It is definitely possible, and is a pattern used in the standard library, especially in the Lambda class:
class Test {
static function main(){
var arr = [0,1,2,3,4,5,6,7,8,9,10];
var newArr = Lambda.filter(arr, function (num) {
return num % 2 == 0;
});
for (i in newArr)
{
trace (i);
}
}
}
(See http://try.haxe.org/#C9dF3)
To define your own methods that take functions as parameters you use the (param1Type)->(param2Type)->(returnType)
syntax:
function test1(myFn:String->Void) { myFn("hi"); }
test1(function (str) { trace(str); });
function test2(myFn:String->String) { var newStr = myFn("hi"); }
test2(function (str) { return str.toUpperCase(); });
function test3(myFn:Int->String->Array<Int>->Void) { myFn(3, "Jason", [1,2,3,4]); }
test3(function(num:Int, name:String, items:Array<Int>) { ... });
b(a)
you substituted the name of a function with its definition. These are almost always the same. I say almost always because if haxe tries to do something funny to the function argument (saya
had a parameter and, upon passing it as an argument the semantics of haxe enforced the argument to be reduced to some normal form...you may end up with non-termination. As it stands I can't think of any difference – Jut