=== g2[ATL] is now known as g2
=== juliank is now known as Guest54569
=== juliank_ is now known as juliank
=== JanC is now known as Guest24786
=== JanC_ is now known as JanC
[07:56] <Mirv> ScottK: thanks for the find! I was also puzzled where the scandir would be - it turns out it's nowhere.
[07:56] <Mirv> and I was looking at only src:python-pathlib, not 2
[08:59] <infinity> xnox: Indeed, nothing has changed, it's always required 3.2 on xenial.
[09:00] <infinity> xnox: It's a miracle if it's working.
[09:01] <infinity> xnox: Oh.  Except I have it set to 2.6.32 on x86, specifically for briandead OpenVZ.  So the preinst is lying.
[09:01] <infinity> xnox: Bother.
[09:03] <infinity> xnox: Err, nevermind.  I see what aurelien did here, and I support him 100%. :P
[09:03] <infinity> xnox: It didn't bomb out, right, it just told you that it might suck?
[09:04] <infinity> xnox: For versions between 2.6.32 and 3.2, we tell you that you're pretty much doing things at your own risk.  3.2 and up is supported.  Less than 2.6.32 and we'll refuse to work at all.
[09:20] <Mirv> I think we have transition completion happening!
[09:20] <Mirv> the downside is that I broke my promise to try to relax on the weekend
[10:07] <acheronuk> Mirv: great :)
[10:08] <acheronuk> + you can relax better later knowing Qt has moved ;)
[10:08] <Mirv> yeah, that's the plus side :)
[10:30] <acheronuk> Mirv: and this zesty box now sees the new Qt in release and can upgrade
=== JanC_ is now known as JanC
=== scottt is now known as Guest83800
=== g2 is now known as g2[ATL]
=== ignacio is now known as DevSome1thing
=== DevSome1thing is now known as ignacio
=== scottt is now known as Guest30925
[21:21] <mwhudson> is anyone planning to do a /usr merge for ubuntu?
[21:25] <jbicha> mwhudson: did you see the first bullet point at https://lists.debian.org/debian-devel-announce/2017/01/msg00004.html
[21:25] <mwhudson> jbicha: oh heh, no i didn not
[21:26] <jbicha> that doesn't directly answer your question but Debian is hesitating a bit
[21:26] <mwhudson> jbicha: it is super related to what i am actually doing though
[21:26] <mwhudson> the /usr merge breaks snapd on debian (if apparmor is enabled)
[21:27] <mwhudson> or at least that's what this bug report says
[23:38] <Son_Goku> mwhudson, really? why?
[23:39] <Son_Goku> SUSE has UsrMerge with AppArmor and it's fine
[23:39] <Son_Goku> is there that great of a policy difference?
[23:39] <mwhudson> Son_Goku: and snapd works there?
[23:39] <mwhudson> i don't know, i'm just saying what the bug report says :)
[23:40] <Son_Goku> well, in both SUSE and Arch, when we manually swap the kernel with Ubuntu ones, it's fine
[23:40] <mwhudson> hm
[23:40] <mwhudson> it may be something about the apparmor in debian's kernel i guess
[23:40] <Son_Goku> as much as I don't like AppArmor, I really doubt it's AppArmor's fault
[23:40] <Son_Goku> I know that snapd does not work on Debian AppArmor because the AppArmor in upstream kernels does not match what snapd expects
[23:41] <Son_Goku> it does not work in SUSE either
[23:41] <Son_Goku> unfortunately, I've not seen much progress on the Ubuntu AppArmor team in upstreaming their changes to the mainline kernels that all other distros use
[23:42] <Son_Goku> mwhudson, if anything, it's most likely a break in Ubuntu's contributed AppArmor profiles
[23:43] <Son_Goku> despite Ubuntu having thought about doing UsrMerge years ago, I doubt anyone really thought about it when the profiles were being written and upstreamed into Debian
[23:44] <mbiebl> you have to update the  individual apparmor profiles to support a merged usr system
[23:44] <mbiebl> I assume suse has done that
[23:45] <Son_Goku> they did it a long time ago, yes
[23:45] <mbiebl> bugs.debian.org/846966 is such an example
[23:45] <Son_Goku> and Arch's AppArmor profiles are derived from SUSE's
[23:45] <Son_Goku> and if Mageia decides to implement AppArmor, I will be importing SUSE's profiles
[23:45] <Son_Goku> just like I import Fedora's for SELinux
[23:48] <Son_Goku> I don't know if you guys talk to the SUSE security team, but you might want to get in touch with them about synchronizing work on the profiles
[23:49] <Son_Goku> mbiebl: I'm disappointed that the latest release of d-i disabled usrmerge by default :(
[23:49] <Son_Goku> I was really looking forward to default usrmerge
[23:53] <mbiebl> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=843073 was the reason why it was reverted
[23:53] <ubottu> Debian bug 843073 in dpkg-dev "dpkg-shlibdeps: broken on i386 with merged /usr" [Important,Fixed]
[23:58] <Son_Goku> but it's fixed...?
[23:58] <Son_Goku> so if it's fixed, it can be unreverted?