QScriptEngine: which expression is faster ?

[dextermagnific]

Hi Everybody,

I'm making a scriptable application, and I'm wondering which kind of expression is faster:

engine.evaluate("object1.property1").toString();
engine.evaluate("object1.property2").toString();
engine.evaluate("object1.property3").toString();

or

QScriptValue object1 = engine.property("object1");
object1.property("property1").toString();
object1.property("property2").toString();
object1.property("property3").toString();

Is the second method faster (I 'save' object1 and I'm hoping that property() does not invoke the interpreter). ? I'm I right ?

Thanks

[A Former User]

Hi. Why don't you just measure and see what's faster?

[dextermagnific]

Possible answer: it's the same.
Either property() invokes the interpreter or object1 is "changed" everytime the underlying script changes it.

Because if I do:

// let's assume object1 is an array
QScriptValue object1= engine.property("object1");

Then I use Javascript to alter the contents of object1 (for example using evaluate())

and I do:

qDebug() << object1.toVariant();

The object1 will contains the latest changes, even if it has been changes outside C++.

This makes me believe that QScriptValue objects are tied to the underlying JS objects. Any change in the JS will immediately be seen in C++.