Conversation
|
I started an darwin-aarch64 build (might take a while because it's building stdenv...) |
|
Building & testing both firefox and thunderbird. Edit: Both are looking fine. |
|
I started an aarch64-linux build as well. |
|
firefox and thunderbird are fine on x86_64-linux |
|
Built |
|
(noise, but I can't hide this comment) These are the ones which are already broken on master for darwin-aarch64. ario (broken on master): cargo-watch (broken on master): coq8.15-InfSeqExt, and indeed large tracts of coq which just seems extremely borked: Lighthouse (not packaged for darwin-aarch64 on master): MozWire (broken on master): cobalt (broken on master): eccodes (broken on master): alfis (broken on master): gbl (broken on master): gssdp (broken on master): gurk-rs (broken on master): jsonwatch (broken on master): a52dec (broken on master): mate.mate-menus: pqrs: (this is in e.g. pqrs-0.1.1-vendor.tar.gz/lexical-core/src/atof/algorithm/bhcomp.rs:62:24, a great deal of "error[E0277]: cannot divide |
|
(likely noise, but I can't hide this comment) Ones which work on darwin-aarch64 master, fail on nixpgs-review, but appear to be nothing to do with this PR. bsnes-hd-10.6-beta (works on master): Lua (works on master): Luarocks (works on master): mikutter (works on master): freeciv (works on master): geoclue2 (works on master): |
|
@Smaug123 many of these packages don't even depend on Rust. Sorry but this is just not useful at all. |
|
Fair enough; there are a few Rust ones in there, but they can presumably be cleared up after the fact. |
|
Useful would be "x does build on master but does not build with this change". If you can re-test all of these on master and filter out those that also fail there we could actually derive some meaning from it. |
|
Certainly, I was intending to do this all over again on master afterwards, but this PR takes many days to build; my M1 Max has been maxed out for two days now, so "afterwards" is not yet. (For what it's worth, the number of times I've been rescued by random snippets of nix build traces on apparently unrelated GitHub issues makes me strongly prefer to default to putting them somewhere, and there's not really any official place to do it.) |
|
The official place for build failures that are not related to any PR would be creating a seperate issue for that (title could be "lua does not build on aarch64-darwin"). |
This is nix-review in action, by the way. To the best of my knowledge, these are all downstream of this change. |
|
Ah, then this might be because this branch is based on staging, but nix-review compared to master? |
|
nixpkgs-review's docs say that it "will merge the pull request into the pull request's target branch (most commonly master)", and it's definitely working against staging because the most recent commit in |
|
A friend reports that fd builds fine from this PR on aarch64-darwin |
|
This is the ones which work on master darwin-aarch64 but were broken according to Ignore if desired. elfcat (works on master): cargo-generate (works on master): More specifically: dotter: These errors all appear to be the same dotenv-linter (works on master, but surely an error of this shape is not caused by a rustc upgrade): |
|
aarch64-linux and aarch64-darwin can build |
|
Regarding cargo-generate, we should first try if simply upgraded to https://github.com/cargo-generate/cargo-generate/releases/tag/v0.13.1 fixes the build again. |
|
elfcat looks like the vendoring in buildRustPackage is broken. The code run there is before any rustc binary is involved. |
|
Parent PR: #175150 |
Description of changes
nix-build -p fdpasses on x86_64-linuxWaiting for Firefox and Thunderbird.
Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)nixos/doc/manual/md-to-db.shto update generated release notes