| 13 Aug 2025 |
 dramforever | * but yes i need -march=mips3 | 04:01:18 |
| dramforever | why the hell does everything have mesonFlags | 05:17:22 |
| dramforever | why the hell is there cmakeFlags and mesonFlags in pkgs/stdenv/generic/make-derivation.nix | 05:18:03 |
| dramforever | i can't fix this, i'll leave this be | 06:08:45 |
 Toma | In reply to @dramforever:matrix.org
why the hell is there cmakeFlags and mesonFlags in pkgs/stdenv/generic/make-derivation.nix I was very surprised too when I first encountered it. | 11:36:27 |
| Toma | I'm guessing it is there so that splicing/cross doesn't make us use targetPackages or other hacky things (because cmake goes into nativeBuildInputs, shifting the offset)
just a guess, though | 11:44:01 |
| Toma | * I'm guessing it is there so that splicing/cross doesn't make us use targetPackages or other hacky things when setting up the cmake vars (because cmake goes into nativeBuildInputs, shifting the offset)
just a guess, though | 11:44:30 |
 Alyssa Ross | I thought it was for typing/validation | 11:46:23 |
| Alyssa Ross | I remember being a bit annoyed about it when it happened | 11:46:31 |
| Toma | interesting...
cmakeFlags was added to stdenv 7 years ago in https://github.com/hercules-ci/nixpkgs/commit/b3041b44552f50109760e7fa41a16636e4b4d5c0
and 3 years ago, some type-checking logic was added to it: https://github.com/hercules-ci/nixpkgs/commit/3173c3b6b61cd06620bda26ac8f0b674866d6101
in any case, I don't want to pollute this room anymore with cursed non-rust stuff, rust has enough cursed stuff itself. | 12:13:42 |
| Toma | * interesting...
cmakeFlags was added to stdenv 7 years ago in https://github.com/NixOS/nixpkgs/commit/b3041b44552f50109760e7fa41a16636e4b4d5c0
and 3 years ago, some type-checking logic was added to it: https://github.com/NixOS/nixpkgs/commit/3173c3b6b61cd06620bda26ac8f0b674866d6101
in any case, I don't want to pollute this room anymore with cursed non-rust stuff, rust has enough cursed stuff itself. | 12:31:08 |
| dramforever | mesonFlags is absolutely rust-relevant | 12:49:21 |
| dramforever | because it depends on *Platform.rust | 12:49:43 |
| dramforever | which explains why changing *Platform.rust rebuilds the world | 12:50:02 |
| dramforever | * because it depends on
*Platform.rust on cross | 12:50:13 |
| dramforever | * because it depends on `*Platform.rust` on cross | 12:50:25 |
|  impqxr set their display name to impqxr. | 12:53:49 |
 mpiechotka | For cargo2nix - is it possible to disable a workspace member? I have a workspace with server, wasm client and common code. Cargo2nix attempts to build various packages on wasm as dependencies because it tries to merge features. | 15:18:29 |
 apyh | you could have a different src/ filter for different target platforms, perhaps? | 15:27:43 |
| apyh | since it'll have to build wasm stuff from scratch anyways | 15:27:57 |
| mpiechotka | In reply to @apyh:matrix.org
you could have a different src/ filter for different target platforms, perhaps? It's for dependencies | 15:51:54 |
| mpiechotka | Indirect dependencies | 15:52:02 |
 rosssmyth | I have been told that the commend in bootstrap.example.toml is a lie. So just put it in there without llvm and apparently it just does it. I have also been warned that it might be unusable slow. | 18:52:17 |
| rosssmyth | * I have been told that the comment in bootstrap.example.toml is a lie. So just put it in there without llvm and apparently it just does it. I have also been warned that it might be unusable slow. | 18:52:22 |
| rosssmyth | * I have been told that the comment in bootstrap.example.toml is a lie. So just put it in there without llvm and apparently it just does it. I have also been warned that it might be unusably slow. | 18:52:32 |
| rosssmyth | The future is now | 18:52:51 |
 emily | that's cool | 18:53:34 |
| emily | I only wanted it to give hope to the next person trying to unbreak mrustc-bootstrap | 18:53:41 |
| emily | after I drop its LLVM 12 dependency | 18:53:44 |
| emily | the idea would be to bootstrap to something that can use a modern LLVM so slowness probably wouldn't matter there | 18:54:04 |
