wasmtime: 0.21.0 -> 0.35.2

[ereslibre]

Description of changes

Supersedes #109571

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 22.05 Release Notes (or backporting 21.11 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
  • (Release notes changes) Ran nixos/doc/manual/md-to-db.sh to update generated release notes
• Fits CONTRIBUTING.md.

[ereslibre]

/cc @happysalada

[happysalada]

Just tested and this won't build for darwin. (The ci shows the failure if you are interested, but I'm afraid those are quite cryptic)

FAILED: obj/libv8_libbase.a
rm -f obj/libv8_libbase.a && "../../../../../../nix/store/cvz17hkb0wpbsyx58l4zcaamqs2bhlmz-clang-wrapper-11.1.0/bin/llvm-ar"  -r -c -s -D obj/libv8_libbase.a "@obj/libv8_libbase.a.rsp"
/bin/sh: ../../../../../../nix/store/cvz17hkb0wpbsyx58l4zcaamqs2bhlmz-clang-wrapper-11.1.0/bin/llvm-ar: No such file or directory

Could you change the platform to only linux ?

Also the ci fails on linux with the following

failures:
    commands::compile::test::test_successful_compile
    commands::compile::test::test_x64_flags_compile
    commands::compile::test::test_x64_presets_compile

test result: FAILED. 16 passed; 3 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.05s

Could you mark those tests to be skipped ?
something like

  checkFlags = [
    # tries to make a network access
    "--skip=sinks::loki::tests::healthcheck_grafana_cloud"
  ];

(I just copied and pasted the relevant part from another test to give you an example in case you hadn't seen it. (more can be found in pkgs/tools/misc/vector/default.nix

oh and last but not least, thanks a lot for the PR!

[ereslibre]

Could you change the platform to only linux ?

Done, although I will check what's wrong, I'd like to have the derivation for Mac OS as well. I would not block this PR because of that though.

Could you mark those tests to be skipped ?

Done, and I have done the same with the patch that I had removing a test, and removed the patch that was removing the test itself. Thanks for the pointers! 🙌

Also the ci fails on linux with the following

I cannot reproduce this and I wonder why. Do you know how can I reproduce this locally, or get more information from the CI run? I would like to go further into this.

oh and last but not least, thanks a lot for the PR!

:)

[happysalada]

Ci passes, thank you!

I think the ci is running in strict eval or something. I have to say, I forgot the exact difference.

If you ever get to the darwin issue, feel free to ping me, happy to test.

Thanks again!

[happysalada]

oh small detail regarding the first commit. It should probably be wasmtime: fix build
Some people like to have the commit message start by the package name to make it easier to search through commits.
Just something for next time.

[nomeata]

Oh, did we just lose wasmtime on Darwin? That's be very unfortunate. I hope it's only temporary. (Pinging @ggreif)

[ereslibre]

Oh, did we just lose wasmtime on Darwin?

I want to check why it's failing and my plan is to restore it as soon as possible. It was also nice to have a more recent version of wasmtime.

[nomeata]

Ok, thanks :-)

[ereslibre]

@nomeata I will keep looking into it. v8 was being problematic so the first approach I wanted was to disable checks only on darwin. Now, I'm getting a linker error. I will open a PR when I have something that works :)

WIP: master...ereslibre:wasmtime-build-darwin

The error, in case is meaningful:

error: could not compile `wasmtime-cli` due to previous error
error: builder for '/nix/store/rjx90jsczzn5v4y74qd79v4yivgl8255-wasmtime-0.35.2.drv' failed with exit code 101;
       last 10 log lines:
       >   = note: Undefined symbols for architecture x86_64:
       >             "_futimens", referenced from:
       >                 rustix::imp::fs::syscalls::futimens::h8cff26252e668e9b in librustix-960c26ac3e1fab35.rlib(rustix-960c26ac3e1fab35.rustix.00f71eac-cgu.4.rcgu.o)
       >             "_utimensat", referenced from:
       >                 rustix::imp::fs::syscalls::utimensat::h92f324ed16adf14b in librustix-960c26ac3e1fab35.rlib(rustix-960c26ac3e1fab35.rustix.00f71eac-cgu.4.rcgu.o)
       >           ld: symbol(s) not found for architecture x86_64
       >           clang-11: error: linker command failed with exit code 1 (use -v to see invocation)
       >
       >
       > error: could not compile `wasmtime-cli` due to previous error
       For full logs, run 'nix log /nix/store/rjx90jsczzn5v4y74qd79v4yivgl8255-wasmtime-0.35.2.drv'.

[sunfishcode]

@ereslibre bytecodealliance/wasmtime#4022 is a PR which addresses those linker errors.

[ereslibre]

Thank you @sunfishcode!

I just did a smoke test (master...ereslibre:wasmtime-build-darwin) and it builds and runs correctly.

I will submit a new PR to nixpkgs as soon a new wasmtime version is tagged.

cc/ @nomeata

[ereslibre]

wasmtime 0.36.0 was released. I opened #169544.

It builds and links in darwin under aarch64, but fails to link under x86_64 (when running with rosetta).

We can follow up over there.

[ereslibre]

@nomeata #173870 is up and brings wasmtime to darwin back again :)

/cc @terlar