[00:26] hm. i can't commit if a completely different file in an other folder is in conflicted state? [00:44] hm. i can't commit a specific file if a completely different file in an other folder is in conflicted state?? [00:50] jelmer: ping? [00:52] Peng_, pong [00:53] jelmer: Little subvertpy bug: subvertpy/__init__.py calls warn() but never imports it. I could file a bug (or send a patch), but since it's really trivial and I'm not sure how you'd want to fix it... [00:53] (And I don't have a browser open ATM.) [00:55] Peng_, hmm [00:55] Peng_, There's nowhere to import from unfortunately [00:55] I guess I'll just change it to ImportError [00:57] any hints for my problem? [00:57] jelmer: Just use warnings.warn()? [00:57] zsquareplusc: Well, you could resolve the conflict. ;P [00:58] zsquareplusc: Why are you even trying to commit when another file is conflicted? [00:58] and when i dont want to? its in a different subfolder [00:58] zsquareplusc: Doesn't that mean you have an uncommitted merge or someting? [00:58] Peng_, Ah, wasn't aware of that [00:58] Peng_, fixed [00:58] Peng_, thanks [00:59] jelmer: Thanks for ifxing it. :) [00:59] Augh, more typoes. [00:59] Peng_ well, i pulled the branch from a server. that generated the conflicts. that ok. now i want to commit locally one changed file [01:01] the file i want to commit is not in conflicted state. it just modified [01:01] Eh, TV to watch. See ya later. :) [01:01] Good luck. [01:03] hmm.. major disapointment :( [01:07] and bzr-gtk isn't working neither :( [01:15] ok the later was because an older copy of the gtk plugin was found in the .bazaar folder and they were mixed up when running olive-gtk :/ === sdboyer__ is now known as sdboyer [01:53] jelmer: mmmmm, just found out about a repository with 250K revisions. Should be fun to branch from that using bzr-svn. [02:39] How do I resolve diverged branches when upstream is svn? [02:39] ☺ [02:44] hrm [02:44] only 'bzr dpush' has the problem [03:31] isn't that the point of dpush? [03:51] bob2: the point of dpush is to not use metadata [03:51] I ended up just using metadata anyway [03:51] and that worked fine [03:52] <.< === TheMuso_ is now known as TheMuso [06:16] Is BB down? [06:23] "502 Proxy Error" says that it is. :P [07:15] Peng_: That has meant yes in the past. [07:15] abentley: ^ [07:26] Peng_: Fixed [07:36] abentley: What happened? [14:06] hello, i have built some packages for bzr for AIX - anywhere I can put them? [14:10] crisb: Ideally they should end up here: http://bazaar-vcs.org/Download [14:11] You should find more info on the wiki itself [14:14] AIX? Wow. That brings back painful memories... [14:18] ;) [14:18] its not exactly speedy for some reason [14:19] unfortunately i had to build most of the deps myself too - so they cant really go on the download page can they? [14:27] crisb: at least you can list yourself as a contact for any interested user ? Or explain how you proceed to get there (may be at least fun :)? [14:27] crisb: did you succeed in running 'bzr selftest' ? [14:27] hmmm :) [14:28] well it works, but so slow [14:28] 30 minutes and it had not yet done 1000 tests [14:28] i think something is up with either python or bzr itself on aix. [14:28] ouch [14:28] im probably to blame as i built python too ;) [14:29] how about the C compile ? ;) [14:29] how about the C compiler ? ;) [14:29] * vila wonders when he will able to make a joke without typo... [14:29] not me, but its the IBM one...so should be fine on this platform! [14:32] bzr rocks takes 3 seconds... [14:32] oooouch: [14:32] time -p bzr rocks [14:32] It sure does! [14:32] real 0.08 [14:32] user 0.06 [14:32] sys 0.02 [14:32] * fullermd is perfectly willing to assign blame to AIX for anything and everything. [14:33] indeed fuller, its soooo.........IBM ;) [14:33] The first system I ever adminned ran aches. [14:33] That was in another millennium. It still makes me twitch. [14:33] aches? [14:33] crisb: wait, I thought IBM stopped to public ennemy #1 with AIX, or am I mixing with something else ? [14:33] That's how you pronounce "AIX" :p [14:34] ah [14:34] hmm [14:34] legacy systems can be a pain [14:35] luckyly im not going to be an admin later [14:35] ronny: fullermd never talked about legacy here, I think the system was brand new :) [14:35] sadly its got a good reputation in financial systems! [14:35] Well, that system would be pretty legacy now :p [14:36] * vila first one was SunOS (not that Solaris gizmo...) :-) [14:36] If it weren't smashed into a kadnillion tiny bits, stomped into the ground, pissed on, dug up so they could be buried upside-down, had runes scrawled and incantations spoken over... [14:37] I mean, not that I have any hard feelings toward it or anything. [14:37] :-) [14:37] gotta love that man with the dumbells though ;) [14:37] hmm [14:38] i grew ~5 files in my repo every 500 commits [14:38] how does bzr bundle up commits? [14:39] Sorta kinda log(10)ly. [14:39] It'll grow ~1 file every thousand commits. Then ~5 every 5 thousand commits. ~1 every 10 thousand commits. etc. [14:40] It groups packs at 10, 100, 1000 etc [14:40] would anyone be able to point me in the right direction if I showed the output of bzr rocks --profile-imports from my AIX machine? [14:41] Each pack comes with a set of index files .rix revisions, .iix, inventories, .tix texts, .six signatures [14:42] crisb: better post it to the mailing list: https://lists.ubuntu.com/mailman/listinfo/bazaar [14:43] but bzr rocks may not be the most interesting, it does really the minimum and doesn't involve any of the important parts [14:44] crisb: for a start did you build celementtree ? [14:44] python 2.5 - so I dont need to right? [14:44] right [14:45] pyrex ? [14:45] no, but I built solaris and bzr rocks is like 0.2 seconds [14:45] on roughtly comparable, if not slower hardware [14:45] yes, not needed for bzr rocks [14:46] perhaps if bzr rocks is slow then that points to python [14:46] crisb: right, start with that [14:46] Well, something lower in the stack than bzr anyway. [14:47] ISTM that AIX has always had rather painful fork()'s. But certainly not 3-seconds-painful on even remotely up-to-date hardware. [14:48] time -p python -c 'print "hello"' [14:48] 0.03s [14:50] hi. I checkout from lp:ideatorrent. but as the project doesnt give me write access, I did "bzr push lp:~sayakb/ideatorrent/devel" to create a new personal branch. But http://bazaar.launchpad.net/~sayakb/ideatorrent/devel/files is giving me error [14:50] crisb: same here, are you using some mounted file system ? [14:51] nop, all local [14:51] btw, commit done about 45 mins back.. [14:52] glade88: https://code.edge.launchpad.net/ideatorrent [14:52] the branch is empty [14:53] try pushing again [14:54] vila: "bzr push lp:~sayakb/ideatorrent/devel" ? [14:55] glade88: It should works, yes [14:56] vila: thanks. and what about future commits? "bzr commit -m 'blah blah' lp:~sayakb/ideatorrent/devel" [14:56] sorry, I am too used to subversion I guess :) [14:57] glade88: If you want your commits to appear in the remote branch, you should bind your local branch to the remote one with [14:57] bzr bind lp:~sayakb/ideatorrent/devel [14:57] but push first [14:59] vila: when I bind, will the fact that I checked it out from lp:ideatorrent matter? [14:59] since I didnt bind it with lp:ideatorrent [15:01] glade88: no it doesn't matter [15:01] the branch you checked out *from and the branch you commit *to* may be (and are here) different [15:02] but until you're more familiar with bzr you may want to avoid 'bind' until you better understand what branches are involved when [15:02] vila: cool. push under progress [15:03] bind transform your branch in a heavy checkout which is generally what users from cvs and svn want, but it can also be a source of misunderstandings when not working in a true centralized workflow [15:03] to avoid that, can I always "push" to my personal branch ~sayakb/... ? [15:03] And here, your're not in a such a workflow because you don't have commit rights in the trunk [15:04] glade88: yes, you can push whenever you wnat to publish your changes, which amy not be at each commit [15:04] will that show me diffs on the remote location? [15:04] push -v will show you what you push [15:04] but push increments the revision number, right? [15:05] bzr missing will show you the revisions you need to pull as well as the revisions you created locally [15:05] push with increment the revno of the remote branch, not the local one [15:05] s/with/will/ [15:06] so what would be the best deal to pull from one branch and commit to another one (with increment in commit revs on both local and remote?) [15:06] exactly what you've done so far [15:07] bzr branch [15:07] bzr push [15:07] when you want to synchronized from do [15:07] bzr merge ; bzr commit -m 'Merging from upstream' [15:07] cool. (as the original project owner would review and merge my revisions to lp:ideatorrent) [15:08] damn, still "pushing" the rev.. [15:08] 'bzr version' ? [15:09] 1.6.1 [15:19] glade88: 1.12 is out, try upgrading, what os/version are you using ? [15:21] vila: kubuntu 8.10 [15:21] maybe it isnt sybce [15:21] *synced [15:22] You may try adding one of the ppas : https://launchpad.net/bzr [15:22] vila - ctypes not built on my AIX python, would that slow anything? [15:22] https://launchpad.net/~bzr/+archive will gives you the official releases [15:23] vila: thanks. wil update as soon as the push is complete :) [15:23] (still pushing :( :( ) [15:23] you're certainly pushing the whole project, more recent bzr versions should allow you to use stacked branches which result in pushing only your additional commits [15:23] ok [15:24] That will not be always true, but launchpad should be configured for stacked branches [15:25] yes. true. that is the longest commit I ever did to a repo :) [15:25] crisb: ctypes is not installed here either, I think it's more important for windows though [15:33] how can i turn a branch + worktree back to just branch? [15:34] ronny: first check you have to modifications pending (bzr st) then 'bzr reconfigure --branch' should do the trick [15:35] hmm [16:03] Hello, I am just starting to learn bazaar. Is it possible to combine branches from multiple shared repositories into a single branch in a new shared repository? [16:11] gskelly: If the branches are related, you can merge back and forth to your heart's delight. [16:13] Where they are hosted, standalone or in shared repositories, local or remote, has no influence on the functionality, only on performance and disk space. [16:24] Lo-lan-do: The repositories were created independently. [16:25] Lo-lan-do: I created a different repo with a single branch for different aspects of a project, (docs, data, calcs) that really should be a single directory structure in a single branch. [16:25] Hmmm... I'm new to IRC also :) [16:31] "oops" [19:25] so I updated my loggerhead - a recent change coflicted with a hack I had to put in to get the links right, so I thought it would be fixed, but it still seems to be completely broken as standar [19:30] all the links are broken [19:30] * Lo-lan-do eyes beuno [19:32] is there any statistic on how the performance incrases over the versions? [19:32] ah take it all back - I had a trailing slash on server.webpath that was the problem [19:32] sorry for the frustrated tone - its been a long day [20:56] jelmer: ping [20:56] nevans, pong [20:56] bzr-svn NEWS file says that svn-push => push requires patched version of bzr [20:56] will everything else work with plain old bzr 1.12? [20:57] nevans, yes [20:57] jelmer: thanks. :) [20:57] nevans, (svn-push is also still there, for those who don't have the patch) [20:57] wonderful [21:01] jelmer: weird message: "Upgrade to Subversion 1.5 or higher for faster retrieving of revision properties." but it only says this when I use the implicit (saved) push/pull location. [21:02] I'm on Ubuntu 8.10. as far as I can tell, everything svn is at 1.5. [21:02] nevans, it says that when pulling from a < 1.5 svn server [21:02] ahhh. [21:02] well then, I suppose the weird thing is that it *doesn't* say that when I use an explicit push/pull location. [21:03] nevans, are you sure you're specifying the same location? [21:03] yep [21:03] nevans, This is all outside of bzr-svn [21:04] nevans, so bzr-svn will get the same URL in both cases [21:04] weird. [21:09] jelmer: a noop push now takes 1min (rather than 16 to 20). not ideal, but MUCH better. thanks. :) [21:10] nevans, cool :-) === thumper_laptop is now known as thumper [22:07] is there a way to edit the commit message of some older commit (and it's even pushed to launchpad :/ ) [22:08] chx: No, revisions are immutable. [22:09] ok [22:09] i did a fix on a latest commit by first uncommitting it :) but sure. [22:15] chx: Uncommitting threw away the old commit, the second commit created a separate commit. [22:15] yes, i understand that. [22:15] OK. :)