On 10/11/18 1:02 PM, aedt wrote:

In the documentation, it has been discouraged to use addOnActivate() function and it also has been suggested that the users should use addOnNotify(). In the D forum 5 years ago, it has recommended that switch signals be handled as follows:

Switch sw = new Switch();
sw.addOnNotify(&notify);

void notify(ParamSpec param, ObjectG obj)
{
	if(param.getName() == "active")
	{
		//Do stuff
	}
}

I am using this Switches like this:

         auto sw = new Switch();
         sw.addOnNotify(delegate void(ParamSpec p, ObjectG _y) {
             if (p.getName() == "active") {
                 static bool on;
                 on = sw.getActive();

                 //do stuf
                 import std.stdio : writeln;

                 writeln(on ? "On" : "Off");
             }
         });

I have two questions -

1. Is it still the latest way to connect activate signals to switches?
2. How is it different that the code below:

         auto sw = new Switch();
         sw.addOnNotify(delegate void(ParamSpec _x, ObjectG _y) {
             if (sw.getActive()) {
                 // Do stuff
             }
         });

I would probably do something like this:

auto sw = new Switch();
sw.addOnNotify(delegate void(_, __) {
	if (sw.getActive()) {
		// Do stuff
	}
}, "active");

When using inline delegates / lambadas the so we don't have to specify
the types of the parameters.
Using _ and __ as the parameter names since we don't use them.
With the "active" at the end the delegate is only called when the
active property of the switch changes.

Whether to use a inline delegate / lambada or a separate function that
will be passed in is personal preference. Personally i would base my
decision on the size of: "// Do stuff".