[05:00] <slangasek> why does the ubuntu kernel's test suite install... exim?
[21:12] <apw> slangasek, erm something must think it wants to send email ... where you seeing that, adt, somewhere else ?
[21:13] <slangasek> apw: it's in the autopkgtest output; when the ubuntu-regression-test test runs, it goes about installing a bunch of packages one by one (nice that the output lists the commands it runs, btw!  Wish other things would do that...)
[21:13] <apw> slangasek, yeah ... it is
[21:13] <slangasek> apw: I looked through the code and it appears to call a sendmail api directly to talk to gmail
[21:13] <apw> *blink*
[21:13] <slangasek> no visible invocations of sendmail or exim in the tree
[21:13] <apw> why on earth would it need to do that ... bjf ^
[21:13] <slangasek> and exim is not the preferred mta :)
[21:14] <slangasek> well I think it's sending mail to Brad
[21:14] <apw> yeah, i'd say it is a hangover from the old trigger mechanisms, we moved to aqpm for most things so its likely redundant
[21:14] <slangasek> ahhhh why is this machine firewalled off of ddebs.u.c!
[21:15] <slangasek> apw: ok. fwiw I'm working on fixing the root cause of the actual kernel autopkgtest failures right now (util-linux stuck in xenial-proposed, and autopkgtest doing silly things with base packages)
[21:17] <slangasek> ahhhh why is binutils failing to create dbgsyms packages!
[22:00] <apw> slangasek, doing silly things?  most of the recent carnage was to do with the more selective package pinning changes ...
[22:00] <slangasek> apw: the autopkgtest infrastructure does a dist-upgrade of the testbed before installing any per-package deps; then it sets the package pins; which means libuuid1 gets upgraded to xenial-proposed, then we drop a pin that says libuuid-dev can only come from xenial
[22:01] <slangasek> according to infinity this is a deliberate misfeature at the moment because it "ensures testing of the base packages".
[22:01] <slangasek> but those packages should get their own autopkgtests declared, and we should set the pin before touching anything in the tetsbed
[22:10] <Ben64> are we going to get new kernels without rebooting now?
[22:24] <apw> slangasek, yep, that does sound confusing indeed