| 3 Sep 2024 |
 emily | I already wrote the comment | 16:23:08 |
| emily | (didn't mean to step on your toes) | 16:23:25 |
 Alyssa Ross | no worries | 16:24:00 |
| emily | FWIW it doesn't look like it's doing anything fancier than last time so I expect it'll break things | 16:24:03 |
| emily | we can probably move the file in the build hook instead, I'm assuming? | 16:24:40 |
| Alyssa Ross | Looks like the other PR this one is a dupe of has been tested properly | 16:24:43 |
| Alyssa Ross | Maybe we should just merge that one? | 16:24:46 |
| emily | aha | 16:25:03 |
| emily | the dupe looks better if we don't care about pre-2019 Rust versions | 16:25:31 |
| Alyssa Ross | Other PR looks very reasonable, as it doesn't touch fetchCargoTarball at all. | 16:25:35 |
| emily | which people can use with rust-overlay/fenix | 16:25:40 |
| emily | er, the non-dupe, rather. the [v2] one. | 16:25:49 |
| Alyssa Ross | I do not think it is our responsibility to support years old Rust compilers | 16:26:27 |
| emily | agreed | 16:26:33 |
| emily | so, I'm a little confused about the FODs | 16:27:49 |
| emily | do we not use the .cargo/config that fetch-cargo-tarball outputs at all? | 16:27:58 |
| emily | it seems like we'd need to mv or ln them to avoid the warning unless they're just totally unused as part of the Rust builds? | 16:28:13 |
| Alyssa Ross | that's what I'm wondering | 16:28:17 |
| emily | yet more reason opaque FOD bundles with computed stuff that can never be changed are bad :'( | 16:30:06 |
| * emily bumps Rust rework a couple places up her to-do list | 16:30:17 |
| Alyssa Ross | although everything in this PR seems good anyway | 16:31:04 |
| Alyssa Ross | worst case it only fixes some of the warnings | 16:31:11 |
| emily | yes, I think we could just merge it and sort out the rest later | 16:31:16 |
| emily | but it'd be good to understand what's going on and @mmlb probably knows that better than me :) | 16:31:46 |
 Ilan Joselevich (Kranzes) | In reply to @ultranix:matrix.org
I guess crate2nix wont be able to build with something like this https://github.com/paritytech/polkadot-sdk/blob/master/substrate/utils/wasm-builder/src/builder.rs#L54 that reaches out to the root workspace manifest
right? https://github.com/nix-community/crate2nix/commit/14d9c68123ae2d7ee94f91da28f3bf278c5cff0b can always do some patching and overriding | 17:33:34 |
| emily | is there any way to .overrideAttrs a Rust package in practice? I've overridden src and cargoDeps but it just complains about the lock file being out of date for reasons I can't fathom. | 20:05:24 |
| emily | doing the equivalent patch to the original package file works. | 20:05:33 |
 ghpzin | How did you override cargoDeps ? | 20:11:39 |
| emily | I tried cargoDeps = prevAttrs.cargoDeps.overrideAttrs { src = <the new src>; outputHash = "…"; } and it was behaving weird. now I'm trying cargoDeps = pkgs.rustPlatform.fetchCargoTarball { … } and it seems like it might be working? | 20:15:03 |
| emily | ah, I think I figured out what was going wrong. (flake input overrides acting weirdly.) | 20:19:30 |
