Specify maximum static size for component value types

[ANAMASGARD]

Fixes #682

Summary

Adds a spec-defined maximum static value size of 2^28 - 1 bytes for Component Model value types.

Validation now:

• checks both i32 and i64 pointer types;
• uses overflow-safe size calculations;
• validates nested/despecialized types recursively;
• rejects oversized types during validation instead of at runtime.

Includes Python and WAST regression tests for fixed-length lists, compound types, nested maps, pointer-width differences, and integer-overflow cases.

The WAST test remains in nyi.txt until wasm-tools/Wasmtime implement the corresponding validator rule.

[lukewagner]

Thanks for working on this! To resolve #682, I had been thinking that perhaps we could reuse the elem_size function as already defined, and simply add validation notes mentioning it in all the relevant places: Explainer.md, Binary.md, and canonopt validation. That should then reduce the size of the PR.

[ANAMASGARD]

@lukewagner Thanks for the feedback

Validation now reuses the existing elem_size definition instead of adding a separate checked_* algorithm. The full rule is in Element Size in CanonicalABI.md (every despecialized AST node, both i32 and i64), with short pointers in Explainer.md, Binary.md, and canonopt validation. WAST tests cover the edge cases; nyi.txt until wasm-tools/Wasmtime implement the validator.

@alexcrichton please review this PR and give your feedback , thanks

[alexcrichton]

This seems fine by me modulo precise spec text yeah, it's what I was imagining on #682 at a high level

[lukewagner]

Overall looks good to me; thanks! A few requested changes and then I expect it'll be good to go.

[ANAMASGARD]

@lukewagner & @alexcrichton Thank you so much for the review .
I have applied the requested wording and placements : Element Size intro + Binary/canonopt notes, Explainer paragraph moved next to defvaltype. WAST now has one valid component (including just-under-bound compounds); assert_invalid covers types whose elem_size(t, i64) is not less than 2^28.

[lukewagner]

Looks good to me, thanks.

[ANAMASGARD]

@lukewagner I have wrapped the Binary.md validation bullet as requested.