[06:04] <dholbach> good morning
[16:34] <jtaylor_> ScottK: any opinions what to do with the numpy multiarch issue for oneiric? bug 818867, ugly patch in linked debian bug
[19:07] <ScottK> jtaylor: I can apt-get install python-enable in a clean chroot with no issue, so I don't think the bug is correct.
[19:08] <jtaylor> enable was fixed
[19:08] <ScottK> OK.
[19:08] <jtaylor> the issue is numpy
[19:09] <jtaylor> btw you told me to ahve a look at that, besides what is in the debian bug I have no more input
[19:09] <jtaylor> no idea what plans there are to upstream multiarch stuff
[19:11] <ScottK> My opinion then is we should do what I always do with multiarch stuff.
[19:12] <ScottK> I pass the buck to slangasek.
[19:12] <ScottK> slangasek: Thoughts on http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=5;filename=numpy-multiarch.patch;att=1;bug=640940
[19:16] <jtaylor> btw installing was never an issue, its just about compiling stuff using numpy.distutils
[19:17] <slangasek> heh
[19:18]  * slangasek makes himself scarce and lets the remaining multiarch problems fix themselves ;p
[19:19] <slangasek> well, that's not upstreamable of course
[19:19] <slangasek> wasn't there a more-upstreamable multiarch search rule being pushed upstream for python itself?
[19:20] <slangasek> http://bugs.python.org/issue12418 -ish?
[19:21] <slangasek> ultimately, we need a non-Debian-specific runtime interface for querying the path, but we're not there yet
[19:21] <slangasek> (config.multiarch or some such)
[19:25] <lfaraone> slangasek: think you could ACK my recent upload of pitos to oneiric?
[19:25] <jtaylor> so are there any objections to adding that non-upstreamable patch to numpy for oneiric?
[19:25] <lfaraone> *pithos
[19:26] <jtaylor> it doesn't break many package (non which havent been fixed via workaroudns to my knowledge) but who knows what user programs it breaks
[19:26] <slangasek> jtaylor: uh... "it doesn't break many packages" is not the right standard for an upload during final freeze
[19:26] <slangasek> lfaraone: looking
[19:27]  * lfaraone is afk, airplane. 
[19:28] <jtaylor> hm yes probably a bit late
[19:28] <jtaylor> hopefully we'll have a proper fix for P
[19:29] <slangasek> lfaraone: I guess support for proto 31 may be dropped some time during the o support cycle, then?
[19:32] <slangasek> oh, I see v31 is already dropped - right then :)
[19:39] <ScottK> slangasek: How about 0 day SRU for the numpy thing.  Then we have an easy backwards path if there are issues?
[19:40] <slangasek> ScottK: that sounds saner to me.  what's the actual bug being fixed?  The patch doesn't enlighten me as to why it's needed
[19:40] <ScottK> The bug is numpy can't find the X11 libs.
[19:40] <ScottK> Which is a side effect of the bigger issue that it thinks it's required to care.
[19:41] <slangasek> but then adding this change breaks other packages?
[19:42] <jtaylor> it should only fix them
[19:42] <jtaylor> its adding an extra library search directory
[19:42] <jtaylor> can that be an issue if there is a possible library in both?
[19:42] <jtaylor> gcc will ust pick the first in the list or?