spivGood morning.01:04
jbowtiespiv: Is PQM happy with bug 596239 now?01:40
ubot5`Launchpad bug 596239 in Bazaar "Bazaar supports https! Update documentation (affected: 1, heat: 4)" [Medium,Confirmed] https://launchpad.net/bugs/59623901:40
spivjbowtie: good question, let's find out...01:40
pooliehi there spiv, jbowtie01:40
jbowtieI'm just wondering if it's actually using sphinx on Python 2.4 - if it's using pure docutils it won't recognize the ref directive even with the better line wrapping.01:42
jbowtieThat might put a crimp in my plan to fix up all the links.01:42
spivI *think* we stopped using pure docutils, but I might be wrong.01:42
spivI'm fairly sure we intended to, at least :01:43
jbowtieHI, poolie, sorry for the lack of bug fixes this weekend, should land some mid-week.01:44
poolienp, no obligation, we appreciate them02:00
spivjbowtie: same error, the Makefile still invokes the 'docs-plain' as part of 'check' target02:02
spivpoolie: any reason why we haven't switched to just Sphinx in trunk?02:02
spivpoolie: I thought that was the plan, maybe we just never got around to flipping the last switch?02:02
pooliespiv, there may be some snags02:05
poolieperhaps that it's not in the pqm chroot?02:05
poolieyou should check with vila02:05
lifelessthere is an rt open for sphinx in th chroot02:06
lifelessthe chroot should be upgraded from hardy too02:06
spivThe Makefile has a comment about how "Post 2.0" we will most likely switch the 'docs' target from 'docs-plain' to 'docs-sphinx'.02:08
pooliespiv, want to chase it with spm etc?02:14
spivspm: ping?  I'm wondering about the status of installing sphinx in the PQM chroot.  lifeless says an RT open, is it blocked on anything?02:21
spmspiv: time probably02:22
spivThat's both good and bad news, then :)02:22
spmmmm :-)02:22
spmspiv: fwiw, it's 9th in priority in the LP queue atm, exclusive of all the other stuff going on.02:23
spivIs that good or bad? :)02:23
=== _thumper_ is now known as thumper
lifelessI have several month long things in there ;)02:37
jbowtiespiv: So should I rework the patch to be docs-plain compliant, or leave it as a useful test case?03:31
spivjbowtie: I think rework it to be docs-plain compliant.  You can leave a XXX comment in the source about how to improve it when we can require sphinx.03:35
jtvThis conflict looks incorrectly bracketed: http://paste.ubuntu.com/501267/04:43
jtvOne line is near-identical between the two conflicting versions (the only difference being punctuation at the end),04:43
jtvbut it is shown as occurring only in one of the two conflicting versions.04:44
spivjtv: I think a case like that came up on the list, or perhaps a bug report, recently...04:44
spivjtv: https://bugs.edge.launchpad.net/bzr/+bug/616749 ?04:51
ubot5`Launchpad bug 616749 in Bazaar "Incorrect merge - Lines missing inside conflict markers (affected: 1, heat: 5)" [Medium,Confirmed]04:51
jtvspiv: thanks—I'll click "this bug affects me."04:51
vilahi all07:33
poolie_hi there vila07:35
=== poolie_ is now known as poolie
vilapoolie: helllooo !07:35
pooliehi there07:36
spivInteresting; running a 'bzr pack' top reports Python is using ~160% CPU (on a 4 CPU laptop), but vmstat consistently shows 25% for the same period.07:40
spivI suspect one of these tools is coping better with a process bouncing between CPUs.07:40
spiv(to be fair I imagine vmstat probably has an easier job here)07:41
spivThat does mean I need to avoid trusting 'top' as an indication of whether a process is productively using more than one CPU (or even using more than one concurrently at all).07:41
spivProbably for per-process information on that I should be using perf anyway :)07:49
pooliethat will tell you about all the cpu transitions, if you want that data07:51
knittlhey poolie *wave*07:52
spivHmm, interesting that sometimes top says Python is on just 10007:53
spivI wonder if that's when it's in pure python code that isn't frequently dropping the GIL?07:53
vilaspiv: quick question: does the news_merge plugin takes releases into account and sort them in chronological order or does it just left them untouched ?08:00
vilaspiv: leaving the unreleased one at top that is... (there should be only one at a time right ? :)08:01
spivThe current version punts as soon as it sees a conflict involving more than sections and bullets.08:05
vilaspiv: ok08:05
spivThere's a work-in-progress prototype of one that can cope with more, but it doesn't try re-ordering releases either.08:06
spivAlthough it probably could if we wanted it too...08:06
pooliehi knittl08:06
vilaspiv: ok, I didn't respect this order when releasing, so I'm fixing it right now and was wondering...08:07
knittlpoolie: do i really have to sign that agreement?08:07
vilaweird weird stuff: a VM suddenly colliding with its own ipv6 address...08:10
* vila suddenly remembers he didn't sacrifice any chiken or goat this morning...08:10
vilaspiv: bah, there *could* be several releases in progress in the same NEWS file08:13
spivvila: yeah, but they'd all have version numbers...08:14
vilaspiv: yes, but no release date08:14
vilaI thought the consensus was to sort them by release date, intermixing series, did I get this wrong ?08:15
vilanow since we can't predict release dates (haha), 'NOT RELEASED YET' should probably stick to the last know one in its series08:16
vilaand be on top if it's the first in a new series08:16
spivI'd thought it was by version, not date.08:18
spivBut I'm not very sure of that.08:18
vilagrep ^:2. NEWS08:19
vilaspiv: at least during the 2.1b era we did sort by date, I think it makes more sense but I'll respect the consensus, so poolie, what's your remembering here ?08:20
vilathough :2.2b4: 2004-07-09 is not very convincing ;)08:21
poolievila i think we should split the news out into one file per series and avoid the issue08:32
pooliealso i think we should not copy news where things are just merged up08:33
pooliejust say "includes everything from 2.0.6" etc08:33
poolieknittl: yes, i see you already did put "copyright Canonical" but we do need the email signature too08:33
pooliespiv istm that regarding network performance, we should file (or just mention) bugs about there being too many startup roundtrips08:34
pooliei think there is one, at least for the graph08:34
poolieand then i guess there's another about the stream holding fulltexts08:35
vilapoolie: so you mean we stop using NEWS in favour of NEWS-2.0, NEWS-2.1, etc ?08:36
vilapoolie: there will be some fallouts regarding the news_merge plugin and the doc generation08:36
vilaspiv: how far from completion is your wip on the news_merge plugin ?08:37
vilapoolie: getting away from NEWS will rejoice jam ;)08:39
knittlpoolie: i didn't do for my last patch08:44
knittland i don't really want to08:44
spivvila: I don't recall off the top of my head, but probably about 50%?08:45
vilapoolie: hmm, thinking a bit more about this proposal, I think the consequences are a bit heavy to switch to it quickly and transparently: we'll need to teach evrybody to use the new scheme, update the news_merge plugin, update the doc generation and fix all the quirks encountered doing that (and who knows what I forget here). I'm not that comfortable doing it without giving it a lot of publicity...08:45
vilaspiv: and can it easily be extended to target NEWS-* instead of NEWS ?08:45
spivvila: yes, with the caveat that the hook is per-file08:46
spivvila: so it can't really cope with situations that need to consider multiple files together.08:47
vilaspiv: does this matter ? If a news entry should now appear only in a single file no ?08:47
vilaor did I forget something ?08:47
spivIt just means it would be hard to make it to implement something like "automatically propagate NEWS entries when merging 2.x -> 2.x+1", if we wanted to do that.08:48
spivvila: as far as doc generation goes, I'm pretty sure it would be a clear improvement08:49
spivvila: currently the doc generation has to basically do the NEWS -> NEWS-* split itself08:50
vilayeah, poolie, I'm not sure the incremental step you're proposing really address the "where did this bug get fixed" need as conveniently as the duplication does today08:50
spivSo if we start keeping the data in that shape in the first place, we can just delete that cruft.08:50
vilaspiv: yup, code wise it will simplify this part, but as a communication mean ?08:51
spivWell, it apparently is how we are communicating — see the generated docs ;)08:51
spivCertainly, as someone that looks at the NEWS file fair frequently to see what changed and/or when something was done, I think I'd find it a little more convenient, if anything.08:52
vilaspiv: no, I mean about duplicating the news entry or not08:52
spivThat's orthogonal, surely?08:53
spivBesides, we aren't consistent on that today, I think.08:53
vilaspiv: right, but part of the same proposal ;)08:53
spiv(Certainly we haven't been in the past)08:53
spivWell, as an orthogonal issue, I'm mildly in favour of it.08:54
spivBut I don't really see it as being related to the split-NEWS proposal at all.08:54
vila<poolie> vila i think we should split the news out into one file per series and avoid the issue08:54
poolieit's true it's mildly related08:55
vilaha, hmm08:55
pooliei don't care strongly about them08:55
spivWell, I don't know why poolie considers it related, but that doesn't change my view :)08:55
pooliei'd like to get the news merge plugin working08:55
vilasorry about the confusion, I ran into it *while* duplicating entries and I mixed the two08:55
vilaand I ran into this while merging up the bugfixes targeted at running the tests during the package build and I try to avoid losing my focus which I seem to failing :)08:57
spiv(Separately, it would be nice to explore how to have nice hooks that deal with situations involving multiple files.  I think actually it might be possible with the current hook after all, at least in some cases...)08:57
vilas/to failing/to be failing at(sp?)/08:57
vilaso, do you (poolie, spiv) mind if I finish my NEWS cleaning *before* looking at changing the scheme ?08:58
vila(which will also leave us a bit of time to at least prepare the new_merge plugin)08:59
spivI don't think any preparation for news_merge is required to maintain the same quality of merging.09:00
vila<vila> spiv: and can it easily be extended to target NEWS-* instead of NEWS ?09:01
vila<spiv> vila: yes, with the caveat that the hook is per-file09:01
vilaspiv: You mean it already accepts a glob ?09:01
poolievila, what do you mean by 'news cleaning'?09:01
pooliei don't think i'll mind anyhow09:01
spivI mean, developers will need to update their configs09:01
spivBut that's not a big deal.09:02
spivIt would be nice to improve, certainly.09:02
spivBut sufficiently minor to be a blocker.09:02
vilapoolie: I'm duplicating the news entries while merging up 2.0 -> 2.1 -> 2.2 ... -> bzr.dev09:02
spivvila: the "per-file" aspect I was referring to isn't about globbing09:02
vilaand sorting them while doing it09:02
spivvila: it's that imagine you have a single semantic conflict that spans two files09:03
=== spike_ is now known as spikeWRK
spivvila: maybe one branch renames a function, the other moves the definition from file A to file B09:03
vilaspiv: sure, different topic, I thought you couldn't use a glob right now09:03
spivI don't think you can, but it's not a big deal.  Update your news_merge_files right now to say news_merge_files=NEWS,NEWS-2.0,NEWS-2.1,NEWS-2.2,NEWS-2.3,NEWS-2.4,NEWS-2.5,NEWS-3.0 and you're sorted for probably the next year at least ;)09:04
spivIt'd be a nice improvement, and relates I think to a feature request for the bzr-changelog-merger, so I'll definitely look at it quite soon.09:05
spivBut I don't see that as a blocker for reorganising NEWS.09:05
vilaspiv: me neither, only as an interruption in another unrelated task09:06
spivWell, changes in general do tend to be a bit disruptive.09:09
spivSo long as we do what we reasonably can to minimise the disruption, I think that's acceptable.09:10
vilaspiv: I'm not saying it's bad, I'm just saying I don't want to do it while I'm already doing something else than can be done without it in less time :)09:12
vila(both elapsed and CPU)09:12
spivvila: well, get that other thing done soon then ;)09:15
spivMore seriously, I'm sure this isn't going to happen overnight.09:16
spivBecause of the need to tweak doc generation, etc.09:16
spivAnyway, EOD for me.  Happy autumnal hacking!09:16
vilaspiv: happy evening09:17
rom1Hi all10:44
rom1Can we rollback packs to obsolete packs ?10:44
vilarom1: that sounds like a really weird and bad idea, may be you can explain the rationale ?11:06
rom1well, i have a corrupted repository (i still look for the cause of the corruption). When a user does any operation in the shared repository, a file nopt found error is returned from an indice which is in obsolete packs.11:08
rom1i guess that an operation has been interrupted while the repositry was packing...11:09
vilarom1: it's generally worse than that: a crash or a file system corruption11:09
vilarom1: os/fs versions ?11:10
rom1That's why i am wondering if we can use the obsolete packas as a rollback11:10
rom1etch/bzr 2.111:10
vilafile system ?11:10
rom1ie ?11:11
vilaext2 3 4 ?11:11
ddaaext3 / xfs / reiserfs / ntfs / bttrfs11:11
vilanfs (urgh)11:11
* ddaa blames cosmic mind-control rays11:13
vilarom1: first, try: md5sum .bzr/repository/packs/*  and look for files whose name doesn't match11:14
vilarom1: you can do the same for obsolete_packs11:14
rom1ok, what if i have some ?11:16
rom1I have made the following test in a copy of my corrupted repository : i copied the expected indices from obsolete to indices & packs11:17
vilarom1: the files are created with a name matching their content and *never* modified by bzr, so the files that don't match indicate a problem with the fs or worse the hd11:17
rom1I could do the blocked operations (ie the previous file not found errors)11:17
rom1then i repacked the repository, all worked fine11:18
vilarom1: we'll come to that shortly, can you do the md5sum check ?11:18
rom1vila : thx, i understand11:18
rom1is has been done, and effectively, i have some differences11:18
vilabad bad, are these files empty ?11:19
vilaeven worse :(11:19
vilarom1: now try: bzr dump-btree .bzr/repository/pack-names --raw11:20
Glenjaminwould it be eaiser to reconstruct the repo from people's branches/checkouts?11:20
vilaGlenjamin: last resort, yes11:20
vilarom1: pack-names is where the pack files *used* by the repo are recorded11:21
rom1ok done11:21
ubot5`For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://tinyurl.com/imagebin | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic.11:22
vilahmpf, a single pack file11:23
vilarom1: now we need a .pack file with this base name and the associated index files11:24
vilarom1: and we need to know whether they are in packs/ dir or the obsolete_packs/ dir11:24
rom1found in obsolete11:25
vilarom1: depending on the format you use there could be .cix .iix .rix .six .tix index files11:25
vilarom1: which ones ?11:25
rom1*ix and pack11:25
vilarom1: then sorry for annoying you, you were just right to begin with :-) Puth them all in the packs/ dir and all should be fine11:26
rom1vila : there's no annnoying. Really thx.11:27
rom1My concern is about the reason of such a failure.11:27
rom1You seem to target the filesystem ?11:27
rom1io error, taht's it ?11:28
vilarom1: now, bzr move these files *before* creating the new pack-names file, so what you're seeing is X-file-esque, and all reports of such result have *always* been tracked down to a fs failure, generally a system crash11:28
vilarom1: it's an ordering issue with the file system, things happened in a way that doesn't match what the fs is telling us11:28
=== jtv is now known as jtv-afk
vilarom1: when operations are buffered and the system crash, *some* operations are lost, somehow :-/11:29
rom1ok, get it. It the first and only repository that causes such a failure. It is a huge one, a really huge one. I suspect the time consuming in packing such a repository and something done/undone during that time...11:30
vilarom1: but if you *also* have f.pack files that doesn't pass the md5sum check... it's an indication that your hd may be dying...11:30
rom1thx vila11:30
vilarom1: even ctrl-C'ing a 'bzr pack' should not lead to such a result11:30
rom1vila : hope so because with 250 developers, i cannot imagine the mess ;)11:31
rom1but the users are always stonger in the worst !11:32
vilarom1: now 2.1... we've fixed some related bugs in the last.. monthS, nothing recent, which 2.1.x are you using ?11:32
Glenjaminin theory, pushing all the tips should leave you with a complete repo.11:32
vilarom1: and your users access the repository with which protocol ?11:34
Kamping_Kaiserare there known problems with pulling LP repos with bzr head? http://paste.debian.net/91961/11:35
vilarom1: the last fix I had in mind seem to be part of 2.1.0 already... so I'm pretty sure you're safe on bzr side.. which leave only the fs/crash or hd theories11:36
vilarom1: now, as Glenjamin hinted, as a last resort, you can ask all devs to push *all* their branches again... hoping that they didn't locally delete some relying on the server...11:37
rom1vila : thx again. I gonna track this disk/repository. If i find something, i keep you informed.11:37
vilarom1: thanks11:38
vilarom1: at least for the record, it would be good to file a bug mentioning the file sizes for the .pack files that failed the md5sum check11:38
vilaKamping_Kaiser: you need to upgrade your local repo if I read this right11:39
vilaKamping_Kaiser: which will be faster if you compile the extensions too ;)11:40
Kamping_Kaiservila: cheers. i could probably have extratged that from the error, perhaps i'll readit a few more times in future :)11:42
vilaKamping_Kaiser: yeah, can be a bit hard to read :-/11:42
vilaKamping_Kaiser: but the reward is a faster bzr ! :-D11:52
Kamping_Kaiserwell yes. but i'd rather it succeeded in doing it automagically ;)11:52
vilaKamping_Kaiser: the rich_root frontier is a no return one, this can't be done lightly when many people are involved11:53
Kamping_Kaiservila: the error implies it tried to though. i wouldn't mind it saying 'manually update', i just object to the 'tried, failed, you do it'11:55
vilaKamping_Kaiser: hmm, that's convincing :) I don't remember the details though, may be you can file a bug with your rationale to get more feedback or luckily a fix ;)11:57
Kamping_Kaisergrrr, i should have seen that coming :p11:58
=== jtv-afk is now known as jtv
Glenjamincan anyone reproduce the :bound alias not working on lightweight checkouts?12:19
Glenjaminexample use case: bzr branch; bzr colo-ify; bzr pull -d :bound :parent12:24
mobbyHi, I wonder if someone could clear something up for me please? It is to do with the "move" command...13:51
mobby If I have "folder_one/FileOne.txt" and I decide to move things around a bit using windows explorer rather than "move" command13:51
mobbyso I end up with "folder_two/blah/FileOne.txt", when I do a "bzr status" I get.. removed: folder_one/ folder_one/FileOne.txt, Unknow: folder_two13:53
mobbythat seems ok13:53
mobbyso I do a "bzr move folder_one folder_two"13:53
Kamping_Kaisermobby: try --after13:54
mobbybzr status gives me - removed: folder_one/FileOne.txt, renamed folder_one/ => folder_two/, unknown: folder_two/blah13:54
mobbyif i try to "bzr move" FileOne.txt I get an error saying it already exists. If I commit I end up with...13:55
mobbyrenamed folder_one => folder_two, missing folder_two/FileOne.txt renamed folder_one/FileOne.txt => folder_two/FileOne.txt13:56
mobbybut folder_two/FileOne.txt doesn't exist in my working copy13:56
mobbyis this correct? I'm not saying it's wrong, I'm just trying to better understand Bazaar :)13:57
mobby*Kamping_Kaiser I did wonder about --after but seeing as the commands were working without it I assumed it was automatically guessing the "--after" flag13:58
sysRPLa new bazaar user here .... Q: how do i add user accounts to a new bazaar project?13:59
sysRPLi.e. i want to allow other people to log into my zerver and check out/check in sources14:00
rubbssysRPL: there is no bzr specific user authentication. bzr respects filesystem permissions, so the best way to allow people access is to create accounts either via ssh/sftp, or ftp. And regulate their permissions with existing user management tools14:02
rubbsI'll see if I can dig up a link14:02
sysRPLso i need to setup an ftp server?14:03
sysRPLi thought bazzar ould BE the server14:03
rubbsbzr can be a server, but there is no user authentication. Also bzr's built in server isn't secure for write access. It's ok for read-only access.14:04
rubbsThis is the admin guide. It may be a little over kill for you, but it covers everything you need. I'll pick out the parts most important to you.14:04
sysRPLokay ... i am on windows here14:05
sysRPLno sshd14:05
sysRPLi'll explore getting it setup to share via ftp14:06
rubbsThat link should help you out.14:06
rubbsthe first one I mean14:06
rubbsthis might help too. it's the command for serving directly with bzr14:08
jamhi jml15:34
jmlwe'd like to import non-master git branches in Launchpad15:34
jmlI heard you guys were doing something that could help us with that15:34
jamjelmer has been working on describing the url format to be able to access it15:35
jamI'm not 100% sure why it can't be done in just bzr-git15:35
jambut he seems to say it still needs changes in bzr itself15:35
jamI think we settled on http://path/to/something,name with something,branch=name allowed to be more specific15:36
jmlmakes sense.15:37
=== deryck is now known as deryck[lunch]
=== Ursinha is now known as Ursinha-lunch
jeremiahHi :)16:17
jeremiahbzr: ERROR: Unknown repository format: 'Bazaar repository format 2a (needs bzr 1.16 or later)\n'16:17
jeremiahI'm getting that error ^^16:17
jeremiahfrom launchpad.16:17
ddaayou're probably not getting that from launchpad16:18
jeremiahAh, okay. But I am trying to download from Launchpad, that's for sure.16:18
ddaainstead, you are probably trying to get a branch in format 2a on launchpad using a version of bzr that's too old16:18
ddaaso you get the error from your bzr client16:18
jeremiahBut my version of bzr is 1.516:18
ddaaso, that's older than 1.1616:19
jeremiahWhich ought to be older than the required 1.1616:19
ddaalater = newer16:19
ddaalater != older16:19
ddaaso you DO need to install a newer bzr to get that branch16:20
jeremiahWell, 1.16 would be released _before_ 1.516:20
jeremiahSo presumably 1.5 is newer.16:20
jeremiahHence 1.16 being the opposite, older16:21
fullermdUm.  No, 16 > 5.16:21
fullermdAt least, it used to be.  Who knows what they've done to math lately...16:21
jeremiahSo 1.50 < 1.16?16:21
fullermd1.5 isn't 1.50.  It's 1.5.16:21
ddaa1.5 < 1.16 < 1.5016:21
fullermd1.5, 1.6, 1.7, 1.8, 1.9, 1.10, 1.11, 1.12, 1.13, 1.14, 1.15, 1.1616:22
ddaabut there is no bzr 1.50 anyway16:22
jeremiahokay, well, I guess I'm off to download a newer bzr.16:22
fullermd(and cpp should learn to handle apostrophes in #define's, 'cuz it's really hard for me to misspell "aren't"  :p)16:23
maxbPerl is the only thing I know crazy enough to treat versions as floats16:31
fullermdI think everything goes through a cycle of trying, with increasingly ugly hacks, until finally giving up and just treating them implicitly or explicitly and n-ples.16:32
maxbThe Java world annoys me16:33
maxbLife would be so much easier if they just adopted debian-style versioning16:33
fullermdI know I went for a while trying increasingly rococo schemes before realizing I was being an idiot.16:33
mtaylorok... what's the way I can fix a branch on launchpad to not be stacked on anything anymore?16:40
fullermdreconfigure has an --unstacked arg.  Whether it works through LP's black magic I don't know.16:40
maxbIt does work16:42
maxbProvided the old stacking branch exists16:42
mtaylorit does. cool16:42
maxbOtherwise, we can tell you about more esoteric solutions16:42
* mtaylor moved the trunk focus branch of a project16:42
mtaylorbut pushing the new branch wound up stacking it on the old focus (of course)16:43
maxbI keep meaning to attempt to write a 'bzr push --unstacked' option16:44
mtaylorfullermd, maxb: thanks! that worked and was _hella_ easier than the last time I had to do that :)16:46
=== deryck[lunch] is now known as deryck
=== Ursinha-lunch is now known as Ursinha
=== beuno is now known as beuno-lunch
nekohayohey there, am I doing something wrong when doing "bzr push lp:~username/project_name/branch_name" ?18:01
nekohayo'cause it returns "bzr: ERROR: Invalid http response for https://xmlrpc.edge.launchpad.net/bazaar/: Bad status line received"18:02
vilanekohayo: sounds like a bad proxy, what os/bzr versions are you using ?18:12
nekohayousing ubuntu 10.04, but I'm in a SSH tunnel right now (SOCKS in gnome's proxy preferences)18:13
vilanekohayo: but if it try to use https instead of bzr+ssh, it's probably because you didn't 'bzr launchpad-login' (once)18:13
maxbI believe the thing that resolves lp: URLs doesn't integrate well with proxies. You might like to try manually replacing the lp: with bzr+ssh://bazaar.launchpad.net/18:14
nekohayomaxb: is there a bug report about that?18:14
nekohayoor should there be?18:14
nekohayovila: nah, I'm logged into launchpad already18:14
maxbprobably, somewhere :-)18:15
vilanekohayo: I'm not sure, it depends on how you configure your proxy... we shouldn't give such an obscure error... but I'm not sure I understand what happens here18:15
vilanekohayo: we had bugs with proxies, I don't remember the exact versions, what does 'bzr version' says ?18:16
vilanekohayo: there is also a 'bzr ping' plugin that could help validate your proxy18:17
nekohayobzr version 2.1.118:18
vilabug #558343 may apply18:19
ubot5`Launchpad bug 558343 in Bazaar 2.1 "NotImplementedError: "should resend request to http://feeds.edge.launchpad.net/bazaar/, but this isn't implemented" during lp name lookup (affected: 17, heat: 105)" [Undecided,In progress] https://launchpad.net/bugs/55834318:19
vilaha, may be not ;)18:19
vilamaxb: that message was totally obscure too but coming from lp :)18:20
mgznekohayo: I'd file a bug, your specific symptom doesn't turn up any other bugs though as vila says one of the wider proxy issue ones might cover it18:20
vilanekohayo: or you can just try to file a bug and see if lp find a dupe... mgz types faster :)18:21
mgzredoing the (failing) command with -Dhpss and putting that in the bug may help.18:21
vilaand -Dhttp since the error message refers to it18:22
nekohayovila, mgz: putting either -Dhpss or -Dhttp between "bzr" and "push" = same output18:26
nekohayono diff18:26
mgzah, sorry, I meant then paste the contents of .bzr.log for that command18:26
mgz(which will have the extra debugging stuff)18:26
mgzvila: I've got a fix for tests not having times on babune, but it's ridiculous18:28
vilamgz: now you got me interested :)18:29
nekohayofiled as https://bugs.launchpad.net/bzr/+bug/64912418:30
ubot5`Launchpad bug 649124 in Bazaar "push to launchpad doesn't work through a SOCKS proxy (SSH tunnel) (affected: 1, heat: 6)" [Undecided,New]18:30
mgzto get started, here's a picture of the test result classes involved when running selftest --parallel <http://float.endofinternet.org/temp/subcomprehensible.txt>18:31
nekohayothanks folks, gotta go now :)18:31
maxbnekohayo: Do you have a moment to try one more command?18:31
maxbDoes bzr push bzr+ssh://bazaar.launchpad.net/~kiddo/recipe-manager/performance work?18:32
nekohayodamn, my top secret url! :P18:32
maxbIf so, we can pin the problem squarely on the lp: xmlrpc resolved18:32
nekohayomaxb: yes, it works18:33
maxbbug 186920 I think18:33
ubot5`Launchpad bug 186920 in Python "bzr xmlrpc client doesn't use http proxy, causing network errors trying to resolve lp: urls (affected: 18, heat: 139)" [Unknown,Confirmed] https://launchpad.net/bugs/18692018:33
vilamgz: shudder18:34
vilamgz: the time is measured on the wrong side of the process boundary ?18:34
maxbnekohayo: According to comments in that bug, apparently bzr 2.2 will handle this better18:35
maxbIf you are able to upgrade and retest, that would be greate18:35
=== beuno-lunch is now known as beuno
mgzwell, subunit doesn't actually time tests, it just has a think that whams timestamps in all over the place18:39
mgzanyway, can fix without touching subunit but need changes in both bzr and testtools18:41
vilawill your changes be *compatible with older testtools ? (Installing 0.9.6 on pqm is still pending for... unkown)18:42
mgzmmm, good crème caramel18:43
mgz^I think so, but working out a way of fixing this without breaking the world (across three projects) was a pain18:44
vilamgz: worth mentioning :-/18:44
barryhi folks.  i'm working on a branch that adds some functionality to a plugin (bzr-builddeb).  obviously i want to do tdd, but i'm not sure of the right/best way to run tests locally.  i'd like to disable all plugins but run the tests from my working branch of the bzr-builddeb trunk.  the online testing guide didn't help too much with this scenario (or i missed it).  any suggestions?18:54
james_wbarry: bzr selftest -s bp.builddeb18:54
james_w"run all the tests with ids starting with bzrlib.plugins.builddeb"18:54
james_wif your local branch is where builddeb plugin is currently being taken from18:55
barryjames_w: cool.  do i need to fiddle with ~/.bazaar/plugins to point to my branch?18:55
barryjames_w: ah, i guess "yes" :)18:55
james_wif not, then BZR_PLUGINS_AT=builddeb@$PWD bzr selftest -s bp.builddeb18:55
barryjames_w: nice18:55
james_wwe're closer to being able to have a test.py in builddeb that runs the tests from where it is invoked from18:56
barryjames_w: that would be a nice addition18:56
barryjames_w: what version of python does bzr-bd have to remain compatible with?19:26
james_wbarry: bzr sticks with 2.4, and I tend to stick with that19:26
barryjames_w: cool, so no ''.format() :)  (and i'll have to install py24 and py25 from source)19:27
james_wbarry: yeah, sorry19:27
james_wbarry: I don't actually test on 2.4 anymore though :-)19:27
barryno worries19:27
james_wit was just that until recently I didn't know any >=2.5 features, so would never write any ;-)19:28
ScottKDo it in a Debian VM and you won't have to build 2.5 from source.19:28
barryScottK: yeah.  james_w: 2.6 is a nice sweet spot.  has some great features (.format among my favorites)19:29
barrybut anyway... thanks19:29
james_wI've been enjoying context managers19:29
james_whaven't tried .format() yet19:30
maxb2.5 for lucid is still lurking in the launchpad PPA19:31
maxbwhich reminds me, I should propose removing that19:31
barryjames_w: with statements are highly awesome19:31
* barry is old skool and builds python from source in his sleep19:32
barryjames_w: so BZR_PLUGINS_AT doesn't override a different version of the same plugin in ~/.bazaar/plugins, right?20:23
james_wbarry: I thought it did20:23
vilabarry: yes it does or it will be useless, or did I misunderstood the question ?20:24
barryjames_w: oh, nm.  i had conflicting registration code in a differently named plugin20:24
james_whi vila20:24
vilajames_w: hey :)20:24
mwhudsonBZR_PLUGINS_AT is awesome, btw20:24
barryvila: no, it's pebkac.  i register ubuntu: in bzr-debuntu but i'm moving that to bzr-builddeb ;)20:24
barryand testing the latter20:24
vilabarry: ouch 8-)20:24
vilabarry: there is also BZR_DISABLE_PLUGINS, just for you :)20:25
barryvila: you guys think of everything! :)20:25
vilabarry: nah, just hacked it furiously and bzr pqm-submit --two-days-ago20:26
agrumanis it possible to have a (binary) file in bzr that there is only one version of (ex update will delete old one completely)?20:47
beunoagruman, no, not possible20:52
git__is there a way to bzr upload to a remote ftp site without having it delete the log file there that's not in the working tree20:59
beunogit__, I *think* we added a bzr-upload ignore command with vila a while back21:01
vilabeuno: hehe almost, there is .bzrupload-ignore file, but don't forget to commit it as it applies to the uploaded revision21:02
vilagit__: ^21:03
beunothat's right21:03
git__let me try21:04
git__works like a charm, thanks !!!21:06
vilabeuno: thumbs up !21:06
roryyhuh. launchpad is pretty awesome.21:29
mgzvila: put up both branches for the test timings fix, if you want to give 'em a run on babune. now I can go to bed and have nightmares about being chased by subunit streams.21:43
vilamgz: testools branch and bzr.dev branch ?21:44
vilaurls ?21:45
mgzlp:~gz/bzr/use_testtools_timings_625594 and p:~gz/testtools/result_timings_forwarding_62559421:45
vilamgz: caveats ?21:47
vilamgz: can use this testtools with an unmodified bzr, can we  use this bzr with an unmodified testtools21:47
vilamgz: I don't want to sound too demanding, I just want to know what will break if mess things up :)21:48
mgzI think it's fine both ways, but the testtools change is more invasive.21:50
vilamgz: any significant subset we can try it against crosses your mind /21:50
vilamgz: any significant subset we can try it against crosses your mind ?21:50
vila-s subset I mean21:50
mgzI like bt.test_ or bb.21:51
vilaoops, wrong branch :)21:54
vilamgz: http://babune.ladeuil.net:24842/view/debug/job/selftest-subset-all/ in progress21:58
mgzoo, is that new?21:59
vilanot quite, but the plan is to make it open to devs22:00
vilaand *that* is not done yet22:00
mgz...I was just pasting that too22:01
mgzlooks good!22:01
vilayup :)22:01
bialixgood sleepless night, evryone22:02
vilafinally... we may get some idea why FreeBSD is so slower than the others (I'm pretty sure it's self specific but why...)22:02
bialix\o_ vila22:03
vilabialix: hey !22:03
vilamgz: http://babune.ladeuil.net:24842/view/debug/job/selftest-subset-all/lastCompletedBuild/aggregatedTestReport/22:03
bialix\o_ mgz22:03
vila^ is the right page to track for completion22:03
mgzhey bialix!22:03
fullermdIt's slower because bialix isn't sleeping?22:05
bialixthat will explain many problems22:05
vila...and the solution will be sooo elegant: Hey ! Bialix, go to sleep22:06
fullermdTo think, AMD and Intel are spending $kadzillions on speeding up CPU's, when they could just spend a few bucks on sleeping pills for you instead.22:06
* fullermd has a revoluationary new business plan.22:06
bialixkill the chicken?22:06
bialixor just me22:07
* fullermd has TWO revolutionary new business plans!22:07
vilabill the kitchen >22:07
vilabill the kitchen ?22:07
vilabialix: nah, fullermd is in the goats business now22:07
fullermd"Your computer systems too slow?  Pay us $2750 a day and provide a cot, and we'll fly bialix out to your location to take a nap!"22:08
bialixman who stares on the goat?22:08
bialixfullermd: that won't scale22:09
bialixor you need to buy a cloning license as well22:09
vilabialix: no, just selling goats to be killed, works better than chickens22:09
fullermdIt doesn't have to scale; I'll split the proceeds down the middle with you, and I can live very happily on $1375 a day.22:09
* ddaa applaudes fullermd for his business acumen22:10
* bialix mutters: a day or a night?22:10
fullermdAnd heck, it'll be great for you; you'll have a boss who never yells at you for sleeping on the job.  What more can you ask?22:10
ddaathe thing, is you'll need to sleep 24/722:11
vilabialix: fullermd doesn't care, he lives in a cave, no light there anyway22:11
ddaalest customers start complaining22:11
ddaaso maybe it's not such a good deal for you22:11
bialixhey ddaa22:11
ddaabut at least it will sustain your wife and children ;-)22:11
fullermdWell, we price for risk.  For an extra $20k, we'll put bialix into an induced coma.22:11
bialixsounds good, but the price is too slow, I think22:12
fullermdWell, that's per day.22:12
fullermdAnyway, if it kills you, I figure we can keep you from smelling for a week or so, and keep charging them for the extra days.22:13
fullermd(I know, it's hard to believe nobody's offering me senior executive positions, isn't it?)22:14
bialixyes, it is22:14
ddaaI figure ukrainians are essentially imputrescible, with all the vodka...22:14
vilamgz: don't forget to add the babune url in your mp22:14
ddaaotoh, the french decay very fast, with all the smelly cheese22:15
fullermdSo, if we cross-bred them, we'd get...   moldy vodka?22:16
mgzvila: which one? :)22:16
vilamgz: the aggregated one I thin, it gives access to all the details22:18
vilaoh, hmm, you need in number in there22:18
mgzI've got a free bsd slowness guess.22:27
mgzSeems to be on writing to the filesystem. Different sort of tmp partition?22:28
maxbabentley: Hi. Could you upload the bzrtools 2.2.0 tarball to Launchpad? Thanks.22:45
vilamgz: tmpfs configured there AFAIR but I haven't checked for... pfew, never since I started hudson22:47
dobeyhow does one do _iter_revisions() on a RemoteRepository, since it doesn't seem to have an _iter_revisions() method?23:15
bialixdobey: it seems the closest method you can use is get_revisions()23:23
dobeyyeah, i'm trying that now23:24
lifeless_iter_revisions is private ;)23:25
lifelessdobey: what are you trying to calculate?23:25
dobeylifeless: i have a graph result, and i am iterating over the revisions in it, to get the apparent authors for all the revisions in the set23:27
dobeyand _foo isn't truly private in python, __foo is though.23:33
bialixit's a convention23:34
lifelessdobey: __foo is no more private.23:36
lifelessdobey: There's no enforcement ever; _ is widely recognised as not-public-thanks.23:36
dobeylifeless: trying to access __foo raises an exception.23:37
dobeyaccessing _foo does not23:37
lifelessthats because __foo is compiled to __ClassName_foo23:38
lifelessor something approximately like that23:38
dobeyanyway, get_revisions() is sufficient23:40
=== Ursinha is now known as Ursinha-afk
bialixdOxxx: hi23:50
dOxxxbialix: hey :)23:50
bialixcan you comment something on bug #505692 ?23:50
ubot5`Launchpad bug 505692 in Bazaar Explorer "Bazaar Launcher (dock icon) for Mac OSX (affected: 4, heat: 20)" [Wishlist,Confirmed] https://launchpad.net/bugs/50569223:50
bialixdOxxx: ^23:51
bialixdOxxx: is it something that Mac installer should provide or?23:51
dOxxxbialix: the mac installer should provide an .app bundle for launching bzr explorer, I just haven't gotten around to doing it.23:52
dOxxxI think there's already a bug for it on the installer project23:52
dOxxxhmmm or not23:52
dOxxxah it's in bzr-explorer project23:53
ubot5`Launchpad bug 505692 in Bazaar Explorer "Bazaar Launcher (dock icon) for Mac OSX (affected: 4, heat: 20)" [Wishlist,Confirmed]23:53
dOxxxthat's the same bug XD23:54
dOxxxI can take a look at it, if you like.23:55
dOxxxIt does seem to be something that should belong to the installer.23:55
bialixI need some comment along the bug discussion23:55
bialixso I can understand what needs to be done23:56
dOxxxok, I'll add something23:56
bialixor change the project to mac-installer23:56
dOxxxI'll do that23:56
bialixor add mac-installer as affected too23:56
bialixdOxxx: thank you23:57

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!