[00:00] more info later [00:00] igc: thank you! [00:00] bbl - few errands to run [00:03] are there plans to add svnmerge.py-like merge blocking to Bazaar? [00:03] 'merge blocking' ? [00:03] spiv_: are you coming down to the vibe too? [00:04] sort of like bzr merge --record-only [00:04] just a little nicer [00:04] what would it do differently [00:05] I'd like to be able to run bzr block [revision] and bzr unblock [revision] [00:06] lifeless, you rock, open_index_branch solved all my problems [00:06] pushing a more usable branch now :) [00:07] thats fantastic [00:07] Jc2k: have you chatted with beuno about what you wanted for loggerhead @ gnome ? [00:11] lifeless: not [00:12] Jc2k, feel free to whenever you want [00:12] ythx beuno, neet a while to come round.. [00:13] lifeless, pushed LH + search changes. It now needs code to index the branches and do something reasonable if it hasn't been indexed yet [00:13] I would reverse the order there :P [00:14] beuno: e.g. do something reasonable is first thing [00:14] mwhudson, now that I got this off my mind, I'm going to start removing revision caches, unless you think there is something else on the top of the list [00:14] bbiab, -> train [00:14] lifeless, right, seems reasonable :p [00:17] beuno: no, that seems like a good thing to do === rockstar_ is now known as rockstar === W_ is now known as _W_ [01:09] 'evening [01:13] hi jelmer [01:36] mwhudson, lp:~beuno/loggerhead/remove_caches [01:36] I think I covered everything [01:36] removed revision and full text caches [01:36] just left the files changes one [01:36] would you like me to send that in as a patch to the list? [01:37] I'm going to go back to the clean url branch and tweak a few places where it still uses revids [02:26] if I want to run bzr update without CD-ing into the directory, how would I do that? [02:28] "bzr update some/directory"? [02:30] hmm, cool [04:00] where is the official 1.4 branch? [04:00] * thumper can't find it [04:17] Bazaar 1.4? [04:17] http://bazaar-vcs.org/bzr/ [04:17] * Peng wanders off. [04:39] Peng: ofcourse, thanks [04:39] I was looking on LP [04:44] thumper: really 1.4? [04:44] back btw [04:44] poolie: I'm looking for the version where pqm stopped working [04:49] and did you find it? [04:49] uh it's not there [04:49] i'll register it [04:49] hey the lp ui for this is now pretty understandable :) [04:49] done [04:51] thumper: interesting, someone still using bzr 0.90 (which is pretty old) got confused in trying to use an lp: url to push to launchpad [04:52] poolie: yeah, saw it [04:53] poolie: somewhere between 1.4 and 1.5 the function that pqm uses in its tests was removed [04:56] thumper: which function? [04:56] spiv: https://bugs.edge.launchpad.net/pqm/+bug/240288 [04:56] Launchpad bug 240288 in pqm "PQM tests do not support bzr 1.6" [Undecided,New] [04:56] cannot import name smart_add_tree [04:56] thumper: you probably want tree.add() [04:57] or tree.smart_add(), actually [04:58] are we going to kill support for older bzr versions by changing to that? [04:58] and do we care? [04:59] they can use old versions of PQM, right? :) [05:00] What jamesh said. === spiv_ is now known as spiv [05:02] thumper: pre 1.0 versions I'd expect [05:02] thumper: MutableTree.smart_add has been present since at least 0.18rc1 [05:03] ok [05:03] I'll look to fix this then :) [05:04] although perhaps tonight [05:09] beuno: so how does it perform sans cache? [05:09] hi guys [05:10] poolie: did you want a call today? [05:11] lifeless, yeap, revision cache was only useful for old versions of bzr [05:11] it's very nice to bring up LH and now have it kill my CPU until it caches all the revisions [05:12] 'not' ? :P [05:12] heh, yes, "not" [05:12] I finally bought a wireless router for my place, so I'm experimenting with new places to work [05:13] hence my increased typos :) [05:13] nice [05:13] as for caches, the only remaining one is one for files changed, which only gets filled when needed [05:14] the difference between having it and not is still too big [05:14] k [05:14] lunch apparently, bbiab [05:46] poolie: am I right to merge the rules-based preferences stuff or do you want to wait until after stacked branches and VFs land? [05:58] igc: is that abentley's stuff? [05:59] thumper: no, part of my work on end of line support [05:59] igc: ah, ok [05:59] igc: how are the stacked reviews going? [06:01] thumper: I did a round of them last week and poolie/jam/others have been reviewing my reviews [06:01] :) [06:01] quite a lot of lines as I understand it [06:01] yeah - and the VersionedFiles stuff is another 17k worth of patch [06:02] many deletions in there though (which is good of course) [06:02] * thumper nods [06:20] back === i386_ is now known as i386 [07:02] more trains... [07:07] from chatswood? [07:14] Canonical needs to start supplying teleporters. [07:38] Peng: that would be handy [07:39] Bazaar t-shirts are nice, but teleporters would be *awesome*. [07:58] night all [07:59] igc: g'night [09:12] bob2: milsons point [09:13] * Jc2k vaguely remembers IRC at sleepy o clock [09:14] bueno: my main interest is theming loggerhead so that it has the GNOME look you can see at http://bzr-mirror.gnome.org/ [09:15] lifeless: would you always suggest bzr smart server over vanilla HTTP if its possible? [09:15] i havent enabled it for bzr-mirror.gnome.org yet [09:16] Jc2k, that should be easy enough. Have you set it up yet? [09:17] I would hold off [09:17] more in a bit, I have to run [09:17] beuno: i've not yet, i've been scared off by earlier mention of its dependencies [09:19] Jc2k, we're trimming them pretty quickly, and I'd expect a lot of improvements to go into it in the following weeks [09:19] awesome :) [09:19] currently, IIRC, it's turbogears, python-sqlite and simpletal [09:19] thats not so bad [09:19] nah, we just want to make it better :) [09:20] :) [09:20] anyway, I'll be happy to help out with getting it running and theming it properly [09:20] so let me know when you take the plunge [09:20] i'm readying a big bzr-mirror.g.o push [09:21] so soon :) [09:22] cool, I'm usually around, so just give me a ping [09:26] beuno: where do the most recent instructions live :) [09:27] Jc2k, hrm, they don't :/ [09:27] updating the setup and README is on my ToDo list [09:28] I'll try and get that done tomorrow and bribe mwhudson into getting that intro trunk, or I'll post it on the bazaar wiki somewhere [09:28] beuno: are you up early or late? :) [09:28] mwhudson, late. Aaaaaaaaaaalways late [09:29] (it's a holiday here tomorrow) [09:29] beuno: ah [09:30] beuno: excellent, i wont start just now then :) [09:31] Jc2k: the dependencies will probably shortly change to python-paste rather than turbogears fwiw [09:31] Jc2k: when you talk about a "big push" what's the timeframe? [09:31] a few days, a couple of weeks, ... ? [09:34] as soon as i can, really [09:35] ok [09:36] mwhudson, trunk + cache removal should work pretty decent for now, no? [09:36] probably should get that in before the clean URL bit [09:37] yeah, i'm not feeling intelligent enough to review the clean url stuff [09:38] because i think the branch changes behaviour in some ways, but i'm not sure whether it makes more sense before or after :) [09:39] ah, right, that happend to me too :) [09:39] I had to go back and forth a few times before I could tell the differences [09:39] beuno: remove_caches seems to be based on the cleaner urls branch? [09:40] mwhudson, uhm, yes. Although it shouldn't really change much from trunk, as it doesn't touch templates at all [09:41] I can send you a patch based off trunk if you'd prefer [09:46] please [09:46] * beuno pets his clean urls branch and starts making use of bzr's magic [09:47] * Jc2k watches [09:51] mwhudson, sent [09:52] Jc2k, I'd say it's worth waiting a few days for these changes to settle in. Does that work for you? [09:53] beuno: sure, i'll need at least that long to work out my jhbuild patch with the maintainers [09:54] Jc2k, cool, I'll make sure I ping you, and help you get through the initial setup :) [09:54] beuno: changecache.py looks a bit mangled [09:55] beuno: in particular i'm pretty sure FakeShelf.close should not look like that :) [09:55] * beuno looks [09:56] I shouldn't of changes anything for FakeShelf [09:57] argh [09:57] vim screwed with me again [09:58] i also think that bits of fakeshelf can go [09:58] like update [10:05] beuno: can loggerhead list all of the repositories in a folder, or am i responsible for that? [10:05] (we have a lot of modules, so (2) is probably better?) [10:06] Jc2k, you can specify a folder and loggerhead looks for al branches in it [10:06] or, manually [10:06] as you wish [10:08] something i am looking to encourage with my rewrite of this end of this things is for serious end users (e.g. gnome) to write little bits of code for this sort of thign === mwhudson_ is now known as mwhudson [10:08] so if we let loggerhead do it, would it show all branches over *all* projects, or will they be like seperate instances? [10:08] so if you want to have a list of repositories in a text file or something, that should be easy [10:09] Jc2k: not sure i understand what you're asking? [10:09] well, if you look at http://bzr-mirror.gnome.org/ [10:09] there are many individual repositories [10:09] and each folder has many branches [10:09] oh right [10:10] how does loggerhead work in this setup, basically :) [10:10] not very well :/ [10:10] :D [10:11] mwhudson, re-sent without the nonsense code and removing update too [10:11] you can point loggerhead at the directory containing the repositories, but the results won't be very pretty [10:12] can i have a loggerhead for each repository? [10:12] * mwhudson hmms [10:12] surely should be possible [10:12] i have a script that is updating the mirrors each time someone checks in to SVN [10:12] can extend it to poke the loggerhead bits too [10:12] and update some kind of master index [10:13] I wonder how hard it would be to fix LH so it *won't* get confused with multiple levels of branches [10:14] oh, would loggerhead get confused by the SVN style layout? [10:14] beuno: easier on my wsgi-ify branch :) [10:14] e.g. $MODULE/{trunk,brances/$BRANCH,tags/$TAG} ? [10:15] yes, basically [10:15] mwhudson, heh, well, even more reason to land that ASAP ;) [10:16] Jc2k: what time is it where you are? [10:17] 10:16 AM [10:17] oh ok, so you'll be around for a bit :) [10:17] yeah :) [10:17] got another 12 hours in me, aside from lunch and commuting :p [10:17] i'll see if i can slap something together you can test [10:18] wicked :) [10:20] beuno: all those lovely -s :) [10:20] beuno: i think i can add a few more though :) [10:23] heh [10:23] go crazy [10:23] the less the merrier! [10:24] man util.Container needs to be stabbed [10:26] ah, get_revision_history_matching should go to [10:26] *too [10:27] ah yes [10:27] what happens now for a text search? [10:28] it always returns "no results found" [10:28] although, at the rate lifeless's plugin is going, I think we should be able to make that plugable pretty soon [10:28] heh heh [10:28] fair enough :) [10:30] beuno: what do you need ? [10:31] lifeless, for now, just time to work it into LH properly [10:31] I didn't want to imply that *you* had work left to do :) [10:31] * mwhudson is confused [10:32] beuno: do you know where fix_year went? [10:32] * beuno looks [10:32] i think i shall use bzr-search to find out :) [10:33] :) [10:33] it seems it's gone in trunk too... :/ [10:34] beuno: right... [10:35] mwhudson, seems revno 128.18.2 [10:35] uh uh uh [10:35] http://pastebin.ubuntu.com/20570/ :( [10:36] -lines in patches are not yet indexed [10:36] though I think they would be interesting to index [10:36] uh, they sure would be here :) [10:37] beuno: what branch is that revno in? [10:38] file a bug :) [10:38] mwhudson, trunk [10:38] revid: michael.hudson@canonical.com-20080310040805-sp0j98ristv9znvm [10:38] lifeless: for the bzr-search thing or the bzr traceback? [10:38] beuno: uh, no? [10:39] - params = { 'file_id': navigation.file_id } [10:39] + params = { 'filter_file_id': navigation.filter_file_id } [10:39] is all in that revid [10:39] well, bzr-search is lying then... :/ [10:39] loggerhead/util.py in revision 'michael.hudson@canonical.com-20080310040805-sp0j98ristv9znvm'. Summary: 'def fix_year(year):' [10:40] mwhudson: for - lines [10:40] lifeless: ok [10:40] ah, it was the date formatting changes [10:41] beuno: bzr cat -r revid:michael.hudson@canonical.com-20080310040805-sp0j98ristv9znvm loggerhead/util.py | less [10:42] lifeless, ah, right, it's in the file [10:42] utils.py was changed at that revision, so it looks in there. Makes sense [10:43] mwhudson, you found where it was changed? [10:43] beuno: yeah [10:43] beuno: r152 [10:44] oh, a while ago [10:44] beuno: it gets false positives where a snapshot occurs [10:46] ok, I'm starting to make even less sense. It's probably a good idea for me to call it a day [10:47] Jc2k, I'll read the backlog and see what mwhudson cooked up for you, and we can talk theming once it's up and running :) [10:47] mwhudson: you had a traceback ? [10:48] lifeless: http://pastebin.ubuntu.com/20570/ [10:48] beuno: good night [10:48] good night, beuno [10:48] g'nights! [10:49] mwhudson: file a bug on that traceback too [10:57] oh, it's in before:$invalid_dotted_revno [10:59] lifeless: https://bugs.edge.launchpad.net/bzr/+bug/240346 [10:59] Launchpad bug 240346 in bzr ""exceptions.ValueError: get_parent_map(None) is not valid" in before:$invalid_dotted_revno" [Undecided,New] [11:04] * mwhudson merges remove-caches to trunk [11:09] mwhudson: "if revid_list and len(revid_list) > 0:" <-- Wouldn't revid_list evaluate to false of its length was 0? [11:09] Peng: uh, yes indeed [11:09] oh well, can't get rid of all the cruft in one go i guess :) [11:15] Jc2k: how much do you mind being a guinea pig for half-baked software? :) [11:16] mwhudson: as long as it runs uninstalled, then i know no limits [11:16] good good [11:17] did i mention the server is running etch.. [11:17] can you get simpletal and paste installed? [11:17] or at least on $PYTHONPATH? [11:18] i havepython-paste 1.0.1-1 [11:19] and simpletal 4.1-6 [11:28] great [11:45] lifeless: about the error i was getting before ("bzr: ERROR: Repository KnitPackRepository('https://myusername@server.org/bzr/.bzr/repository/') is not compatible with repository KnitPackRepository('file:///home/tro/code/asdf/.bzr/repository/')"), yes, i ran bzr info -v on the branch repo i'm pushing to. it's as i posted before [11:46] tro: thats really strange then. [11:46] tro: I'm a little tied up now, perhaps you could file a question on launchpad and I'll put together some python to try and diagnose this in more detail [11:48] lifeless: will do. thanks for helping so far. :) [12:01] grr [12:02] time to add DontZap to the x config again... [12:09] Jc2k: ready to test something? :) [12:10] mwhudson: go on then [12:11] Jc2k: grab this branch https://code.edge.launchpad.net/~mwhudson/loggerhead/wsgi-ify [12:11] Jc2k: and run it from the directory that contains the repositories [12:11] then browse to port 9876 [12:12] hi, I'm running a bzr checkout of an svn repo, I made 2 local commits and then updated. Is there a way to "reapply" the pending merges that it has created so that I can do an svn-push or similar. [12:13] basically I want to save the commits and their logs as separate entities [12:13] mwhudson: i need turbogears? [12:13] Jc2k: i hope not :) [12:14] pkg_resources.DistributionNotFound: TurboGears [12:14] File "/home/john/wsgi-ify/start-loggerhead.py", line 4, in ? [12:14] pkg_resources.require("TurboGears") [12:14] Jc2k: oh [12:14] * Jc2k removes said require :) [12:14] don't run start-loggerhead [12:14] sorry [12:14] oh [12:14] run path/to/loggerhead/wsgitest.py [12:14] sorry, it's getting a bit later here [12:14] hehe [12:15] my paste is too old [12:15] i guess [12:15] File "/usr/lib/python2.4/site-packages/paste/translogger.py", line 100, in make_filter [12:15] from paste.deploy.converters import asbool [12:15] ImportError: No module named deploy.converters [12:15] or rather [12:15] my paste is broken [12:15] maybe you need paste.deploy installed? [12:15] but yeah [12:16] ahh, python-pastedeploy :) [12:17] does it only serve on localhost? [12:17] hm, looks like it :) [12:17] change the host='127.0.0.1' to ... something [12:17] at the end of the file [12:17] kk :) [12:17] i guess host='' will work [12:18] hm maybe an external ip would be better [12:19] i had to put host='bzr-mirror.gnome.org' [12:19] http://bzr-mirror.gnome.org:9876/ [12:19] or host=* perhaps [12:19] Jc2k: oh, you want to remove the EvalException wrapper btw [12:19] wassat? [12:19] sorry meant to remove that before committing... [12:20] Jc2k: insecure debugging aid :/ [12:20] well, its plain, but its there :P [12:20] yeah, the file listing is, uh, simple [12:21] what branch is Jc2k running? [12:21] http://bzr-mirror.gnome.org:9876/yarrr/trunk/changes?q=walters didn't do much :P [12:21] lifeless: https://code.edge.launchpad.net/~mwhudson/loggerhead/wsgi-ify [12:22] mwhudson: where is this wrapper i need to delete :) [12:22] oh, i didn't merge remove-caches either [12:23] Jc2k: http://pastebin.ubuntu.com/20594/ [12:24] ahh [12:24] :) [12:25] cool :) [12:25] so this is all totally horrible code, but it works, eh? [12:26] seems so :) [12:26] * mwhudson is practising his kiwi dialect [12:26] i can kill it, though :) [12:26] http://bzr-mirror.gnome.org:9876/tracker/branches/indexer-split/revision/svn-v3-trunk0:13f93fa3-e725-0410-a08d-a391da2ecc01:branches%2Findexer-split:1537?start_revid=svn-v3-trunk0:13f93fa3-e725-0410-a08d-a391da2ecc01:branches%2Findexer-split:1537 [12:27] hm, urlquoting, what's that? [12:28] hmm? [12:28] * Jc2k pops out for lunch [12:28] i need to be more careful when constructing urls [12:29] :D [12:34] hm, maybe that's not it [12:40] Jc2k: try revision 175 of my branch === weigon_ is now known as weigon [12:43] or rather 176 [12:46] Jc2k: i'm going to bed soon, but please send any comments to michael.hudson@canonical.com [12:51] fwiw: kilner (who is sitting next to me) solved his problem with the pending merge by committing the merge, then branching off the resulting tree at the revno just before the merge (the last commit that was a part of the merge), which created a new tree with all the merged revisions as part of the same branch, which meant that an svn-push didn't lump the revisions involved in the merge together. [12:53] jelmer: ^;2~ [12:58] lifeless, hi [12:58] lifeless, what's up? [12:59] see semslie's note is all, thought it might be relevant to you :) [13:00] semslie, what's it exeactly that you were trying to work around? === sven_|away is now known as sven_ [13:04] jelmer: when an update comes in on top of local commits, those commits are converted to a pending merge, and you can then commit the merge. That would be fine if everyone in our team was using bazaar - they would still see each individual revision and its changelog as part of a merge. However, when an svn-push happens our subversion colleagues only see the merge-commit: ie: the results of the merge, not each individual change that made it up. [13:04] hi! after 'bzr merge', when there are conflicts, i get file.THIS, file.BASE, file.OTHER. is there a way to figure out which revision file.BASE is from? [13:05] semslie, this is docemented in the FAQ. you may want to use rebase [13:05] jelmer: we effectively wanted to 'replay' the changes in the pending merge on top of the tree after it had been updated, and we found that this was one way to simulate that behaviour. [13:06] jelmer: rebase is indeed an excellent tool, but we got worried that we were stuck with a merge once those local revisions had become a pending merge. Would a rebase have resolved the pending merge do you think? [13:06] semslie, no, but you can throw away the pending merge with "bzr revert" [13:08] jelmer: ah, so would rebase replay the changes in the pending merge on top of the new base as well (I had thought that these would not be included)? [13:09] semslie, yes [13:14] jelmer: thanks! giving that a try now [13:21] sven_: "bzr find-merge-base . .../other_branch" should do it [13:25] do I need to use a specific verion of bzr to utilise launchpad ? I have v0.11.0 and am getting... bzr: ERROR: Unknown branch format: '' [13:26] matholio: which branch? [13:26] this is the command I used to start one : bzr push sftp://matt-joyce@bazaar.launchpad.net/~matt-joyce/scrabbly/dev [13:28] and that gives tou the error? [13:28] It does. [13:28] 0.11.0 is pretty ancient! [13:29] Debian Etch. [13:29] Happy to add an apt source. is there one? [13:29] matholio: backports [13:29] matholio: that branch works fine for me [13:30] * mwhudson goes to bed [13:30] matholio: was it the second push that gave the error? [13:30] the first push. [13:30] matholio: can you run again with "bzr -Derror" please? [13:31] 'no such option: -D' [13:32] ah, sorry, too old for that [13:32] can you look in your ~/.bzr.log then, and fish out the relevant section please? [13:33] you'll find a stanza for each invocation, and near the top is the arguments [13:33] I'd like the traceback, but the whole staza would probably be enlightening [13:35] woo, theres a fair amount of stuff... [13:35] pastebin ? [13:36] AfC: http://nixos.org/nixos/ <- you might find that interesting [13:37] looking [13:37] Hi ! Is there any bzr executed commands history available ? (per working tree ? ) TIA [13:37] As yes, nix. I was there when this was first presented at LISA 4 years [13:38] I have a lot of respect for what they did. [13:38] It solves a lot of packaging problems. [13:38] I did point out to them a) that they had effectively taken on the challenge [ie, hassle] of becoming a distro, and [13:38] http://pastebin.com/m635c8448 [13:39] b) suggested that they might have a look at some of the things the Gentoo does along similar lines. [13:39] [specifically as related to package description formats] [13:40] Turns out they had studied Gentoo (as it then was, this is some years ago) and agreed that their biggest problem was the fact that they had to reinvent packges from scratch [13:40] which is a shame. Otherwise, their idea has a *lot* of appeal from the systems operations don't fucking hammer my linking side of things. [13:41] cool [13:41] matkor: not as such [13:41] james_w, thanks! [13:41] matkor: there is ~/.bzr.log whicn you could post-process [13:43] [a major theme in large systems being the need to exactly recreate specific environments. Embedding link targets in content addressing style leads to an impressive degree of robustness in this regard] [13:50] james_w: oh dear. I installed bzr from backports, v1.5 deleted the old .bzr and bzr init a new one. same error. [13:53] http://pastebin.com/m214aa475 [13:54] lifeless: thanks ! [13:57] Is it possible to checkout/branch only part of other branch ? Like checkout only one directory ? Or checkout only files matching given pattern like *.spec ? [13:58] bzr split ? [13:58] * fullermd blinks. [13:59] Not really the same thing... [13:59] Possibly not, but the closest I could think of [14:04] * matkor reads about bzr split ... [14:05] not really ... [14:06] I still want keep files in sync ... [14:07] Any VCS (except cvs) allows checkouting only subsets of branches ? [14:07] svn [14:07] and thats it AFAIK [14:07] we have some plans to do so [14:07] I presume p4 can. [14:07]