2012/09/20/#bzr.txt

[00:04] <fullermd> bzr ci foo.c bar.pl baz.pl quux.sgml [...]
[00:07] <fullermd> bjp: Not via the UI, I don't think.  And it may not be specifiable as a range period, since there may not be a unique starting point.
[00:15] <bignose> how can I view the log entries which affected a file that is now deleted or moved, when I don't know when that was?
[00:15] <bignose> ‘bzr log foofile’ just gives “bzr: ERROR: Path unknown at end or start of revision range: foofile”.
[00:17] <fullermd> Think you gotta find the last rev that has the file first, then use that as the end.
[00:17] <lifeless> you could use bzr search foofile to find the file in your inventory /somewhere, then bzr log -r ..lastrev foofile
[00:17] <lifeless> or bzr log -v and grep/search for foofile.
[00:18] <bignose> bzr: ERROR: No matches were found for the search [u'foofile'].
[00:19] <bignose> I remember this being asked a while ago but can't find it.
[00:20] <lifeless> huh
[00:20] <lifeless> I may not have finished my path indexing spike. I thought i had :(
[00:31] <bjp> fullermd: really?  theres a before: seems like there should be an after: :)
[00:43] <fullermd> Well, for one thing, ancestry is like a single-linked list, so finding the after: revs would mean shuffling through the whole repo, rather than just reading an id.
[00:43] <fullermd> But there can be multiple revs in after: none of which are related to each other, so which would you pick to start from?
[00:46] <fullermd> (well, end with, more accurately...)
[00:47] <bjp> well, what does -rtag:sometag.. show for its second tag?
[00:47] <bjp> starts at sometag and goes to the end right?
[00:48] <fullermd> Roughly speaking, but close enough, yes.
[00:49] <bjp> and i'm using it with bzr log
[00:49] <bjp> so i would figure -rafter:tag:sometag.. would show all the same revs except the first one with the tag? :)
[00:50] <bjp> i'm using it to print all the changes/commits since it was last tagged
[00:50] <fullermd> But that leaves you with a range with no unique start [end].  How to display that without misimplication isn't quite as easy as it sounds.
[00:50] <fullermd> (in the general case, which is what the tool has to look at)
[00:51] <fullermd> It's not an insurmountable issue, but it's more complicated than you'd think at a glance, and nobody has put in the time to answer all the ambiguities and figure out a good display and whatnot.
[00:51] <bjp> so is there a better way to "show the log changes since sometag"? or is the easiest way to just "bzr log -rtag:sometag.." and just ignore the first entry
[00:52] <fullermd> 's probably the best.  More accurately, figure out which rev tag:sometag is, and ignore that one.
[00:52] <fullermd> (you definitely don't want to ignore the _first_, since that's the head; lists backward, remember  ;)
[00:53] <bjp> right
[00:54] <bjp> i'm parsing it all in python and spitting it out in wiki syntax anyway, so its not hard to add
[00:54] <bjp> i just thought there was a builtin way of doing it that i was missing
[00:54]  * fullermd nods.
[00:54] <fullermd> I think log shows the tags by default too, so you don't even have to figure out the revno/revid for the tag yourself to exclude it.
[00:54] <bjp> and i usually look for the method with the least amount of work :)
[00:56] <fullermd> If you're doing it in python, you could probably bypass a lot of the UI level stuff and get the data more directly from bzrlib.
[00:56] <fullermd> Though that's almost certainly the opposite of "least amount of work" unless you already know the API top to bottom.
[00:56] <fullermd> And probably a long cry from it even if you do...
[00:58] <bjp> yea i was messing with bzrlib in another script
[00:58] <bjp> it's nice, but not nearly as well documented as the tools
[00:58] <bjp> and if i was doing more than just 'bzr log' i'd consider it
=== vednis is now known as mars
[02:33] <delinquentme> is there a way to drop an individual file back to the way it was at a particular commit?
[02:59] <fullermd> delinquentme: See revert
[07:43] <mgz> morning btw
[08:12] <vila> morning mgz ;)
[08:12] <mgz> morning vila!
[08:21] <vila> mgz: care to re-review https://code.launchpad.net/~vila/bzr/mergetool-doc/+merge/125150 ?
[08:21] <vila> jelmer: I know I've been far too long to propose https://code.launchpad.net/~vila/bzr/514301-auth-config-unicode/+merge/124627 but I also hope you will review faster than me ;)
[08:24] <vila> darn, tyop in bug # again (bug #388725 instead of bug #388275)
[08:25] <ubot5> Launchpad bug 388725 in Duplicity "Report when duplicity is 'done' with restoring a file" [Wishlist,New] https://launchpad.net/bugs/388725
[08:25] <ubot5> Launchpad bug 388275 in Bazaar "want configuration option to control progress bars" [Medium,In progress] https://launchpad.net/bugs/388275
[09:17] <emeric> Hi folks,
[09:18] <emeric> I have a bzr repository checked out somewhere, and would like to check out the same repository on another computer. Is there a way to find out the URL used in the first checkout ? bzr info does not display it.
[09:19] <mitya57> I hope I can ask loggerhead-related question here:
[09:19] <mitya57> Does this error (http://anonscm.debian.org/loggerhead/pkg-grub/trunk/grub/revision/2550.1.1)
[09:19] <LeoNerd> If bzr info doesn't say, then likely the information doesn't remain. But you can always checkout from the first computer.
[09:20] <mitya57> happen because of old loggerhead version?
[09:20] <mitya57> or is that a known bug?
[09:21] <emeric> LeoNerd: I can checkout/commit from the first computer, but if I want to code from another one, I'm out of luck if the URL is not displayed in bzr info ?
[09:21] <fullermd> If it's a checkout, info will certainly show it, otherwise it couldn't work.  As an independent branch, you _could_ have cleared out the info so it's no longer there, but why would you?
[09:21] <LeoNerd> If you're committing, then you are just committing to that local branch then
[09:22] <LeoNerd> Otheriwse, brz info would be showing the commit branch
[09:23] <emeric> Output of bzr info.http://www.privatepaste.com/f2817d362b
[09:24] <LeoNerd> Yup, so there's no upstream checkout or submit branch there
[09:25] <emeric> Which means I can only work from the place where the first checkout has been made ?
[09:27] <emeric> (Sorry, I sound like complete helpless , but that's what I currently am regarding bzr)
[09:30] <emeric> bzr checkout bzr+ssh://root@ctrl/root/atta_dev lctests
[09:30] <emeric> This one seems to work, yay !
[09:35] <emeric> Note to self: also install python*-gobject*. Thanks alot for the help folks.
[09:36] <fullermd> I'm not sure what value of "works" corresponds to "can ssh in as root"...   :p
[09:39] <emeric> Automated tests having to run as root, written on a test computer, in root home, versioned using bzr. If I want to develop more tests from a workstation, it seems the only option I have is to bzr+ssh as root, unless I missed something.
=== slank` is now known as slank
=== slank is now known as Guest43410
=== yofel_ is now known as yofel
=== Guest43410 is now known as slank