[00:00] as in branch a subtree of an existing branch? you can try "split", but I don't think it's production-y yet [00:05] okay, so the correct fix is to s/map.// [00:05] * mathrick hugs bzr blame [00:07] bob2: the option is to copy the dir to another place and init is as a new bzr branch, right? [00:08] yeah, but that loses history [00:08] you could also branch the whole thing, delete everything else and then move the files in this one dir to the root [00:08] bob2: oh, that works too - thanks! [00:43] is there some trick to get bzr blame to find where a particular line has disappeared? [00:46] use gannotate and click around versions [00:46] is what most people seem to do [00:51] aha [00:55] http://bazaar.pastebin.com/GaJP7cRb [00:55] any idea what's wrong? [01:27] yarr [01:33] mathrick: this looks like an already fixed bug, are you up to date? [01:33] poolie: 2.1.0 [01:34] and also in bzr-git? [01:36] bzr git is r721 [01:37] sorry [01:37] suggest you file or look for a bzr-git bug then [01:37] ok [01:47] when I do bzr cat it changes my remote directory setting how do I stop it? [01:48] what do you mean? [01:50] I think I misunderstood the error [01:51] I got a "permanently redirected" report [01:51] << is tired [01:51] btw if you are interested in how I am using bzr for squeak/cuis http://smalltalkers.pbworks.com [02:07] thanks keithy [02:08] how are things working out? [02:08] in what respect? [02:08] just with bzr and lp [02:08] I am suffering form spiderman syndrome...."with great power comes great responsibility" [02:09] very well indeed [02:09] some quibbles with the squeak licence [02:10] I dont think people in the OSI community realise how much trouble they caused [02:11] i heard about some of that [02:11] it has taken something like 10 years to relicence squeak, all because of two paragraphs in the licence that arent relevant any more [02:11] for example [02:11] the licence says, you cant distribute apples fonts [02:12] but readers of the licence dont think to ask if there are any apple fonts in the product to distribute [02:12] the last remnant was removed 6 years ago [02:13] Apart from that [02:13] right [02:13] In a week or so [02:13] i see your point about the difficulty of updating things that predate OSI [02:13] exactly, it was very disingenious of them [02:13] especially when squeak was a pioneer of OS for commercial purposes [02:14] the SqueakL was actually designed to enable reuse for commercial purposes [02:14] who is 'them'? [02:14] unlinke Gnu etc [02:14] squeak was developed by a team in the Apple Technology Group [02:14] this team was originally at Zero parc [02:14] xerox [02:14] parc [02:15] ok [02:15] The system was released by xerox to the public by publishing the full spec in a book, the famous blue book [02:15] so, i hope we work something out [02:15] rebooting, biab [02:15] i know, i read it years ago [02:15] it's pretty cool that it's out there [02:15] its on the net now [02:16] rebooting, biab [02:30] i'm thinking about just removing the attempted fix for bug 456077 from 2.0 because of the fallout it caused [02:30] Launchpad bug 456077 in bzr "MYSQL/BZR P3: bzr doesn't explain it's doing a slow cross-format fetch" [High,In progress] https://launchpad.net/bugs/456077 [02:34] lifeless: ping, teddybear wanted on this [02:34] poolie: voice? [02:35] irc will do [02:35] just about whether to revert the cross-format fetch warning in 2.0 [02:35] what was the fallout [02:36] warnings during upgrade [02:36] and the message is not always very clear, eg bug 513157 [02:36] Launchpad bug 513157 in bzr "warning about on-the-fly upgrades to "(remote)", and when upgrading" [Medium,Confirmed] https://launchpad.net/bugs/513157 [02:36] right [02:36] uhm, I think the upgrade one is easy enough; squelch the warning before starting upgrade. [02:37] the clarity one, seems like polishon the format %s that spiv worked on in france [02:37] so, I would attempt to finish it off rather than revert it, myself. [02:37] i guess the fallout is not so severe that it needs to be immediately pulled [02:37] yeah [02:37] ok [02:37] thanks [02:45] poolie: no probs [03:13] poolie: are you doing https://bugs.edge.launchpad.net/bzr/+bug/515356 too? [03:13] Launchpad bug 515356 in bzr "unnecessary message about conversion during upgrade" [Medium,Confirmed] [03:13] yes [03:13] cool; was wondering as its not assigne [03:15] i'm not doing it right now [03:15] i'm going to fix the underlying fetch thing first [03:19] is bug 388269 a dupe of bug 435048? [03:20] Launchpad bug 388269 in bzr "getting a branch into an empty shared repository takes a long time to figure out revisions to send" [High,Triaged] https://launchpad.net/bugs/388269 [03:20] Launchpad bug 435048 in bzr "many get_parent_map calls during walk to common revisions" [Medium,Confirmed] https://launchpad.net/bugs/435048 [03:20] I think so [03:20] if not a dupe then remarkably similar [04:40] Hello, based on the discussion on bug 503670 'bzr grep', I have created a bzr-grep plugin explained here https://bugs.launchpad.net/bzr/+bug/503670/comments/11 [04:40] Launchpad bug 503670 in bzr "bzr grep should be builtin" [Medium,In progress] [04:40] Launchpad bug 503670 in bzr "bzr grep should be builtin" [Medium,In progress] https://launchpad.net/bugs/503670 [04:41] I would appreciate any review comments this as I am still new to bzrlib API. Whats a good launchpad mechanism for this? Probably the comments can be left on the bug listing? [04:42] Another question, I am planning to work on the test cases, any pointers would be much appreciated. [04:45] parthm: there wasn't a bzr-grep already? [04:45] I could find anything. I saw bzr-search and ~vila/bzr/grep which was unix only. [04:46] ok [04:46] so i don't think there's any particular way to say please review this from scratch [04:46] other than making an empty trunk branch [04:46] then proposing a merge of all your code into that [04:46] which would work ok [04:47] Sounds ok. So for the purpose of review, I would just branch the trunk and put in the grep folder so the comments can be tracked. [04:49] What would be a good way to create tests? Basically, I need to create a working tree, add files and file content. then run "bzr grep" on it pattern matching on output. [04:53] ...I think I just figured out a set of process I can use to let my team treat our non-versioned ftp servers at work as if they are under version control with all the people who use raw ftp as a sort of "ghost committer." [04:55] I'm not sure though. Gotta think about that. It may require either more steps to follow when committing, or finding somewhere I can put a bzr smart server and getting rather creative with hooks. === mordred_ is now known as mtaylor [05:59] Hi vila [06:02] hi GaryvdM, vila === mordred_ is now known as mtaylor [06:03] Hi poolie [06:08] Where is the bzr documentation? [06:08] The "Get Help" link on the website doesn't seem to lead there. [06:09] http://doc.bazaar.canonical.com/en/ [06:09] Thanks. How does a normal user get there? [06:09] which website? [06:09] http://bazaar.canonical.com/ [06:10] "Documentation" link at the top left [06:10] or at the bottom [06:10] Ah, so to get help I don't actually use the "Get Help" link. [06:10] easily fixed [06:10] if you say so [06:11] with sudo_sarcasm [06:13] wgrant: howzat [06:16] poolie: Looks good. Thanks! [06:20] * wgrant wonders if 90000 users is really a stat that wants to be widely advertised. [06:39] wgrant: a often quoted figger for world wide developers is 9 mil. Based on that, bazaar is used by 1% [06:39] Hmm, dose not sound to good. [06:39] it's very hard to estimate usage of open source [06:40] So it's probably a bad idea to try. [06:44] is that on our homepage now, or elsewhere? [06:44] vila: I wanted to talk to you about releasing bzr-gtk. I pop back later. [06:46] poolie: It's on the homepage. [06:47] Otherwise the website is pretty nice now. [06:47] Particularly with the compelling linked logos at the bottom. [06:48] thanks [06:49] wgrant: how about the project count [06:49] and the link to the popcon page? [06:49] Popcon needs to die, not be linked to from anywhere :/ [06:50] Bazaar is used by thousands of people around the world on both open [06:50] and closed projects, including
[06:50] +1 [06:50] gary? [06:50] maybe s/thousands of// [06:50] I considered that. It's hard to say. [06:51] Bazaar is used by thousands of projects, both open and closed, [06:51] including
[06:51] That works. [06:51] And is far more likely to be accurate. [06:52] And cannot easily be interpreted as being a very low number. [06:54] k [06:54] cron will update it in a bit [06:54] Thanks. [06:54] not at all, thanks for the review [07:01] hi all [07:18] poolie: Oh, I see that the number wasn't actually pulled out of nowhere. [07:19] poolie: But popcon is completely the wrong order of magnitude to use. [07:19] cause it undercounts? [07:19] BzrPopularity oddly states the numbers as facts. [07:19] poolie: It's not enabled by default, and the option is well hidden. [07:19] I don't know anybody who has it on. [07:19] well, it's a fact (i presume) that's what it reports [07:19] right [07:19] me :) [07:19] but feel free to update the wiki page [07:20] Ah, it is a wiki page. I see. [07:21] See, the most popular popcon package has only 1.4 million installations. We have quite a few more users than that. [07:22] sure, you're quite right [11:28] Hi all [11:30] I made the mistake to branch manually (copy-paste of directory). And I have commited some changes in both branches now. How can I make the branch aware that it diverged from the trunk? [11:31] copying branches is fine [11:31] yeah, but merging them will be a problem [11:31] why? [11:31] bzr merge /the/other/one [11:31] well, they don't know where they diverged no? [11:31] sure they do [11:31] ah [11:32] does bzr compare the history or something? [11:32] yes [11:32] oh I see [11:32] branch doesn't do anything inherently magic [11:32] there was no need to worry then [11:32] you can use 'cp -a' if you prefer [11:32] ok [11:32] I was afraid to have done something dumb :) [11:32] thank you for your help bob2 [11:32] well, back them both up first if you're worried :) [11:33] ok [11:33] thanks a lot [13:37] jaьЖ рш [13:38] jam: hi [13:38] jam: about next windows installer: please use qbzr 0.18.1 for it, not 0.18.2 [14:11] bialix: why is that? [14:14] ah, found your email [14:14] bug #524483 [14:14] Launchpad bug 524483 in qbzr "qrevert: Select all checked even if there is some files unchecked" [High,Fix released] https://launchpad.net/bugs/524483 [15:01] офьЖ https://bugs.launchpad.net/bugs/526933 [15:01] Launchpad bug 526933 in qbzr "qadd inadvertantly adds all my ignored files " [Critical,Confirmed] [15:01] jam: https://bugs.launchpad.net/bugs/526933 [15:01] sorry === gnomefreak76 is now known as gnomefreak === salgado is now known as salgado-lunch [16:12] yo, bzr people. Imagine, hypothetically, that I'd got a branch merged into the trunk of my project and everyone hated me for it, so I now want to submit a new branch which removes the changes that the previous one made. What's the easiest way to create this new branch? [16:13] do I just branch trunk, do bzr diff -r -5..-6 > patchfile (or whatever revision my branch landed in), then patch -p0 < patchfile and bzr commit? [16:13] that seems a bit long-winded [16:14] aquarius: branch trunk [16:14] bzr merge -r -6..-5 . [16:15] bzr commit [16:15] bzr push lp:... [16:15] haha! you can merge from yourself? [16:15] propose merge [16:15] smart. I didn't know you could do that :) [16:15] yep [16:15] I knew you'd know a better way than making a patch file. Cheers, pal :P [16:15] and doing the revision range backwards applies the reverse diff [16:21] wait. -5..-6 *is* the backwards diff [16:23] aquarius: err yeah, what IslandUsurper said :-) [16:23] james_w, I knew what you meant. negative numbers. It's the opposite of the opposite of what you think :) [16:23] vds, ^ :-) [16:23] aquarius: yup [16:38] how can I show which files were edited at a specific revision [16:39] CaMason: bzr st -c 123 [16:39] CaMason: Probably [16:40] CaMason: try bzr annotate filename [16:40] er... [16:40] yeah I miss understood your question [16:40] peng's answer might be better. [16:40] My answer is the best! :D Probably. [16:40] heh [16:42] thx :) === IslandUsurper is now known as IslandUsurperAFK === radoe_ is now known as radoe [17:46] Hi, is it possible to remove a tag? [17:47] oh ok. tag --delete works. i was using -d and it wasnt === _TiN_ is now known as TiN [18:12] stefanlsd: -d means --directory === IslandUsurperAFK is now known as IslandUsurper [19:27] channel trolls? really> [19:27] * bialix summons garyvdm [19:28] Eh? Trolls? Where? [19:29] banner said: [freenode-info] channel trolls and no channel staff around to help? [19:30] only channel staff has weapon? [19:30] evening Peng, btw [19:31] Hi. :) [19:37] anybody seen new Joel Spolsky's project? http://hginit.com/index.html Looks like very nice tutorial [19:38] Hehe, "Mercurial tutorial" rhymes. [19:41] I wonder why people started bashing svn [19:42] Because it's terrible, like Boxbot. [19:43] It's ugly, and its mother dresses it funny. [19:47] Peng, Boxbot is http://gunnerkrigg.wikia.com/wiki/Boxbot ? [19:47] bialix: Yes. [19:47] nobody likes boxbot [19:48] but Martin Fowler does rate svn highly [19:48] and he don't like a Feature Branches [19:48] is not it's funny? [19:50] i'd like to make a co a branch from launchpad, then branch that co, make changes to the new branch, and push the changes to a new branch on launchpad (without effecting the original branch on launchpad). Is that sane/o-able? [19:50] er, do-able [19:51] gregcoit: almost [19:52] well, almost if close. :) what am I missing? [19:52] Sure, though within the confines of that the co isn't doing anything. [19:54] gregcoit: it's a bit hard to read your slang [19:54] bialix: yeah, i'm new to bzr terminology [19:55] gregcoit: you can omit the step of co [19:55] bialix: what I'm trying to do is take a part of a branch and make it into a new branch with history preserved while keeping the original branch unaffected [19:55] co wo nt he lp yo [19:56] kk - I can see that [19:56] Hey! No Swahili in #bzr! [19:57] I hear you the white master [19:57] When you run 'branch', you're creating a new independent branch. Things done to it don't affect the original, unless you do things like 'push' that are designed to doing things inter-branch. [19:57] gregcoit: to extract part of the tree and preserve history.. it's not very easy [19:58] bialix: well, copy might be a betterterm than extract - I don't want to affect the orginal branch contents [19:59] every branch is separate copy [20:00] bialix: ahh, ok, perfect [20:00] bialix: thanks for talking me through this [20:01] may I advice you to read tutorial? [20:01] I'm not really sure I understand your intent [20:06] not a bad idea.... === gregcoit is now known as gregcoit_lunch [20:15] hmmm, Kiln is really attempting to be launchpad competitor [21:13] Except Launchpad doesn't even attempt to enter the "here it is, you can run it on your own server if you like" market === salgado is now known as salgado-afk [21:20] moin [21:35] hi jam? [21:35] morning poolie [21:35] how are things? [21:36] going pretty well === gregcoit_lunch is now known as gregcoit === jelmer__ is now known as jelmer === jelmer is now known as Guest75587 === Guest75587 is now known as ctrlsoft [22:57] hi all [23:04] 'morning poolie [23:10] Morning. [23:15] vila: ping [23:23] hi ctrlsoft [23:28] poolie: fwiw I managed to get colocated branches in a git repo partially working, I can now list and access them using the patch I proposed to bzr.dev and an additional small change to bzr.dev [23:29] ctrlsoft: YOU ARE MY HERO [23:30] RAOF: :-) Please note the "in a git repo", this doesn't work in a native Bazaar branch yet and it's still a prototype at this point. [23:31] The *only* reason I want colocated branches is so that I can switch to bzr as my full-time git client. [23:31] Which means being able to pull & push non-master branches. [23:31] * RAOF imagines a world where reading “git help add” doesn't happen quite so often [23:41] that's great [23:44] jelmer: nice [23:45] Peng: was Peng taken on twitteR? [23:47] hi - I'm a newbie at version control and at bazaar, and have bzr set up from the latest package on an os x server. Can I ask: when I set up a central repository, what are people doing? Setting up a single repository (e.g. organization name "acme") and putting projects under that, or, are people setting up multiple repositories... ? [23:48] RickCogley: repositories are just for storage optimisation: when making a new branch in a repository from another branch in the same repository no data needs to be copied around [23:48] RickCogley: this has several implications: many repositories are the norm - every user usually has a repository on their machine [23:48] secondly, you don't need to worry about them until you are up and running, they are an optional facility. [23:49] lastly, whether you make one for all your projects, or one per project, is totally up to you. [23:49] Although for performance and disk-space reasons they rapidly become less optional if your project is large [23:50] Thanks lifeless and maxb. We are thinking we'd like to have a single central repository so that we can back it up, but I can see that perhaps if you were doing dev for clients, you might want multiple ones to make sure things are kept separate. [23:50] maxb: yes, but that users having to get it 'right' at the start leads to analysis paralysis [23:51] lifeless, ok, so it's kind of "ready, fire, aim" ? [23:51] RickCogley: the use or not of repositories won't affect backups (except for the size of data: as a new user that is unlikely to worry you for a while) [23:51] RickCogley: yup. My advice: *completely ignore* repositories for at least a week [23:51] get used to branching, pushing, pulling, merging [23:52] I'd agree with that for a user. I think a server admin setting up a central server might want to have a think about them up front [23:52] maxb: ONLY if the admin is already fluent with bzr. [23:52] The *only* reason I want colocated branches is so that I can switch to bzr as my full-time git client. <-- oooooh [23:52] maxb: otherwise they don't have the context to make the right decisions. [23:53] I guess the idea is to just get used to it and you can merge things into a repository later... ? [23:53] maxb: and as repositories are just storage optimisation, its easy to add them in later, or remove them. [23:53] and yes, git help whatever takes up way too big a chunk of my time spent interacting with git [23:53] lifeless, ok, I see. [23:54] RickCogley: 'repository' is just a database. If you don't make one, bzr makes one when you run any of 'init' 'branch' 'checkout' 'push' [23:54] RickCogley: let bzr do that while you learn the user facing concepts and concerns. [23:54] RickCogley: /none/ of the every day bzr commands manipulate repositories. They all work with 'branches' and 'trees'. [23:56] lifeless, so in other words, you make a branch anyway when you check it out of a repository (not yet sure how to do that but...) and then the commands just work on your local copy. ? [23:56] you can't check out of a repository; its just a database [23:56] you check out of a branch [23:56] delete the work repository from your vocab; its not needed. [23:56] s/work/word/ [23:56] mathrick: but do you will miss index? [23:56] not at all [23:57] index just gets in my way [23:57] I use git when I have to, not by choice [23:57] bialix: For me, the index is an incovenient version of “bzr shelve” [23:57] it seems there are 2 kinds of people: who loves index and who's not [23:57] *check out of a branch* *check out of a branch* *check out of a branch* :-) [23:57] RickCogley: yup :) [23:57] RAOF: is not it's reverse thing? [23:58] bialix: Which is one of the reasons it's inconvenient :). Mostly I want to commit my changes. [23:58] That's the default case. [23:58] bialix: index is used so rarely it doesn't make sense to make it the central element of everything [23:58] ok ok [23:58] If I want to take some of my changes out, I'll *explicitly* do so, using a (more powerful) explicit command. [23:59] just too many people around me buzzing how's cool index [23:59] actually I use the interactive plugin to do most of my change management [23:59] so I just go bzr ci and pick logical sets of changes until I have zero left [23:59] poolie: pig [23:59] sorry! [23:59] poolie: ping [23:59] lifeless, may I ask: what is the strategy if you want to let visitors to a site download the latest version of a script, like a perl or shell script? [23:59] lifeless: !