[00:18] hi from Argentina, does anybody know how to make this post_commit hook for trac work? : https://trac.m0n5t3r.info/oss/wiki/BzrTracHook [00:32] * awilkins needs sleep [00:35] markh: ping [00:36] jam: hi. Let me guess - asking about binaries? :) Launchpad is giving me grief :( [00:36] just curious [00:36] you mentioned you built them yesterday [00:36] "Sorry, there was a problem connecting to the Launchpad server. " - about 7 times now :( [00:36] ouch [00:36] last time I asked on #launchpad, the advice was basically "keep trying"! [00:36] this is *after* it spends 30 minutes doing the upload :( [00:37] MapMan: I meant to say: from the cmdline, consider doing 'bzr qadd' - that way you get the same GUI as TBZR, but can avoid that bug :) [00:38] thanks man [00:40] markh: is there somewhere else you could upload them? [00:40] yeah - lemme do that. [00:40] I don't know what resources you have available, but I would be happy to try to upload them if I could get to them. [00:40] markh: that helps a lot [00:41] This is only the second time I've struck this launchpad problem [00:41] MapMan: cool - many commands have a 'q' implementation - that is what tbzr uses under the covers. [00:41] bzr help will list them [00:42] err - bzr help commands [00:42] k [00:42] i figured that on my own [00:42] markh: I wonder if it happens because you are in AU, and something times out while uploading [00:42] qcommit etc. [00:42] You're uploading larger files than the rest of us, too [00:43] markh: ima make some bath files for my friends so they can access verything easily and quickly [00:43] jam: likely - maybe my isp struggling. Everything else seems fine. Last I asked on #launchpad though, others were having identical problems at the same time. Who knows... :) [00:47] of course, now I have the l/p upload and the other upload competing the the bandwidth that does exist! I'll wait though, as if I cancel the l/p upload, that then obviously would have turned out the be the one that worked :) [00:53] hello [00:53] hi markh, jam [00:53] morning poolie [00:53] hi poolie [00:53] jam, should i merge bzr.1.7 back to trunk? [00:53] guilhemb was just wonering why lca-shape-merge was not there yet [00:54] poolie: if you want, I'm planning on getting to it, but had about 4 other approved things that didn't get in yet [00:54] poolie: actually, I'm set up here, give me 2 secs and I'll submit it [00:54] because of pqm delay? [00:54] oh, ok, thanks, [00:54] i'll reply to tell him so [00:56] I mentioned it to him on IRC. When do you want to release 1.7.1 final? Mon or Wed? [01:02] somewhere in that range :) [01:08] hi people, i need help with post_push / post_commit hooks, anyone experienced could help? thx in advance [01:14] i'm trying this one: https://trac.m0n5t3r.info/oss/wiki/BzrTracHook but does not get run after i commit local and push to a remote ssh+bzr repo using versions 1.6 ... any clues? [01:21] jam: http://starship.python.net/crew/mhammond/bzr-setup-1.7-1.exe and ...asc - I'll do 1.7.1rc1 now [01:21] rama sorry i need to got in a sec, can you post to the list or a question on launchpad.net/bzr [01:21] markh, thanks! [01:22] rama, are you sure the hook is configured to run for that push location? [01:22] i did setting it on ~/.bazaar/location.conf [01:23] (jam - obviously that means the launchpad upload failed yet again! Not even bothering to try with 1.7.1) [01:26] poolie, i have this on my locations.conf: [/usr/local/bzr/vodo-net] [01:26] post_push = bzrlib.plugins.trac_hook.post_push_hook [01:26] post_commit = bzrlib.plugins.trac_hook.post_push_hook [01:26] where bzr is setup as a repo and vodo-net a branch [01:26] markh, what happened with the launchpad upload? [01:27] last few days been getting an error from l/p after the upload saying there was trouble connecting. [01:27] happened once before and I asked in #launchpad, and they basically said "keep trying" [01:27] :-/ [01:27] so I did :) last time it eventually worked after about 6 goes. [01:27] that sucks, did you get an oops or something? [01:27] or is there a bug? [01:28] rama, i get an error loading that trac page atm [01:28] markh: uploading now [01:28] no oops - just a plain html message asking me to try again, or if it keep happening to ask on #launchpad [01:28] I even checked the html source for clues :) [01:29] (I've closed that tab now - jml helped me out on #launchpad last time) [01:30] s/few days/24 hours/ :) [01:31] and no other indications of network problems [01:31] poolie, strange, it opens here.. anyway, i've been fiddling with its code and found in fact it was done for bzr 0.15 and reading the doc on the site found how to adapt it [01:31] but it does get "registered" into bzr, despite not printing any sign of being run [01:34] i'm doing this: install_named_hook('post_push', post_push_hook, 'trac_hook') [01:35] and post_post hook is the function to be called upon a post push on the remote bzr repo pushing from my box.. if i get it right [01:50] jam: bzr-setup-1.7.1rc1-1.exe and bzr-setup-1.7.1rc1-1.exe.asc there too. Let me know when I can remove them [01:56] hey there... i want to use bzr for my project but i already have 61 revisions in svn.... most of the migrators on bzr's site are for using svn through bzr... i just care about the history.. is there a doc for this ? [01:57] alperkanat: you can just do a "bzr branch" from svn and throw the svn away, or you could use something like bzr-fastimport to do a conversion [01:58] they're pretty much equivalent [01:59] james_w: but i have http authentication at the svn side...? [02:00] markh: did we end up filing a bug about that? [02:01] alperkanat: hmm, I think that may work with bzr-svn, but I can't be sure, sorry. [02:01] jml: no - we just agreed it was transient - someone else was having the same problem at the same time too - then it worked for both of us, so we all forgot it :) [02:02] markh: :) [02:03] markh: for things like this, it's good to ask a question at https://answers.launchpad.net/launchpad/ -- it's easy enough to turn into a bug and means that people who *really* know what they're doing can answer it. [02:03] of course, asking on IRC is also wonderful [02:03] I speak mostly to remind myself :) [02:05] the error page specifically directs you to #launchpad :) the error message is a very generic "sorry, something went wrong" page. [02:08] * jml finds the page in the source code. [02:10] markh: actually, I'll have to follow this up later. [02:10] jml: no probs - its only the 2nd times it happend and jam is doing the upload for me - thanks tho! [02:11] np [02:15] anyone here used svn2bzr before ? [02:15] I was under the impression that it was pretty much superceded by bzr-svn. [02:18] pff i just don't want to lose the history of the svn repository.... and the documentation of all plugins are bad.. [02:18] :S [02:21] alperkanat: bzr-svn won't lose any history [02:21] james_w: i can't even make bzr-svn work [02:21] james_w: make complains about compilation errors at the end.. and bzr co http://repo doesn't do anything :S [02:22] probably because it didn't compile [02:23] james_w: Make sure the directory name is 'svn'. => which directory does it mean for instance ? [02:24] james_w: the directory i copy into the plugins or..? [02:26] you want the bzr-svn plugin directory to be called "svn" [02:26] e.g. ~/.bazaar/plugins/svn [02:26] sorry I can't help you further, I must go to sleep now. [02:27] ok thanks anyway [02:30] any ideas for this error ? => bzr: ERROR: Invalid http response for http://svn.raptiye.org/raptiye/trunk/.bzr/branch-format: Unable to handle http code 500: Internal Server Error [02:32] someone please ? [02:32] alperkanat: it seems that the server is using mod_svn or mod_dav, and can't serve the url (possibly it not exists) [02:33] but there's no .bzr directory inside the repository ? [02:33] if it's a svn repository no [02:34] Verterok: then how am i supposed the import the svn project into a bzr project ? [02:34] s/the/to [02:35] alperkanat: I assume you want to use bzr-svn to do that [02:35] Verterok: yes [02:35] Verterok: i just care about the history.. [02:36] alperkanat: if you have bzr-svn installed, juts: bzr branch http://svn.raptiye.org/raptiye/trunk [02:36] alperkanat: to check if bzr-svn is installed, execute: bzr plugins -v [02:37] Verterok: http://screencast.com/t/0xsL59q8a [02:38] i didn't think it'd be this much painfull :( [02:39] alperkanat: it shouldn't [02:39] alperkanat: please try with: 'bzr -Derror branch' and pastebin the output [02:40] alperkanat: do you want to keep you bzr branch in sync with the svn one? [02:40] no [02:40] i'll delete the svn one [02:41] i don't want any .svn or any other related thing in the project and want to use bzr as vcs [02:41] Verterok: http://dpaste.com/80835/ [02:42] alperkanat: you could try using svn-import (I'm not an bzr-svn expert, but it sounds reasonable) [02:42] i'd just wipe out the svn repo and start over but i don't want to lose the history [02:42] sure, makes sense [02:43] Verterok: svn-import ? [02:43] there's no such command [02:43] alperkanat: it seems that bzr-svn is doing absolutely nothing :(, so it's bazaar trying to find a branch [02:44] it's a bzr-svn command [02:45] alperkanat: try running the bzr-svn tests with: bzr selftest svn [02:45] ok [02:51] Verterok: this will take up long... i'm going to sleep now.. thanks any way.. i'll try my chance tomorrow.. [02:51] np [03:54] How can show a file splitting into several pieces? I wrote a module that I now want in 3 - 4 parts. I did not want the module to be just removed without expressing this change. [04:28] ausimage, about plugins i can't setup the post_push hook to get run when pushed via ssh.. any clue? [04:30] not sure about that one :/ [04:31] * ausimage is only a user, and a recent one at that... [05:34] * ausimage wanders off, but hopes someone could give him insight into how keep history when splitting files up. PM him with answers ;) [06:58] Hi - I'm getting the following error in one of my repositories: [06:58] bzrlib.errors.BadIndexFormatSignature: 8b44a161cf3c0ccead798d1f111c8379.rix is not an index of type . [06:58] How do I fix this? [06:59] I have no idea how it happened :-( [07:39] I think this might have happened because this repo was on a flash drive, which I removed with out doing the safe remove thing. [07:40] I've recreated the repo form other sources. [08:50] hmm [08:51] say i have a bzr repo on a remote host [08:51] the directories/files must always have the same owner [08:52] is there any method i can use to ensure the files are always owned by the same owner:group? [08:52] assuming i'm using bzr+ssh:// [09:01] hi! [09:03] hello [09:04] I'm just running "bzr branches http://code.python.org/python" and it's been hanging for two minutes or so, is that normal? [09:21] I wouldn't expect branches to be able to work over http... [09:24] I'm rather new to Bazaar, can you tell me why? [09:25] and, is there a different way to see what branches are present at that location? [09:29] Well, it has to be able to list directories to search for branches. You can't do that over http. [09:29] Maybe it has enough smarts to process some of the forms of dir listings web servers send out, but I'd be a little surprised. [09:30] You can just pull the URL up in a browser and look; at least some presumed branches are visible there. [09:39] ah, I see. [10:56] I'm trying to wrap my brain around the bzr workflow and need some help. When I 'bzr branch' something, does it also create a repository? [10:57] I'm also not sure I understand the difference between a branch and a repository... [10:57] there is no difference [10:57] there is a huge difference [10:57] :| [10:58] doomster: branches are the entities you work with most, repositories are just a storage area for revisions [10:59] hmmm... [10:59] doomster: wether branch makes a repository or not depends on if you are branching into a shared repository, if not, you're making a standalone branch and it will store it's own revisions in a private repository for that branch [11:00] doomster: but in your workflow, you usually don't care about repositories [11:00] doomster: branches are where the action is [11:01] doomster: this is in bzr terminology, in other systems branch and repository can mean slightly different things [11:01] yes, I guess my SVN background is giving me troubles understanding this.... [11:02] right, svn doesn't really have a branch concept, it's just convention by the users [11:04] doomster: is there a specific piece of workflow you're confused about? [11:04] I just did "bzr init-repository test" and then "bzr log test", but that complained that it is not a branch. Doesn't a repository contain changes? [11:04] LarstiQ: no, I guess I'm just confused by terminology. [11:04] doomster: not as such [11:05] doomster: when you make branches under 'test', those branches will use 'test' as the place to store their revisions [11:05] doomster: that is also true for unrelated branches, so the set of revisions in the repository can form multiple disjunt revision DAGs [11:06] doomster: there isn't a sense of a 'line of development' for a repository, it's just a big bucket of revisions [11:06] doomster: whereas a line of development is exactly what a branch is [11:08] okay, so when I'm working on some project, I have two options. One is to just "bzr branch $URL", then I get a branch with an integrated repository where I can make changes. [11:08] The other option is to first create a separate repository and then create branches within. [11:09] doomster: you can always decide to switch from the one to the other later [11:09] doomster: what a shared repository gives you is storage advantages and faster branching within (since it doesn't need to copy revisions it already stores) [11:09] The idea is that you always "bzr branch $URL". If you happen to be putting it inside a shared repository, it will use that. Otherwise it'll create an internal repository. [11:10] doomster: functionally, there is no difference [11:10] doomster: what fullermd said [11:10] okay... [11:10] You don't [superficially speaking] really care either way, because everything will act the same. [11:10] Okay, let's get to a real-world example... [11:10] (and it's easy to switch things from being shared to being standalone if you come to care in the future) [11:11] This is very different from SVN, where the boundaries of the repository determine what you can do between branches. [11:11] I'm trying to port Python to MS Windows CE. On the way, there is quite some cleanup to do with the 'normal' win32 port. [11:12] So, I guess I will have a handful of separate branches, each for one small step. It would be useful to create a shared repository for those, right? [11:12] doomster: correct. [11:12] It will save you space and time, yes. [11:12] okay. [11:14] I think I'm getting it now. ;) [11:15] LarstiQ: Hi, BTW. Good to see you alive :) [11:15] doomster: and if compiling your project takes a long time/has a big working tree, you might consider using lightweight checkouts and `bzr switch` [11:15] fullermd: thanks, you too :) [11:16] qi [11:16] * fullermd slaps his mouse. [11:16] doomster: this will decrease the cost of the working tree, which when significant is, well, significant ;) [11:17] And it should be a reasonably familiar workflow if you're used to svn. [11:18] It is, however, a somewhat less common workflow with bzr, so you're more likely to snag some rough edges in the implementation. It may be better to skip that part until you get comfortable with bzr. [11:18] (said with no real authority other than my offhand opinion) [11:19] * LarstiQ nods [11:19] when ~/repo is my repository, I will typically have ~/repo/branch-x and ~/repo/branch-y. Can I also create branches somewhere else and connect them to that repository? [11:19] No, they have to be inside it. [11:19] Now, you can (as in the just-described workflow) have the branches there, but have the working tree elsewhere. [11:20] Branches don't "point to" their repository, it's searched for. When you need the repository (to find a revision, or save one, or something), bzr will look for it right with the branch, then in $BRANCH/.., then $BRANCH/../.., and so on. [11:20] doomster: branches have to be within the directory hierarchy, but deeper levels are fine (repo/trunk, repo/win32/cleanup, repo/win32/CE, etc) [11:21] It'll use the first repo it finds along that walk. [11:22] the repository is just the .bzr directory? [11:22] Technically, the repository is the .bzr/repository/ directory. [11:22] The .bzr/ is... I dunno what you'd call it. 'bzrdir' probably. [11:23] yup [11:23] It contains the internal files for the repository, and/or the branch, and/or the checkout, depending on which is at that location. [11:23] so with branches in a repository, there will be repo/.bzr/repository, and repo/branch-X/.bzr/branch/ [11:23] It's sorta a combination in SVN terms of the repository and the .svn/ dirs scattered through the tree. [11:24] okay, so both the repository and each branch have a .bzr dir [11:24] doomster: yes, and each working tree [11:25] Right. It basically means "there's something bzr here, and I've got all the control files" [11:25] btw: can I move a repository around using the shell or do I have to notify bzr about it? [11:26] It depends on whether you have something external referencing it (like a checkout somewhere outside it, or other people trying to look at it) [11:26] If not, you can move it around or tar it up or whatever. [11:26] good. [11:26] Similarly, you can move branches around inside it, as long as they stay in it. [11:27] (or if they're standalone branches, you can move them around anywhere) [11:33] doomster: if you move a branch outside of it's repository, it will do the walk upwards again to find it's repository. If it doesn't find one, it's unhappy, but if it finds one that doesn't contain it's revisions, it's not so happy either. [11:33] doomster: move it back and it's fine again. [11:33] poor unhappy branch!!1! O_O [11:33] (or provide a repository that has the same set of revisions) [11:34] :) [11:34] comparing to SVN, doing a lightweight checkout from a branch is similar to SVN's working copies, right? [11:35] doomster: yes [11:35] doomster: a lightweight checkout will have to contact the branch it is bound to for all operations concerning history (diff, commit, log, etc) [11:36] doomster: but if the branch lives on the same harddisk (~/repo/branch and ~/project/checkout for example), that's ok. Whereas doing that over a 300ms latency link to the other side of the word is less snappy. [11:37] In contrast, a heavy [default] checkout will only have to contact the branch for a few operations (commit, update, etc). The better choice if it's across a network somewhere. [11:38] (You'll find discussions in the docs of more details regarding the low-level implementation of heavy checkouts. Join with me in narrowing your eyes in mild displeasure, and ignoring them :p) [11:43] well, sometimes understanding the internal motivations helps a lot using the tool [11:43] It doesn't in this case. [11:50] bzr is now working more than five minutes on a normal checkout from a local branch... [11:51] How big is it (both checkout and history size)? [11:51] 275MiB for the branch itself [11:53] 40k revisions [11:53] A normal [heavy] checkout makes a full copy of the history. So if it's big, that can take a while. === Spaz is now known as Kittens === Kittens is now known as jmyers === jmyers is now known as Spaz === a_c_1 is now known as a_c_m [14:59] beuno: there has been quite some api changes on Tree I notice. [17:15] what package do I have to install to be able to use the "bzr vis" command? [17:16] bzr-gtk [18:07] LarstiQ, hi [18:07] changes since when? [18:46] is it a known issue that the password prompting for bzr-svn doesn't inject the password into subversion's cache? [18:48] as a side note, the password prompting did work (awesome!!) as did a branch, edit, commit, push cycle, depsite using weird and broken code.google.com svn!! rock!! [19:40] Any loggerhead devs around by chance? [19:40] abadger1999, sure, what's up? [19:41] I almost have loggerhead working under mod_wsgi . [19:41] I'm having one issue with the redirect from /project/ =? /project/changes [19:42] In the code that's [19:42] raise httpexceptions.HTTPMovedPermanently( [19:42] self._url_base + '/changes') [19:43] Instead of redirecting it's giving me this exception: [19:43] [Sat Sep 27 11:04:03 2008] [error] [client 127.0.0.1] TypeError: expected string object for header value [19:44] abadger1999, no tracebacks? [19:45] For some reason, mod_wsgi doesn't give me a traceback there. [19:45] beuno: I found the place in the code by inserting debug statements. [19:46] It's in app/branch.py [19:46] so, _url_base isn't a string? [19:46] can you check to see what it's getting [19:46] ? [19:46] Let me print [19:48] ugh. firefox is being a dog [19:49] it's barking and licking your face? [19:51] chewing on my swap and growling faintly. [19:55] beuno: url_base:/bzr//packagedb/0.3.x: [19:55] bzr is my webpath [19:55] packagedb is the project [19:55] 0.3.x is the branch [19:56] URL I'm hitting is http://localhost/bzr/packagedb/0.3.x [19:56] Oh... type is unicode. [19:56] ah [19:56] odd [19:56] Not sure if that would make a difference [19:57] try and convert it [19:58] Yep, that's it. [19:59] so, any ideas how we ended up with unicode? [19:59] I wonder if this is something that needs fixing in paste's httpexceptions.... [19:59] Nope. There's no unicode in the URL. [19:59] Let me look at the wsgi env [20:00] 'SCRIPT_NAME': u'/bzr//packagedb/fchiulli' [20:00] (different branch) [20:00] * beuno pokes mwhudson_ ^ [20:01] it's 8am-ish for him, but he'll read the scrollback :) [20:01] So mod_wsgi is adding that as a unicode object instead of a byte string. [20:01] Cool. [20:01] we should probably add a check [20:01] could you file a bug for it? [20:02] Ah ha. [20:02] http://pythonpaste.org/news.html#id2 [20:03] Exceptions with unicode messages don’t cause the collector to fail. [20:03] That was added in paste 1.7 and I'm using paste 1.6 ATM [20:03] beuno: k. I'll file a bug. [20:04] abadger1999, interesting [20:04] hardy has 1.6 [20:04] beuno: Are you guys interested in my script to start loggerhead using mod_wsgi? [20:04] abadger1999, thanks for debugging it! [20:04] abadger1999, yes we are! [20:04] It's mostly removing things from the start-loggerhead script. [20:04] Cool :-) [20:04] upload the branch, file a merge request [20:04] make sure it doesn't break any tests [20:04] Okay. [20:06] How do I run the tests? [20:06] Ah nose [20:07] beuno: Thanks! I'll get the bug and the branch up today. [20:07] abadger1999, fantastic, thank you :) [20:36] * beuno is off to catch another plane === rocky1 is now known as rocky [21:26] hello === astrobunny|afk is now known as astrobunny === astrobunny is now known as astrobunny|afk === astrobunny|afk is now known as astrobunny === astrobunny is now known as astrobunny|afk === astrobunny|afk is now known as astrobunny