[00:11] Yeah, we be hardcore. But how come it's not listed in Gutsy's /etc/services? [01:28] how do I revert to a specific revision? [01:30] revert -r revno [01:31] where would that be documented? :) [01:31] bzr help revert [01:31] bzr help revert [01:31] as well as the man page and IIRC users guide too [01:31] "Come join #bzr, the channel with attitude for the tool with attitude" [01:32] :) [01:32] cool :) [01:38] thanks guys, not sure why I couldn't find it [02:59] I have some questions. [03:00] Is it costly to push from a repository (init-repo) to a branch not in the repo? [03:01] Rotund: you mean, push a branch which is sitting in an existing repository to somewhere else? yep [03:01] Rotund: you can push it anywhere you want, including other shared repositories [03:02] oh. I should read *all* the words [03:02] Would that be more expensive than pushing from one branch not in a repo to another that is not in a repo [03:02] Rotund: so, if you're pushing to an existing branch, only the new revisions will be sent [03:02] so it can be very fast [03:02] if the other side has most of the revisions already [03:03] I'm not finding that to be the case [03:03] igc, hi? [03:03] and you can even push into remote repositories, so even if you're creating a new branch it can be fast if the repository has most of the revisions. [03:03] hi poolie [03:03] I have a collection of drupal sites I'm managing with bzr [03:04] I have a "base install" and each site gets a branch. [03:04] I make changes to the base. Go to each website and merge [03:05] FAST great (they are in a shared repository). [03:05] igc, would 9am tomorrow be ok for a standup meeting? [03:05] Now pushing those changes to an already created repository is pretty costly [03:05] actually, 9am daily starting tomorrow? [03:05] poolie: my 9am or your 9am? [03:05] yours [03:05] yes [03:06] cool [03:06] I was expecting one today :-) [03:06] I've tried sftp, bzr (running a server), and bzr+ssh. All seem to take 2+ minutes to push just a module change [03:06] so you've got something like site-repo/{base,website1,website2,website3} ? [03:06] poolie: this channel? [03:06] radix: exactly [03:06] Rotund: and it gets slow when you push what to where? [03:07] I now need to push from my local (my computer) branch to the server's branch [03:07] Rotund: ah, do you have this structure mirrored on your local computer and on the server? [03:07] yup [03:08] I don't have it bound (is that the term?) [03:08] Rotund: ok. and are you extra-sure that both sides have the shared repository in the parent directory? there should be a .bzr there [03:08] would that help? [03:08] Rotund: binding shouldn't matter [03:08] i.e. site-repo/.bzr [03:08] only one side has the shared-repo [03:09] Rotund: I'm not exactly sure what kind of operation you're doing, but making both sides shared can vastly increase speed of some operations [03:09] is there an added penalty for having one side shared and the other not? [03:09] notably, pushing up new branches [03:09] I'm just pushing changes though (version 8->9 for instance) [03:09] hm. to an existing branch? [03:09] in that case, it should be fast no matter what your shared repo setup [03:10] That's what I would've thought [03:10] 2 minutes definitely sounds weird, for just one revision. are the revisions large? [03:10] 933K [03:10] Rotund: is it made up of many small files? [03:11] it is 42 files [03:12] 933k isn't exactly a light revision, but I guess this is the point where my knowledge reaches its limit [03:12] I've done things like 1 big file and it'll end up 20 min + [03:12] Rotund: what kind of bandwidth and latency do you generally get between these hosts? [03:13] big = ~3 MB [03:13] I'm not sure. It's dreamhost [03:14] gonna test... one sec [03:15] ouch. 50KB. [03:15] Hmmm. That's awful. [03:15] that's plenty fast for uploading 3MB :) [03:16] way faster than 20 minutes, anyway [03:16] Though, I'm still getting the file in ~6 min [03:16] (Testing a 25 MB one) [03:16] I just walked away for that [03:17] It's typically the fetch stage that kills me [03:17] (stage 1) [03:18] Is that the actual transfer? [03:19] If so, I guess I could live with it. [03:19] I assume it is [03:19] okay. Well, thanks [03:20] BTW: I'm all hyped for PyCon this year =) [03:20] Mph. I guess I'm probably not going to be there. [03:20] They have a week before and after for sprinting, right? [03:20] Why not? [03:21] How good is the bzr to CVS export tools? [03:21] Can they relied upon in an automated way... [03:22] You mean converting from bazaar to CVS? [03:22] Yes. [03:22] That was my question [03:22] Rotund: a friend of mine (who would also be at pycon were it not for this) is getting married :) [03:22] radix: were they there this year? [03:23] Okay, well, I have no idea about that. [03:23] Man, I may have to find someone else to bug all the time now =( ;) [03:23] Rotund: glyph was. [03:24] Rotund: heh. it was pretty good convo last year. [03:24] glyph's getting married? [03:24] righto. [03:24] Rotund: I hope your game projects are going well, btw :) [03:24] radix: Wow. And he chose PyCon to do it during? Why am I hearing the cracking of whips over and over ;) [03:25] heh heh [03:25] radix: so is most of the Twisted guys gonna be MIA? [03:26] I imagine there will be a weakened presence this year. [03:26] radix: I wish... unfortunately, real world work is taking all my time [03:26] Rotund: stupid real world. [03:26] radix: Though, I've been doing quite a bit of Python training and evangelism [03:27] radix: Many of our customers are switching to python [03:27] excellent [03:27] warren: are you saying the conversion tools, or the one that lets a CVS user checkout files from a Bazaar branch [03:27] jam, conversion [03:27] (AFAIK they are readonly, but reasonably good) [03:27] ah, didn't know the latter existed [03:27] I don't know of any specific converters that start with Bazaar and go to CVS [03:27] oh [03:27] I suppose Tailor would be an any-to-any [03:28] radix: Honeywell, Rockwell Collins, (testing for both) and most of the tools used for the 787 are Python (though I don't think I've seen Twisted) [03:28] https://launchpad.net/bzrcvsserve [03:29] radix: well, that real world thing is calling. have a good one. [03:29] you too man. take it easy. [03:30] Rotund: fun to hear about Rockwell Collins, I interned there for a summer for the Large Format Displays on the 767 [03:31] IIRC they were looking at Python because its license was better than Perl [03:31] (being BSD rather than a form of GPL meant it was more obvious that it wouldn't invade accidentally) [04:04] * igc lunch [07:10] New bug: #163995 in bzr "Re: bzr register-branch is undocumented in manpage" [Undecided,New] https://launchpad.net/bugs/163995 [08:23] So notwithstanding the email thread on the topic, should I be telling people to `bzr branch` or `bzr clone` when downloading the sources (for the first time)? [08:23] branch [08:23] here's how I describe it [08:23] 'if you want a mirror that you won't edit, do 'bzr checkout'' [08:24] 'if you want to start a new line of development, to hack on the code yourself, do 'bzr branch'' [08:25] (for the former, use bzr update to keep the mirror up to date, for the latter use bzr merge to incorporate changes from the origin) [08:33] I wonder if people would complain if "bzr fork" was an alias for "bzr branch"? [08:40] fork should do more than just branch, though. It should send of a fiery "screw you guys, I'm going home" mail to the project's mailing list. [08:40] Obviously another address we need to store in branch metadata... [09:03] lifeless: (actually I was asking about `clone` vs `branch`, not `checkout` vs `branch`) abentley was pretty adamant that clone and branch behave differently, which was rather an interesting surprise. [09:05] Meanwhile, you know how you look at the generated Index listing on a web server for a directory with a Bazaar Branch present, but no Working Tree, and it looks to a lay person like nothing is there? I had some fun this afternoon and scratched an itch: [09:06] http://research.operationaldynamics.com/bzr/java-gnome/mainline/ [09:11] AfC: neat [09:11] for launchpad we were thinking of running codebrowse from the same url space as the branch [09:12] mwhudson: yeah, that works too [09:12] AfC: abentley said that the internals differ for the 'sprout' and 'clone' apis. [09:12] mwhudson: I had a darcsweb installation before, but at a different URL space [09:12] AfC: at the ui clone is an alias for branch. [09:13] lifeless: oh. Ok. I will stick with clone, then. [09:13] AfC: saying 'branch' is really better IMO [09:13] lifeless: (he really made it sound like there was some difference having to do with Working Trees or soemthing) [09:13] AfC: it indicates the semantic action that takes place. [09:14] 'get' is the one that always throws me for a loop. [09:14] lifeless: personal reasons: 1) I find the term branch getting overloaded, especially with new users: [09:15] "run the bazaar branch command to get a branch of the bazaar branch which is then a branch that you can branch yourself" [09:15] AfC: I did a mini-rant on that Way Back When... [09:16] lifeless: personal reasons: 2) many of my new potential hackers at the moment are Git aficionados; making the very first command they see a bit more familiar helps get them going. It's a little thing, but it seems to help. [09:16] [and I'm hyper conscious of trying to present them a good impression] [09:17] fair enough [09:18] lifeless: had a really interesting run through that with Carl Worth via phone the other day. He'll be down at LCA in Melbourne; I'd like to encourage you and Martin and whoever else to have a chat with him then using him flailing at [as it happens my project] as a case example for someone who is _really_ deep into Git but not a shit when dealing with other people. [09:18] i alternate between get and branch randomly i think [09:18] more likely to use 'get' on a remote branch [09:19] mwhudson: that makes sense too [09:42] lifeless: pong [09:42] LarstiQ: where is your subtrees branch ? [09:43] LarstiQ: It keeps coming up, and no-one knows where to go to get your current state. [09:43] LarstiQ: so I'd really love it if you made it visible - it doesn't matter if its not functional, just making it accessible++. [09:44] lifeless: https://code.edge.launchpad.net/~larstiq/bzr/nested-trees may not be entirely up to date, just a sec [09:44] thanks! [09:44] * LarstiQ needs to fix the pending work and actually commit something [09:45] small commits++ [09:45] lifeless: I'm spending 2 hours on that now [09:45] LarstiQ: yay! [09:45] I'm about to sleep, so gnight! [09:45] lifeless: I cognitively entirely agree with small commits++ [09:45] lifeless: good night [09:47] i think bazaar don't have many starting level tutorial ! [09:48] i really struggled to start with [09:49] jayesh_: what are you having trouble with? [09:50] to understand DRCS [09:50] i am an svn user === weigon_ is now known as weigon === mrevell is now known as mrevell-lunch === mw|out is now known as mw [13:24] Hi. Is it possilbe to downgrade branch/checkout to given revision ? How can I do it ? [13:24] What do you mean by 'downgrade'? [13:25] Change the files? Change the basis tree? Discard later changes semi-permanently? [13:25] fullermd: checkout source is now under -r 100 [13:25] I did bzr update [13:25] but no I want have checkout at revison 50 [13:25] becouse sth went wrong [13:26] http://bazaar-vcs.org/ConfiguringBzr#head-8b89e546875b2447e015c0b83e423ef4d99ce6e0 refers to "when-required" "Sign newly committed revisions only when the branch requires signed revisions." [13:26] how does one set a branch to require it [13:26] I can't find it [13:27] matkor: Well, still, kinda a question of what you're trying to accomplish. Are you wanting to just look at/play with the files as of rev 50? Are you wanting to craft up a rev 101 that looks just like r50? Are you wanting to throw away r51-100 and start over from 50? [13:29] I want temporarly have rev 50 [13:29] until I fix problems and commit more [13:29] revisions [13:30] lets say 105 which fix poblems [13:30] Well, sounds like using 'revert' is probably your closest bet. That'll just change your WT files, not anything else. [13:31] anyway what about the signing? [13:31] :) === mrevell-lunch is now known as mrevell [13:32] AnMaster: Search me. I think there's some branch config option that may do it, but last I heard the gpg code had a lot of exposed innards and spiky edges in those places. [13:33] I see, I searched all docs, googled and so on [13:33] I really need to enforce signing however :/ [13:33] Might try asking jam when he shows up (couple hours, probably); I think he knows that part of the code reasonably well. [13:34] hm, couple of hours, argh. not here then probably [13:34] i think you edit gpg_policy or something in locations.conf [13:35] hm? isn't locations.conf for the old way of storing parent branch and such? [13:35] * AnMaster thought it was stored in the branch nowdays [13:35] if it matters: I'm using a shared repo for all my branches with the format "dirstate-tags" [13:36] (and I do use tag feature) [13:36] fullermd: Thank you very much ... revert seems ok [13:36] fullermd: To get latest reviosion again. I do : bzr revert ; bzr update, right ? [13:37] matkor: Just the 'revert' by itself should do it (unless more new revs showed up in the branch of course) === cprov is now known as cprov-lunch [15:01] i just created an experimental branch in bazaar.. and i am trying to push to that.why it is taking long time? [15:02] is it not pushing a diff ? === bigdo1 is now known as bigdog [15:49] fullermd, hm you there? [15:49] how long until that "jam" shows up do you think? [15:58] hi, can use a command to use my version of all conflicts? === cprov-lunch is now known as cprov [19:14] moin [19:33] hey lifeless [20:32] morning bizarre dudes [20:33] hi === cprov is now known as cprov-away [21:34] is the push-and-update plugin still the "right way" to maintain a working tree on a sftp remote location? [21:35] floam: yes [21:36] ok. I thought maybe it was replaced by something smarter since I've been getting deprecation warnings every time I use it since 0.90 and I've seen no updates to it [21:36] perhaps you could file a bug on it then ? [21:36] well, it works fine [21:36] I just get this every time: [21:37] /Users/floam/.bazaar/plugins/push_and_update/push_and_update.py:119: DeprecationWarning: bzrlib.transport.split_url was deprecated in version 0.90. [21:37] yup, please file a bug. [21:37] ok. [22:05] check_signatures = always is invalid, right? [22:05] Arrgh. [22:05] I mean create_signatures. [22:05] check_signatures uses always but create_signatures uses "require". [22:07] Oh. I do mean check_signatures. [22:08] Apparently I needed more sleep. [22:09] Anyway, I was asking because the configuration.txt doc gives an example of check_signatures = always once but usually uses "requires" and only lists "requires" as valid. [22:09] Err, "require". no s. [22:22] morning [22:22] igc: good morning [22:23] hi jam-laptop === jam-laptop is now known as jam [22:23] silly Colloquy won't let me change the default username [22:23] ... :( [22:24] I blame Mac's "don't have an Ok button" design philosophy === asak_ is now known as asak [22:57] lifeless, igc: so are we getting ready to do the conference call? [22:57] i don't see poolie [23:00] jam: there he goes ^^ [23:00] hi poolie [23:00] bye everyone [23:02] thumper: will you be joining us? [23:02] jam: no, got to go and collect my daughter [23:02] pfff, weak [23:02] you make a big deal about sitting in [23:02] :) [23:02] and then back out :) [23:02] I was just checking [23:13] poolie, igc, lifeless: Are we meeting now? It was supposed to start 10 minutes ago [23:13] well 13 by my clock [23:13] jam: we did :-) [23:13] shoot [23:14] i pinged at 3minutes till and nobody said anything [23:14] I'm glad it went quickly, though [23:14] it was a conference call - just finished [23:14] sure, but I don't like calling phone numbers cold... [23:14] when I know there are people online [23:15] anyway, sorry I missed it [23:15] poolie is waiting for lifeless to arrive so there may be a follow-up shortly if you're still around [23:16] was there a conference code number? [23:16] jam: anyhow, for me today: 1. blackbox tests for switch; 2. User Guide [23:16] or is that just Martin's room [23:16] # [23:16] I think we need to address the community pushback about 1.0 [23:16] yes [23:17] which probably isn't a 15-minute phone call [23:17] agreed [23:20] anyway, I actually need to get going [23:20] I'll thank you in advance for all your help [23:20] (since I'm gone for Thanksgiving :) [23:21] * jam => away [23:34] ola! [23:36] jamesh: enjoy! [23:36] jam: Enjoy! [23:40] enjoy what?