Parsing and validation of atomic instructions.

[resulknad]

Looking for early feedback on general direction, content and style. Thank you.

[ia0]

Thanks! Looks good so far. I didn't review in details though, I'll do so once it's closer to mergeable.

I'm not sure what's the best solution for the tests is. I think the simplest (I think that's what I did in my dev branch when I wrote the interpreter initially), is to add #[ignore] on the tests that fail, like atomic, and just remove the #[ignore] locally to know where you are, and remove it definitely once the full test passes. This way you don't need to modify the tests and can use them directly from the wasm/threads repo. You can always modify them there locally for testing purposes.

[ia0]

I've merge #467 which let's you write test!("threads", atomic); or actually test!(#[ignore] "threads", atomic); initially.

[ia0]

I didn't review in details because the PR is not up-to-date with the branch (let's merge the submodule PR first). Once done, I'll do a thorough review. But from a quick glance over the PR I'd say the overall approach is sound, so it's just about me checking the spec on one side and the PR on the other.

[ia0]

Thanks! It seems almost there.

[ia0]

Thanks!

[ia0]

Thanks! Looks good enough to merge. The details can be handled when we'll merge this branch into main.

[ia0]

The changelog test is a bit noisy for a dev branch, let me remove it in #479.

[ia0]

I've resolved the comments, but 2 things you could do in main are:

• Always enable the mutable globals since we always support them.
• Move the FC parsing to a parse_instr_fc similar to the _fe you just did.