Nix Rust | 698 Members | |
| Rust | 158 Servers |
| Sender | Message | Time |
|---|---|---|
| 25 Mar 2025 | ||
which would meet the "no upstream Cargo.lock" use case and ~obsolete importCargoLock | 14:08:53 | |
| I guess very much like with fetchYarnDeps | 14:09:44 | |
but for historical reasons almost everyone uses ${src}/yarn.lock instead of inherit src there (it's not IFD, since it only uses that in the build process) | 14:10:37 | |
| https://github.com/NixOS/nixpkgs/blob/15f3d37c73c8c1090f8fef7b8508675c9260eab6/pkgs/build-support/rust/import-cargo-lock.nix
isn't a trivial maintenance burden, so if 25.05 is doing breaking changes anyway, it might be best to eliminate it | 14:11:05 | |
| * https://github.com/NixOS/nixpkgs/blob/15f3d37c73c8c1090f8fef7b8508675c9260eab6/pkgs/build-support/rust/import-cargo-lock.nix + https://github.com/NixOS/nixpkgs/blob/15f3d37c73c8c1090f8fef7b8508675c9260eab6/pkgs/build-support/rust/replace-workspace-values.py isn't a trivial maintenance burden, so if 25.05 is doing breaking changes anyway, it might be best to eliminate it | 14:11:09 | |
ah I guess we use the latter in fetchCargoVendor too | 14:11:34 | |
| so it's not so bad | 14:11:35 | |
| but it still seems better to just consolidate | 14:11:45 | |
| if still thinking of how we could replace depsExtraArgs with a better alterative | 14:12:27 | |
| * if still thinking of how we could replace depsExtraArgs with a better alterative (semi related) | 14:13:03 | |
it looks like we could add lockFileContents ? null and lockFile ? null to fetchCargoVendor, and simply replace ./Cargo.lock with ${if lockFile != null then "${lockFile}" else if lockFileContents != null then writeText … else "./Cargo.lock"} | 14:13:47 | |
| I guess the downside is for "local" flakes with Rust packages though | 14:14:20 | |
| so maybe it is not a good idea, even if it is better for Nixpkgs | 14:14:32 | |
| btw we're still using replace-workspace-values.py in fetchCargoVendor | 14:14:42 | |
| yeah I saw :) | 14:14:48 | |
| anyway, ok, this is a distraction. if we want to do that we can do it another release anyway | 14:15:16 | |
| but I think we should do this now before we forget :P | 14:15:26 | |
I'm reproducing system-syzygy before merging | 14:15:38 | |
| fwiw, the thing about supporting old lock files isn't so much about existing packages but the fact that I find it hard to imagine a PR to add a new one that we'd want to accept | 14:19:04 | |
so given the existence of a fallback path in the form of importCargoLock it seems bad to add any complexity to our current infrastructure for it | 14:19:31 | |
| yeah, not worth rebuilding so many packages for those | 14:24:01 | |
| I know i've already sent this in a few times, but: https://github.com/NixOS/nixpkgs/pull/390171 | 14:25:49 | |
| not sure I understand this PR enough to approve it without reading up more on the issue | 14:27:52 | |
| Toma: are you up for doing this or would you like me to have a go at it? | 15:13:08 | |
| I will be away thursday-sunday when I won't really be able to respond, so I'd appreciate if someone else did it | 15:14:19 | |
| 17:42:02 | ||
| 17:44:37 | ||
| What's the desired avenue for discussing and submitting problems found within
| 18:05:21 | |
| * What's the desired avenue for discussing and submitting problems/fixes found within
| 18:05:32 | |
| IMO opening an nixpkgs issue with some details is fine btw I had a TODO for fetchCargoVendor to add a way to deduce the download URL schema from the registry url. (but first we have to figure out how under what directory should we store files fetched from non-standard registries) | 19:01:23 | |