phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
<gchristensen>
ekleog: kind of funny, but yeah
<ekleog>
any idea why "x86_64-linux" and not even builtins.currentSystem?
<gchristensen>
hmm not sure builtins.currentSystem works in hydra?
<gchristensen>
"In pure mode, the Nix evaluator forbids access to anything that could cause different evaluations of the same command line arguments to produce a different result."
<ekleog>
oh
<ekleog>
so it's basically “well anything really”
<gchristensen>
what do you mean?
<ekleog>
but at the same time it's used to generate symlinks… so it needs to be the right arch to be able to build the `ln` & co tools, doesn't it?
<gchristensen>
Nix will ensure those build jobs are run on a system which supports x86_64-linux
<gchristensen>
are you having a problem based on this?
<ekleog>
no, I'm just refactoring the region of code around (to make the tests attrset reusable), and wondered what that meant
<gchristensen>
cool
<ekleog>
thank you :)
<ekleog>
(now, the fun stuff: `nix-instantiate --eval --strict nixos/release.nix -A tests` doesn't evaluate :D)
<arianvp>
ugh
<arianvp>
scripted networking sure is a headache
<arianvp>
when are we swithcing to networkd?
<arianvp>
found another bug
orivej has quit [Ping timeout: 252 seconds]
drakonis has quit [Ping timeout: 272 seconds]
<gchristensen>
depending on how complicated your network is, you can switch now
drakonis has joined #nixos-dev
<gchristensen>
(networking.useNetworkd = true)
<arianvp>
alright nice
<arianvp>
but basically, I just want to use "systemd.network.<blah>" directives
<arianvp>
and not deal with the "networking" dsl at all
<arianvp>
when using networking.useNetworkd = true is used, I guess all the "networking" stuff is translated to "systemd.network" statements?
<gchristensen>
Mic92 has a config somewhere using systemd.network.blah directives
orivej has joined #nixos-dev
init_6 has quit []
orivej has quit [Ping timeout: 272 seconds]
<gchristensen>
ekleog: by referencing nixos tests from nixpkgs, is the nixos configuration evaluated for each evaluation of nixpkgs?
<gchristensen>
for each test referenced by each package *
<ekleog>
gchristensen: it's the point Sonarpulse mentioned, I'm currently working on the test refactor to avoid re-importing nixpkgs every time :)
<ekleog>
(then, meta attributes aren't evaluated by default, so there's always this)
<gchristensen>
I'm worried more about memory and cpu for nix users
<ekleog>
it's not really often that `meta` attributes are all evaluated, is it?
<ekleog>
I mean, apart from the case of JSON-exporting
<gchristensen>
and when attributes are type-checked or using nix search
* ekleog
doesn't know about nix search
<ekleog>
isn't it just checking the package name and description?
<gchristensen>
might be, not sure
<ekleog>
then it'd need benchmarking I guess, no one thought of this issue :°
<ekleog>
(well, no one raised it until now, at least)
<gchristensen>
it is hard to remember until you're on a t2.micro and can't run `nix-env` because nixpkgs is too big to fit in ram :P
<ekleog>
:D
FRidh has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos-dev
drakonis1 has joined #nixos-dev
drakonis has quit [Ping timeout: 244 seconds]
drakonis_ has quit [Ping timeout: 252 seconds]
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 252 seconds]
zarel has joined #nixos-dev
drakonis1 has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-dev
drakonis_ has joined #nixos-dev
drakonis has quit [Ping timeout: 260 seconds]
aszlig has quit [Quit: Kerneling down for reboot NOW.]
<samueldr>
infinisil: in case you want a local noisy failure for issues like in #44076, you can use `env -i NIXPKGS_ALLOW_UNFREE=1 nix-build ...` (where NIXPKGS_ALLOW_UNFREE isn't necessary, but useful in those cases)
<LnL>
gchristensen: we could also provide a <nixpkgs> that aborts with a more detailed message
<samueldr>
and any <nixpkgs/other/files> would fail since only default.nix would exist I presume
<LnL>
hmm yeah, it would only help for a toplevel jmport
<samueldr>
well no, it would also help for e.g. <nixpkgs/nixos> by failing (though not with a nice message)
<LnL>
sure but not having a nixpkgs entry would also fix that
<samueldr>
yeah
<samueldr>
(and I confirmed that using
<samueldr>
`NIXPKGS=" =..." adds a __nixPath entry with the prefix " ", so it's a hack, but since <> has a stricter definition, one would need to find the entry in __nixPath to import it)
<samueldr>
(and I do understand not wanting to use it! just shared the findings)