On 02/22/2014 12:17 PM, Frustrated wrote:
> Again, the whole point of why it is illegal because you can pass
> a RogueButton... BUT I DON'T INTEND TO PASS THEM!
Thinking that way, many rules of a statically type-checked language like
D would be unnecessary. ;)
> WindowsGui only uses WindowsButton which is a iButton type.
WindowsGui is totally free to use any type it wants. However, it cannot
both claim to implement an interface without actually obeying its
requirements.
> Why can't I relax the condition to use the base type?
The users of iGui don't even know what a WindowsButton is:
interface iGui
{
@property iButton button(ref iButton button);
}
Imagine the following that I write:
void foo(WindowsGui gui, iButton b)
{
gui.button(b);
}
Is the call legal or not? How would I know and why should I care? I have
a WindowsGui, which happens to be an iGui and I have an iButton.
According to the contract of the interface I should be able to call it
without fear of type problems.
Even, I can create my own iButton and pass it to that WindowsGui:
gui.button(new MyButton());
I should be able to do all of that just because WindowsGui promises to
be an iGui.
Getting back to what I said above: WindowsGui is totally free to use any
type it wants. That's why both Steven and I made it take its button as a
constructor parameter. Then, you also mentioned a setter. That's fine
too. As long as WindowsGui allows me to give it an iButton when I call
button(), it is all fine.
Ali
