The last examples showed, how a function, that is not available as such, will be added to the generated Python sourcecode.
In Python there are functions to convert variable types, too. To convert a variable to the String type there is the function str(). To convert a variable to an Integer there is the function int(). How can we use these native function directly?
One idea could be to just write the Python function into the Haxe sourcecode like so:
package;
class Main {
static function main() {
var ten = "10";
var eight = 8;
var str_eight = str( eight );
var int_ten = int( ten );
}
}
If you try this the Haxe compiler gives you two error messages:
src/Main.hx:10: characters 18-21 : Unknown identifier : str
src/Main.hx:11: characters 16-19 : Unknown identifier : int
That's because the functions str() and int() are not known to the compiler.
Well, but we know, these two functions will be valid in Python. For a case like this there is the keyword untyped, which we include before the function call:
package;
class Main {
static function main() {
var ten = "10";
var eight = 8;
var str_eight = untyped str( eight );
var int_ten = untyped int( ten );
}
}
The compiler is pleased, and the generated Python code looks like this:
class Main:
@staticmethod
def main():
ten = "10"
eight = 8
str_eight = str(eight)
int_ten = int(ten)
Main.main()
Perfect! Just what we wanted.
When using the untyped keyword, keep in mind that the compiler doesn't check the rest of the line for errors.
It will not be checked
Other examples for untyped are when functions in your own Extern classes are not defined, or for matrix multiplications with NumPy. More about that in a later part.
The compiler still checks if the code after untyped is valid Haxe code. Something like this won't go through, because array access with : and , is not possible in Haxe.
var is_value_empty = untyped data[:,4] == "";
For something like this there is another method.
Next part: Syntax.pythonCode() (coming soon)