[spec] Incoherent type for memory.grow

[4y8]

In the validation rules, memory.grow has type at -> at where at is an address type which can be either i32 or i64. In the index of instructions, it has type i32 -> i32.

I can make a PR if that helps.

[4y8]

This actually seems to be the case for most memory instruction - at being replaced by i32.

[rossberg]

Indeed, somebody forgot to update the instruction index when we added memory64. The same issue with table instructions. A PR would definitely be welcome!

[4y8]

Here it is #1948. I may have missed some however