endocrimes has quit [Quit: running from the computers]
<colemickens>
Is there a way to express flakes overrides in a file and then specify that file to a nix command?
<colemickens>
I've got a "nixup" script that is mostly around to manage adding --override-input for all of my local checkouts of various nix repos. I'd like to have an overrides.nix file that specifies overrides. I might file an issue, curious if anyone had thoughts or if I'm missing an obvious problem or better option
<samueldr>
and what about the first bootstrap LnL? ;)
<infinisil>
ryantm: Nice!
<LnL>
samueldr: ah back in time, at some point you'll end up in a nixpkgs that didn't have it's own compiler yet :)
<samueldr>
niksnut: to rephrase the question as a whole: how was our current root of trust seeded bootstrap-wise?
<samueldr>
LnL: yeah, exactly
<samueldr>
I'm tangentially curious, I think it was siraben that wanted to know more, not sure if it was about the historical facet or the current implementation
<siraben>
samueldr: Yes, it's related to the "trusting trust issue" as it pertains to the Nix ecosystem.
<LnL>
to be fair nothing that's in the current bootstrap tools ends up in the next so having a previous influence the new one seems pretty impossible
<niksnut>
they were generated by me on suse linux
<samueldr>
LnL: I guess there's always the thought experiment about having the compiler trojaned deeply enough that it will always produce compilers that produces tools that will bypass logins in a specified pattern
<samueldr>
but that's, as far as everyone knows, a though experiment :)
<samueldr>
I wasn't sure if it was directly "reflections on trusting trust" or an extension of it, looks like it might be an extension of it
alp has joined #nixos-dev
<LnL>
also, I doubt many other distros can trace the chain back to 2006
<samueldr>
yep, though it'll be neat to have a from-nothing bootstrap whenever that'll happen; and would be extra-neat if it ended up being bit-for-bit equal :)
<LnL>
oh right, if it's completely reproducible you could tie the loop
<das_j>
I hope this is the right channel for that, but what's up with the header of https://status.nixos.org/ ?
<samueldr>
still unfixed casualty of the website redesign
<samueldr>
it was (and still is) directly including styles from the nixos.org domain
<gchristensen>
I wonder why nixos-rebuild is failing with this
<gchristensen>
ehhh I can't seem to build my previous version either ...
<gchristensen>
nixos-rebuild isn't pure -- using an old version uses my system's PATH, and doesn't change the Nixversion in the PATH. maybe we can resholve that ... :)
<gchristensen>
:/
<gchristensen>
I don't understand
<LnL>
the assert output isn't very nice, but at least it shows something now :)
<gchristensen>
yeah... I'm just not sure why it is tripping the assert after an nixos-rebuild switch
<gchristensen>
and now I can't undo it
<LnL>
not exactly sure but seems like the assert is trying to prevent conflicting system definitions
<LnL>
I know you can specify some of those in your configuration
<gchristensen>
ah ha I had an unstaged diff in my config
<gchristensen>
heh
<gchristensen>
this caused that error! environment.etc.nixpkgs = pkgs.path;
<LnL>
hmm very strange
<gchristensen>
interesting way for that mistake to materialize
<LnL>
I think that should be environment.etc.nixpkgs.source since it's a path
<gchristensen>
yeah
<LnL>
almost seems like it tried to import that or something
<gchristensen>
yeah...
costrouc has joined #nixos-dev
justanotheruser has joined #nixos-dev
alp has quit [Ping timeout: 272 seconds]
alp has joined #nixos-dev
<arianvp>
the mother is a snapshot from a SUSE box by eelco I think
saschagrunert has quit [Quit: Leaving]
orivej has quit [Ping timeout: 258 seconds]
<samueldr>
[07:29:13] <niksnut> they were generated by me on suse linux
rajivr has quit [Quit: Connection closed for inactivity]
<raboof>
jtojnar: ah, bad timing on my part I guess :D
orivej has joined #nixos-dev
* aminechikhaoui
thinks `nix-info` should be part of the new nix command
<aminechikhaoui>
random thought based on the pain of investigating stuff with coworkers :D
<cole-h>
Would need to support flakes then (aka not rely on NIX_PATH / <nixpkgs>), but I'm also +1 on that
raboof has quit [Changing host]
raboof has joined #nixos-dev
<andi->
would need to support *also* flakes not just flakes :)
<niksnut>
aminechikhaoui: maybe it could be integrated into nix doctor
<niksnut>
or just part of nix --version
<aminechikhaoui>
niksnut right, I actually thought nix doctor already did that at first then discovered it doesn't
<LnL>
is there some way to define a flake as a standalone file?
<LnL>
I often work in repos where I don't really want to commit nix stuff but still use project specific expressions for development
<cole-h>
I thnk `nix --version` would be good place. `nix doctor` doesn't seem like the right place, IMO (though I did think of that, as well)
<cole-h>
andi-: :P
<aminechikhaoui>
did anyone ever succeed in fully statically linking protobuf including protoc using musl, I can't make libtool and friends behave so far :/
{`-`}_ is now known as {`-`}
justanotheruser has quit [Ping timeout: 260 seconds]
justanotheruser has joined #nixos-dev
__monty__ has quit [Quit: leaving]
justanotheruser has quit [Ping timeout: 260 seconds]
<aminechikhaoui>
so libtool basically insists on adding /path/to/gcc-lib/<..>/libstdc++.so and throwing away my -static from g++, if get into the nix-shell and fixup the g++ command it produces protoc as I expect
<aminechikhaoui>
don't know if libtool or the complicated protobuf build setup, although maybe that's not complicated and it's just me lost inside :D