[00:09] how to write a multi-line commit comment in command line? [00:11] RenatoSilva: That's a question for your shell, nothing bzr-specific. [00:13] It works automagically in bash. You can just hit Enter while writing a string. But I dunno about any other shells. [00:13] if you do not trust that, you can use a file for the message: bzr commit -F msg.txt [00:13] ok [00:14] but actually I would prefer using a real editor [00:15] * RenatoSilva trying to find how to do that in windows [00:19] cygwin? [00:20] I use MinGW [00:42] morning [00:43] hi mneptok! [00:53] hey guys. i'm having a hard time pulling a new reversion: [mlimonciello@devel mythbuntu-weekly-build]$ bzr pull [00:53] Using saved parent location: http://bazaar.launchpad.net/~mythbuntu/mythbuntu/mythbuntu-weekly-build/ [00:53] bzr: ERROR: Cannot lock LockDir(http://bazaar.launchpad.net/~mythbuntu/mythbuntu/mythbuntu-weekly-build/.bzr/branch/lock): Transport operation not possible: http does not support mkdir() [00:53] any suggestions? [00:54] igc: hi [00:54] superm1: What version of bzr? [00:55] Peng, should be a freshly installed 2.0.0 [00:55] superm1: Just for fun, try "bzr --no-plugins pull". [00:55] Peng, same thing [00:55] superm1: Did you 'bzr co' from an HTTP URL? It sounds like you have a branch bound to an HTTP URL. [00:55] (in which case 'bzr up' is what you want, not 'bzr pull') [00:56] igc: good news! I finally got a working installer for PyQt+sip, but only for OS X 10.5...I need to work on the 10.4 version [00:56] Ooooh. [00:56] wgrant, well this c/o was from ages ago. since then the branch got upgraded to 2a, and i had to upgrade bzr, and then it stopped working [00:56] * Peng_ goes /away [00:56] superm1: It shouldn't be related to the upgrade. [00:56] wgrant, bzr up does the same thing :( [00:56] superm1: pastebin 'bzr info -v' [00:57] wgrant, it looks like it might have been something related to being bound to HTTP though. i just did bzr unbind, and it's doing more stuffs [00:57] yup, that worked swimmingly! thanks! [00:58] Strange. [00:59] hello all [00:59] hi igc, are you at work today? [00:59] hi poolie [01:00] Hi igc. [01:00] hi verterok. Very well done. Even getting the Leopard one out will be huge step forward! [01:00] hi poolie. Are you back in Oz today? [01:00] hi garyvdm! [01:00] hi, i am [01:00] i'm just cleaning up my receipts and office atm [01:00] poolie: I'm well thanks [01:01] glad to hear it [01:01] poolie: just arrive back this morning? Must be tired I'd imagine [01:01] yesterday morning [01:01] slept a lot, ok now [01:01] for the moment :) [01:02] :-) [01:03] superm1: This is a guess... Maybe you have the webdav plugin, but that location does not support webdav. [01:04] superm1: If you have a launchpad account, with a ssh key set up, then set bzr launchpad-login, and run bzr pull lp:~mythbuntu/mythbuntu/mythbuntu-weekly-build/. [01:05] poolie: a new qbzr (0.14.3) is out which fixes a bad bug in TortoiseBzr apprently [01:05] Tha will then use bzr+ssh [01:05] poolie: I'm not sure if it's serious enough to warrant rebuilding the Windows installer for. GaryVdM may know [01:06] poolie: otherwise, I'm hoping you're keen to make the big announcement today. [01:06] poolie: we don't want it coming out after the 2.1.0b1 one :-) :-) [01:08] igc, i am [01:08] verterok: did you see the os x tour? [01:09] superm1: This is a guess... Maybe you have the webdav plugin, but that location does not support webdav. [01:09] superm1: If you have a launchpad account, with a ssh key set up, then set bzr launchpad-login, and run bzr pull lp:~mythbuntu/mythbuntu/mythbuntu-weekly-build/. That will then use bzr+ssh rather than http. [01:09] igc: nope [01:09] igc, btw it's a holiday here so spiv will be away [01:09] i'm swapping it for next week [01:09] and lifeless is in montreal of course [01:11] igc: btw, I've been trying to get a app bundle working, I have a working bzr.app bundle (with some hacks to py2app)...but only for CLI, I'm having some problems with qbzr/explorer [01:11] Hi Poolie. [01:11] hi gary [01:11] and verterok [01:11] GaryvdM, poolie hey ho folks [01:11] poolie: ah - didn't know about your public holiday [01:12] poolie: I sent my gpg key :-) - I'm not sure if I sent the correct part. [01:12] Hi pygi! [01:12] k, i just got back from a launchpad meeting in london yesterday [01:13] still catching up on stuff [01:17] ok - sure. [02:01] what's up when some viewers show 9 revisions and some other 400... (there are 400 revs merged into that tree) [02:02] well, the actual problem is that bzr qlog shows e.g. rev 398 and when i try to remove that with merge -r 397..398 it says so such revision [02:03] bzr log even shows negative revsions :/ [02:06] is there a way to repair that? [02:16] zsquareplusc: oh that bug [02:16] zsquareplusc: i think reconcile will fix that [02:19] mwhudson: indeed, it did. thanks [02:20] now that merge command just did nothing. bzr merge . -r 397..398 says "all changes applied" and changed nothing [02:37] how do i remove the changes introduced by a revision? i thought the command above should work. [02:38] Y..X [02:39] ah.. heh, the example in bzr revert has negative numbers, translating that to positive rev numbers isn't easy ;-) === statik` is now known as statik [02:52] zsquareplusc: you need to reverse the numbers [02:53] mwhudson: yep. i understood the hint by bob2 [02:53] ah cool [03:10] I'm getting a "file id ... is not present" error with 2.0.0. [03:10] This is a branch with a single committed revision. [03:10] I'm doing a diff before committing a second revision. [03:11] Standalone tree (format: pack-0.92) [03:11] Is this a known issue? [03:12] Doing bzr init in a new dir followed by bzr pull gives me a working branch. [03:12] Format is 2a in that branch. [03:13] I'd guess it's bitrot in the support for pack-0.92 ... [03:13] I would have guessed that even before you tried 2a [03:13] SamB_XP: Not sure which bzr version I was using when I *created* that pack-0.92 branch. [03:13] I'm using bzr from the PPA, so I get updates... [03:13] well, if it pulled okay, it must not be too bad! [03:13] Guess so. [03:14] I'd rather stick with pack-0.92 for some reason. I like to sit on the new formats for a bit usually. [03:14] so, I'd make a bug and attach that repo to it [03:14] This is an internal project, though, so not a big deal in this case. [03:14] as a tarball [03:14] ... if it's not a big deal ... [03:14] SamB_XP: Hm, can't publish this particular repo. [03:14] oh, not so good :-( [03:17] Maybe a log would be okay? [03:18] I don't mind creating a bug. [03:20] Hm, I just created a new pack-0.92 branch and pulled into that. Works fine. [03:21] Not sure what would've been wrong with my other branch... (looks like an issue with the WT, not the branch itself, for whatever that's worth). [03:28] * igc lunch [03:32] abentley: If I have an addition to bzrtools, should I just submit it as a merge proposal to lp:bzrtools? [03:48] Any idea why a process would hang on process.stdin.write? [03:48] I'm trying to pipe to tar for a bzrtools extension [03:48] and tar runs properly, but the code never moves beyond process.stdin.write [04:03] davidstrauss: maybe it needs stdout read ? [04:03] or possibly stderr? [04:04] SamB_XP: I've already decided to move to tarfile, anyway [04:04] SamB_XP: but thanks [04:04] probably wise ;-) [04:04] that's more portable anyway [04:07] davidstrauss: if i was there, i'd run strace on the tar process and see what it's doing rather than reading from stdin === timchen1` is now known as nasloc__ [04:09] davidstrauss: Sure. [04:09] abentley: This utility downloads a tarball and replaces the current stuff in the branch with the tarball contents. [04:09] abentley: Candidate for bzrtools? [04:10] davidstrauss: Sounds like the import command. [04:10] abentley: where is that defined? [04:11] abentley: Neither code nor bzrtools seems to have one [04:11] abentley: nevermind [04:11] davidstrauss: It's part of bzrtools. It's mainly defined in the file upstream_import.py [04:12] abentley: looks perfect. thanks. [04:13] abentley: And it does exactly what I want with directory stripping, etc. === abentley1 is now known as abentley [05:06] Hi Everyone - [05:08] igc, please don't push to the web site branch for the moment while we're deploying it [05:09] poolie: ok, I'm done in any case [05:10] ok, it's done [05:10] poolie: awesome! [05:10] igc, search isn't working well though - i don't know if i broke it, or if it was just not working? [05:11] * igc takes a look [05:11] Getting an error with v 2.0.2 - when trying to commit bzr: ERROR: Invalid url supplied to transport: "file:///home/developm/public_html/xyz/": Win32 file urls start with file:///x:/, where x is a valid drive letter [05:12] I think it might be a bzr upload plugin issue [05:12] igc, do you think we should redirect the /Welcome page too? [05:12] i think maybe so [05:13] poolie: no, please leave Welcome there as the Wiki home page [05:13] CameronP: 2.0.2? [05:13] yeah 2.0.2 [05:13] poolie: buttons aren't rendering properly either btw [05:13] CameronP: could you please file a bug with the traceback from bzr.log, then paste the number here [05:13] the ones next to the left hand links? i'm fixing it [05:13] sure, how do i do the traceback from bzr.log [05:14] poolie: also Get Xxx buttons [05:14] CameronP: open it in notepad and copy from it [05:15] ah should i do it from the server (linux) or the client im comitting from (winblows) [05:16] the client [05:17] kirkland ping [05:18] igc, the buttons should be ok now [05:18] the default font still looks a bit small to me? [05:18] poolie: edit it in css/html-elements.css maybe [05:19] poolie: maybe search.html has to move to the end directory as well? [05:21] poolie: also, the wiki pages aren't rendering correctly in case you didn't know [05:22] i hadn't seen that but you're right [05:24] igc, looks ok again now? [05:24] poolie: much better [05:25] poolie: btw, clicking on the Bazaar logo in the Wiki goes to the home page, not Welcome [05:25] is that what we want? [05:25] (trying to fix search) [05:25] hm [05:26] probably yes? [05:26] I guess we need a link in the header to both somehow? [05:27] i'd like to eventually have the top nav bar include a link to the wiki and then be common to both [05:28] that would be good [05:28] so i propose to change the search thing to have it just present results in a google page [05:28] like [05:29] http://www.google.com/cse?cx=009852948614216791564%3Adr5xzrczfnu&ie=UTF-8&q=checkout&sa=Search [05:29] i think ej is trying to have it present them within this page [05:29] which would be nice but i can't see how it would work with this current code a [05:29] and last time i tried it put it in the wrong place in the page [05:29] also i might pay for the ads to be removed from that page [05:30] ok. I know kirkland helped her get it working but he's probably asleep now [05:30] igc: not sleeping, just not paying very close attention [05:31] hi kirkland. we've just make the website live but search isn't working [05:32] igc: url? [05:32] probably because index.html moved to en/index.html [05:32] igc: yup, should be trivial to fix [05:32] hello kirkland [05:32] http://bazaar-vcs.org/ [05:32] can you tell me how you're setting this up? [05:32] poolie: another minor quibble - the bazaar icon is missing from the tab in ff now [05:33] this is the iframe approach? [05:33] favicon.ico? [05:33] ok [05:33]
[05:34] add action="/en/" [05:35] ok, i see [05:36] poolie: something has gone weird btw [05:37] mm? [05:37] well done? We couldn't get that effect if we tried :-) :-) [05:37] which effect? [05:38] try refeshing your page [05:38] wow [05:38] lots of strange spacing? [05:38] yep, something about links [05:38] I suspect [05:39] all mine are yellow and on newlines [05:39] missing a close [05:40] poolie: should i file this bug under the upload plugin or under bazaar - i have confirmed when the auto upload feature is turned off of the plugin on the server, I no longer get the error [05:40] ok, bzr-upload then [05:41] igc, kirkland, ok, search and links are ok now istm [05:41] poolie: well done [05:41] poolie: much better [05:42] search button on line below box though? [05:42] on the results page? i know [05:42] main page [05:46] ok, i see why [05:46] yay version control :) [05:47] we need to switch this to some kind of templating system [05:47] verison control ftw! [05:48] https://bugs.launchpad.net/bzr-upload/+bug/442798 [05:48] Launchpad bug 442798 in bzr-upload "error with --auto feature turned on with 2.0.2" [Undecided,New] [05:48] can anyone tell me what the progress of nested trees is? [05:49] CameronP: the thing is, there is no 2.0.2 [05:49] realbadapple: it's still pending; the best option atm is probably scmproje [05:49] scmproj* [05:50] anyone? nested trees! how soon till this work flow will be supported? [05:51] sorry poolie [05:51] realbadapple: it's a feature under development [05:51] did not notice the response [05:51] realbadapple: it won't be in production for another 5-6 months given we're now on 6 months releases [05:52] thanks can't wait! [05:52] realbadapple: it *may* appear earlier in a beta but use scmproj for now if you can [05:52] this would really help with useing eclipse for development [05:52] scmproj? [05:52] is it a plugin? [05:52] yep [05:53] http://bazaar-vcs.org/ScmProj [05:53] ok thanks I'll give it a look [05:53] k, search looks good now [05:53] realbadapple: if you haven't already, see http://bazaar-vcs.org/BzrPlugins [05:53] that page has just been reorganised so it ought to be easier to find plugins now [05:53] i'm just going to handle the noscript case a bit better by falling back to the google-hosted version [05:54] ya I've glanced at the plugins page but have not looked closely [05:55] poolie: search is good for me now. It's pretty damn cool actually [05:55] indeed [05:55] thanks a heap kirkland! [05:55] though it gives me sex ads when i search for 'test' [05:55] igc: cheers [05:55] :-( [05:55] so i'd like to get them off [05:55] like many things in life, a small matter of handing over money :} [05:56] poolie: maybe we want a Search and an "I'm feeling lucky" button in the nav bar ... [05:56] the latter gives the ads :-) [05:58] poolie: so favicon.ico is the only issue I can see now [06:00] well, and noscript search [06:00] igc: heya! (late) [06:03] hi mneptok - great to hear from you! [06:04] hello mneptok [06:12] igc, i think the favicon is now/soon fixed [06:12] pooie: yep - working now [06:13] do you have any opinion about how to do templating? [06:13] with pizza. [06:13] i tested in lynx and it looks pretty good, including search [06:13] ? [06:13] templating is always more enjoyable when you have pizza. [06:34] poolie / igc: i've been sort of incomunicado the past 10 days while my wife and i moved to a new house we just bought. there's a guest bedroom if any bzr devs feel like seeing New Mexico before/after North American sprints. :) [06:43] poolie: how about this ... [06:44] With true rename tracking for files and diriectories, [06:45] you can refactor code without fear of merging. [06:45] still too sales-ish? [06:46] oh thanks very much mneptok [06:46] poolie: I haven't though about templating [06:46] no plans to go there atm [06:46] igc, i might try to convert it to cheetah tomorrow [06:46] i'm going to stop for today now [06:46] sphinx uses jinja fwiw: http://sphinx.pocoo.org/templating.html#working-the-the-builtin-templates [06:46] ok [06:47] consistency++ [06:47] Before I go down any more rabbitholes, is there a streamlined way to take a bzr branch and check in the current working copy to CVS? [06:47] there are dozens of libraries implementing "string % kwargs" :_) [06:47] davidstrauss: just have one directory which is both a bzr and cvs checkout? [06:47] poolie: More like git-cvsexportcommit [06:49] davidstrauss: is the export command useful to you for that? [06:49] Is there a file under .bzr which will always be updated when some change has been made? (a commit, a tag created, changes have been pulled, etc?) [06:49] igc: Not really. The complexity is in the CVS/ metadata. [06:50] DaffyDuck_`: .bzr/branch/last-revision goes close [06:50] I'm looking for some kind of "last change" timestamp, which convers _all_ types of changes. [06:51] davidstrauss: what is it that you want to have happen in cvs? [06:51] I'm not sure you ought to rely on a given file. Can you use a hook instead? [06:51] (filenames can always change over time as formats evolve) [06:52] DaffyDuck_` ^^ [06:52] see you later igc [06:52] poolie: I need to take the working tree from Bazaar and push it to CVS as a commit on a specified branch. [06:52] see ya poolie [06:52] poolie: Well, I don't need to take the working tree. I can use a full-on commit. === davidstrauss_ is now known as davidstrauss === vila is now known as vila-dentist === vila-dentist is now known as vila [08:45] hello bzr 2.0 [08:47] hi igc [08:48] hi bialix [08:48] congrat on new webpage rollout. I have one comment though, about BzrExtra [08:48] bialix: ru tranlsation of explorer is looking complete btw [08:48] hi all [08:48] as is the fr one [08:48] how can i get a rss feed from launchpad [08:48] is there a public hosting site for that? [08:48] that page listed plugins in the same end. It's a bit obscure [08:48] I was wondering whether a 0.8.3 was worth it with [08:49] or do i need my own server? [08:49] Phurl: what do you mean? [08:49] i would like a rss feed from commits [08:49] i found some scripts and such to do that [08:49] but i am missing this in lauchpad [08:50] you mean a rss feed of commits to mainline? [08:50] in general I don see any rss feed option [08:50] yes or to my branch [08:50] Loggerhead has one. [08:50] or anything from the project [08:50] Phurl: do you mean RSS feed for branch hosted on lp? [08:50] yes bialix [08:50] exactly [08:50] Phurl: I'm not aware about such feature [08:50] ok [08:51] you may ask in #launchpad though [08:51] Launchpad has one too: http://feeds.edge.launchpad.net/~bzr-pqm/bzr/bzr.dev/branch.atom [08:51] * Peng_ points at the feed button in Firefox's location bar. [08:51] nice!!! [08:51] thanks [08:52] hmm, I'd prefer to have explicit URL on the page [08:52] bialix: It is explicitly in the page, just in the section. [08:52] bialix, perfect [08:52] http://feeds.edge.launchpad.net/~kosova/+junk/openstreetmapkosova/branch.atom [08:52] explicit in the terms of I can copy-paste it [08:52] bialix: what do you find confusing about BzrExtras again? [08:53] igc: it's hard to find plugins now [08:53] igc: I mean there is no more short way to plugins from main page anymore [08:53] Phurl: you mean Peng_ [08:53] bialix: would it help if the plugins link came at the top? [08:54] bialix: see the footer links too btw [08:54] igc: IMO it will be nice [08:54] igc: this is just observation [08:54] plugins is under Extras [08:54] at the bottom of the main page [08:54] Peng, thanks!!! [08:54] observation of monday morning mind, my personal wtf for today [08:55] bialix: also, there's a "Browse the plugins registry" under the introductory text [08:55] igc: and btw content of BzrPlugins page is a bit long now ;-) I dunno is it possible to workaround this fact [08:55] I mean list of contents [08:56] not that I know of [08:56] I guess we could limit the depth of the contents but it kind of defeats the purpose [08:56] you're right, it is (Browse plugins). hmm, but other (main links) at top and big icon don't imply this [08:57] anyway, congrat for new site rollover, I really hope to see BIG announce in next few days [08:57] bialix: right, so we could move plugins higher up on the BzrExtras page [08:57] it's a bit wrong situation IMO [08:57] bialix: I thought poolie was going to do it today but it might be first thing tomorrow now [08:58] it seems so [08:58] bialix: I suspect he wants people to test the website overnight first [08:58] may be [08:59] igc: re translations, I'm a bit digressed. Do you want me to sync translations with lp? [08:59] he's only just back from London btw [08:59] bialix: it was just an fyi [08:59] I'm wrapped that the ru and fr translation are so complete! [08:59] igc: ah ok, I'm a bit busy at work [08:59] and don't check it regularly [09:00] bialix: if we need to repackage the windows installer to include 0.14.3, then ... [09:00] perhaps your new welcome page require translations template update [09:00] I can do it right now [09:01] igc: jam said he think that 2.0.1 is more preferrable [09:01] maybe it's time to release 2.0.1 and do big announcement with 2.0.1? [09:01] cool. So maybe we'll include an explorer 0.8.3 in that as well - just updated translations [09:02] igc: if you have access to http://pypi.python.org/pypi/bzr [09:02] there is still 1.17!!! [09:02] it require update [09:03] it does. I can't do it right now but I'll put it on my list [09:03] please [09:04] this delayed announce stuff is open can for such nit problems here and there [09:04] * bialix believes release should be very focused on day job, not something that lasts for weeks [09:05] * bialix believes release should be very focused one day job, not something that lasts for weeks [09:05] agreed [09:06] grrrrr [09:06] new lp ui is.... [09:06] grrr [09:06] translations ui becomes worse and worse in last few month IMO [09:07] bialix: I hoped you like the explorer tour btw. It should help get some more Windows users on board which will be good for all of us [09:07] got to go now [09:08] * igc dinner [09:08] igc: it seems quality of russian translations is below my quality standard. I have to review new stuff before importing it into trunk [09:08] ok [09:08] Is there an unrevert in bzr? A couple of times now I've forgotten to add the filename that I want to revert and accidentally reverted all changes. [09:08] igc: the tour is FANTASTIC [09:09] arnarl: your changes backed up in *.~1~ files [09:09] Is there some cure for my stupidity? Not that I have lost much work this time, but still :-) [09:09] there is no unrevert, but changes are usually saved in ~N~ files [09:10] a, nice I. [09:10] you may want to use qrevert GUI dialog (from QBzr plugin) as safer revert [09:10] I thought those were some emacs tmp files... [09:10] I just try to commit often [09:10] commit often is ok [09:11] but, I did not know about the *.~n~ files [09:11] bialix: so thnx [09:11] :-) [09:11] glad to help === doko__ is now known as doko === quicksil1er is now known as quicksilver [10:43] hi, quick question about hosting a bzr server on ec2, is there a way to pass an identity file to bzr+ssh ? [10:43] similar to invoking the -i option on ssh [10:50] not currently [10:50] you can configure things in .ssh/config [10:50] or add the identity to your agent [10:55] lifeless: cheers, i guess i have to dig around and see how can i add identity info to my ssh agent [10:58] nedosa: ssh-add /path/to/identity.pem? [11:01] mwhudson: fantastic, thanks ! === mrevell is now known as mrevell-lunch === Noldorin_ is now known as Noldorin === mrevell-lunch is now known as mrevell === vxnick_ is now known as vxnick [14:13] Hi All [14:13] trying to use bzr patch command - should it work fine with a .diff file, in the directory of the file I want to patch? [14:15] I dont think Im meant to use it in thatway ;) oops [14:15] Hi ! Is there any way to find when file containing given text was deleted from given file ? in bzr or bzrtools ? [14:19] Like: for every rev listed in bzr log do bzr diff -c | grep [14:20] there's a bisect plugin [14:24] Tak: "Use a binary search algorithm to apply a test of the source code for each revision", test may be simple grep ? [14:25] the test may be whatever you wish [14:25] basically you say "yes" or "no" for every iteration [14:31] Tak: Ok. thanks === Adys_ is now known as Adys === deryck is now known as deryck[lunch] [16:31] matkor: bzr-search does good indexing, too [16:32] matkor: bi-sect may not work if the file's been deleted more than once [16:34] davidstrauss: Thanks. bzr-search is another module ? [16:34] Has any work gone into making bzr less resource hungry in 2.0? (see bug 408526 and bug 408531). [16:34] Launchpad bug 408526 in bzr "bzr commit on large projects require large amounts of memory (dup-of: 109114)" [Undecided,New] https://launchpad.net/bugs/408526 [16:34] Launchpad bug 109114 in bzr "[master] bzr holds whole files in memory; raises MemoryError on large files" [Medium,Triaged] https://launchpad.net/bugs/109114 [16:34] Launchpad bug 408531 in bzr "bzr branch on large projects require vast amounts of memory" [Undecided,New] https://launchpad.net/bugs/408531 [16:34] matkor: yes [16:35] DaffyDuck_`: bzr 2.0 with the 2a format holds a lot less in memory. [16:36] DaffyDuck_`: IIRC, it's down to file text + compressed version when committing. [16:36] Peng_: Would that specifically be 2.0, 2a or a combination of both? I already use 2a, but with 1.7/1.8, but they both exhibit the problem. [16:36] DaffyDuck_`: It requires the 2a format, but I don't know the version of bzr. [16:37] Ah, it's 1x file bytes + 2x compressed bytes: https://bugs.edge.launchpad.net/bzr/+bug/109114/comments/15 [16:37] Launchpad bug 109114 in bzr "[master] bzr holds whole files in memory; raises MemoryError on large files" [Medium,Triaged] [16:38] Peng_: That's promising! Thanks! [16:42] #ubuntu-desktop, #ubuntu-devel, #ubuntu-doc, #ubuntu-installer, [16:42] #ubuntu-website, #whatwg [16:42] (oops - tried a /join command that didn't work - sorry) [16:56] can someone help me translate this into bzrlib API terms? i want to build a pre-commit hook that greps through the incoming diff for a keyword. [16:57] would this be checking through bzrlib.delta.report_changes(oldtree.iter_changes(newtree), ???) [16:58] hmm or maybe better bzrlib.diff.show_diff_trees(oldtree, newtree, stream_to_string), and loop through that looking for my keyword === deryck[lunch] is now known as deryck === beuno is now known as beuno-lunch === beuno-lunch is now known as beuno [19:26] when I get a patch created with bzr, how can I know which file revisions the patch was made against? [19:26] cvs diff adds the file revision number to the file, so it's easy to know [19:27] but bzr diff only adds the date the file was edited, it looks like [19:27] Is it a bundle generated with bzr send, or just a diff generated with bzr diff? [19:28] Lo-lan-do: with bzr diff [19:29] Not sure you can then. [19:29] :( [19:29] LpSolit: Why not use "bzr send"? [19:30] Peng_: 1) because I never heard about it, and 2) because people create patches using bzr diff [19:31] does bzr send alter my installation? [19:31] or does it only generate a patch? [19:31] LpSolit: why you need to know revision then? [19:31] It generates a patch with extra metadata [19:33] bialix: because in Bugzilla (it's a bug tracking system developed by Mozilla), we want the ability to display patches with their context, i.e. patches most of the time have 3 lines of context, and we have the ability to display patches with as many lines of context as you want [19:33] bialix: so you can select context=50 and you have the same patch redisplayed with 50 lines of context [19:33] this works for years with CVS [19:34] and I have to update Bugzilla internals to work with bzr [19:34] You can't do that with bzr, though it likely wouldn't be hard to add. [19:34] * LpSolit wonders if what I said is clear enough [19:34] `bzr send --no-bundles` is roughly equal to plain diff but with revision id [19:35] let me try bzr send [19:36] Peng_: it can be done with plugin [19:38] $ bzr send --no-bundles test_config.t [19:38] bzr: ERROR: no such option: --no-bundles [19:39] Peng_: would it be reasonable to request to append the revno to the file name? [19:39] --no-bundle sorry [19:39] so instead of --- process_bug.cgi 2007-10-12 07:21:59 +0000 , we would have: [19:39] --- process_bug.cgi 2007-10-12 07:21:59 +0000 344 [19:39] or something like that [19:40] maybe revno:344 [19:40] LpSolit: revno is local info for the branch [19:41] It's still valuable info. Not everyone uses branches everywhere :-) [19:41] in the a diff context you should use revision id [19:42] IIRC svn shows revno in the diff comment [19:43] yeah, it does [19:44] same for cvs [19:44] * LpSolit really needs this info [19:45] bzr send --no-bundle -o toto.diff [19:45] Using saved submit location "test_config.t" to determine what changes to submit. [19:45] bzr: ERROR: Not a branch: "/var/www/html/selenium/bugzilla/bzr/3.4/t/test_config.t/". [19:45] what is it trying to do? :) [19:47] send requires trunk branch available and passed as 1st argument [19:48] and send works with committed changes [19:48] perhaps it's not what you need [19:48] no, it's not [19:48] before committing changes, the patch has to be approved by reviewers first [19:49] IMO you can file a bug asking for this feature or write plugin [19:49] > before committing changes, the patch has to be approved by reviewers first [19:50] in bzr we working another way [19:50] commit your changes, then send a patch for review [19:52] LpSolit: bzr is DVCS [19:52] commit won't clutter trunk until it will be merged [19:53] yes, I know that. Being a CVS user for many years, the change is not obvious :) [19:53] oh, right [19:54] LpSolit: look https://code.launchpad.net/bzr/+activereviews [19:55] abentley: can you restart bundle buggy please? [19:56] bialix: done. [19:56] abentley: error 503 [19:58] bialix: Bundle Buggy is currently plowing through a big backlog of email, but it works for me. [19:58] now and for me [19:59] LpSolit: look also http://bundlebuggy.aaronbentley.com/ [19:59] this one works with the output of bzr send command [20:00] abentley: thanks [20:00] where is the metadata? [20:00] * LpSolit is looking at http://bundlebuggy.aaronbentley.com/project/bzr/request/%3C49F22B67.8020703%40aaronbentley.com%3E [20:00] bialix: np. btw, launchpad also works with send. [20:01] abentley: IIRC the e-mail should be GPG-signed, right? [20:01] oh, I can see it when downloading the patch [20:01] ok [20:01] bialix: Yes, and you need to have a Launchpad account with that GPG key. [20:02] I was not aware about assigning GPG to account [20:02] Hello all. I've come up dry on google so I'm wondering if anyone on here knows. Is there any way to delete a file from your revision history? [20:03] * bialix forgot about GPG [20:03] I was able to do it in Subversion a while back through an annoying process of exporting and importing the repository. I'm hoping there is at least some similar way, if not easier. [20:04] lgoodrich: if you mean to completely remove history for the file then no [20:05] not even through some sort of export, filter, then import process? [20:05] this will works [20:05] bzr log test_config.t [20:05] ------------------------------------------------------------ [20:05] revno: 48 [20:05] timestamp: Mon 2008-10-20 01:41:26 +0000 [20:05] works [20:05] but after re-import you'll have completelly unrelated history [20:05] but: bzr cat -r date:2008-10-20,01:41:26 test_config.t returns: [20:05] bzr: ERROR: The file id "test_config.t-20081020014126-eujmy2eg2qvr-1" is not present in the tree what's wrong with that? [20:06] I'm trying to use the timestamp when the revision id is not available [20:06] bzr cat -r date:2008-10-20 test_config.t [20:07] maybe [20:07] hmm, I'll see if its worth the hassle, thanks [20:07] bialix: same error [20:08] is it possible the file was renamed? [20:08] no, the file is still here: http://bzr.bugzilla.org/selenium/3.4/annotate/head%3A/t/test_config.t [20:10] this page http://bzr.bugzilla.org/selenium/3.4/changes?filter_file_id=test_config.t-20081020014126-eujmy2eg2qvr-1 [20:11] hum, so it matches what the error message was reporting [20:11] test_config.t-20081020014126-eujmy2eg2qvr-1 [20:12] says the date should be 2008-10-19 [20:12] perhaps TZ issue [20:12] it's a bug though [20:12] see 20081020 in the file-id [20:12] doesn't work with 2008-10-19 either [20:13] it's the timestamp when file was added [20:16] if I type: bzr cat -r 48 test_config.t, I get the file content [20:16] there is plugin file-revno which can print revno number of the last change for the file [20:16] so it only fails with date: [20:16] file a bug please [20:17] bialix: URL to your bug tracker? [20:17] I know you don't use Bugzilla :) [20:19] https://bugs.launchpad.net/bzr [20:35] bialix: thanks! https://bugs.launchpad.net/bzr/+bug/443419 [20:35] Launchpad bug 443419 in bzr "bzr cat -r date:YYYY-MM-DD,HH:MM:SS filename doesn't work" [Undecided,New] [20:36] thanks [21:45] Hi. I've been considering moving a project from cvs to bzr. After deliberation, I think the sticking point is bzr's worldview of having a branch, not a repository, as the primary object. Specifically, I'll end up losing history (or at least having it cumbersome to access) as parts of it become confined to branches which no one remembers to look at [21:46] I was wondering if there was a planned binary Mac OS X 10.6 install for bzr-explorer? [21:46] I'm wondering if anyone has any bzr documentation or evangelism that might help me understand how this isn't a problem for more people. [21:51] eferraiuolo: I suspect there is [21:51] maxb: bzr log --include-merges shows that "lost" history [21:51] maxb: if that's what you mean [21:51] maxb: (or qlog or whatever) [21:51] Only if it actually got merged [21:51] maxb: what do you mean by 'more cumbersone'? [21:52] As in, in any other VCS, I would point at the repository and explore all the history [21:52] well, I dispute the 'any other' phrase :) [21:53] in any DVCS it would only be 'all the local history', in other ones like codeville, they behave like bzr [21:53] however, qbzr can do viz on multiple branches [21:54] I think a bug on it to allow qbzr to work would be well received [21:54] maxb: anyhow, I think it isn't a problem because people don't need to do it [21:55] I suppose my consternation is magnified by this being a conversion from CVS, and knowing that I'll have to coax people away from CVS practices [21:57] * mwhudson has almost entirely forgotten that software development is even possible with CVS, let alone how it is done [22:00] maxb: so, one use case is 'find a brnach name' which is replaced by 'ls' [or 'bzr branches'] [22:01] mwhudson: I wish I _could_ - but I'm forced to use it at work :-) Hence the attempt to plot a way out. [22:01] lifeless: A use case I'm concerned about is "Find a tag. I don't know which branch it's in" [22:02] maxb: so, what bzr does is merges release branches to trunk [22:03] ls / bzr branches are somewhat problematic unless you have *all* branches locally [22:03] maxb: so all tags are in trunk [22:03] maxb: yes, but see under distributed [22:03] I think you are right, using bzr is going to entail convincing cvs2bzr to merge branches to trunk whereever possible, for the past, and convincing developers of the same, for the future [22:04] maxb: well if a feature isn't merged to trunk, then it would stand to reason that the tag isn't there, no? [22:04] what are all these unmerged branches in cvs being used for? [22:05] Well, some of them probably have been merged. (but being CVS, you can't tell). And some are ad-hoc "just getting a release out there" branches [22:07] jam, hi? [22:15] I wonder how plausible it would be to write a tool which can add an extra parent to an EXISTING revision? [22:16] I realise this can only work if you know you hold the only copy of the repository [22:16] But it would be an interesting option for patching up a cvs conversion before making it public [22:16] maxb: filtering the output of 'cvs2bzr' (the fast-import file) is the perfect time to do that [22:16] before it gets imported [22:17] jam: Well.... sort of. It's a lot easier to find the joinpoints if you can 'bzr viz' [22:17] maxb: you can iterate over it [22:17] convert, filter, convert again, etc [22:18] this is true. I'm just a bit unsure how you'd inject the information back into the conversion [22:18] maxb: you might look at fastimport's 'filter-branch' stuff [23:17] morning [23:26] hi [23:27] i have my email in Bazaar set to "meoblast@aol.com" [23:27] if i change my email address, that might become a problem [23:27] is it safe to just set it to "meoblast" [23:35] hi, I have a repository being served with the bzr+http protocol. I recently upgraded the server from 1.9 to 2.0, and now commands to it die with "An attempt to access a url outside the server jail was made". I'm not entirely sure how to fix this. [23:51] meoblast001: Some applications (eg. Launchpad) identify you by email address, so changing it to just 'meoblast' would result in Launchpad no longer linking your commits to you. [23:51] ugh [23:51] i don't like aol though [23:51] i plan to switch [23:52] and when i do, everything will recognize me as a new person entirely [23:52] Is it really a problem if some of your old commits have an old email address? [23:54] wgrant: well, services, such as ohloh, detect developers by email [23:54] new email makes you a different developer [23:55] meoblast001: those services allow you to have multiple emails per user [23:55] Right. [23:55] You can easily merge the users. [23:55] I am certain of this for at minimum, LP and Ohloh. [23:55] do pretty much all services like that have those features? [23:55] meoblast001: also, changing to 'meoblast' would cause a discontinuity now, anyway [23:55] meoblast001: just the same as moving to a new email address would [23:55] bob2: well, if you do it the normal way ;) [23:56] keep in mind pretty much no one but me and one other developer has a checkout of my project [23:56] so now is the best time to fix screw ups [23:58] hello igc [23:58] hi poolie