justanotheruser has quit [Ping timeout: 250 seconds]
jonringer has joined #nixos-dev
justanotheruser has joined #nixos-dev
LnL has quit [Ping timeout: 260 seconds]
LnL has joined #nixos-dev
LnL has quit [Changing host]
LnL has joined #nixos-dev
LnL has quit [Quit: exit 1]
AlwaysLivid has joined #nixos-dev
cjb has quit [Quit: brb]
cjb has joined #nixos-dev
Taneb has quit [Quit: I seem to have stopped.]
Taneb has joined #nixos-dev
ma27[m] has quit [*.net *.split]
aaronjanse has quit [*.net *.split]
zuh0 has quit [*.net *.split]
jtojnar has quit [*.net *.split]
kalbasit[m] has quit [*.net *.split]
angerman has quit [*.net *.split]
hexa- has quit [*.net *.split]
angerman has joined #nixos-dev
hexa- has joined #nixos-dev
ma27[m] has joined #nixos-dev
kalbasit[m] has joined #nixos-dev
jtojnar has joined #nixos-dev
aaronjanse has joined #nixos-dev
zuh0 has joined #nixos-dev
supersandro2000 has quit [Remote host closed the connection]
orivej has joined #nixos-dev
justanotheruser has quit [Ping timeout: 250 seconds]
cole-h has quit [Ping timeout: 252 seconds]
cjb has quit []
srk has joined #nixos-dev
<LinuxHackerman>
At that point, why not make them systemd services?
<symphorien[m]>
systemd depends on some of them to start, no ?
<LinuxHackerman>
hm yeah. IMHO think anything that isn't required for systemd to run shouldn't be an activation script though
<symphorien[m]>
I agree
<LinuxHackerman>
And, maybe more controversially, I don't think there should be that many things required for systemd to run that it needs to be exposed to all nixos modules and all users.
evanjs has quit [Read error: Connection reset by peer]
<gchristensen>
it makes 2 nixos servers incompatible w.r.t. transferring data between 2 machines
jonringer has joined #nixos-dev
<sterni>
anyone around who has worked on freebsd nixpkgs?
<eyJhb>
gchristensen: What would you propose then?
<gchristensen>
not sure exactly
<ajs124>
I'm sure this can be solved with PAM and LDAP somehow
<eyJhb>
gid/uid that is just the _name_ instead.
<gchristensen>
yeah, it can be :P but how do solve it in the small case I dunno
<gchristensen>
maybe a tool to take a machine's gid/uid allocation and emit users.users.name.uid = ...; for each one
<gchristensen>
or a module that asserts every defined user has an ID
<gchristensen>
maybe systemd's DynamicUsers will save us
<eyJhb>
But, that's also what is stated in the RFC that services that can should use DynamicUsers, which can change the permissions, on start.
<gchristensen>
right
<eyJhb>
If I read correctly.
<gchristensen>
but services that don't use that and unpin UIDs
<eyJhb>
But there are many services that needs to be rewritten to support this..
<eyJhb>
*transmission*
<gchristensen>
anyone seen this before in systemd-resolved? systemd-resolved[xxx]: malloc(): smallbin double linked list corrupted :)
<das_j>
gchristensen: My experiences with resolved were mostly resolved breaking for no reason and being fixed after I restarted it
<das_j>
so… restart it maybe?
<ScottHDev>
I don't understand, lsmod shows that kernel kvm module is loaded, but nix says that it only founds those system features : features {benchmark, big-parallel, nixos-test}
<das_j>
stackoverflow hints this might be an incorrect use of the strcat() function
<gchristensen>
it is in a crash loop
<das_j>
ScottHDev: You need to set the features yourself
<das_j>
gchristensen: Classic resolved :D
<ScottHDev>
I thought that it was the builder which figures out what features are available ?
<gchristensen>
is this a remote builder ?
<gchristensen>
nix will determine if the local builder has kvm, but nix won't go querying remotes for their features -- you have to configure the machines file to tell it
<ScottHDev>
No, it's a build in a nix docker container
supersandro2000 has joined #nixos-dev
<gchristensen>
docker may not be exposing kvm to the container?
<ScottHDev>
I ran lsmod in docker ? is that not enough ?
<das_j>
you seem to need `docker --privileged` for kvm to work
<ScottHDev>
Ok, will try that
<das_j>
`docker run --privileged` of course
<supersandro2000>
docker run --device=/dev/kvm
Synthetica has joined #nixos-dev
rj has joined #nixos-dev
rj has quit [Ping timeout: 240 seconds]
rj has joined #nixos-dev
rj has quit [Quit: rj]
<ScottHDev>
The both commands fixed my issue but they also brought a new error : machine # Could not access KVM kernel module: Permission denied
<ScottHDev>
machine # qemu-system-x86_64: failed to initialize kvm: Permission denied
rj has joined #nixos-dev
rj has quit [Ping timeout: 240 seconds]
<symphorien[m]>
call for comments about making isSystemUser or isNormalUser mandatory when declaring users to prevent easy mistakes: https://github.com/NixOS/nixpkgs/pull/115332
<{^_^}>
#115332 (by symphorien, 5 weeks ago, open): nixos/users: require one of users.users.name.{isSystemUser,isNormalUser}
<sterni>
siraben: overlays but they don't get applied to buildPackages roughly
<Ericson2314>
sterni: oh cross overlay**s**
<Ericson2314>
not the pkgs/top-level/cross-overlay.nix thing
<sterni>
Ericson2314: yes!
<sterni>
I wasn't even aware that the latter existed
<siraben>
it doesn't
<Ericson2314>
yeah I am pretty neutral on that
<siraben>
I only see pkgs/top-level/release-cross.nix
<Ericson2314>
i forget the exact name
<Ericson2314>
`pkgs/top-level/static.nix`
<Ericson2314>
that one
<siraben>
huh what's that do
<siraben>
Is that pkgsStatic
<Ericson2314>
that is used as a crossOverlay to make package static
<Ericson2314>
but the overlay bussiness is a bit awkward if you want to replace package but not through away it's static fixes
<sterni>
oh yeah then the order would be wrong
<Ericson2314>
so we have `stdenv.hostPlatform.isStatic` (which might need a rename) means we shouldn't need the overlay
<Ericson2314>
but having a cross-only overlay in general is still a fine idea
<Ericson2314>
vs adding a bunch of a conditional stuff to a regular overlay
<sterni>
crossOverlays makes the pkgsMusl, … things a bit more convenient to make I think
<Ericson2314>
anyways yeah crossOverlays seems to document
justanotheruser has joined #nixos-dev
plumm has joined #nixos-dev
__monty__ has joined #nixos-dev
cole-h has joined #nixos-dev
<gchristensen>
I'm not sure, but I wonder if hydra's scheduler progressively deprioritizes big projects by accumulating how many seconds it has taken over time and never resetting those values. So a week or more after it has been running for 3082610 seconds and gnu:hello-master has been running for 700 seconds, and the shareUsed is 1027 vs. 7, so gnu:hello-master would be prioritized
<gchristensen>
maybe that is a feature / good thing
cole-h has quit [Quit: Goodbye]
plumm has quit [Remote host closed the connection]
cole-h has joined #nixos-dev
rj has joined #nixos-dev
AlwaysLivid has quit [Remote host closed the connection]
plumm has joined #nixos-dev
AlwaysLivid has joined #nixos-dev
tomberek has joined #nixos-dev
rajivr has quit [Quit: Connection closed for inactivity]
rj has quit [Quit: rj]
rj has joined #nixos-dev
kini has quit [Ping timeout: 258 seconds]
krkini has joined #nixos-dev
rj has quit [Ping timeout: 240 seconds]
rj has joined #nixos-dev
pmy has quit [Ping timeout: 240 seconds]
pmy has joined #nixos-dev
pmy has quit [Ping timeout: 252 seconds]
pmy has joined #nixos-dev
rj has quit [Ping timeout: 240 seconds]
rj has joined #nixos-dev
{^_^} has quit [Remote host closed the connection]
{^_^} has joined #nixos-dev
justan0theruser has joined #nixos-dev
justanotheruser has quit [Ping timeout: 260 seconds]
rj has quit [Ping timeout: 240 seconds]
rj has joined #nixos-dev
orivej has quit [Ping timeout: 252 seconds]
<sterni>
hmm the stdenv.hostPlatform == stdenv.buildPlatform check for “not cross compiling” is actually a bit flawed
<sterni>
because as soon as you pass crossOverlays != [] nixpkgs will take all the codepaths as if it was cross compiling
<sterni>
but host and build platform can still be the same
<sterni>
however buildPackages and pkgs will be different due to the crossOverlays and this breaks e. g. pythonPackages spectacularly
rj has quit [Ping timeout: 240 seconds]
rj has joined #nixos-dev
srk has quit [Ping timeout: 260 seconds]
__monty__ has quit [Quit: leaving]
tomberek has quit [Quit: Connection closed]
<samueldr>
is there a libfaketime, but for predictable random() calls?