[05:24]  * broder sighs
[05:25] <broder> i think i just lost at the game of code review chicken
[05:25] <broder> which is to say, i just opened up https://code.launchpad.net/~jderose/ubuntu/precise/couchdb/1.1.1-low-delta/+merge/85301
[05:48] <ajmitch> broder: what's the problem with that? that diff looks like it should only take a few minutes to review :)
[05:49]  * broder glares at ajmitch
[05:49] <ajmitch> with that many conflicts?
[05:50] <broder> i mean, whatever bzr dance jderose was trying to pull off clearly didn't work
[05:50] <broder> that's on my list of pending feedback
[05:50] <ajmitch> yeah, I can't remember what's required to get it to base the diff off the debian branch
[05:51] <ajmitch> at least the diff in the description looks a bit more manageable
[05:52] <ajmitch> still, if the couchdb split is done differently than it was previously, it may need some fixing
[05:52] <broder> right, that's what i'm trying to figure out
[05:52] <broder> it does look like there are differences in the resulting package, which would have also been helpful information to include
[05:53] <broder> bah. also, debian bug #573061 refers to a VCS commit with a potential package split for debian to a repository which no longer exists
[05:53] <ajmitch> what still depends on couchdb these days?
[05:53] <broder> so i can't really evaluate how similar this split is to what debian is going to do, if the maintainers ever do actually do anything
[05:53] <broder> all of novacut's stuff, appraently
[05:53] <ajmitch> will it affect U1 in precise?
[05:53] <broder> also channel-server, libnode-cradle, and desktopcouch
[05:54] <broder> ENODATA
[05:54] <broder> i thought U1 was mostly transitioning away from couch
[05:54] <ajmitch> I see it's been dropped back to univere anyway
[05:55] <broder> yeah, and ubuntuone-client at least hasn't
[06:09] <broder> i do feel like i'm struggling not to blame the contributor for our lameness in my writeup. there's a balance here that's hard to find
[06:19] <broder> :-/ the diff he posted in the MP description isn't accurate
[06:26] <ajmitch> despite the version not having changed?
[06:27] <broder> i guess? the branch changes debian/rules differently than the description suggests
[06:27] <broder> anyway, it looks like the discussion is actually happening on the bug, not the MP
[06:31]  * achiang struggles to add an apport hook to a package... http://pastebin.ubuntu.com/837476/
[06:31] <achiang> the idea is that i'd like to run dh_apport only on ubuntu
[06:31] <achiang> but i don't know where to put that. make an override_dh_apport: target?
[06:32] <achiang> one thought i had was to see who rdepends on dh-apport to find an example, but i guess rdepends doesn't work on B-D
[06:33] <broder> i don't believe there's any way to do that
[06:33] <broder> you're not allowed to change the build-depends as part of the build process
[06:34] <ajmitch> but you can prefix a command with - so that it doesn't error out when built without it
[06:34] <broder> yes, but if there's no way for you to get dh-apport on the build-dep list in the first place...
[06:34] <ajmitch> I don't know if that's the best way to do it here, as it also hides other errors
[06:34] <broder> see "CDBS" on http://ftp-master.debian.org/REJECT-FAQ.html
[06:35] <achiang> hm
[06:35] <broder> i think your best bet is to do what dh_apport does by hand, conditionally, in the rules file
[06:35] <broder> or possibly just ship the hook unconditionally
[06:36] <achiang> yeah, by hand is what bluez (and other examples) do
[06:36]  * ajmitch would prefer apport to be in debian :)
[06:38] <achiang> ok, thanks for the help
[06:47] <broder> does anything good happen when a conffile moves from one package to another? i would assume that triggers a conffile conflict more or less automatically
[06:48] <ajmitch> http://wiki.debian.org/DpkgConffileHandling has info about moving location, but not about it being moved between 2 packages & keeping the filename
[06:49] <broder> right, i know about the stuff on that page
[06:49] <broder> maybe if you add a Replaces
[06:55] <ajmitch> http://lists.debian.org/debian-devel/2006/12/msg00647.html implies that Replaces wasn't enough, at least back then :)
[06:56] <broder> oh god, so *that's* why openssh's maintainer scripts are such a mess
[06:56] <broder> "once we can expect everyone to have etch's dpkg installed we can move conffiles between packages more or less like any other files"
[06:58] <ajmitch> right
[06:58] <ajmitch> I think we're a bit past that point now
[06:58] <broder> given that half of the upgrade is going to pre-depend on a dpkg with lzma and dpkg-maintscript-helper...
[06:59] <broder> ooh, speaking of which, http://lintian.ubuntuwire.org/tags/preinst-uses-dpkg-maintscript-helper-without-predepends.html has built up a collection of tags
[07:05] <broder> i'm not sure what to do with the changelog for the couchdb merge
[07:05] <broder> jderose didn't make any attempt to merge in the ubuntu history
[07:05] <broder> but he also (deliberately) didn't attempt to merge in the rest of the packaging
[07:05] <broder> so dropping the ubuntu history doesn't seem completely unreasonable
[07:15] <broder> phew. that was a doozy
[10:43] <jtaylor> how do I resolve conflicts in .pc/* with bzr?
[13:12] <jtaylor> the numpy dh_python2 was sure done carefully ._.
[13:12] <jtaylor> no wonder debian is so relucatant to take patches
[13:15] <tumbleweed> :P
[13:23] <jtaylor> so numpy 1.6 branch including multiarch fix ready
[13:23] <jtaylor> anyone want to sanity check it?
[13:31] <jtaylor> lp:~jtaylor/ubuntu/precise/python-numpy/merge-1.6
[13:32] <jtaylor> there are dangling symlinks, but they are also present in debian
[13:32] <jtaylor> they don't look like a reason to diverge further