<civodul>jonsger: indeed, i'll merge the two entries
<jonsger>civodul, don't forget to do it in flashing-tools.scm
<civodul>i'm not doing anything other than telling people to work together :-)
<janneke>sneek: later tell pmikkelsen: not running gnome, but with latest master i get upon sudo -i: sudo: pam_open_session: System error; sudo: policy plugin failed session initialization. user is in group wheel
<efraim>In the interest of throwing ideas into,the wind, the kernel, busybox, uboot, and I'm sure others, accept a config file for building, we should probably generalize this type of build option
<civodul>efraim: i think we should start by generalizing that of the kernel
<civodul>adding an API to generate a kernel .config file essentially
<efraim>It might be enough of a start to make the config an input and add a phase like (copy-file (assoc-refs %inputs "my-config") ".config") so 'make config' will read the config in its default location
<efraim>But I do like the 'add these flags to the bottom' thing we have going for the kernel
<janneke>civodul: is the elogind/libcap causing my sudo breakage?
<ng0>hi! I've been meaning to fix "make dist" for gnURL so that the receipe can drop autotools. We had this issue many months back where shebangs were left in some autotools (or otherwise build-process related) files. Has this been fixed?
<ng0>*actually more correct: /gnu/store/… shebangs
<ng0>I don't know very much about mingw32. are those binaries really executable on native Windows (which versions?) or does it require some tools in windows to run them?
<civodul>janneke: i don't know if it's related, i couldn't reproduce the sudo problem you mentioned
<civodul>efraim, janneke: i've just pushed the elogind patches
<ng0>oh. /gnu/store/q87cg01m7sksk17yjyahjxrmic86mqvj-make-boot0-4.2.1.drv for i686-mingw32 failed its build. I wanted to test this crossbuild guix pack.. is this failure new or is someone looking into it already?
<reed_>So I added guile-git to my system.config because guix was telling me I needed it to run a `guix pull`. But after I ran the system reconfigure, it is still saying that I need to install guile-git
<pmikkelsen_>civodul: i had a problem yesterday where i couldnt log in too, not even in tty1. the update to 232.4 fixed it for me, but i havent tried master yet
<sneek>Welcome back pmikkelsen_, you have 1 message.
<sneek>pmikkelsen_, janneke says: not running gnome, but with latest master i get upon sudo -i: sudo: pam_open_session: System error; sudo: policy plugin failed session initialization. user is in group wheel
<civodul>"Cannot determine cgroup we are running in: No data available"
<cbaines>I think I've got a bit confused with derivations... I have multiple machines, with supposedly the same code running, but they generate different derivation files. I'm looking at two .drv files, for the same package, with different contents, but the same value for the output, which I didn't think was possible?
<civodul>cbaines: it's possible due to fixed-output derivations
<civodul>with fixed-output derivations, there's an infinite set mapping to a single output store item
<civodul>e.g., there are many ways to fetch a file with a given hash
<cbaines>I think I'm looking at the derivation for a package build, would the above still apply?
<cbaines>Ok, maybe things are not as odd as I was thinking, I seem to be getting consistent output paths on multiple machines
<ng0>q: are .so files, for some reason, created in a phase after "build" with gnu-build-system?
<ng0>i get .lo .la right after build when I insert a "breakpoint"
<slyfox>normally 'make' is supposed to link .so files from .lo files