digitalmars.D.learn - Calling un-overridden class method
- H. S. Teoh (43/43) Oct 12 2012 Today I ran into a bit of a bind. I have a class hierarchy in which a
- Timon Gehr (3/8) Oct 12 2012 I'm sure it does :).
Today I ran into a bit of a bind. I have a class hierarchy in which a
base class B defines a method eval, which returns a forward range struct
whose save method consists of a delegate that simply re-invokes eval
with the same arguments. Then there's a derived class C, which overrides
B.eval, but uses B.eval as part of its implementation of C.eval.
The problem is, I can't seem to specify that I want it to _statically_
bind the save method to call B.eval; even though I specify B.eval
explicitly, it still ends up in C.eval, thus causing infinite recursion:
class B {
// This is a forward range
static struct ResultRange {
ResultRange delegate() saveImpl;
property auto save() { return saveImpl(); }
... // other range methods
}
ResultRange eval(T[] args) {
return ResultRange(
// calls C.eval(), even though B.eval is
// explicitly specified!
/* this is saveImpl */ () => B.eval(args),
...
);
}
}
class C : B {
ResultRange eval(T[] args) {
auto orig_range = super.eval(args);
// infinite recursion that eventually overflows
// the stack.
auto saved_range = orig_range.save;
auto modifiedRange = ...;
return modifiedRange;
}
}
Why does D still generate a dynamic call to the overridden eval method,
even though I explicitly asked for B.eval? How do I get a static binding
to B.eval? Is there a way to work around this?
Perplexed,
T
--
PNP = Plug 'N' Pray
Oct 12 2012
On 10/13/2012 04:22 AM, H. S. Teoh wrote:... // infinite recursion that eventually overflows // the stack. ...I'm sure it does :). topic: yes this is a bug.
Oct 12 2012








Timon Gehr <timon.gehr gmx.ch>