/srv/irclogs.ubuntu.com/2008/07/12/#bzr.txt

Pengbeuno: bzr-search_integration + bzr.dev are working. :)00:00
beunoPeng, yay!00:00
beunonow, really off to bed00:01
beunog'night00:01
Pengbeuno: (Well, trunk + bzr-search_integration + my serve-branches.py changes.)00:01
Pengbeuno: G'night. :)00:01
james_w[cliff]: you can merge it with "bzr merge /path/to/bundle", is that what you did?00:02
LaserJockPeng: hmm, are there any suggestions for good ways to keep branches out of the way?00:02
[cliff]james_w, not an absolute path, no. let me try that00:02
LaserJockbzr tends to make me not want to branch much00:02
james_w[cliff]: relative will work as well.00:02
[cliff]james_w, right... same result00:03
PengLaserJock: No, I'm just crushed under a pile of separate branches. Sorry.00:03
* Peng wanders off.00:03
[cliff]the NotBranchError doesn't appear to be used a lot. I was trying to figure out if there is a way of tracing it00:03
james_w[cliff]: see "bzr help debug-flags" for one thing that can help. There's also plenty of information in ~/.bzr.log00:04
james_wLaserJock: you can simulate it with one checkout and switch00:04
james_wLaserJock: there is loom, which is like quilt00:04
[cliff]james_w, says no help could be found for debug-flags... maybe I'm missing a package00:04
LaserJockjames_w: hmm, I'll have a look at those00:05
james_wLaserJock: and it's possible that a plugin for multiple branches in a dir will turn up next week00:05
LaserJockI just feel like I'm losing the power of a DVCS if I'm not doing any branching00:05
james_w[cliff]: sounds like you have an old bzr, which would explain the error. What version are you using?00:05
[cliff]but .bzr.log does the trick! nice one. checking it out00:05
[cliff]james_w, 1.3.1, ubuntu's standard I think00:06
james_w[cliff]: I'd check with --version00:06
[cliff]james_w, right, 1.3.100:06
james_w[cliff]: ah, sorry, "bzr help global-options" has a debugging flags section.00:07
[cliff]just out of curiosity, how do you manage to output those nice traces?00:07
james_wthe traceback module00:07
[cliff]nice one, cheers00:07
[cliff]it appears the error is raised on "/usr/lib/python2.5/site-packages/bzrlib/bzrdir.py", line 1426, in find_format00:08
james_wtraceback.format_exc() will print the last one, there are other functions for being more specific00:08
james_w[cliff]: can you pastebin the relevant stanza from ~/.bzr.log please, I'll see if I can spot anything00:09
[cliff]certainly00:09
[cliff]james_w, http://pastebin.com/d65b306ce00:10
james_wah, I meant the whole section, but the traceback is a start, let me look00:10
james_whmm, it does know that it is a bundle apparently00:12
[cliff]I've updated the pastebin00:13
james_w[cliff]: ah, got it I think, can you read the top of the bundle and see what submit_branch is?00:16
[cliff]none, I only have a target_branch00:16
james_wis that a URL?00:17
[cliff]nope, a relative path00:17
[cliff]that doesn't exist on my filesystem00:17
james_wah yeah, target branch00:17
[cliff]don't know if it's relevant but I managed to successfully merge another bundle on a different branch00:17
james_wI was reading it wrong00:17
james_wyeah, the bundle has been created wrong00:18
james_wif you could ask the person that sent it to tell you how they created it it would be useful00:18
james_wthis should probably be caught00:18
[cliff]how should we do it? I'm going to smack the other guy on the head hehe00:19
[cliff]I can ask him, but not now, he's off for the weekend00:19
james_wso, a merge directive has a submit branch, which is the branch that you intend it to be merged in to.00:19
[cliff]I'll try to reach him still00:19
james_wah, I see something else as well00:20
james_wyou can specify this as the first argument to "bzr send"00:21
[cliff]james_w, hmmmm although I understand (or at least I think so :-) ) how come the target_branch can't be ignored when merging the bundle to the current workbranch?00:21
james_wif you don't then it uses the branch's parent00:21
[cliff]so for example the submit branch should be .00:22
[cliff]?00:22
james_wit stores the branch location in the merge directive as you saw, but if the target branch has a "public_url" set then that will stored instead.00:22
james_wrunning "bzr send" creates a bundle (the base64 encoded bit at the bottom) that has all of the revisions that are in your current branch, and not in the submit branch.00:23
james_wso "." is pretty much always wrong, as this will be the empty set, which isn't what you wnat00:23
james_whe should have a mirror of your branch (or just use the location of it if it is public) and point to that00:24
james_wnow, I see this in the code:00:24
james_w                except errors.RevisionNotPresent:00:24
james_w                    # At least one dependency isn't present.  Try installing00:24
james_w                    # missing revisions from the submit branch00:24
james_w                    submit_branch = _mod_branch.Branch.open(self.target_branch)00:24
james_wwhich is where it fails00:25
[cliff]I see. let me shed some light on this matter. the way we're working is like this: a master-branch is stored on a server and all devs have a local branch off it. when someone wants to work on something, they branch their copy of the master-branch into a new branch00:25
[cliff]so the way the dev created the bundle was by bundling the differences between his work branch and his local copy of the master-branch00:25
[cliff]so the target_branch reflects his own fs directory layout00:25
james_wyep, that should work correctly.00:26
james_wyou should set "public_url" on the local copy to be the url of the master branch.00:26
james_wis your copy of the branch out of date compared to master?00:26
[cliff]hehe which one? :-) the local or the remote?00:26
[cliff]yup00:26
james_wok, so what should happen here is that it would fetch these missing revisions from the master branch, but as the location of that was set to a local path in the merge directive it cant get to it00:27
[cliff]hhhhmmmmmm00:28
james_wI'm working up an example00:29
[cliff]got the guy out of the bed :-) I've asked his revno of the master00:29
james_wheh00:30
james_wunfortunately I can't find documentation on public_url00:30
james_wso, if you have http://server/master00:31
james_wand a local ~/dev/master00:31
james_wand your feature branch to submit ~/dev/feature00:31
[cliff]james_w, that's it. my master is at a different revno of this master00:31
[cliff]james_w, that's it. my master is at a different revno of his master00:31
james_wand you set "public_url = http://server/master" in ~/.bazaar/locations.conf for ~/dev/master00:32
james_wthen "bzr send ~/dev/master" should do the right thing from ~/dev/feature00:33
james_wand if the parent of the feature branch is master then "bzr send" should work00:33
james_w[cliff]: does that match your set up?00:34
[cliff]we don't set the public_url actually00:34
james_wyeah, that's why I think this fell over.00:34
james_wI'm going to see if I can write a test for this so that we can provide a better error00:35
[cliff]hmmmm maybe insert a note in the documentation, that should be sufficient (and easier to do I think)00:37
[cliff]I reckon the bundle parts of the documentation could use some love00:37
james_wyeah, it does seem to confuse people sometimes.00:38
james_wit works really well when it works00:38
james_wWe can spot this problem quite easily I think, so providing an informative error would be good, even with improved documentation00:38
[cliff]true00:39
james_wI think I see a similar problem just below as well, so we can catch that as well.00:39
[cliff]he sent me a proper bundle now (branched off the same revno I have and created the bundle based on that). I'm checking out, it's close to 1am :-) thanks a lot for your help james_w00:41
chandlerc[g]jelmer: you about? i'm hitting weird errors with bzr-svn01:03
chandlerc[g]jelmer: http://pastebin.com/d1e4f1b4201:08
chandlerc[g]happy to help any way I can w/ debugging.01:08
james_wchandlerc[g]: https://bugs.edge.launchpad.net/bzr/+bug/24663501:16
ubottuLaunchpad bug 246635 in bzr "svn+ syntax may be deprecated but it is still essential" [Undecided,New]01:16
chandlerc[g]james_w: yep, that explains the first one. thanks, i'll add myself to the CC of that one01:17
chandlerc[g]still no clue about the second one...01:19
chandlerc[g]its just a segfault, w/o any traceback or anything01:19
chandlerc[g]maybe i'll try and to a gdb on it01:19
chandlerc[g]jelmer: I did some gdb poking: http://pastebin.com/d6e9e66cf  I'll rebuild ra.so with debugging symbols tonight, and let you know what I find.01:24
=== Pieter is now known as Pieter|Paris
lifelessjelmer: Bug 24778706:59
ubottuLaunchpad bug 247787 in bzr-svn "assert len(tview) == tview_len" [Undecided,New] https://launchpad.net/bugs/24778706:59
* ToyKeeper hopes his vis changes get approved and merged10:32
beunoToyKeeper, approved  :)10:40
beunonow you just need jelmer to pitch in10:40
ToyKeeperI'm just sending a summary + merge directive, to make things clearer.10:41
ToyKeeperI'm not too familiar with bundle buggy.  Does it take the first merge directive in a thread, or the last?10:42
beunoToyKeeper, always the latest one10:42
beunoeach time you send a new one, it overrides the previous one10:43
ToyKeeperOkay, good.10:43
beunoToyKeeper, http://bundlebuggy.vernstok.nl/bzr-gtk/request/%3C20080712094212.GA15283@mutt.xyzz.org%3E10:43
ToyKeeperI had the wrong public branch the first time too, thinking it was supposed to be my public branch, not trunk.10:44
ToyKeeperJasper's message in the olive bookmarks thread cleared that up for me.10:44
beunoToyKeeper, patch looks great, really, thanks for working on it10:45
beunoah, we have 2 bundlebuggys now10:46
beunojelmer, which are we using?10:46
ToyKeeperI feel dirty about building labels in an unrolled loop, but didn't see an obviously better approach.10:53
ToyKeeperAnyway, vis, gci, and mdiff (diff with meld) are the only GUI bits I really use.10:54
ToyKeeperHmm, maybe gblame, once in a while.10:54
* beuno is trying to pack before check-out time10:56
apolwhere can I find the bzr python api?10:58
apoli've been looking for it for a while and I can't find it10:58
beunoapol, you mean, documentation for it?11:00
beunohttp://starship.python.net/crew/mwh/bzrlibapi/11:00
apolbeuno: yes, yesterday i found a page that showed how to add files, remove files and so using python11:00
beunoapol, http://doc.bazaar-vcs.org/latest/developers/integration.html11:01
apol(i'm a the kdev+bzr test thingie)11:01
apolyes that's what I meant11:01
apolthanks a lot11:01
beunoapol, ah, Aleix, yes, hi!11:01
apol;)11:02
beunohow's that going?11:02
apolI should be able to have something working today11:02
beunooh, yay!11:02
beunoplease send to the list11:02
beunoeverybody loves screenshots11:02
apolstill working on the multilanguage plugin integration though11:02
apolbeuno: I'll consider that ;)11:03
beunoapol, and, also, bzr-gtk may be a good place to find out how to get some things from bzr11:05
apolbeuno: nice, I'll get it then...11:06
beunoalright, I'm off to a 16 hour flight11:08
beunoapol, I look forward to your screenshots  ;)11:09
Stavroshello11:34
Stavroshow can i have bzr run pylint on files and aborting the commit if the score is too low?11:35
awilkinsPre-commit hook?11:36
Stavrosawilkins: this is interesting11:36
Stavrosis there anything ready-made?11:36
awilkinsI don't know. You know, the docs don't even make it clear if hooks can abort steps.11:40
awilkinsBut you can probably get them to stop by throwing an error...11:40
james_wyeah, pre-commit can abort the commit I believe11:41
Stavrosah, good, thanks11:43
james_whttps://bugs.edge.launchpad.net/bzr/+bug/24784711:48
ubottuLaunchpad bug 247847 in bzr "hooks documentation doesn't clearly state which hooks can modify or abort the operation" [Undecided,New]11:48
james_wjust filed that for you11:48
lifelessjames_w: read more closely please; you're describing bugs incorrectly13:17
james_wsorry, which ones?13:18
lifeless244310 was wrong13:18
lifelessI'm checking the other recent updates now13:18
lifelessseparately, I'm not sure the platform is relevant for many of the bugs13:19
james_woh yeah, I don't know why I put that13:19
lifelessno prob13:19
james_wI'm not sure it is for all of them either13:20
james_wI just got fed up of hunting through each time to dupe them.13:20
james_wI can pull it out of the ones that probably aren't if you like13:21
lifelesswell13:21
lifelessnot really worth the time I guess13:21
lifelessjust note it for future :)13:21
james_wsure13:22
james_wp.s. I have an archive import running at the moment, about 14 hours on main and it's up to kdebase13:22
james_wdebian+ubuntu13:22
lifelesscool13:23
lifelesswe'll want that to play with Monday :)13:23
james_wyep, Colin asked me to make sure we had some branches for others to play with next week. There's a bunch of failures, but I think two things account for 90% of them.13:24
james_ware you still in Istanbul?13:24
lifelessyup13:24
lifelesstill sunday 4pm13:25
lifeless(as in 4pm I hit london)13:25
james_whow was GUADEC?13:25
lifelesspretty good13:25
lifelesshave you seen the playground?13:25
lifelessbzr-playground.gnome.org13:25
james_wno, I haven't looked yet13:25
lifelessuhm battery; bbiab13:25
lifelessback13:49
lifelesshi markh, how went the sprint ?13:52
markhlifeless: not too bad - I'm still here - but I'm all alone in the bzr world13:52
lifeless:)13:52
lifelessjames_w: do look at the playground13:52
markha couple of people have promised to send a couple of patches - hopefully they will turn into contributors13:52
lifelessawesome13:53
markhhow is the middle-east (whereever it is you are!)?13:53
markhsuccessful?13:53
lifelessistanbul :)13:53
lifelesspretty intense13:53
lifelesslots of passionate users of both bzr and git13:53
markhwith the gnome guys, right?13:53
lifelessyes13:53
james_wlifeless: it doesn't seem to work too well for me. I just get the project lists, but they don't seem to go anywhere for the people ones, and the trunk ones give a proxy error13:54
markhyou winning their hearts and minds? ;)13:54
lifelessjames_w: the people ones have not created branches of their own yet13:54
lifelessjames_w: the proxy error - needs looking at; its LH again:(. but just refresh and it should come good13:54
mark1stooopid wireless networks...14:02
lifeless:)14:02
=== mark1 is now known as markh
awilkinsWoo, lots of bzr-svn branches14:22
apolwhy doesnt WorkingTree.has_filename() work? it is always returning True to me... :S15:42
lifelessapol: really?15:42
apollifeless: yes, it is the first time i use it so maybe I'm doing something wrong... :S15:43
lifelessapol: it only tells you if the _path is versioned_, not if its on disk15:43
lifelessapol: er, hang on - was looking at wrong code15:43
lifelessapol: it does a lexists15:43
apoli just want to know if a path is on a working tree15:44
lifelessapol: in what manner, do you mean 'is versioned', or 'is present on disk', or 'is versioned and present on disk'15:44
apollifeless: is versioned15:45
apollifeless: i'm working on the kdev4 support15:45
lifelessapol: for that, use path2id15:45
apollifeless: i just want to know if the user still has to add it or he can work with it (eg check the log or commit)15:45
lifelessapol: tree.path2id('README.txt') != None15:46
apolcool15:46
apollet me check15:46
apolthanks a lot lifeless15:47
lifelessno problem :)15:49
lovebug356Is it possible to merge with another branch but only with one command? ( like git merge --squash)15:51
lovebug356s/command/commit15:51
lifelesslovebug356: yes15:51
lifelesslovebug356: merge FOO && bzr revert --forget-merges && bzr commit -m 'squashed!'15:52
andrea-bshow about a cherrypick?15:52
lovebug356lifeless, thanks, will try that.15:52
lifelessandrea-bs: bzr merge -r x..y FOO && bzr commit -m 'This is a cherrypick'15:53
andrea-bsoh, thanks lifeless15:53
lifelessnp15:54
lovebug356And does somebody know how to work with a git branch? tried bzr-git but that crashed15:54
lifelesslovebug356: your best bet to day is git-fast-export ++ bzr fast-import15:54
lifelesshttps://launchpad.net/bzr-fastimport15:55
lovebug356lifeless, thanks, but can you give a real-life example of that? the recipe is not clear to me15:56
lifelesslovebug356: I can't, but I am quite sure there are details on the list :).15:57
lifelesshttp://bazaar-vcs.org/BzrFastImport15:57
lifelessthat wiki page has a howto15:57
lifelessbzr init-repo .15:58
lifelessfront-end | bzr fast-import -15:58
lifelesswhere front-end is the exporter from git15:58
=== oleavr_ is now known as oleavr
lovebug356lifeless, ok, thanks it works16:10
lifelesslovebug356: cool16:11
lifelessI'm going offline for a while; hope someone else can help on an further questions :)16:11
lovebug356lifeless, ;-)16:13
antoranzHi guys! I had problem with 1.5 with filenames with non-ascii characters17:04
antoranzA fix has been issued for this... but I found another kind of problem related to non-ascii characters in 1.6b217:04
antoranzI posted a report in the same bug in launchpad... should I file a new bug for it?17:05
antoranztake a look: https://bugs.launchpad.net/bzr/+bug/135320/comments/2717:05
ubottuLaunchpad bug 135320 in bzr "dirstate updating fails if there are symlinks and non-ascii filenames" [Medium,Fix committed]17:05
antoranzyes... that's it17:05
antoranzso? File a new bug or leave it like that?17:25
james_wantoranz: yeah, yours seems to be a different issue, so different bug report please.17:43
antoranzk17:44
james_wthough is it https://bugs.edge.launchpad.net/bzr/+bug/244360 ?17:44
ubottuLaunchpad bug 244360 in bzr "UnicodeError running "bzr st"" [Undecided,New]17:44
james_wit looks like it to me, so you could just subscribe to that one17:45
Necorohmm ... I need to run bzr under windows ... but using the cmdline seems not to work... or am I missing something?17:53
enobrevNecoro, have to make sure bzr is in your path17:57
Necorohow do I do this? have never really used windows17:57
enobrevNecoro, on my system, I had to add "C:\Program Files\Python25\Scripts" to my system path17:57
enobrevRight click on "My Computer", go to Properties, Advanced Tab, Environment Variables (button), Look on hte bottom where it says "System Variables", Click on "PATH", then "Edit", and add a semi-colon and the path to your Python Scripts folder to the end of that line.17:59
Necorook - worked... thanks :)18:00
enobrevsure :)18:00
apolcould somebody tell me why this doesn't work? http://rafb.net/p/bNXtm645.html18:10
apolor how to add something to a workingtree... 8-)18:10
antoranzI just added my stuff to that bug.18:11
foobydooWhat's the appropriate URL to file a bug-report for bzr-svn clobbering commits if multiple users are editing their commit messages simultaneously?18:16
luksapol: smart_add takes a list of paths18:18
lukswhat you did is basically wt.smart_add("/home/kde-devel/testbzr/aaa")18:18
lukser18:18
lukswt.smart_add(["/", "h", ...])18:19
apolaaah18:19
apolluks: thanks a lot18:20
matthewlmccluredoes bazaar have a "copy file" command?19:20
* Foskasse http://failblog.files.wordpress.com/2008/06/fail-duck-writing1.jpg LOL!19:24
=== abadger1999 is now known as abadger_afk
apol_how can I initialize a directory from python?19:56
james_wmatthewlmcclure: no, it doesn't20:03
james_wapol_: checkout "bzrlib/builtins.py" for "class cmd_init"20:03
james_wapol_: that handles a few cases, so you may be able to strip it down depending on what you need20:04
apol_james_w: cool, thanks20:04
james_wit's not a simple one-liner though, as you have to deal with formats20:04
apol_really? :/20:04
apol_well, I'll check it out20:04
apol_thanks a lot20:04
james_wit's not one line, but it is about three I think, if don't need to handle all cases of existing dir, no parent dirs, etc.20:05
matthewlmcclurejames_w: thanks20:34
libwilliamI have a question... With bzr send... Is the --message option only used with the email option of send?21:43
libwilliamNevermind, I guess it works with the --output also... I thought it didn't but I wanted to ask first, but the test I did that made me think I didn't was wrong.21:46
libwilliamSo ignore these message :)21:46
libwilliammessages*21:46

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