In reply to @hsngrmpf:matrix.org

Hello, I would like to become a member of the NAT.

As maintainer of dream2nix, improving the architecture of nixpkgs is in my interest because much of my project is based on things in nixpkgs and the quality of integration and UX depends on the architecture of nixpkgs. Given that my work is focused on alternative ways to use nix outside the nixpkgs bubble, I think I can bring a valuable perspective to the team.

While my work in the NAT itself is not directly sponsored by a company, dream2nix is sponsored by NLNet and customer projects, and therefore I have an ongoing interest being involved here, as this potentially simplifies my work.

Specifically I was recently focusing on alternative uses of the nixos module system, and would like to continue working on this while getting feedback from the team.

adisbladis:
I now did some benchmark using the current dream2nix core modules for builtins.derivation and mkDerivation.
source: https://github.com/nix-community/dream2nix/tree/main/misc/benchmarks

grmpf@grmpf-nix ~/p/g/d/m/benchmarks (benchmarks)> ./benchmark.sh

Benchmarking 1000x builtins.derivaton via pkg-funcs

real	0m0.542s
user	0m0.207s
sys	0m0.102s

Benchmarking 1000x builtins.derivaton via modules

real	0m2.855s
user	0m2.089s
sys	0m0.324s

Benchmarking 1000x mkDerivation via pkg-funcs

real	0m1.107s
user	0m0.456s
sys	0m0.181s

Benchmarking 1000x mkDerivation via modules

real	0m7.478s
user	0m5.902s
sys	0m0.646s

Is there any effort to replace substitute in place scripts with something like scripts + a json file

My particular use case is wanting to easily use python developer and unittesting tool's on the systemd-boot installer script which isn't exactly nice when its hard coded on evaluated instead of loading metadata

In reply to @hsngrmpf:matrix.org
Note that this benchmark tests with almost empty derivations and therefore there is almost no overhead in evaluating any input to the derivations which is unreal.
In real world scenarios, there should be a constant overhead for the input that must be applied to all benchmarks and therefore should improve the ratio in favor of the module system.
Though I don't know if this will be significant or not. We'd have to port something like firefox over to modules in order to test it

In reply to @hsngrmpf:matrix.org
Note that this benchmark tests with almost empty derivations and therefore there is almost no overhead in evaluating any input to the derivations which is unreal.
In real world scenarios, there should be a constant overhead for the input that must be applied to all benchmarks and therefore should improve the ratio in favor of the module system.
Though I don't know if this will be significant or not. We'd have to port something like firefox over to modules in order to test it