IRC channel logs

2018-03-23.log

back to list of logs

<temporaryusernam>hello
<temporaryusernam>I would like to ask: Is it possible to play proprietary games on Guix ?
<temporaryusernam>Like GOG games?
<pkill9>yeah ofc
<pkill9>just won't be available in the Guix repo
<pkill9>they typically get distributed as bundles anyways
<temporaryusernam>of course!
<temporaryusernam>Well it's great news because I love freedom but I love gaming too :D
<temporaryusernam>And of all software I don't mind games being proprietary the most
<temporaryusernam>from privacy perspective at least
<CompanionCube>also beware that you may or may not need to provide your own GPU microcode somehow
<temporaryusernam>What do you mean CompanionCube ?
<temporaryusernam>If I use nouveau, for example, that should cover it?
<CompanionCube>not sure about nouveau
<CompanionCube>but I know radeon GPUs would
<marusich>Anyone know how to watch the livestream videos for LibrePlanet with GNU IceCat built with Guix?
<marusich>When I visit this page, I get an error: https://libreplanet.org/2018/live/
<marusich>"No video with supported format and MIME type found."
<Gamayun>I don't think it's streaming yet, is it?
<efraim>You might need some gstreamer plugins, or I've had luck in the past with streamlink or mpv instead
<efraim>I have low RAM so I try to keep my browser closed
<civodul>Hello Guix!
<xshumeng>hi, i just installed guix in CentOS7 reference 《GNU Guix Reference Manual》. When i run command 'guix package -i hello', i seen an error "guix package: error: build failed: cloning builder process: Operation not permitted". how to fix it?
<xshumeng>the section 2.1 Binary Installation.
<jlicht>hi guix!
<snape`>o/
<rekado>Hi Guix
<rekado>In light of the new glibc problem on RHEL6 I think the best course of action is to switch to an older glibc on the “rhel6” branch.
<efraim>That does seem like the best option at this point
<xshumeng>is there have any way to fiex it on RHEL7?
<civodul>hi rekado!
<civodul>rekado: what are the new libc issues?
<civodul>xshumeng: is this in a VM or in a container?
<xshumeng>yes, in docker container.
<civodul>you need to run docker with --privileged for this to work, IIRC
<rekado>civodul: glibc 2.26 needs the syscall prlimit64 to be implemented. The patched RHEL6 kernel does not implement it.
<civodul>where is it used?
<rekado>civodul: this means that the JVM fails initialisation and dies with “no memory”.
<civodul>oh
<xshumeng>ooh, thanks a lot.
<rekado>yeah :(
<civodul>rekado: is there something we can do in core-updates?
<rekado>I don’t know yet.
<civodul>ok
<rekado>g_bor pointed me at a glibc commit that might be revertable, but I really don’t know.
<rekado>commit 695d7d138eda449678a1650a8b8b58181033353f
<rekado>“Assume prlimit64 is available.”
<rekado>I’m confused by these lines in that commit:
<rekado>-/* prlimit64 is available in 2.6.36. */
<rekado>-#if __LINUX_KERNEL_VERSION >= 0x020624
<rekado>-#endif
<rekado>
<civodul>it does seem to be revertable
<civodul>actually we only need to keep the code that deals with ENOSYS
<civodul>the bits about removing __ASSUME_PRLIMIT64 are fine
<civodul>seems doable
<xshumeng>civodul you are right. i can unpackage now ...
<rekado>we are really unlucky. “prlimit64 is available in 2.6.36”, but we use 2.6.32. So close yet so far.
<rekado>I’m actually already one week into my two-week vacation and still haven’t been able to leave things be.
<civodul>bah :-(
<civodul>maybe the easiest thing is to revert to glibc 2.25 in 'rhel6' for now, and let "others" take care of the prlimit64 issue in core-updates
<civodul>supporting those old kernels is going to be increasingly difficult though
<rekado>indeed.
<rekado>IT will update the OS on our cluster nodes to CentOS 7.x “in the next few weeks”, but I don’t know when they start counting.
<rekado>RHEL6 is already in the stage of its life cycle where it only gets important security fixes, so my hopes are that the MDC cluster is an outlier and that RHEL6 will soon disappear completely.
<civodul>that would be nice
<civodul>we need to decide whether to work on the prlimit64 problem, and subsequent problems if we find any
<rekado>I’ve reverted to glibc 2.25 on the rhel6 branch.
<rekado>I’ll look at the prlimit64 problem in a week if nobody beats me to it.
<snape>marusich: if you want to copy the content of our private emails, it's fine too
<snape>but I think it's better to have the discussion on the ml, so that other people can participate
<civodul>rekado: sounds good, try to enjoy your vacation now!
<rekado>heh, I’ll try harder :)
<mbakke>Is there an alternative to "copy-file" that does not dereference symlinks?
<mbakke>I wanted to fix #30006 and was surprised to find that I got three copies of the same library, instead of one lib and two symlinks.
<mbakke>Not a problem in practice due to deduplication, but somewhat surprising.
<az`>hi there all
<snape>mbakke: you can look at copy-recursively, it checks if the target is a symlink, and if it's the case it does (symlink (readlink file) dest) instead of (copy-file file dest)
<mbakke>snape: Thanks!
<snape>yw :-)
<nckx>mbakke: In practice the files get turned into hard links (which is why I'm not as bothered by this being the status quo for everything using the Meson build system), but yeah, not great.
<nckx>*not as bothered as I should be otherwise.
<nckx>Do git updates go on staging or core updates? (I.e. should I trust ‘guix refresh -l git’ to get it right), follow-up: is staging open for bidness?
<swati>Hi I'm having trouble installing the guix package can someone help me out
<mbakke>nckx: AFAIK updating git won't change source derivations. But I'm not 100% certain, should be easy to test though.
<mbakke>Also, staging is eagerly awaiting your commits.
<civodul>anyone knows how to use the guix completion under zsh?
<nckx>mbakke: Thanks. Judging by the guix-commits archives (is there a git-native way to find out which branch first received a commit), master's fine indeed.
<rekado>swati: could you please give us additional details? How do you try to install it?
<jlicht>The new release of shepherd is quite nice :D
<civodul>glad you like it, jlicht :-)
<civodul>mange did nice stuff in there, if you use it as an unprivileged user
<jlicht>that is exactly how I am using it :-).
<jlicht>Is it normal that a socket (e.g. `/run/user/1000/shepherd/socket') is still around after stopping the shepherd?
<swati>I was running an installation script which failed and now I'm stuck at keyring created
<swati>can someone help me out?
<civodul>jlicht: i think so, it's just that nobody listens to it anymore
<civodul>swati: what part of the installation instructions you are referring to?
<wigust>swati: Are you talking about https://git.savannah.gnu.org/cgit/guix.git/tree/etc/guix-install.sh ?
<swati>yess
<wigust>swati: Could you paste a script output to http://paste.debian.net/ ?
<swati>Sorry, my college internet connection is erratic. I posted the code. Please have a look
<rekado>swati: could you tell us where it has been posted?
<swati> http://paste.debian.net/1016246
<rekado>swati: did you run “gpg --keyserver pgp.mit.edu --recv-keys” or “gpg --keyserver pgp.mit.edu --recv-keys 3CE464558A84FDC69DB40CFB090B11993D9AEBB5”?
<rekado>it should be the latter.
<rekado>the goal here is to fetch the public key of the person who signed the source code release.
<rekado>once you have that key you can check that the source code release really was signed by that person.
<rekado>(it also ensures that the release archive has been fully downloaded)
<swati>I tried that but then I'm getting this. http://paste.debian.net/1016249
<swati>Oh I got it! sorry that was a very silly thing to ask. I'm a newbie so please excuse me. And thanks for being so patient
<snape>swati: I imagine you ran the script with sudo, because the output refers to your home. You can't to that because of this bug: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=30728. You need to log in as root first, for example with "sudo su -".
<PotentialUser-79>Hi all, does the usb installer include some utilities to connect internet via pppoe?
<janneke>trying to install with full disk luks encryption...
<janneke>after booting i now get Device /dev/nvme0n1p1 doesn't exist or access denied
<janneke>it's the luks mapped-device source, and it does exist...