• Jacob Carlborg (21/21) Dec 05 2009 Would it be a good idea to allow overloading on opDispatch to make the
• Tim Matthews (6/31) Dec 05 2009 Do you think it would be a good idea? or just want to know what others

Jacob Carlborg <doob me.com> writes:

Would it be a good idea to allow overloading on opDispatch to make the 
following compile, what do you think?

import std.stdio;

class C
{
     void opDispatch(string s, ARGS...) (ARGS args)
     {
	writefln("S.opDispatch('%s', %s)", s, typeof(args).stringof);
     }

     void foo (string s)
     {
         writeln("in asd");
     }
}

void main()
{
     C c = new C;
     c.foo("bar"); // calls C.foo
     c.foo("bar", "bar"); // calls C.opDispatch
     c.foo(3); // calls C.opDispatch
}

Tim Matthews <tim.matthews7 gmail.com> writes:

Jacob Carlborg wrote:
 Would it be a good idea to allow overloading on opDispatch to make the 
 following compile, what do you think?
 
 import std.stdio;
 
 class C
 {
     void opDispatch(string s, ARGS...) (ARGS args)
     {
     writefln("S.opDispatch('%s', %s)", s, typeof(args).stringof);
     }
 
     void foo (string s)
     {
         writeln("in asd");
     }
 }
 
 void main()
 {
     C c = new C;
     c.foo("bar"); // calls C.foo
     c.foo("bar", "bar"); // calls C.opDispatch
     c.foo(3); // calls C.opDispatch
 }

Do you think it would be a good idea? or just want to know what others 
think for no obvious motive? If the latter: absolutely not. It would be 
a major source of bugs. That code would be much safer if C.foo was a 
private method that opdispatch called when it determines its arguments 
met the requirements.