[11:59] <doko> ginggs, jtaylor: python-numpy autopkg tests now down to python-scipy failure.  the one on amd64 is y pathon2 test called test_gzip_py3. should that one even run for python2? the i386 one ... I don't know
[15:42] <ginggs> doko: tomorrow i'll try to check whether upgrading scipy to 1.0.0 helps
[15:44] <doko> ginggs: the amd64 issue is gone by testing against -proposed. only the i386 one left
[15:47] <ginggs> doko: ok, so it's just the precision errors then?  should we get scipy marked badtest for i386?
[15:51] <doko> from my side fine ... python-numpy in main doesn't show any failures anymore
[17:27] <NikTh> Hi, any suggestions about this build failure ? https://goo.gl/ZcTpmh
[17:34] <NikTh> ping apw
[18:07] <ginggs> doko: i *think* this patch to scipy will do it: https://paste.ubuntu.com/25998603/ - i can't test it because i can't reproduce the failure in my VM
[18:12] <LocutusOfBorg> doko, https://launchpad.net/ubuntu/+source/rubberband/1.8.1-7ubuntu1 this is making mpv uninstallable
[18:13] <LocutusOfBorg> and then this fails https://launchpad.net/ubuntu/+source/gnome-mpv/0.13-1ubuntu1/+build/13746173
[18:13] <LocutusOfBorg> should I no-change rebuild mpv?
[22:39] <TJ-> anyone on 17.10 can check what "keyctl show" reports for their regular user? is the top-level session keyring  owned by UID 0 ?
[23:02] <Faux> zsh: command not found: keyctl
[23:06] <TJ-> Faux: it's part of the keyutils package; should be installed if ecryptfs-utils was installed I think
[23:07] <TJ-> Faux: trying to get to the bottom of a rather serious bug on 17.10 for a user who's lost access to encrypted home, and it seems to be down to the keys being inserted into the wrong keyring (again ... been an on-and-off problem for many releases)
[23:13] <Faux> https://paste.debian.net/hidden/733a3e67/ is my desktop, if it's useful to you at all. I do not have ecrypt.
[23:13] <Faux> /away
[23:15] <TJ-> Faux: is that 17.10? in your's the session keyring is owned by the UID 1000 but it's got root's user keyring linked :)