Module bytecode_interpreter::concrete::ty
source · Expand description
The type system in move-model is a fat type system that is designed to cover all cases that can possibly appear in the whole bytecode transformation pipeline. Natually, this means that some types are no longer applicable when the Move program reaches the end of the transformation.
The type system for the interpreter is a strict subset of what is offered in the move-model.
In other word, it is slimmed down version of the type system in move-model and a very restricted
set of types that are only applicable to the interpreter. Doing so enables us to write code in a
more precise way. For example, a type argument can only be a BaseType and never a reference.
Therefore, BaseType is preferred over Type for if a struct field/function argument holds
a type argument (e.g., struct FunctionContext {ty_args: Vec<BaseType>, ...} is preferred over
ty_args: Vec<Type>, as the former is more descriptive and less error prone).