[00:45] hi all [00:49] hi ;) [00:49] anyone alive? [00:50] maybe [00:51] maybe not [00:51] theory has it that we can't really tell [00:52] ;] [00:52] ok guys :) [00:52] I need to setup multibranch env [00:52] with bug tracking system... [00:52] the question is how to do it "good" [00:52] and what do you recomend ;-) [00:53] For now I cannot use launchpad... [00:53] not yet.. [00:53] pigmej: I don't really know what you want [00:54] I think he wants a bug system integrated with bzr [00:54] lifeless: I have project... [00:54] We have multiple branches [00:54] pigmej, you might be interested in cluemapper then? [00:54] We want to "manage" bugs etc with web interface [00:54] I'm trying to do it with Trac [00:55] pigmej, http://projects.serverzen.com/pm/p/cluemapper [00:55] :) [00:55] also look into ClueBzrServer [00:55] (its there on that page) [00:55] hmmm [00:55] thanks.. [00:55] but I'm off to sleep now :p [00:55] I will look ;) [00:55] Cu ;) [00:55] night [01:00] Good morning. === arjenAU2 is now known as arjenAU [01:05] jml: are you back? [01:06] lifeless: i don't think he will be [01:06] according to my record === Kilroo1 is now known as Kilroo [01:12] hmmm [01:15] pigmej: why can't you use launchpad? [01:16] I need to store code in private plase [01:16] until first public release [01:16] GPLv3 licence [01:16] launchpad can do that if you get a subscription [01:16] lifeless: yah payed one [01:16] (private hosting is a for-fee service) [01:17] lifeless: lauchpad hasn't private "projects".... [01:17] or am I wrong? [01:17] hmm, I don't remember [01:17] btw error: Could not find suitable distribution for Requirement.parse('TracWysiwyg') [01:17] I suggest you ask on #launchpad about that [01:18] from easy_install cluemapper [01:21] ok I've fixed it [01:21] lifeless: launchpad hasn't private "space" in 100% [01:21] but it's open source now.. so maybe that's the way.. [01:26] pigmej: if you ask and explain the situation they might give you free temporary private hosting [01:26] i'm not sure [01:27] poolie: https://bugs.edge.launchpad.net/bzr/+bug/375013, see comment #1. That's the hpss blackbox test that fails with 2a. [01:27] Launchpad bug 375013 in bzr "Commit doesn't honor stacking invariants" [Critical,Triaged] [02:23] Hi [02:24] spiv: i pushed lp:~mbp/bzr/default2a [02:24] nice idea [02:24] While I'm trying to checkout a branch on my Windows machine, I keep getting a' bzr: ERROR [Errno 2] No such file or directory: u'.....'' while bzr is downloading the tree. any ideas why? [02:24] GungaDin: not unless you provide more information [02:24] GungaDin: look in the log - bzr --version will tell you where the log is [02:25] there should be a backtrace; could you pastebin that thanks. [02:28] if not hasattr(test, '__call__'): [02:28] raise TypeError("the test to add must be callable") [02:28] I sometimes despair about upstream [02:29] lifeless: have you tried cluemapper with bzr ? [02:29] how to do it? [02:30] http://pastebin.com/d220d4a35 [02:30] pigmej: I don't know what cluemapper is - so no, sorry. [02:31] oops [02:31] wrong one [02:31] sorry [02:31] GungaDin: no problem [02:37] http://pastebin.com/d4f2617a2 [02:40] I think you might be running into a path depth issue [02:40] .. ? [02:40] why? is there a limitation? [02:40] either that, or one of the elements in C:/cygwin/home/me/Sources/myproj/.bzr/checkout/limbo/new-100/A/B/C/D/E/a.b.c.d.e.f.datasource is missing [02:40] GungaDin: windows has many limitations [02:41] I hardly believe that's the case. [02:41] is that the actual string, or did you edit it somewhat ? [02:41] I just changed that directory and file names. Not the depth. [02:42] I need to know how long the original string was [02:42] just a sec, I'll let you know. [02:42] You think it's greater than 255? [02:42] 127 [02:42] the limit is 127? [02:43] there are several different limits [02:43] per-dir and total length [02:43] I'm just refreshing my memory [02:43] the path string length is 262 [02:44] ok [02:44] the limit is 260 [02:44] http://msdn.microsoft.com/en-us/library/aa365247(VS.85).aspx [02:44] "In the Windows API (with some exceptions discussed in the following paragraphs), the maximum length for a path is MAX_PATH, which is defined as 260 characters. " [02:46] there are some things that can be done to use longer paths, but they have various different problems [02:46] such as causing Windows Explorer to fail [02:46] http://msdn.microsoft.com/en-us/library/aa365247%28VS.85%29.aspx [02:46] yeah [02:50] if you can make a temp dir a c:\F [02:50] and checkout into that it may work better [02:50] please file a bug; there are some things we can do to workaround this, if we can detect it even half decently [02:55] MSDN also mentions there are functions to use an extended path name. [02:55] up to 32768 chars. [02:57] yes. - as I said above in fact :). "11:46 < lifeless> there are some things that can be done to use longer paths, but they have various different problems" [02:58] spiv: are you sending to pqm from lp successfully now? [03:07] lifeless: not yet, but I don't think I've tried recently. [03:07] My locations.conf still uses http:// for public_location as a workaround. [03:14] ok, submitted a bug [03:14] thx. [03:16] lifeless: subunit is making me install libtool; ewww [03:18] you do get points though for not shipping the standard useless INSTALL file [03:19] hi [03:19] poolie: you probably don't need the C bindings; you can just add ./python to your path; or use one of the debs [03:19] i have the BZR-CIA plugin installed, by any chance does anyone know how to remove CIA reporting from a project (i'm going to be pushing to a seperate branch [03:20] meoblast001: I don't know, sorry. [03:21] fooding [03:21] you have debs now? [03:21] yes [03:22] as I said to jam :P [03:22] https://edge.launchpad.net/~subunit has two PPA's [03:22] releases and snapshots [03:22] you should link to them from the subunit home page..., [03:22] if you're on a distro-release that isn't represented, let me know and I'll upload targetting it for you [03:23] i'm on jaunty and on karmic [03:23] it would be nice if lp would join these things a little more nicely. [03:24] anyhow, right now. _food_. I shall return shortly. [03:33] fools! [03:33] ? [03:33] using Ubuntu ;-P [03:34] rather than? [03:34] Debian, duh! [03:35] oh, i thought you were going to say GEOS [04:01] poolie: subunit 0.1 (a snapshot between 0.1 and 0.2 actually) is in karmic, no ppa needed there at all [04:02] k [04:03] lifeless: does subunit, eg subunit-stats, print out the errors but not the failures? [04:05] don't you need to use subunit-filter to get that? [04:05] that's what I'm doing [04:06] * SamB is using 0.0.2~bzr68-1 [04:06] poolie: subunit-stats consumes the entire stream, just reports the aggregate; if you're seeing other stuff, its being passed through [04:06] which can be caused by badly escaped exceptions (I have to check for bugs there), *or* bzr corrupting the stream by e.g. top level network progress output [04:07] when I run bzr selftest --subunit it tells me *everything* [04:07] SamB: right; you can pipe that through subunit2pyunit for a pyunit display; trunk has subunit2gtk, subunit2pyunit --progress to get a bzr progress bar [04:07] if you want to read the raw stream, subunit-filter --no-skip is probably what you want [04:11] lifeless: bug 408186 [04:11] Launchpad bug 408186 in subunit "subunit-stats prints some errors" [Undecided,New] https://launchpad.net/bugs/408186 [04:11] I usually end up redirecting/teeing to a file and doing different things [04:13] ... interspersed with cursing unittest for not supporting traceback-format overriding [04:15] poolie: checking [04:15] SamB: I have some thoughts/plans on how to tie multiple filters together better [04:15] a shell Y should utility would be nice [04:15] as would a subunit interleaver [04:16] lifeless: how is that going to help me convince unittest to print tracebacks with variable names, without feeling dirty because of the inevetable monkey-patches? [04:17] SamB: it won't, it will just make it easier to deal wiht ;) [04:17] this would be far more tolerable if I BZR_PDB=1 would catch test failures [04:18] oh, variable values too ;-) [04:27] poolie: thanks [04:27] this is an important unittest model issue, as it happens [04:42] lifeless: also, does it have any understanding yet of the various shade-of-grey results? [04:43] poolie: hmm, I don't think using shades of gray is a good idea -- it might cause confusion between the light-on-darkers and the dark-on-lighters ;-P [04:43] mm thanks [04:43] poolie: it knows skip [04:43] it doesn't have wire representation for unexpected success or expected fail, IIRC. [04:43] subunit-filter --error includes UnavailableFeature [04:43] istm it should not [04:44] lifeless: those probably ought to be added, at least [04:44] poolie: yeah, you'd think those should count as skipped! [04:48] I think my dirstate iter_changes branch will land happily now [04:49] sending it in a bit === Xavura is now known as Xavur[a] [05:08] poolie: in the spirit of getting small things done as they happen - https://edge.launchpad.net/subunit - I've touched it up. [05:14] looks good [05:36] lifeless: would it be expected that the text_index for a 2a format holds one more key than in previous formats? [05:36] maybe for the root directory [05:36] re bug 408199 [05:36] Launchpad bug 408199 in bzr "blackbox.test_check fails in 2a format" [High,In progress] https://launchpad.net/bugs/408199 [05:36] yes, apparently for the root [05:37] poolie: all rich roots will, yes. [05:37] thought so, thanks [05:39] * lifeless enqueinates [05:42] I love it when people look sophisticated looking words ... that get zero Google hits [05:43] :) [05:44] And, of course, just when you're doing what should have been an oh-snap put down, you realize you totally blew it with a editing mistake. {sigh}. I hate the universe. [05:44] oh boy [05:44] s/a/an/ [05:44] * AfC slinks away in shame [05:44] AfC: irc needs a feature that allows you to scrub stupid things you say from your readers' minds [05:45] if it makes you feel better, I checked the spelling on the made up word thrice before using it. [05:45] mwhudson: the very earliest versions of ICQ chat at sent a-few-characters-at-a-time and backspace worked. [05:46] hudson really is very pretty [05:46] My spelling got much better when I added an "Add all [red underlined misspelled words] to dictionary" feature to the spell checker. [05:47] My typing is awesome now. I love it. [05:52] lifeless: i really wish that project was called something else [05:53] mwhudson: yeah, I can see that ;) [05:53] (my brother uses it at work too) [05:54] Does he call his server, Michael? [05:54] i hope not [05:54] it would be childishly entertaining [06:08] poolie: yes, the mapping to subunit is partly subunit not being as rich as bzr is (fixable) but also partly the way bzr does its subunit output [06:09] we should probably put a decorator TestResult to change missing deps to skips or something [06:28] poolie: bug 123688 - use stopTestRun [06:28] Launchpad bug 123688 in bzr "selftest leaves the last test showing" [Low,In progress] https://launchpad.net/bugs/123688 [06:28] I think its called 'done' in bzrlib at the moment [06:28] but upstream python merged the feature as 'startTestRun' / 'stopTestRun' [06:28] that's called on the TestResult? [06:28] yes [06:28] thanks [06:29] if it isn't, we should make it so [06:29] because its the right design :P) [06:36] mm [06:36] so it still feels a bit like whack-a-mole [06:37] but it's definitely good to have a standard interface for it [06:38] lifeless: you can read the diff in https://code.edge.launchpad.net/~mbp/bzr/selftest/+merge/9571 if you like [06:38] well, if you're lucky and the diffmonster put it there [06:43] i'd like, in a subunit gui or command line, to say [06:44] get all the tests whose failures look ~like this~ and run them again [06:53] poolie: I wouldn't add the comment under done() [06:54] about the standard name? [06:54] that just adds conflicts to resolve when merging a fixup to the right method name [06:54] yeah [06:54] i'd care more about people understanding they're the same than getting one tiny conflict when changing the name [06:55] in that case I'd use a docstring not a comment [06:56] for the sake of the generated api docs etc [06:57] or rather just because it's public documentation? [06:57] ok [06:57] help(foo.done) / pydoc etc [06:57] looking at the source is less common than looking at the objects, or the help on them [06:57] really? [06:57] not for me :) [06:58] anyhow i'll move it [06:58] i find i use pydoc much less than i did use similar things elsewhere [06:58] i'm not sure why [06:59] because folk write comments ? :) [07:00] mm :) [07:00] i kind of feel it's because there's less of a static model [07:00] but, there is enough to generate some api docs [07:00] anyhow, it's a beer qn [07:00] heh [07:00] so, what are you thinking of subunit so far? [07:01] flaky but intriguing [07:01] in a nutshell [07:01] well, thats better than horrible ;) [07:08] poolie: isn't TestUIFactory new anyway ? :) [07:10] no, it's old [07:10] CannedInputUIFactory is new [07:10] I thought tests used to use SilentUIFactory [07:10] imbw but i think they were both there [07:10] there is some cruft [07:11] thus the bug asking for more thinking and more cutting [07:11] huh, 2007 [07:11] there we go [07:12] so, I've spent most of today playing whackamole on iter changes [07:12] getting the last kinks out [07:12] k [07:12] I hope to land it tomorrow am [07:12] well you know what i've been doing :) [07:12] so far all of them are pretty shallew [07:12] shallow* [07:12] but this may not last [07:12] would you like to have lunch tomorrow? [07:13] otoh i haven't particularly been starting with the easy ones, so ... [07:13] i have lunch most days :-) [07:13] so yes :) [07:13] oh, with you? sure. [07:13] o/~ welcome to pedanical o/~ [07:14] where? [07:14] I was thinking chinese in epping [07:14] but anywhere reachable is fine [07:14] sounds dangerous :/ [07:14] but actually that sounds fine [07:15] feel free to file more bugs on subunit btw [07:15] i'll see you around 12 then [07:15] I like bugs, they are a sign of users [07:17] i wonder how spiv got on? [07:17] re 12 - cool [07:17] poolie: I have lunch most days too ;) [07:21] that said, EOD time for me. [07:22] hi all ! [07:23] hi vila [07:23] hello vila, welcome back [07:24] lifeless, poolie : thanks ! anything urgent ? [07:24] getting 2a stable :) [07:24] stable-er* [07:25] other than working on the 2.0 targetted bugs [07:25] ok [07:25] you could have a look at the six-month-release thread [07:26] it's not urgent as such but i want to know you're ok with it [07:28] you might enjoy the subunit2junitxml + hudson stuff I did ;) [07:33] spiv, re bug 407834, do your changes add (or remove) a warning for inter-format fetch [07:33] Launchpad bug 407834 in bzr "Branching MySQL too slow" [Undecided,Incomplete] https://launchpad.net/bugs/407834 [07:33] i think i put one in at one point... [07:35] when i mean 're that bug' i'm not implying it's assigned to you [07:35] poolie: I don't think I've added or removed any warning there, not sure that there is one atm though. [07:38] How can I check the parent of a rev id? [07:39] or parents [07:39] lifeless: is bug 407834 a bug in its own right? ie can it plausibly be faster? [07:39] Launchpad bug 407834 in bzr "Branching MySQL too slow" [Undecided,Incomplete] https://launchpad.net/bugs/407834 [07:42] it should be; the bugs on IDS, network deltas will make it faster [07:42] hundreds or thousands of times faster I suspect [07:42] spiv has timing data [07:48] I'm not sure if my network deltas patch will make much difference to the CPU cost, but it should massively reduce the network traffic (both bytes and roundtrips). [08:42] hm, cd'ing to a svn checkout inside a bzr branch makes zsh's prompt stuff do bad things [08:45] I just did a really stupid thing... [08:45] and I need help. [08:45] I merged my local branch into trunk... [08:45] then uncommited that revision [08:46] then someone else committed something into trunk [08:46] and now I want to commit a change into my local branch and then merge into trunk... but I can't [08:46] when I try to merge into trunk I get: "bzr: ERROR: Tags not supported by BzrBranch5('file:///home/winnt/TIBRA/yaron.hirsch/Sources/tibra/COM-226/'); you may be able to use bzr upgrade." [09:22] any help please? [10:11] GungaDin: ehm [10:11] GungaDin: the tags error is unrelated to what you previously did [10:42] lifeless: bzr.dev revno 4581 breaks --parallel=fork, is there some subunit pending merge I need that isn't found in lp:subunit ? [10:42] lifeless: breakage is about expected failures and unavailable features\ [10:56] Hey everyone [10:56] Has anyone tried ClueBzrServer and/or ClueMapper with bzr ? [11:04] lifeless, I'm in Dublin [11:06] jml: welcome to my TZ :-) [11:07] vila, thanks :) === Xavur[a] is now known as Xavura [11:14] hi! help me! why uploading data to master branch generate big traffic(over 10 MB) when commiting changes of little text file? [11:17] why by uploading data to master branch does so much traffic generate transfer when I commit changes of short text file [11:18] mityaj: what format/transport/version of bzr? [11:19] LarstiQ: bzr 1.16 [11:19] mityaj: you only answered 1/3rd of my question :P [11:20] LarstiQ: sftp [11:20] LarstiQ: Bazaar Branch Format 6 [11:21] mityaj: the repository format is more importance, just `bzr info` should tell you [11:21] LarstiQ: format: pack-0.92 [11:21] mityaj: k [11:22] mityaj: did you commit a merge? Or where you previously out of date with master? [11:22] mityaj: sftp is a so called 'dumb' transport, so potentially reading the index files from the remote end is going to be a sizeable transfer too [11:24] LarstiQ: What should I use instead of sftp? [11:24] mityaj: or if a pack got performed [11:25] mityaj: bzr+ssh://, but that requires bzr installed on the remote end [11:26] LarstiQ: thank you, I'll try [12:30] vila: no, e verything should be landed [12:30] vila: (everything in subunit relevant to bzr parallel that is) [12:30] lifeless: weird [12:31] jml: theres a subunit patch up if you wanted to review it it would rock [12:31] lifeless, sure thing. I'm probably going to stay in tonight, so I'll try to look at it then. === You're now known as ubuntulog [12:37] you might like the talk I gave at slug too [12:37] theres an announce on the subunit lp page [12:41] lifeless: http://paste.ubuntu.com/245975/ "fixes" the problem for me but it also lose the /nnnn in the progress bar :-/ [12:52] vila, hey, since I'm in your timezone... [12:53] vila, can you have a look at this? https://bugs.launchpad.net/bugs/408342 [12:53] Launchpad bug 408342 in bzr "'bzr up' fails against a read-only remote, trying to create a write lock" [Undecided,New] === mrevell is now known as mrevell-lunch === jurg_ is now known as jurg === jurg is now known as jurg` === jurg` is now known as jurg [13:07] lifeless: did you see http://paste.ubuntu.com/245975/, digging deeper, I can confirm that ConcurrentTestSuite and CountingDecorator are incompatible because TextTestRunner.run explicitly tests isinstance(test, testtools.ConcurrentTestSuite) around line 600 === Xavura is now known as xavur[a] [13:21] kk === mrevell-lunch is now known as mrevell === Edwin is now known as Guest36657 === beuno_ is now known as beuno === Guest36657 is now known as EdwinGrubbs === EdwinGrubbs is now known as Edwin-lunch === beuno is now known as beuno-lunch === deryck is now known as deryck[lunch] [18:13] I exported the netbsd sources, then ran "bzr init --2a" in the directory, "bzr add" and finallt "bzr commit -m 'Initial commit.'". I got a memory error. [18:14] So I increased process memory usage to 512MB, then I could add and commit the files. [18:14] ...but I can't branch it. I get some internal error. 512MB doesn't seem to be enough. [18:15] bazaar doesn't scale very well. :( [18:16] Is there anything one can do to make bazaar use less memory? [18:17] define "The netbsd sources" ? === deryck[lunch] is now known as deryck === beuno-lunch is now known as beuno [18:45] DaffyDuck_: prod the developers? [18:46] DaffyDuck_: I think it's a bit worse than usual at the moment ... some kind of major rejigging of how transfers are done? [18:47] DaffyDuck_: oh, and anyway that sounds like kind of a dumb thing to do -- wouldn't you want to import the history? [18:48] DaffyDuck_: btw ... what were the file count and total size of the tree you tried to import? [19:01] SamB: I don't really need the history. I'm developing something on a specific branch. [19:02] ..and it's about 99138 files. [19:02] It's just a plain "cvs export -r netbsd-5-0-RELEASE src", and nothing else. [19:02] DaffyDuck_: so, what is that exactly? (ref Kinnison's "define netbsd sources") [19:02] DaffyDuck_: kernel/base system/ports? [19:03] The entire system sources. [19:03] The src module, not including x or pkgsrc. [19:03] ok [19:03] DaffyDuck_: keep in mind we don't (necessarily) know netbsd :) [19:03] None the less, it's an excellent testcase for the bazaar developers. :) [19:04] right, people have been known to version the gentoo/netbsd ports or entirey of Debian before [19:05] DaffyDuck_: the thing about history would be that each commit is more reasonable than one humongous change [19:05] DaffyDuck_: having said that, I think SamB is right about specifcally 2a needing some bits to be ironed out [19:06] The first problem appears to be a simple "out of memory" error. Increasing memory size helped. The second problem was a branching problem, which may also be a memory limit issue. The third actually looks like a mini-server (bzr+ssh) problem, but it may be a memory issue as well. [19:06] Should I file a complete bug report? [19:08] DaffyDuck_: that sounds like 3 different ones, but yes please. The bug filing process should prompt you with similiar bugs (some of yours might already be filed) [19:08] and if not, it is much easier to set dupes than to split a bug in pieces [19:09] Ah, good point. I'll file three bug reports. [19:09] Just out of curiosity, are there any (semi-)hidden options which dump relevant information useful to the bazaar developers? [19:10] DaffyDuck_: what kind of information? `bzr info` and possibly `bzr -Derror` (info otherwise in ~/.bzr.log) for general information [19:11] DaffyDuck_: `bzt dump-btree` is a hidden one specific to some formats, but I doubt you'll need that [19:13] OT: Is there a goo way to link from one PR to another in bug-thingie in launchpad? [19:16] DaffyDuck_: linking bugs together? you can say 'bug 12355' and it will show up in the ui as a link [19:17] Launchpad bug 12355 in xorg "Dell Optiplex GX100 only get 640x480 resolution with live cd (dup-of: 12248)" [Medium,Invalid] https://launchpad.net/bugs/12355 [19:17] Launchpad bug 12248 in xorg "[i810] ddc sync ranges get lost in mode validation" [Medium,Fix released] https://launchpad.net/bugs/12248 [19:17] also, saying it in here will let ubottu give some info ;) [19:18] Aha! Excellent. Thanks! [19:30] sidnei: just to let you know, I used 1.0rc8 on Kerguelen and everything seems to build [19:31] jam: cool, thanks for letting me know [19:32] jam: iirc you've done things about not holding huge files in memory at once, how about large inventory(deltas) like with DaffyDuck_ trying to commit all of netbsd src in one go? [19:33] LarstiQ: so we should be holding a single copy of file content during commit for 2a format repos [19:33] As for the large tree shape, etc [19:33] I haven't really done memory profiling there [19:34] I'm writing a bug report with complete instructions on how to repliacte the problems I've encountered. [19:36] sidnei: can you think of an easy way to have it put a build number into the final executables? [19:36] so I can get "bzr -setup-1.17-1.exe ? [19:37] jam: you could pass it on the command line? [19:38] given the command is "make installer-all" I'm not sure what you would pass [19:38] I'm happy to say at build time "this will be the -1" installer [19:38] I just don't know how to get that info passed around [19:39] jam: 'make installer-all BUILD_REVISION=-1' would work? [19:39] yeah, probably be fine [19:40] jam: ok, that's not too hard to do. [19:40] bug #408526, in case anyone wants to take a look at it. [19:40] Launchpad bug 408526 in bzr "bzr commit on large projects require large amounts of memory" [Undecided,New] https://launchpad.net/bugs/408526 [20:14] * SamB wonders why jelmer has all these imports at the beginning of cmd_svn_import.run ... === elmo_ is now known as elmo [20:26]