Working on custom interpreter of IL code I've noticed that the following code can be compiled by ilasm.exe and executed well:
.locals init (int32 x, int64 y, int64 z)
ldloc x
ldloc y
add
stloc z
But in the ECMA-335 (unfortunately I can't insert a link to html, only pdf seems to be available) I see that combination of int32 and int64 is not supported for binary operations (as shown in the table Table III.2: Binary Numeric Operations).
Moreover adding float32 and int64 also can be compiled and executed.
Is it a mistake in the .net runtime?
How to check what combinations are really correct and should be supported?