You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Dec 22, 2021. It is now read-only.
The reference-types proposal introduced a 'number type' [1] set of the value types. The practical use of this is in restricting the set of allowable types to the 'untyped select' instruction. This was done in reference-types#42 to avoid needing to compute a non-trivial lub/glb between two types.
SpiderMonkey implements the SIMD proposal, but does not allow v128 in untyped select as it was not in the initial set of number types. I've been told that V8 does allow this.
So I'm wondering if we should expect v128 to be a member of the number type set? This seems very reasonable to me, I just want to ensure there's no one strongly opposed.
The reference-types proposal introduced a 'number type' [1] set of the value types. The practical use of this is in restricting the set of allowable types to the 'untyped select' instruction. This was done in reference-types#42 to avoid needing to compute a non-trivial lub/glb between two types.
SpiderMonkey implements the SIMD proposal, but does not allow v128 in untyped select as it was not in the initial set of number types. I've been told that V8 does allow this.
So I'm wondering if we should expect v128 to be a member of the number type set? This seems very reasonable to me, I just want to ensure there's no one strongly opposed.
cc @rossberg @kripken
[1] https://webassembly.github.io/reference-types/core/syntax/types.html#syntax-numtype
The text was updated successfully, but these errors were encountered: