[00:43] * igc out for a few hours - bbl [00:49] hi [00:50] i was in here the other day with a truly messy bzr repo [00:50] so i was wondering how could one clean said repo of cruft? [00:51] where cruft = unwanted branches? [00:51] if by branches you mean revisions then yes [00:51] I don't [00:51] there is like a complete git repo and binary files [00:52] it's scary [00:52] and it's some 300mb with anon bzr branch [00:52] but the program is less when 1mb with icons [00:52] *then [00:53] could i take a good revision that is high and simply drop anything below that? [00:55] btw the repo is here: https://code.launchpad.net/gloobus [01:06] zekopeko__: if it's really messy i'd just drop the history and re-import a snapshot of it [01:06] then rename the old one to trunk-ol/d [01:06] trunk-old* [01:06] i was thinking along those lines [01:07] is there some guide or something that could help me? i never used bzr but hate seeing this cruft [01:09] i can look at the branch if you like but basically [01:09] well, [01:09] actually, i should check first - is this your project? [01:10] no [01:10] but i have somebody that has commit access [01:15] ok [01:15] have you talked to somebody on the project about cleaning it upL [01:15] i don't mind helping you [01:16] but it will be smoother if it's coordinated with others [01:16] anyhow, what i eventually suggest you do is [01:16] get a checkout of the branch [01:16] poolie: I'm on the Gloobus Project :) [01:16] that was 1 [01:16] 2- 'rm -r .bzr' to remove the old history [01:16] 3- remove the junk [01:17] 4- 'bzr init; bzr add; bzr ci -m snapshot' to start a new history [01:17] 5- rename the old trunk on launchpad [01:17] 6- push this new branch and make it the trunk [01:17] you got that DanRabbit? :P [01:17] yep [01:18] .bzr directory here is 88 MB [01:18] lol [01:18] does that make sense? [01:18] that it's 88 MB? not to me [01:18] what to do? [01:18] yep [01:21] sweet, thanks poolie! [01:22] yes, thankyou poolie [01:22] happy to help [01:22] hello spiv? [01:22] poolie: hi [02:11] just saying hi really [02:11] going to close mail and start piloting in earnest soon [02:21] poolie: hooray for piloting :) [02:21] poolie: I'm busily replying to the deluge of review comments on the per-file merge patch [02:21] poolie: it's great to get so much thoughtful feedback on a patch from so many people, but also a bit overwhelming. [02:40] poolie: btw, I've been enjoying new progress reporting [02:41] poolie: it was a bit disconcerting at first because my eyes had been trained to expect to see the info about 30 columns further right than it is now [02:41] poolie: but now I've got used to that I've found I don't miss the bar at all === timchen119 is now known as nasloc__ [03:06] poolie: what time would you like beer? [03:23] jelmer: man, wtf? [03:23] (about codeplex) [03:24] mwhudson: ECONTEXT [03:25] lifeless: https://bugs.launchpad.net/bugs/498925 [03:25] Ubuntu bug 498925 in bzr-svn "assert svn_revprops.has_key(properties.PROP_REVISION_DATE) can fail" [Low,Triaged] [03:25] Codeplex has a custom subversion implementation (on top of Visual [03:25] SourceSafe or something?) that is kinda dodgy and inconsistent. [03:25] my sense of order in the world didn't need to know about that [03:27] ah [03:27] just like google then ? [03:30] Just for the record, I want to make it clear that I find it highly amusing that git and bzr can both take an existing subversion repository and push it to a new blank svn branch without needing shell access at either end, and svn itself apparently can't. [03:31] lifeless: Google's svn implementation is dodgy? [03:31] Peng: "non standard" [03:31] OK. :) [03:41] lifeless: say six? [03:41] spiv, that's good [03:51] poolie: ok [03:53] lifeless, Peng: and a bit dodgy [03:56] :D [03:56] Aside from the odd responses to .bzr requests, and downtime, how is it dodgy? [03:57] Peng: large cscvs imports never succeed [03:57] bzr-svn seems to tickle the pain points much less [03:57] Oh. Maybe they subcontracted it to SourceForge. ;-) [03:58] heh [04:22] spiv, would it be appropriate to raise NoSmartMedium if we can't connect to the smart server over http? [04:22] ie we could potentially connect but we can't actually connect at this url [04:23] poolie: Off the top of my head I think so, yes. [04:24] I [04:24] yes, it seems like it [04:33] when I do: [04:33] bzr info bzr+ssh://ClausenStrub@c-st.net/~/econ_bib.bzr [04:34] I getbzr: ERROR: Not a branch: "bzr+ssh://ClausenStrub@c-st.net/~/econ_bib.bzr/". [04:34] what does it mean? [04:35] $ ssh ClausenStrub@c-st.net ls -a "~/econ_bib.bzr" [04:35] . [04:35] .. [04:35] .bzr [04:36] clausen: it means you probably don't have bzr 2.1.0b3 or newer on the server [04:36] clausen: which is needed for /~/ to work in bzr+ssh URLs [04:37] oh [04:37] the client can't ask the server the version? [04:37] clausen: try bzr+ssh://ClausenStrub@c-st.net/home/ClausenStrub/econ_bib.bzr/ (or whatever your home dir is) [04:37] It can, but that's not really the issue. The meaning of '~' is very much dependent on the server. [04:37] It's not something that clients can calculate without help from the server. [04:38] anyway, it's a completely unhelpful message [04:38] it should say "file not found" or something [04:38] but thanks for your help [04:38] I really appreciat eit [04:38] Yeah, that's true. [04:38] we spent hours on it [04:38] "Not a branch" really should have more explanation about why in it. [04:38] yeah [04:38] I'm fairly sure there's a bug about that somewhere. [04:39] Sorry to hear it cost you so much time! [04:40] is there an easy way to intercept the traffic? [04:40] it would have been helpful to see the data on the wire [04:40] with git, I can wrap ssh [04:40] which I've found very helpful in the past [04:40] "bzr -Dhpss" is often helpful [04:40] As in, add -Dhpss to what ever bzr command you're running. [04:40] It logs a bunch of stuff about the smart server conversation to ~/.bzr.log [04:41] I've also used strace occasionally to trap read/write calls, which is how bzr will be communicating with the ssh subprocess [04:42] And you can change the SSH implementation bzr tries with the BZR_SSH environment variable (and there's a config setting too). [04:42] ah, same as git then :) [04:42] (although it's not documented?) [04:42] Which SSH wrapper do you use for git? [04:43] I wrote a shell script that does "echo $*; ssh $*" [04:43] but you could also use vee [04:43] It should be documented somewhere, 'bzr help configuration' for instance mentions BZR_SSH [04:43] bzr(1) doesn' [04:44] bzr help isn't searchable? [04:44] Hmm, that's odd, I wonder how the man page is generated. [04:44] There's 'bzr help topics', but all of that content is in the HTML docs too. [04:45] python in general has poor documentation tools [04:45] (although I could just be ignorant) [04:45] e.g. http://doc.bazaar.canonical.com/bzr.2.0/en/user-reference/index.html#configuration-settings [04:46] R has a similar philosophy, but seems to work better [04:46] (eg: help.search()) [04:46] how did you navigate to that page? [04:47] ah, you clicked on User Reference [04:47] I guess I was expecting that to be many pages, and hence not searchable [04:47] good to know :) [04:48] clausen: there's a pretty functional search box at http://doc.bazaar.canonical.com/bzr.2.0/en/ too [04:48] but it searches the whole site! [04:49] you get the same result for many versions of the same document [04:49] actually, I'm wrong about that [04:49] it's doing something else [04:49] anyway, searching "ssh" didn't find it [04:50] (item number 33, actually) [04:50] I guess the thing that's nice about man pages is you can search for "ENVIRONMENT" [04:50] and you have a complete list [04:51] Yeah, I don't know why the man page has an incomplete list. [04:51] 'bzr help environment' ought to be up to date. [04:51] ok, I'll submit a bug on that then :) [04:51] (And ought to be what the man page uses for that section) [04:51] clausen: thanks :) [04:51] $ bzr help environment [04:51] bzr: ERROR: No help could be found for 'environment'. Please use 'bzr help topics' to obtain a list of topics. [04:52] you mentioned "configuration" before? [04:52] Oh, 'env-variables' is the topic name. [04:52] it includes BZR_SSH [04:52] Yeah, that's a help topic too, but 'bzr help env-variables' is specifically just the environment variables. 'configuration' is a broader topic than just the environment variables, unsurprisingly :) [04:53] is there a URL where i can browse the bzr source? [04:53] (it's ironic that I can't find it!) [04:54] There's a link on code.launchpad.net/bzr [04:54] thanks [05:00] it's hard-coded! [05:00] http://bazaar.launchpad.net/~bzr-pqm/bzr/2.1/annotate/head%3A/bzrlib/doc_generate/autodoc_man.py [05:00] no wonder it's unsynchronized [05:05] the "bzr help env-variables" command is also hard-coded [05:05] in _env_variables, which is defined in bzrlib/help_topics/__init__.py [05:06] I must confess, I'm kind of freaking out about this [05:06] bzr doesn't seem to have the rigour and elegance of a solid system [05:06] (but, I suppose it will get better) [05:09] Well, env-variables is as close to the canonical version as we have. [05:09] that's a bit of a bug in that page [05:09] it should just use the help topic [05:10] Right. [05:10] i'll file it [05:13] https://bugs.edge.launchpad.net/bzr/+bug/498952 [05:13] Ubuntu bug 498952 in bzr "autodoc_man.py should use env-variables, not duplicate it" [Medium,Confirmed] [05:15] $ bzr info blah [05:15] bzr: ERROR: Not a branch: "/home/clausen/free-software/bzr/blah/". [05:15] ^ should say "path not found" [05:15] Yes. [05:16] "Not a branch" is techinically correct, but not as helpful as it could be. [05:16] agreed [05:19] note that a missing parent dir doesn't imply not a branch, it could be a missing plugin (e.g. bzr-svn) [05:19] lifeless, of course, but the error message should explain [05:19] lifeless, i.e. there should be different error messages for different reasons [05:19] clausen: also we search upwards [05:20] I suppose the kosher way to do this would be to subclass errors.NotBranchError [05:20] I agree it could be improved, but I don't think its as shallow as perhaps you think it is [05:20] and have things such as errors.NotBranchError.NoPlugin [05:20] errors.NotBranchError.NoPath [05:20] etc. [05:20] clausen: hmm, I think actually just providing an extra string when instantiating NotBranchError [05:20] a typical branch lookup will catch 5 or 6 NotBranchError [05:20] spiv, could be [05:20] which one is shown to the user will influence the ui [05:20] clausen: (or the exception that lead to it, and have some common code to turn that into a string) [05:20] it doesn't currently. [05:20] spiv, I suppose the question is: would anyone ever need to catch a special type of exceeption? [05:21] spiv, I guess not [05:21] clausen: right [05:21] NoPlugin woul be a prove-a-negative case. [05:21] lifeless, yeah, I noticed that a lot of these exceptions are getting caught [05:22] It's more providing an extra hint to the user, not a definitive diagnosis: e.g. 'Not a branch: "/foo/barr/baz/qux". ("/foo/barr" does not exist)' would most of the time tell you what you wanted to know [05:23] (although actually that particular case would be unlikely to have that much information without doing extra probing) [05:24] (it would certainly have helped me!) [05:25] (eg: I didn't know if I needed to point at .bzr) [05:30] anyway, you've been very helpful [05:30] thanks a lot! [05:30] sleep time! [05:30] bye! [05:31] oh, nobody filed the bug yet... I guess I should before I go to sleep [05:34] ah, it's #52865 [05:34] https://bugs.launchpad.net/bzr/+bug/52865 [05:34] Ubuntu bug 52865 in bzr ""Not a branch" errors could be better when branch is missing." [Wishlist,Confirmed] [06:11] spiv, is https://bugs.edge.launchpad.net/bzr/+bug/410332 like something you fixed recently? maybe a dupe? [06:11] Ubuntu bug 410332 in bzr "accessing a bzr:// branch with a revisionspec causes bzr to crash" [Medium,Confirmed] [06:15] poolie: hmm, maybe [06:15] poolie: I'll try to reproduce quickly [06:23] poolie: ah, yep, still current, but a bit obscure [06:24] poolie: oh, oops, sorry, it is fixed :) [06:27] fail! no, win! [06:27] :) [07:01] jelmer: morning, i opened a ticket for my yesterday's 'bzr dpush to github' problem - https://bugs.launchpad.net/bzr-git/+bug/498962 [07:01] Ubuntu bug 498962 in bzr-git "KeyError(sha)" [Undecided,New] [08:24] Is it possible to remember two different pushes? [08:24] Eg every time I push, it pushes to both my own branch somewhere, and to a lp: branch [08:25] Adys: you can probably use the automirror plugin to do something like that. [08:25] I don't think there's anything builtin though. [08:26] mm kay [08:27] I'll be lazy then :) [10:46] Adys: spiv: bookmarks plugin [10:47] might be close-enough [10:47] Meh, as i said I'll be lazy, was hoping for a quick&dirty command, its a temporary lp branch === salgado is now known as salgado-afk [13:34] hello all [13:35] hello bialix. [13:35] hi rubbs [13:36] how are you today? [13:40] there is cold today [13:41] hope you're better [13:49] It's cold here too. getting snowy as well. [13:49] but I"m pretty good other than that [13:54] that's fine :-) === salgado-afk is now known as salgado [14:33] hi, I'd like to revert undo a revision at some point in the past (10 revisions away from the current one), apparently the command would be: [14:33] bzr . merge -10 [14:33] is that correct? [14:33] (I'm scared to make a mistake...) [14:34] or is it the number of the revision, like [14:34] bzr . merge -3826 [14:34] ? [14:35] An possibly easier way to think about it (more steps but easier to understand) is this: [14:35] bzr branch project -r 3826 ./projectUndo [14:35] then [14:36] bzr merge projectUndo [14:36] I think that's correct... you may want to back things up just in case [14:37] that will say "nothing to do" [14:37] you want "bzr merge . -10..-11" [14:38] merge the changes between the revision and the one before it [14:38] i.e. undo those changes [14:38] "bzr merge . -r -10..-11" of course [14:39] ah, thanks james_w. I haven't had to do it myself. Thanks for the correction [14:43] james_w, the right command for me was "bzr merge . -r -11..-12", thanks! [14:45] i just installed bzr and qt on mac but i cant find out where bazaar explorer and qbzr are installed [14:50] Solipsist: run `bzr plugins` === salgado is now known as salgado-lunch [14:52] now that was totally obvious :) [14:52] bzr: ERROR: [Errno 2] No such file or directory: '/Users/username/.bazaar/explorer' [14:53] Solipsist: what is explorer version? [14:53] I think this bug was fixed [15:00] version is 0.9.0 [15:07] got it running [15:08] a note from a new user: bzr is supposedly easier to use than git, my experience so far has been the opposite [15:10] so im cloning a remote repos as id like to work distributed, pushing changes from my local repos back to the remote is "bzr push", right? local commits "bzr commit"? [15:13] Solipsist: yep [15:15] Solipsist: as a member of doc team, I'm interested in the problems you had with bzr. Maybe I can get our team to address holes in the documentation [15:17] rubbs: the first roadblock i encountered was to figure out how to run bzr explorer, as it was mentioned nowhere, even after extensive googling [15:17] i tried using spotlight in osx to locate it or qbzr, no luck [15:18] Solipsist: I'll be sure to raise that issue up. We have had a few requests on better plugin documentation. [15:18] a very concise guide for those of use new to bzr would be appreciated, i dont need a GUI but after having used git with gitx, ive come to find them extremely useful when looking through commits [15:19] its not the plugin itself but "now that's bzr is installed, what do i do? where did all these applications go?" [15:19] Solipsist: you've mixed package problems with ease of use the tool itself [15:19] the mac os x package needs better documentation [15:19] where explorer is the question to the maintainers of Mac installer [15:20] Solipsist: I'll have the mac people look into making it easier to get started from install. [15:20] windows installer already provides the user with start menu item, desktop icon and quick launch item [15:20] if this page explained what the package contained and how to launch and use each plugin it would have made things easier: http://wiki.bazaar.canonical.com/MacOSXDownloads [15:21] rubbs: awesome! [15:22] The leader of the Doc team is also the main dev for Explorer, so I'll see what I can do to help him understand the problem. He's a good guy and easy to work with. [15:23] My suggestion is to actually raise a bug on launchpad and tag it with "doc." you can give a very brief description of the problem you had and we'll work hard to change that experience === zekopeko_ is now known as zekopeko__ [15:53] hmm, why is bzr giving me such weird error: $ bzr mv 01_notes.php 01_notes.sql [15:53] bzr: ERROR: Could not move 01_notes.php => 01_notes.sql: UPGRADE/patches is not versioned. [15:53] bzr log 01_notes.php shows it's revision log, meaning it is versioned [15:54] glen: not sure what's going on there, but you can as a work around do a system mv first and then use the same command you were getting the error on. [15:55] so do : $ mv 01_notes.php 01_notes.sql [15:55] then do : $ bzr mv 01_notes.php 01_notes.sql [15:56] that will cause bzr to see that 01.notes.php doesn't exist, and there is a new file called 01_notes.sql. It assumes you moved it and updates the versioning accordingly === menesis1 is now known as menesis [15:56] now even worse: [15:56] $ bzr mv 01_notes.php 01_notes.sql [15:56] bzr: ERROR: Could not rename 01_notes.php => 01_notes.sql: Path(s) do not exist: upgrade/patches/01_notes.php UPGRADE/patches/01_notes.sql [15:57] what the heck it capializes the path component named "upgrade" ? [15:58] $ bzr st [15:58] removed: [15:58] upgrade/patches/01_notes.php [15:58] unknown: [15:58] upgrade/patches/01_notes.sql [15:58] says it [15:58] after the last action [15:58] bzr-0:1.18-2.x86_64 is the version [15:58] any possibility of updating to 2.0? [15:59] i cloned: bzr clone lp:eventum, and the mv i want to do in upgrade/patches [16:00] ok, trying to upgrade [16:00] Hmm.. not sure what's going on here. I'm just a user, but the devs check this channel frequently, so just sit tight. someone should be able to help you better === salgado-lunch is now known as salgado [16:00] ok, cool [16:00] I can help more with the 2.0 series. (only a little more mind you). [16:01] any clues is 2.0 having some different dependencies of external packages? [16:02] not sure. [16:02] what type of system are you on? [16:02] linux [16:02] ubuntu? [16:03] pld linux [16:04] 32bit/64bit behave the same, if that's any matter [16:04] i just suspect "upgrade" has some special meaning somewhere [16:05] i.e the mv command works ok elsewhere [16:05] could be [16:05] $ bzr mv xmlrpc_client.php xmlrpc_client.sql [16:05] misc/xmlrpc_client.php => misc/xmlrpc_client.sql [16:05] I'm wondering if someone did something in Windows, and the case insensitivity is causing problems [16:05] can you test with your bzr? [16:05] sure just a sec. [16:06] anyway, 2.0.3 behaves the same way :( [16:07] maybe this is (also) the reason: [16:07] $ l upgrade UPGRADE -d [16:07] drwxr-xr-x 24 glen glen 4.0K 2009-12-21 17:49 upgrade/ [16:07] -rw-r--r-- 1 glen glen 8.1K 2009-12-03 11:49 UPGRADE [16:07] i.e the repository has lowercase and uppercase objects present [16:08] weird... [16:08] clone is taking a while. so I haven't tried it yet. but it's getting there === weigon__ is now known as weigon [16:10] anyway, getting rid of the uppercased file "UPGRADE" in root "solved" the problem here [16:11] ok cool. I'm still going to test it to see what's going on [16:13] it's solved for me, likely bugreport should be filed, but i'm lazy this time so won't go trought the hassle of registering accounts for bugreporting etc [16:14] I'll go ahead and file one for you. I can reproduce the result [16:15] nice [16:15] if need some reference, then my commit is: http://bazaar.launchpad.net/~eventum-developers/eventum/trunk/revision/4012 [16:16] thanks. I might put that in there. [16:30] glen: finally got around to submitting that bug and found that someone already has put one up: https://bugs.launchpad.net/bzr/+bug/368931 [16:30] Ubuntu bug 368931 in bzr "Rename may fail when file and directory have the same name differing by case" [High,Confirmed] [16:33] i see [16:40] can I put multiple bug numbers in --fixes? [16:41] you can pass --fixes as many times as you like [16:41] thx [16:41] commit --fixes foo:1234 --fixes foo:2345 --fixes bar:4567 === deryck is now known as deryck[lunch] [17:33] Can I reassign existing revisions to a new person? I've changed my whoami to match launchpad... [17:33] I was using chm@buna as my email, which is not valid, so I can't validate that address on launchpad to link those revisions to me [17:46] chmac: Not without redoing the commits [17:47] jelmer_: Ok, thanks === beuno is now known as beuno-lunch === deryck[lunch] is now known as deryck [19:08] yeah, I had the same issue as chmac [19:09] Tak: Did you find a reasonable solution? It's only 10 or so commits that are not attributed to me on launchpad, so it really isn't a big issue for my case. [19:09] no, I never did === beuno-lunch is now known as beuno [19:55] jam: ping from iowa city [19:55] hey phinze [19:55] keeping warm ? [19:56] our radiators ensure that we're always either sweating or freezing, yes :) [19:56] i was wondering if you had a moment to walk me through a particularly tricky (but relatively small) bzr dance [19:57] Iowa City? I'm originally from Decorah. [19:58] rubbs: i'm a transplant, but i bet there are several in my office who would enjoy that... will pass along [19:58] jam: msging you [20:10] Is there a faster way to download a branch than just straight `bzr branch lp:wordpress/trunk`? It's taking an age to download all the history. [20:10] I considered a lightweight checkout, but I'm going to commit back to a different branch, so I don't think that would work. [20:11] I guess it's the price I have to pay the first time to get all the history... :-) [20:11] chmac: if you haven't downloaded it before, then not really... You could try bzr+ssh:// but it requires you to have an account and to have set up keys [20:11] chmac, if you have your launchpad-login specified you will use the [20:11] "smart server" [20:11] which should be faster [20:12] I've run launchpad-login already, so it might be using the smart server already... [20:12] chmac: you may also be running into auto-conversion issues. [20:12] I'm online in Mexico City, it's not the fastest connection ever... :-) [20:12] It looks like wordpress is in --pack-0.92 [20:12] if you have 2a locally, it will convert on the fly [20:12] which could be slow [20:13] jam: I think my bottleneck is the download speed, I'm seeing 20-40kbs... [20:13] it is a ~50-60MB download [20:13] Judging by http://bazaar.launchpad.net/~vcs-imports/wordpress/trunk/.bzr/repository/packs/ [20:13] I would expect it to be smaller in --2a, if/when it gets converted [20:14] jam, I though we where going to convert all imports to 2a? [20:14] mwhudson, did that ever happen? & [20:14] ^ [20:14] beuno: no, not yet [20:15] mwhudson, is it still on the roadmap? [20:16] beuno: yes, but in a pretty sparsely populated area of the map i guess [20:16] heh, ok [20:16] beuno: i just added it to https://dev.launchpad.net/CodeTeam#Code%20imports fwiw [20:17] mwhudson, thanks [20:17] I'm having trouble breaking a lock, I tried the provided command and a few variations with no luck (http://amro.pastey.net/130496) [20:18] amro, try: bzr break-lock bzr+ssh://sigrie/home/sigrie/projects/xflight/ [20:20] beuno: that worked, thanks. I thought I tried that [20:21] what's up with the line provided by bzr? usually it's correct === chmac is now known as chmac_away [20:21] amro, there are some cases where it doesn't work well, I fail to remember in which ones :) [20:21] what version of zbr are you using? [20:21] and both server and client [20:22] 2.0.0, repo is 2.0a I think [20:22] something like that [20:22] on both sides? === salgado is now known as salgado-afk [20:25] Repo is 2.0a yeha [20:25] chmac_away, beuno, mwhudson: wordpress does shrink to 20MB after 2a packing... [20:26] that said, how do i upgrade from branch format 6 to branch format 7? [20:32] Adys: 'bzr upgrade' [20:32] or 'bzr upgrade branch' [20:32] as fmt 7 is the latest format [20:39] cheers [20:40] I have a repository which after 400 revisions had a huge amount of unneeded binary data committed into by accident. Another 20 or so commits have happened over that last few months before anyone noticed. What's the best way to excise the binary files without losing our rich revision history? [20:41] halstead|wccls, you would have to rewrite history [20:42] there's a plugin calles bzr-rewrite (IIRC) [20:42] *called [20:42] not sure how to use it [20:42] and, you need to know that you will break compatibility with all existing branches [20:42] I'll search on those terms. Thats for the lead. I thought maybe I could start a new branch and pull the first 400 revisions and then layer ont he later changes somehow without pull everything in. [20:43] Thank you for the lead. [20:44] halstead|wccls, you can also branch from the last good revision [20:44] Braking compatibility with existing branches is not a problem for us. All our contributers can make a fresh branch. [20:45] and merge in the good revisions [20:45] will make history look wonky [20:46] beuno, I will try that. [20:46] beuno: (lurking for learning) would there be a way to "replay" the revisions rather than merging them all with one revision? [20:46] a la 'bzr pull' would tack them onto the end? [20:46] phinze, well, I'm sure there is a way, but I don't know how I would do it [20:47] jam, is the expert on rewriting history :) [20:47] indeed, he saved the day for me once again today [20:49] i believe bzr-rebase would be the tool that could do it, but how to use it, i know not [20:50] so bzr branch -r$REVISION_BEFORE_BINARIES trunk trunk_sans_bin [20:50] cd trunk_sans_bin; bzr merge ../trunk -r$REVISION_AFTER_BINARIES.. [20:50] bzr rebase --pending-merges [20:50] i _think_ that might do it [20:51] --pending-merges "Rebase pending merges onto local branch." [20:51] that sounds about right, yes [20:51] Thank you. [20:52] halstead|wccls: np -- a bit of the blind leading the blind :) but i hope it works for you [20:53] halstead|wccls, http://doc.bazaar.canonical.com/plugins/en/rebase-plugin.html [21:10] hello jam === poolie changed the topic of #bzr to: Bazaar version control | try https://answers.launchpad.net/bzr for more help | http://bazaar-vcs.org | http://irclogs.ubuntu.com/ | Patch pilot: mbp | bzr 2.1.0b4 and 2.0.3 released [21:12] hey poolie [21:12] beuno, phinze: replay is probably what you want, but jelmer is the expert [21:13] shall we talk? [21:13] phinze, What effect is the double dot at the end having? bzr merge ../trunk -r$REVISION_AFTER_BINARIES.. [21:13] sure [21:14] skype/pots? [21:14] pots is always easier for me, but i've got skype, too [21:14] halstead|wccls: it means "$REVISION_AFTER_BINARIES and every revision thereafter" [21:15] k [21:20] jam: strace yes|sleep 60 === chmac_away is now known as chmac [21:49] jam: Not sure what conversion to 2a means. From what you said I'm thinking that it'll happen automatically at some point... :-) [21:49] Plus, the new branch I request 2.9 will have a lot less history than trunk, so that'll be quicker [21:52] chmac: 2a is the current repository format, it became the default in bzr 2.0. "bzr upgrade" will upgrade an existing repo to it. [21:53] spiv: Ok, so I could upgrade repos I've created myself. lp:wordpress/2.9 is an auto-imported repo so I'm guessing I can't modify it any more, or can i? [21:53] https://code.launchpad.net/~chmac/wordpress/2.9 [21:56] If I rename the repo, or assign it to a team instead of me, will the import process continue to work? [21:56] I noticed some other auto imported branches are ~vsimports/blah [21:57] chmac: I think you can file a request (at https://answers.launchpad.net/launchpad) to get the import upgraded\ [21:57] I think it makes sense for the branch to be owned by a generic user rather than me, it's the wordpress/branches/2.9 release rather than my own version of it, my own version will be based off it [21:57] chmac: ^ mwhudson [21:57] I love the responsiveness of people on this channel, launchpad has a very human feel to it... :-) === khmarbaise_ is now known as khmarbaise [22:06] Oh, I'm on the wrong channel, DOH! [22:15] Sweet, just updated bzr thanks to https://launchpad.net/~bzr/+archive/ppa [22:18] spiv, chmac: hmm? [22:37] mwhudson: see #launchpad, looks like intellectronica has taken care of things [22:38] spiv: ah, ok [22:38] Oh, yeah, sorry, I didn't follow up on that, intellectronica did indeed take care of it for me :-) [22:39] I ran `bzr branch lp:~vcs-imports/wordpress/2.9 wp29` then `bzr rm wp29/wp-includes/js/tinymce/` then `bzr commit -m "blah"`, it gave me the next available revision number [22:39] That might get confusing when a new commit comes into 2.9 as the numbers might be the same [22:40] Is that the way things are supposed to work, my version number will only apply to my branch, or is there an alternative I'm missing? [22:40] Can I specify my own version numbers as rchmac-1 or something? [22:44] chmac: revision *numbers* are a branch local number [22:44] jam: Ok, so I can ignore it, the numbers might be the same but that doesn't matter, is that correct? [22:47] chmac: right [22:47] jam: Ok, thanks :-) [22:47] numbers are a useful shortcut in the context of a branch [22:48] so you can say "did you get rev 10 of my branch", and someone can see that they are only at 8 [22:48] etc [22:48] internally we have unique revision-ids [22:48] but they don't have obvious ordinal semantics [22:48] (bzr log --show-ids if you want the details) [22:48] Ok, they're specific to the branch. I've seen the long gnarly revision ids, I'm guessing they're the globally unique id [22:48] yep [22:50] Can I safely `cp -r branch1 branch2` and then work on them separately? [22:51] I tried `bzr branch branch1 branch2`, then branch2 shows its parent as branch1, which is accurate. I really want to branch from lp but safe myself downloading it all again... :-) [22:51] chmac: create a shared repository [22:51] bzr init-repo . [22:51] bzr reconfigured --use-shared branch1 [22:51] then all branches created underneath the repo [22:51] will share the storage [22:51] and you won't redownload everything [22:52] sorry' [22:52] 'bzr reconfigure ...' [22:52] jam: Ok, I'll read up on those, sounds like what I want, thanks [23:14] I had local commits, which apparently got reverted when I wanted to revert the conflicts from "bzr up" (which I've thought would have been a "bzr merge"). Any chance to get those internal merge back from the internals of bzr? [23:15] bzr heads [23:15] bzr push --stacked-on=lp:~vcs-imports/wordpress/2.9 bzr+ssh://chmac@bazaar.launchpad.net/~wpflavours/wordpress/wpflavour-no-visual-editor [23:15] bzr: ERROR: Not a branch: "/home/chm/W/Code/wpflavours/wpflavour-no-visual-editor/lp:~vcs-imports/wordpress/2.9/". [23:15] bound branches can be awkward [23:16] especially if you forget to commit before bzr up [23:16] Am I overcomplicating maybe? Can I just `bzr push lp:~wpflavours/wordpress/wpflavour-no-visual-editor` and bzr will know that the code came from lp:~vcs-imports/wordpress/2.9 originally? [23:17] chmac: I think you need to give the full bzr+ssh URL instead of lp: to --stacked-on [23:17] spiv: Oh crap, I expanded the lp: in the wrong part of the command, DOH! [23:18] bob2: thanks so far.. have found a revision-id.. how can I apply this back now? [23:18] spiv: Thanks, I'll try that, that's what I meant to do [23:18] blueyed: first, back up everything so you don't lose what you have [23:18] Took me long enough to figure out what lp: expanded to, then I put it in the wrong place! [23:20] blueyed: then 'bzr branch -r revid:afdsdfasdfggsfg@rftwertert-243241234 . ../recovered' [23:22] spiv: Looks like it's working, thanks :-) [23:23] bob2: 'bzr merge . -r revid:revid:afdsdfasdfggsfg@rftwertert-243241234' is easier [23:24] hm [23:26] this is f*cking b0rked.. I'm getting blogs.moved.moved etc.. also when trying to merge the dead parent. [23:26] will have to resolve this by hand I fear. [23:27] this probably conflicts so much again, since I'm merging into a branch, where I had local changes made (and cherrypicked from), but the parent's ancestor had been changed. [23:29] Can I forbid bzr-svn to kick in when I only want to add a symlink (which points at at directory with .svn in it)? [23:30] blueyed: to add a symlink just run 'bzr add', not 'bzr add symlink' [23:30] lifeless: but I do not want to add all that .moved stuff.. [23:31] blueyed: so clean it up first [23:31] ok, but the workaround does work.. will do add "add $DIR" then revert. [23:40] blueyed: you can pass --no-recurse to add