === jelmer is now known as Guest64049 === Guest64049 is now known as jelmer === lifeless_ is now known as lifeless [17:56] vila: what's up with: https://code.edge.launchpad.net/~vila/bzr/cleanup/+merge/28315 [17:56] ? [17:56] though you may already be gone [17:58] jam: almost but not quite, I just need to look at it, I'm looking into .conf files needs to be locked and ran into yes-but-lock-info-use-config loop :-) [17:58] vila: sure, just trying to drive the queue, and that is a 'tweak' entry === Pilky_ is now known as Pilky [17:59] jam: I'll look into them as soon as I submit this one [18:16] jam: hi [18:28] hey parth [18:29] bbiab [18:33] hello [18:33] is there a way for a colocated branch switch to store uncommitted changes? [18:35] anyone? [18:37] Stavros: i haven't used colocated branches but shelve command stores uncommitted changes. maybe you can check if they work together. [18:37] i can't find it anything in the docs, but it would be great if colocated branches could store changes a la shelve [18:37] i could have two branches in the same dir and work on them at the same time [18:39] Stavros: yes. i understand the colocated concept in bzr but just haven't tried it :) ... so i don't know how well it works with shelve. [18:39] as far as i know, it doesn't work at all [18:39] i'll add a feature request, thanks [18:39] Stavros: oh. yup. in that case its probably a good idea. [18:43] back [18:44] Stavros: I'll also note that 'bzr-pipeline' does something similar and stores a related group of 'shelved changes' on a given branch [18:44] Added: https://answers.launchpad.net/bzr-colo/+question/116044 [18:44] I don't know the details, though [18:44] jam: oh, let me check that out, thanks [18:44] abentley might be able to give more details (author of bzr-pipeline) [18:45] pipeline looks great, i should be able to figure it out from the docs, thank you very much === maco2 is now known as maco [19:51] jam: I think you perhaps got tricked by the new feed-pqm [19:51] jam: did you mean 'e' ? [20:01] morning lifeless [20:02] well, right now it is hard to test as it says "No remaining merge proposals matching status ['Approved']?" [20:02] however, if there is a single proposal marked approved ,then running 'n' will repeatedly give you back the same proposal [20:02] (go to next) [20:03] lifeless: also, I submitted martin's proposal, but pqm doesn't seem to be running, nor did I get an error message about it. [20:03] though it did go into the Queued state [20:03] and it didn't seem to write a 'sent to pqm message' [20:04] jam: run with --help [20:04] theres an option to access queued ones too [20:04] then use e to send it via email [20:04] lifeless: so is 's' useless now? [20:04] (until we get pqm reading queued messages again) ? [20:04] I sent email a while back - thumper asked us to stop using queued [20:04] yeah [20:04] k [20:05] it was certainly the obvious step [20:05] yeah [20:05] I would love to fix that [20:05] and Queued gives: [20:05] File "c:\Python26\lib\site-packages\lazr.restfulclient-0.9.10-py2.6.egg\lazr\restfulclient\resourc [20:05] e.py", line 852, in __getitem__ [20:05] raise KeyError(key) [20:05] KeyError: '--queued' [20:05] dah wtf [20:05] apparently you need to put the project name before the arguments [20:05] *blink* [20:05] nope [20:05] that doesn't work either [20:05] it tells me that I need to only supply 1 argument [20:06] sigh [20:06] looking [20:06] the __name__ == __main__ section is borked [20:06] overlapping merges I think [20:06] yeah [20:06] checks length and then aborts [20:06] vs using OptionParser for it [20:06] I'll fix it up [20:07] overlapping merges [20:07] oh, and it still uses argv[1] after parsing into args :) [20:07] joy [20:07] http://pastebin.com/Zc6PrRhy is most of it [20:07] oh lol [20:07] that think is going to be me [20:08] the other bit I dunno what happened [20:11] lifeless: it also doesn't work particularly well if you initially gave hydrazine readonly credentials [20:11] :) [20:12] thats surprising [20:12] surely the queue command would have blown up, then ? [20:12] well, it read it ok, but it failed with big tracebacks when I tried to set something [20:12] sure [20:12] it just failed in hard-to-debug ways [20:15] lifeless: anyway, 's' is much more obvious for submit (IMO), but hey, as long as we get something working. === _thumper_ is now known as thumper [21:51] I've accidentally added my email to my name in all my commits, so I'm seeing something like: 'Josh Brown josh@host.com ' - is there anything I can do about this this? [22:00] JoshBrown: you can try to reset your whoami by doing bzr whoami Josh Brown [22:01] alternatively you can edit the "email" line in your bazaar.conf file. On linux that is located at ~/.bazaar/bazaar.conf [22:01] rubbs: I've got it set fine now ('Josh Brown '), it's just that it looks messy in the logs. [22:01] oh, of that I'm not sure. [22:02] unfortunately I must go, I wish you luck in finding your answer [22:02] rubbs: Bye. [22:02] bye [22:02] :) [22:10] JoshBrown: you might look at using something like 'bzr rebase', but I think that tries to preserve the original author info [22:12] jam: Thanks, I'll look into it. [22:14] Actually it only seems to change the actual content, while preserving author info. [22:41] heh, bzr transports ftw [22:41] * jelmer__ runs bzr.dev pack sftp://charis.local/~/samba3.git [22:49] jelmer__: interesting [22:49] now does that end up doing a git pack on the remote side? [22:50] seems like it would be pretty inefficient, but still interesting to have it work [22:50] jam: Since it's all local it's not that inefficient [22:50] jelmer__: well, other than the sftp protocol overhead :) [22:50] local as in LAN I mean [22:51] jam: right [22:59] jelmer__: \o.