In the section about the validation of instructions (https://webassembly.github.io/spec/core/valid/instructions.html#instructions), there is a note that contains an example using instruction select with f64 consts. The example is rendered as:
(f64.const +64) (f64.const +64) (f64.const +64) (select)
but it should contain 1, 2 and 3 instead of the three +64, as it is suggested by the previous example (which uses i32) and by the fact that the source code contains instructions $fnat(64, 1), $fnat(64, 2) and $fnat(64, 3). Maybe the problem is in the rendering of these instructions.
Moreover, I think that the third parameter should still be an i32 instead of an f64, as the type of the select instruction is t t i32 -> t.
In the section about the validation of instructions (https://webassembly.github.io/spec/core/valid/instructions.html#instructions), there is a note that contains an example using instruction
selectwithf64consts. The example is rendered as:but it should contain 1, 2 and 3 instead of the three +64, as it is suggested by the previous example (which uses
i32) and by the fact that the source code contains instructions$fnat(64, 1),$fnat(64, 2)and$fnat(64, 3). Maybe the problem is in the rendering of these instructions.Moreover, I think that the third parameter should still be an
i32instead of anf64, as the type of theselectinstruction ist t i32 -> t.