[01:07] <mdeslaur> pitti: I tested them today, and will publish them on monday. Thanks!
[01:08] <pitti> mdeslaur: cool, thanks
[01:09] <pitti> mdeslaur: ah, you tested already; I ran into some failures which were due to the recent "nologin as default shell for system users" change
[01:09] <pitti> mdeslaur: (schroot copies the host's /etc/passwd)
[01:10] <pitti> easy enough to fix, but not entirely obvious at first sight
[01:10] <mdeslaur> pitti: on what releases?
[01:10] <pitti> mdeslaur: just with trusty as the host
[01:11] <pitti> mdeslaur: I edited /etc/passwd in the schroot to have /bin/sh as nobody's shell, then it works again
[01:11] <pitti> mdeslaur: that's fixed in trusty's postgresql-common but that's not backported to earlier releases
[01:11] <sarnold> what is user 'nobody' used for?
[01:12] <pitti> sarnold: for some test cases about per-user clusters; that user is always available
[01:12] <sarnold> 'nobody' is supposed to be used only for nfs quashing, not actually running programs...
[01:12] <sarnold> pitti: test cases only?
[01:12] <sarnold> corners can be cut for testing, fair enough..
[01:13] <pitti> sarnold: yes, of cousr
[01:13] <sarnold> thanks pitti :)
[01:15] <pitti> good night everyone, cd /airport and then cd ~ :)
[01:15] <mdeslaur> pitti: good night, have a nice trip!
[01:18] <sarnold> pitti: safe travels :)
[06:15] <sarnold> is there an easy way to pull build logs from launchpad?
[10:51] <xnox_> sarnold: there is launchpad api to fetch them.
[11:20] <bdrung> bdmurray: sounds good to have architecture information in distro-info, too
[11:21] <bdrung> but distro-info is for mapping devel & co to a distro.
[11:21] <bdrung> once you have a distro you could pull information about this distro from some other tool
[13:17] <doko> infinity, https://launchpadlibrarian.net/167193519/buildlog_ubuntu-trusty-i386.python-cffi_0.8.1-1ubuntu3_FAILEDTOBUILD.txt.gz
[13:23] <jtaylor> doko: known 2.19 bug
[13:24] <jtaylor> also breaks scipy
[13:26] <jtaylor> hm its not exactly the same
[14:29] <jtaylor> wow the math functions have really been screwed up in 2.19 :(
[14:30] <jtaylor> there seems to be some global state added between calls
[14:30] <jtaylor> probably rounding mode
[14:51] <jtaylor> doko, infinity: filed a bug https://sourceware.org/bugzilla/show_bug.cgi?id=16625
[15:43] <infinity> jtaylor: Is this i386-only?
[15:45] <jtaylor> yes
[15:48] <infinity> jtaylor: I'll see if I can get jsm to look at it this weekend.  He's the mostly like to (a) know what's broken and (b) have been the person who broke it in the first place.
[15:48] <infinity> s/like/likely/
[23:37] <infinity> ampelbein_: Dude.  Don't disable --as-needed to fix things, actually fix the bug. :/
[23:37] <infinity> ampelbein_: Have you done that for more than ufsutils (the one I just noticed)?
[23:41] <infinity> ampelbein_: Anyhow, solution is simple for ufsutils, since it's changed arch to kfreebsd-any, I'll just remove it. :P