I’d like to declare a gene开发者_如何学运维ric class Simple where a method called Addition would take two generic variables and add them together using + operator. I thought I could accomplish this by giving class Simple the following constraint:
class Simple<T> where T: int
{
public T Addition(T firstInt, T secondInt)
{
return firstInt + secondInt;
}
}
I suspect error has something to do with generics only having the following five types of constraints - ClassName, class, struct,InterfaceName, new()? Thus, why don’t generics also support StructureName constraint type? That way we would be able to apply arithmetic operators on generic variables?!
thanx
A StructName
constraint wouldn't make sense, since you cannot inherit from value types.
If you had something like class Simple<T> where T : int
you could only instantiate Simple<T>
with T = int
, no type inherits from int
or any other value type for that matter.
C# (the CLR) lacks what other languages know as type classes, one popular mechanism to handle operator overloading without hard-coded compiler mechanics (like in C#).
There is no point in constraining a Generic type to a specific struct, since struct types cannot be inherited. This would be identical to writing:
class Simple
{
public int Addition(int firstInt, int secondInt)
{
return firstInt + secondInt;
}
}
Unfortunately, with support for an IArithmetic interface, generic types with mathematical operations are difficult to support. If you want to see one approach for generic-based mathmetical operators, look to the Generic Operators in MiscUtil.
You constrain T
to int
. This is no sensible thing to do, as there is only one int type; so you could directly want to write int
if you want int
.
If you do not constrain T
, you cannot use the +
operator, since there is no +
operator defined on object.
C# generics are not C++ templates.
精彩评论