From D doc:

int test()
{
Foo f;

f.data = 3;		// same as f.data(3);
return f.data + 3;	// same as return f.data() + 3;
}

So how about:

f.data += 3;   // same as f.data( f.data()+3 )
f.data -= 3;   // same as f.data( f.data()-3 )
f.data *= 3;   // same as f.data( f.data()*3 )
f.data /= 3;   // same as f.data( f.data()/3 )

>>=
<<=
..

> So how about:
>
> f.data += 3;   // same as f.data( f.data()+3 )
> f.data -= 3;   // same as f.data( f.data()-3 )
> f.data *= 3;   // same as f.data( f.data()*3 )
> f.data /= 3;   // same as f.data( f.data()/3 )

The op= operators are guaranteed to evaluate the lvalue only once. If you rewrite it the way you propose, this would not be true anymore.

Ciao
uwe

Uwe Salomon wrote:
>> f.data += 3;   // same as f.data( f.data()+3 )
>> f.data -= 3;   // same as f.data( f.data()-3 )
>> f.data *= 3;   // same as f.data( f.data()*3 )
>> f.data /= 3;   // same as f.data( f.data()/3 )
> 
> The op= operators are guaranteed to evaluate the lvalue only once. If you  rewrite it the way you propose, this would not be true anymore.

As far as I can see, (T prop(T)) is still only being called once.  Is there some case wherein the Settor (T prop(T)) needs to emulate a Gettor (T prop()), thereby causing an internal change before the operation?  If not, then I don't see how its no longer true.

-- Chris Sauls