pie_ changed the topic of #nixos-rust to: buildRustPackage is wrong. | the ops are those who op those who donut op themselves
<qyliss> I'm wondering if anybody would be explicitly unhappy if I just opened a PR that added something like import-cargo (but more Nixpkgs-friendly) to Nixpkgs and converted a bunch of rust programs away from buildRustPackage
<qyliss> because it's pretty straightforward to do (especially compared to other approaches that have been floated), and I can't see any way it would be significantly worse than what we have
<cole-h> What are the downsides of buildRustPackage?
<cole-h> e.g. what makes this desirable?
<qyliss> cole-h: FOD
<cole-h> Hm. I'm not opposed, per se, but doesn't it rely on flakes (experimental)?
<cole-h> And would there be a (thin) wrapper around it so we don't have to specify a `buildPhase` ourselves? (The README on that repo seems to necessitate a manually-added buildPhase)
<cole-h> Oh, I guess that plays into the "but more Nixpkgs-friendly"
<cole-h> I'd definitely be interested in seeing what it would look like (with a few examples of conversion from bRP -> import-cargo).
<qyliss> cole-h: I wouldn't be doing it with flakes
<qyliss> there's nothing in it that requires flakes
<qyliss> it also uses builtins.fetchGit but there's no reason it couldn't depend on nix-prefetch-git being run instead
<qyliss> cole-h: for buildPhase and stuff I think ideally we'd want to use setup hooks but for now could just hook into buildRustPackage
<qyliss> my dream is that eventually you could build Rust programs with normal mkDerivation but that's a long way off
cole-h has quit [Ping timeout: 264 seconds]
evanjs has quit [Ping timeout: 260 seconds]
evanjs has joined #nixos-rust
evanjs has quit [Ping timeout: 260 seconds]
evanjs has joined #nixos-rust
cole-h has joined #nixos-rust
cole-h has quit [Ping timeout: 246 seconds]
qur70 has joined #nixos-rust
cole-h has joined #nixos-rust
qur70 has quit [K-Lined]
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-rust