<mark_weaver>we have a phase now that checks for missing rpaths, so I'm not sure why it couldn't find its libraries. <codemac> sneek: later tell mark_weaver: no surprise, it was a systemd problem. they didn't expect the process to fork away (-D means don't fork), systemd would kill off the entire cgroup (strace & bitlbee+child) before it had finished printing the library lookups. so it *definitely* found it's libraries, it's just that the log I was looking at was artifically truncated. <sneek>Welcome back mark_weaver, you have 1 message. <sneek>mark_weaver, codemac says: no surprise, it was a systemd problem. they didn't expect the process to fork away (-D means don't fork), systemd would kill off the entire cgroup (strace & bitlbee+child) before it had finished printing the library lookups. so it *definitely* found it's libraries, it's just that the log I was looking at was artifically truncated. <codemac>hahah, I was just happy to have figured out sneek :P <sbidin>is there a way to specify windows within the grub-configuration? <davexunit>not sure, we should allow for some method for dual booting. <davexunit>I sure do love how I can keep a definitive list of all the packages I want installed in my user profile in a version controlled text file. <sbidin>agreed. personally I have a repo with various config files and an "install" script that links subsets of them depending on the computer I'm on <sbidin>but with packages, it's the next level :) <davexunit>mark_weaver: would be opposed to me updating the guix-devel snapshot to the latest commit on master and re-enable container tests? <davexunit>I pushed a commit awhile ago that made the guix-devel build pass with container tests enabled. <sbidin>I'll have conky packaged soon. Any suggestions where to place it? <sbidin>Currently I'm thinking conky.scm... <remi`bd>where should the shared code between two scripts, publish and publish-gnunet, go? <remi`bd>at first, I’ve placed it in a `guix/scripts/publish-utils.scm` file, but now `publish-utils` is listed as a command, thus it’s obviously not a good location. <sbidin>can I see the state of the build dir after a failed build? <davexunit>I believe it will owned by one of the guix builder users <sbidin>yes, chown is needed if I want to run stuff there manually <sbidin>is there a way to refer to the dir containing all inputs? <sbidin>I find myself needing to set a search path for a cmake file <sbidin>placing the 5 different paths work <sbidin>but I assume they're also all linked within some directory <bavier>sbidin: there's an "environment-variables" file, that might provide what you want <bavier>between 'guix environment foo' and that file, you should be able to recreate the build environment <sbidin>bavier: thanks! I'll have a look <davexunit>the environment will be a bit different, since you'll have network access and won't be in a chroot <bavier>correct, I was using 'build environment' loosely <davexunit>I just wanted to clarify for sbidin. I know what you meant. :) <sbidin>just using --annotate doesn't work, as git send-email expects the From: etc. headers to be on the top <sbidin>using --compose adds an introductory email, which I don't want <sbidin>more precisely: how could I do this while using git send-email? <davexunit>sbidin: this was probably done by writing an email with the patch as an inline attachment ***sbidin` is now known as Guest93813
<mark_weaver>davexunit: I'd prefer to get the security updates merged before updating the guix snapshot <davexunit>feel free to remove the phase that disables those tests <ifur>davexunit: whoa, that fs is the best news all years ***Guest93813 is now known as sbidin
***sbidin is now known as Guest23043
<mark_weaver>sneek: later tell remi`bd: guix/publish.scm might be a good place <mark_weaver>but there are some broken packages that we might want to fix first <mark_weaver>sawfish's configure script doesn't seem to like the updated gdk-pixbuf <davexunit>okay, no rush on my end. just trying to get my ducks in a row so I can get the rest of the container patches in. <mark_weaver>and we need an updated gnurl, although I didn't like it because it was signed by a person who I've never heard of, nor have I heard of anyone who signed his key (which is almost no one), and someone who has no commits in the gnurl repo <mark_weaver>probably we need a newer version than they have anyway <mark_weaver>curl-7.43.0 failed its test suite against the new gnutls, but curl-7.44.0 worked. <mark_weaver>anyway, I think it's okay to update guix-devel after we get sawfish fixed ***davi_ is now known as Guest72060
***wgreenho` is now known as wgreenhouse