[00:15] Ack! [00:15] Paste's CLF-like log lines don't go in the log file either. [00:15] Paste-- [00:20] I should just use 'tee'. [00:20] That actually worked. [00:38] tee is love [00:45] abentley: around? [00:46] abentley: I think that what I'm proposing can be done using 'mpdiff' at least in part; if the accumulation of A,->B,->C etc is done without including the diff instructions - just including the new lines [00:48] then I can hold basis line #->(parent #, line #), and output mpdiff instructions; reconstruction will only ask for the lines included in that set so we can avoid creating intermediary objects beyond creating a ParentText object that has a view on the aggregate basis lines. [00:49] this will mean that one of the incremental deltas is a well formed mpdiff delta, and we can grab it own its own and reconstitute it locally (the first in such a chain we'd need to do scatter-gather IO to put together its base texts, but after that we'd be laughing) [00:50] the only problem is that the parents list will grow, so I wouldn't really want to include it per se with each diff, I'd want to have it at the front of some region [00:57] * Peng replaces sys.stdout and sys.stderr with a file. Beat that, Paste! [00:59] Peng: os.write(1, ...) ? [01:00] Peng: which log lines are you complaining about? [01:00] :( [01:00] (oh, and you're using some prehistoric paste anyway, right?) [01:00] mwhudson: The ones formatted like an Apache log file. [01:00] mwhudson: Probably. [01:00] Peng: they should go to the 'wsgi' logger by default [01:00] I'm using Ubuntu Hardy's Paste 1.6-2. [01:00] oh [01:00] no then really [01:01] Should I upgrade? [01:06] I was just using bzr-builddeb (from trunk; with bzr 1.5.0) and did "merge-upstream". This caused a "ConflictsInTree" error (basically because of a wrong, empty upstream file). So, I did "bzr revert" to get back to the old state, but this reverted the tree back to revision 1! (I just had committed r8 before). [01:07] james_w: ^ [01:08] Isn't there some idiom for "copy this file"? [01:09] "bzr info -v" says "8 revisions" in "Repository", but only "1 revision" in "Branch history" [01:09] mkanat: we don't support copies-with-history yet [01:09] lifeless: Ah, okay. === blueyed_ is now known as blueyed [01:09] I was thinking of svn, which has the reverse idiom (one for mv). [01:09] mkanat: so 'cp + bzr add' === Linnk is now known as Tank|Away [01:09] lifeless: Fair 'nuf. [01:10] mkanat: unless its between branches, then we can treat it as a mv after all ;) [01:10] Heh. [01:16] Peng: no, the messages should go to the 'wsgi' logger [01:17] mwhudson: What about exceptions? [01:18] Peng: dunno, sorry [01:18] utsl? :) [01:18] Heh. [01:20] So...logging.getLogger('wsgi')? [01:21] yeah [01:21] look at what start-loggerhead.py does for one approach [01:21] * mwhudson is no longer here [01:22] * lifeless pokes mwhudson [01:23] Hmm. [01:26] mwhudson: Thanks for the help. [01:27] lifeless, james_w: I've filed bug 245718 about the issue above [01:27] Launchpad bug 245718 in bzr-builddeb ""merge-upstream" + ConflictsInTree + "bzr revert" reverted to r1" [Undecided,New] https://launchpad.net/bugs/245718 [01:28] * blueyed should not use "bzr revert" anymore.. ;/ [01:29] Any chance that the data can be restored (from the "8 revisions" mentioned for the repository)? [01:29] usre [01:29] bzr heads [01:29] its a plugin [01:29] displays nothing [01:29] try -a [01:30] erm --all [01:30] ..this displays the revision, which I've committed last. [01:31] so [01:31] I have to go catch a plane [01:32] what you need is 'bzr pull -r revid:XXXX --overwrite .' - this will reset you to an arbitrary revision id [01:32] you just need to figure out what that id is [01:32] bye! [01:32] lifeless: thx, bye. [01:33] I guess the revid would be the one displayed by "heads --all"? [01:34] yes, it worked. phew. [01:53] * Peng crosses his fingers. [05:05] Can anyone link a (graphical) benchmark comparison of recent versions of bzr and hg? Everything I'm finding seems to be pretty ancient. [05:19] Well, thanks anyway. Bye. === Tank|Away is now known as Linnk [09:25] _damn_, bzr is faster on Linux [09:26] *way* faster :) [09:26] * awilkins goes to try out the 34,000 file tree that is usually quite sluggish on Windows [09:27] Hmm, that seems to be slightly more challenging for it [09:28] real1m8.739s [09:29] I suppose it was off a USB stick [09:29] Whee, cold status, 4 seconds, hot status, 0.9 [09:30] I almost feel like going in on Monday and claiming that Windows has a fatal data corruption error in it [10:37] hi [10:37] bzr-svn does not save the location to pull from for an import via "bzr get $URI"? [10:54] blueyed_: yeah, sorry, it's the result of a silly decision in the past. [11:47] How do I change file permissions in bazaar? I've changed the file permissions from -rwxr-xr-x to -rwxr--r--, tried to remove and re-add the file, but the file is still there with the 755 permission. Is there another way to do it? [11:49] Laibsch: Not in earlier versions [11:49] OK [11:57] savvas: AFAIK it tracks the x bit but not the other permissions [11:59] savvas: And not group permissions, from the look of things [11:59] hmm thanks [12:00] savvas: If you remove the public "r" access, then I presume they can't execute it? [12:08] awilkins: so if I remove the file, commit, (wait a bit), re-add and commit, it won't change it? [12:09] let me try it again, doesn't hurt heh [12:09] savvas: I think the permissions are the default ones for your folder, except the x bit, which is either set or unset for all three groups or [12:09] ah [12:10] savvas: It's not entirely unreasonable behaviour ; if you can read the file, you can copy it to a location where you have permission to set the x bit, and run it there, so it possibly doesn't make it very much more secure [12:12] .. to be able to set the x bit on a per-group basis [12:13] ok thanks [12:27] awilkins: so if I got it right, I can add or remove the x bit, but not group-specific? either the file is executable for all groups or it's not, correct? [12:34] well.. I tried removing the file and re-added it with no x bit [12:34] better this way, at least during development :) [12:35] thanks a lot for your help! === jdahlin_ is now known as johan [15:53] is anyone having problems with `bzr log bzr+ssh://...' with bazaar 1.5? [15:54] bazaar terminates with a lot of debug messages after I run that command [15:55] here's the error: http://paste.debian.net/9426/ [15:56] All I can say is that it WFM with bzr.dev. [15:57] WFM? [15:57] works for me [15:57] ah [15:57] hmm [15:58] tried 1.6b2? [15:59] no, i haven't [15:59] try it :) [16:00] hmm, is it stable enough? [16:01] The Bazaar developers do a very good job of making sure it always works. [16:01] ok [16:01] * pygi is using it :) [16:01] "WFM" isn't widely-known, but I just got lazy, and it's not like I make sense anyway. ;P [16:12] yup, 1.6 works for me :) [16:24] vgeddes, great [16:27] hello, wha't's the best way for a 1-time import of a svn repo into bzr ? [16:30] tacone: bzr-svn has a svn-import command tat works well [16:31] I am manually installing right now, the repo version doesn't supporto bzr 1.50 and seems not to be present in bzr ppa [16:33] Unable to load plugin 'svn' from '/home/tacone/.bazaar/plugins' [16:34] mh, pysvn bindings :-) [16:40] tacone: I'm told that 0.4 is the right branch of bzr-svn to use; it has its own bindings and is faster and less problematic [16:40] tacone: jelmer here is the author, and the FAQ has info [16:41] I have to catch the next leg of my flight; ciao :) [16:41] I'll check the faq [16:41] I am using 0.4.10 [16:42] from the tar.gz which is likely to be the problem [16:42] I'll try to check out from the branch. thanks [17:18] hi folks, I am trying to migrate from tortoisesvn to bazaar. One nice feature of tortoisesvn is that it can diff .doc files. any similar feature for bzr? [17:47] I am trying to push local changes back to launchpad using bzr and get this error: "bzr: ERROR: Cannot lock LockDir(http://bazaar.launchpad.net/%7Ekubuntu-members/guidance/powermanager-ubuntu/.bzr/branch/lock): Transport operation not possible: http does not support mkdir()". [17:48] ScottK-laptop: looks like you need authentication [17:48] Would someone please point me in the right direction for how I do this? [17:48] OK. That makes sense. [17:48] have you registered an ssh key for that laptop ? [17:48] otherwise no https, so fallback to http happens and bla bla bla [17:48] No. Are there alternative methods? [17:48] Right. [17:49] with launchpad, not to my knowledge, but I asked the same question here some day ago :-) [17:49] OK. [17:49] Thanks. [17:49] np [17:50] * ScottK-laptop marks bzr + lp off as still to hard and figures maybe next year. [18:40] hello [18:40] how can i check which revision my WC is in? [18:41] bzr revno ? [18:41] ah [18:41] thanks === blueyed_ is now known as blueyed [23:20] ahoy [23:21] word on the street is that bzr-git is back from the dead [23:21] is that true? [23:21] I heard that too. [23:22] * mathrick hopes it is === Linnk is now known as Tank|Away