=== r0bby is now known as robbyoconnor === 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: jelmer [00:42] i'm going to finish a coupel of the pending ones though [00:50] poolie_: btw, the subtree detection stuff would mean that adding files in a subtree will require changes to status to support it. [00:52] true [00:52] (no opinion on desirability, just flagging the impact) [00:52] i'm not going to do it now [00:53] it will need some careful testing [00:53] today after piloting some stalled bzr things i hope to work on the incoming mail feature flag for dkim configuration [00:53] cool [01:24] Good morning folks [01:29] hi spiv [01:32] poolie_: removing the hardy chroot on pqm slipped until today, removing it now unless I hear screaming otherwise. :) [01:32] * poolie_ does not scream [01:33] oh for interest, i submitted back-to-back merges the other day and the run time was pretty stable at 40m [04:16] I have a bzr repo retrieved by "bzr branch" command. How to I change the working directory to a certain revision? Doing bzr checkout -r x yields a file exists error for .bzr [04:16] evanton: 'bzr update -r whatever' [04:16] oh, it's update in bzr [04:16] poolie__: thanks, I'll read the help for update now [04:17] checkout makes a new checkout [04:17] update updates an existing working tree from the repositoyr [04:17] poolie__: I'm just used to switch revisions using checkout in another DVCS === poolie__ is now known as poolie [04:29] i'm going to look at https://bugs.launchpad.net/bzr/+bug/764188 [04:29] Ubuntu bug 764188 in Bazaar "test_report_bug_legacy fails on trunk" [Critical,Confirmed] [04:29] hopefully shallow [04:30] ok [04:31] doctestmatchers really do kind of suck [04:31] i think it was broken by wordwrapping [04:32] Wow, that's a difficult error to parse [04:35] that's another bad thing about it [04:35] launchpad wrapping it again is not helping [04:36] how are you? [04:40] Pretty good. I've got a touch of the cold V had, but so far only a touch of it. [04:40] I've got to the point of having some failing tests for the pack concurrent writes [04:41] oh, nice [04:41] Oh, and http://twistedmatrix.com/trac/ticket/5011 has gone through another round of review and improvements. [04:41] hopefully clean non-thread-based tests? [04:41] Yes :) [04:51] review pls? https://code.launchpad.net/~mbp/bzr/754188-apport-test/+merge/58064 [04:59] Reviewed (and approved) [04:59] almost needs a tag +should-have-known-better === 16WAAA4BV is now known as wallyworld1 [06:26] spiv i thought there was a function to raise knownfailure when a particular assertion fails [06:26] but i can't find it [06:26] oh maybe in testtools [06:26] yes it is [09:35] hm so bellany running make check is about twice as slow as a buildd building the whole binary package (including make check and other stuff) [09:36] maxb: i added some stuff to http://wiki.bazaar.canonical.com/DailyPpaStatus#preview and i think fixed the trunk builds [09:40] hi poolie [09:40] poolie: so the buildd is significantly faster than our pqm... interesting [09:40] so we should be submitting it to buildd first [09:40] :) [09:41] :) [09:41] there is a drawback that you may wait quite a while to actually _get_ to build on the buildd [09:41] if you include that, it's probably slower [09:43] anyhow, it broke a little bit and i think now it's ok again [09:43] did you ever touch the rules code? [09:47] I've touched it very slightly. [09:57] it seems like we need a better api than an iterator of every relevant value [09:58] surely you normally just want one? [09:58] i think i'll take a detour to do it === nixness is now known as foocraft [10:42] Hi all [10:43] Is there a way to ignore tabs and spaces conflicts while merging ? [11:56] rom1: not that I know of. Though if there were conflicts, what would you put in the output? [11:57] jam: well, in fact, it is to avoid that it produces conflicts (for instance in java or c programs). I have noticed some reported conflicts with bzr-builder from spaces versus tab. [11:58] But i will have a look at the content-filters i guess [11:58] rom1: sure, I understand the concern, but what do you actually put since they disagree on what the content would be [11:58] rom1: using content filters you can help enforce the form stored in the repository. However, merging is done on the repository form, AIUI [11:58] so existing content won't be helped [11:59] in my idea, i keep lines from OTHER [11:59] rom1: I know you can create a custom merge function, too [12:00] thx jam, i gonna have a look at that === mrevell is now known as mrevell-lunch === mrevell-lunch is now known as mrevell === Ursinha-afk is now known as Ursinha [14:35] hey guys [14:35] how can i check out only a specific subdir of a repo? [14:37] cheater : you cannot. You can only checkout the whole repository [14:38] hmm [14:38] Revisions are revisions of repository and not revisions of files as in svn or cvs [14:38] gotcha [14:38] s/repository/branch/ to be precise. [14:38] that's fine with me, i just want the file system to only have just that subdir [14:38] There exist 'views', which are a primitive that could be used toward building such a thing. But nobody's done it. [14:38] i wish bzr would keep the whole repository just in .bzr somewhere [14:39] fullermd: aha [14:53] Greetings. Does anybody know of a way to rename a pipeline? [14:54] briandealwis: pipes basically work by setting next_pipe / prev_pipe in branch.conf [14:54] ah, thanks jelmer — I hadn't started poking [14:54] briandealwis: so if you rename the branch dir and update those settings in the previous and next pipe you /should/ be fine (I haven't tried) [14:56] jelmer: yup, that seemed to work. Thanks! [15:41] jam: having fun with trees and inventories too ? [16:52] hi, how can I tell bzr to exclude some extensions? [16:54] gypsymauro: you can uninstall them or set the BZR_DISABLE_PLUGINS env variable [16:54] gypsymauro: why do you want to disable some plugins? [16:55] I mean file extensions for example I qant to forgive .log or .ini files [17:00] gypsymauro: read up on the .bzrignore file [17:14] tanx briandealwis :D === deryck is now known as deryck[lunch] === tremolux_ is now known as tremolux === deryck[lunch] is now known as deryck === beuno is now known as beuno-lunch === beuno-lunch is now known as beuno === imback is now known as JasonO [22:16] I'm using bzr to work with svn repository. Today suddenly I found I can not correctly update it by "bzr update" [22:17] ablmf333: can you elaborate? [22:17] When I try to commit, I got " bzr: ERROR: Bound branch BzrBranch7(file:///home/peter/Work/repo/cms/) is out of date with master branch SvnBranch('svn+ssh://peter.cai@svn_repo/sites/svn/cms/site/branches/sprints/V000006'). [22:18] But when I run "bzr update", I got "Tree is up to date at revision 296 of branch svn+ssh://peter.cai@svn_repo/sites/svn/cms/site/branches/sprints/V000006" [22:19] I found similar problem this afternoon, so I removed local repo and re-checked out the code. [22:19] But again, I met this problem. [22:19] svn works fine. [22:19] ablmf333: what version of bzr-svn are you using? [22:19] ablmf333: what does "bzr missing" report? [22:19] bzr missing says "bzr: ERROR: No peer location known or specified." [22:20] ablmf333: and with the master branch specified? [22:20] bzr-svn version : svn 1.0.5dev [22:20] Support for Subversion branches [22:21] bzr missing . [22:21] Branches are up to date. [22:21] ablmf333: did you commit any merges recently? [22:22] Yeah, I pushed some changes. Then I got found local repo could not correctly update. So I re-checked out. [22:22] So my earlier commits messed up the repo on svn? [22:23] ablmf333: you did "bzr push" to push changes, rather than committing to the checkout? [22:23] did any of those involve merges? [22:23] Yeah, it used to work fine. [22:23] Yeah, I remembered I merged from trunk. [22:24] Did some revision on branch. [22:24] Push [22:24] that's probably what triggered it. Recent versions of bzr-svn should cope with this better. [22:24] Then I found local repo can't update anymore. [22:24] Is there any reason you're using an snapshot rather than a release? [22:24] You mean bzr-svn? [22:24] ablmf333: yeah [22:24] Can't really remember. [22:25] I will try a release version then. [22:25] I don't think a release will work better, the related fixes are only in newer snapshots [22:27] jelmer: So should I try a snapshot of bzr-svn? [22:27] ablmf333: you're already running a snapshot, but a fairly outdated one. [22:28] ("dev" in the version string indicates it's a snapshot rather than a release) [22:28] k, I will try to find a newer snapshot [22:34] jelmer: I checked out bzr-svn by "bzr co lp:bzr-svn" [22:34] Is that right? [22:35] ablmf333: yep === Droid is now known as Guest66278 [22:36] then I coped it to "/home/peter/.bazaar/plugins/" [22:36] but "bzr plugins" still tell me it's 1.0.5dev [22:41] it should say 1.1.0dev [22:41] ablmf333: you need to rename it 'svn' rather than 'bzr-svn' [22:41] jelmer: thanks. I got it, but new problem: cannot import name CommittedQueue [22:41] Unable to load plugin 'svn' from '/home/peter/.bazaar/plugins' [22:42] ablmf333: what version of subvertpy does ~/.bzr.log say you have? [22:42] let me check [22:44] In [2]: subvertpy.__version__ [22:44] Out[2]: (0, 7, 5) [22:45] ablmf333: that's too old; the newer bzr-svn requires subvertpy 0.8.0. It should warn about this, I'll check why it doesn't. [22:51] jelmer: 0.8.0 of subvertpy reports "Unable to load subvertpy extensions: No module named client" [22:52] jelmer: Oh, I see, that's because I am still in subverpy source code dir [22:52] ablmf333: subvertpy has a C extensions that need to be built [22:55] jelmer: opos, svn plugin crashes : http://pastebin.com/TtG9MRyR [22:55] ablmf333: that's because your subvertpy is too old [22:56] hi poolies :) [22:59] hi jelmer :) [22:59] how are you? [22:59] thanks for the review [22:59] i'm going to try to finish off sergei's patch today [23:00] a full version with tests and docs will be a bit more than 5 lines or whatever it was [23:00] poolie: very well, thanks :) [23:01] poolie: is that his patch related to diff ? === Ursinha is now known as Ursinha-afk [23:03] yes, to showing the function name === wizonesolutions2 is now known as wizonesolutions [23:08] you're pp this week btw [23:10] i should read vila's patches too... [23:12] poolie: Yep [23:14] I'm trying to break a lock but I seem to be doing something wrong. when I try to run bzr pull on a svn-import [23:14] i'm getting [23:14] ablmf333: lp:bzr-svn should now warn properly on older versions of subvertpy [23:15] jelmer: Ah, now the problem seems to be bzr is too old [23:15] ablmf333: what error are you getting? [23:16] Unable to obtain lock file: [23:16] http://pastebin.com/C2mK7cBP [23:18] ablmf333: I've pushed a fix, can you try again? [23:18] jelmer: k [23:19] I think I may have figured it out. [23:22] jelmer: k, the plugin work. But it doesn't solve my original problem - I can't commit [23:22] It asks me to update, but when I update, it says my repo is up to date. [23:23] I think I'm missing a step when working with a central svn server. [23:23] Question.. Whenever I have 2 repos with a shared history, they are synced, and then I modify file "a" in repo "a" and file "b" in repo "b".. Now, when I want to pull repo "a" into "b", I have to perform a merge.. Since different files were modified, it basically could have been like file "a" and file "b" were modified in repo "b".. Isn't it possible to make push/pull like this that merge without problem just possible without having to merge? and only [23:23] in case that there is a conflict, that a merge is required? it would shave off some of the needed work... [23:25] ablmf333: do you have any pending merges? [23:25] sven_oostenbrink: No, because you don't merge files; you merge revisions. And you have new revisions in both sides. [23:25] jelmer: Yes. [23:25] ablmf333: (the last few lines of "bzr status" should show you) [23:25] I have some pending merges [23:25] that's why I need to commit [23:26] ablmf333: is your current head the same as the head of the master branch, or is the pending merge the current master branch? [23:26] So far I've been able to do a svn-import and create a branch. Next I was able to pull from the central svn repo into my local repo (where I don't have a working tree) next I guess I want to my branch with the repo ( my head isn't getting how to do that without a working tree). But where i'm *most* confused with is how do I push back up (assuming I did all my actual work in my branch wich points to the repo) [23:27] jelmer: thx, but I will use svn for this now. It already wasted me too much time. [23:27] knighthawk: you can either create a working tree in the branch that was created by "bzr svn-import", or you can create a clone of that branch with a working tree [23:28] knighthawk: you can use "bzr push" to push your changes made in the local branch back to svn [23:29] jelmer when I ran bzr branch did it create a clone of the branch in the repo or do I need to use a different command to have my pushes go up? [23:29] ablmf333: It'd be great if we can narrow it down a bit so I can fix the bug in bzr-svn [23:30] knighthawk: I'm not sure I follow, what sort of "bzr branch" command did you run? [23:30] jelmer: I've delete the repo folder. But I still have another one that couldn't update correctly. [23:30] That one when I run update, it says it's up to date. [23:30] but it's different from what is on svn [23:31] jelmer: What do u want to know? [23:31] ablmf333: so "bzr up" doesn't update to the actual last revision in svn? [23:32] yes. [23:32] ablmf333: are there any merges involved? [23:32] jelmer so after I did the init bzr svn-import (I didn't really want a big working tree in that dir) I ran 'bzr branch ' into another dir. I've been calling that my working tree. I've been doing my work their. Then thought the workflow I needed was to update my "working tree" to what's in the svn repo to go into my local repo(created for svn-import) then run a pull there and merge that with my branch with my working tree [23:32] jelmer: The story is : 1) i merged trunk to branch. 2) work on branch 3) push back to trunk. 4) trunk couldn't update anymore. [23:33] knighthawk: you should be able to just run "bzr push " from wherever you've been working [23:33] ah thanks jelmer [23:33] jelmer: That makes me can't not fetch some reversions on server. [23:34] knighthawk: is the trunk you're trying to update a svn working copy or a bzr checkout? [23:35] I'm going to say a svn working copy. [23:35] knighthawk: sorry, that was for ablmf333 [23:36] ah [23:36] ablmf333: is the trunk you're trying to update a svn working copy or a bzr checkout? (iow, does it have a .svn or a .bzr dir?) [23:36] jelmer: It's a bzr checkout. I used 'bzr co svn_ssh://....." to check out. [23:37] ablmf333: thanks [23:39] ablmf333: what do you mean with push back to trunk exactly? "bzr push svn+ssh://.../trunk" ? [23:39] jelmer: No, I created a branch by `bzr branch trunk_dir branch_dir` [23:40] ablmf333: you mentioned you push the work you did on the branch back to trunk; what did you do to do that exactly ? [23:41] in 'branch_dir' I run 'bzr push :parent' [23:41] Apparently, those changes I made on branch is committed to svn. [23:42] But the problem is, after that, the trunk is not able to synchronize to svn server anymore. [23:43] ablmf333: thanks, I think I've narrowed it down [23:45] ablmf333: so "bzr log" doesn't show the revision that was pushed? [23:45] Is this a fix everyone is going to need? (he asks just before pushing to his svn trunk) [23:45] knighthawk: which fix? [23:46] what's going on with ablmf333, when you said you' narrowed it down it sounded like you were fixing something. but I see now that may not have been how you meant it. [23:46] ablmf333: sorry, "bzr log" in the trunk doesn't show the revision that was pushed? [23:46] knighthawk: no, that's related to "bzr push" to a branch bound to a svn branch [23:47] knighthawk: you're not using bound branches at all [23:47] k thanks [23:47] jelmer: bzr log show the are pushed. But one single file is different to svn sever, even if bzr tells me the working copy has not changed. [23:48] ablmf333: different as in not yet updated? [23:49] jelmer: Yeah. I can't update that file to its newest version. [23:49] I can not see one svn reversion someone else made. [23:50] ablmf333: revisions that have happened on the svn trunk since the push of your revisions ? [23:50] jelmer: I can't really tell because I've removed the bzr repo. [23:51] jelmer: but let me check. [23:51] jelmer: maybe I can find it by svn [23:54] jelmer: It's a reversion before I pushed. [23:54] I should have merged it to my bzr branch this morning. [23:54] But somehow it was lost. [23:55] And I can never get it back anymore [23:55] ablmf333: thanks, that helps [23:55] ablmf333: I think I've also reproduced the commit issue [23:56] jelmer: Great, at least my bad luck might help someone else. [23:56] jelmer: Is there anyway to avoid this problem? [23:56] jelmer: Should I avoid using push on svn? [23:57] ablmf333: the commit issue is something I'm looking at fixing now [23:57] fullermd: I know I merge revisions, not files.. But wouldn't it be nice if.. bzr could recognize that the revisions could be merged without problem? And then just pull, merge automatically, and commit... just with bzr pull? [23:57] ablmf333: the second one is already fixed (between 1.0.5dev and 1.1.0dev) and can be avoided by fetching into a clean bzr repository from svn [23:58] jelmer: k, I will try it tomorrow and let u know the result.