Andreas Rossberg

Indeed, for Java all refs are shared. Isn't that what you would expect? Partitioning the global heap into disjoint logical "worlds" would be possible by extending the "shared" attribute with...

@Horcrux7, in the extension I described `anyref` and `shared anyref` are different types (the latter is a subtype of the former). JS values can only be given type `anyref`, not...

I agree, and would go even further. As with the simd tests, we should have a subdirectory `threads`, and all related tests should go there. Similarly for other non-trivial and...

FWIW, removing redundant type index annotations is orthogonal. I believe I mentioned before that I'd like to remove them on GC instructions, per discussions like WebAssembly/function-references#27 and consistent with what...

My assumption for function references always was approach (2), i.e., coerce at the boundary, like you do for imports. With that, `call_ref` should be no slower than a regular `call`...

@RossTate: > #160 has been open for nearly a year, with no particularly concrete suggestions on how to address the issue That's not unexpected, because it's a Post-MVP question, and...

> Just to be clear, it's space proportional to the instructions where annotations would now be missing (not unrelated, existing instructions). Ah, okay, that's what I would have hoped. Thanks...

> Another thing to consider is what should be used as the "author(s)" of the citation. Based on what other communally developed languages do, I would suggest "The WebAssembly Community...

I think citing a GitHub repo is fine. We don't have proper papers for proposals.

Hm, it doesn't strike me as great API design to conflate the interface for globals with SIMD-specific stuff. If we want to provide SIMD accessors in JS, then let's introduce...