execute0

[Tyler Wright]

Has anyone done performance tests on execute0..2 vs. just using execute.apply?

From my tests the switch-statement actually slows it down enough that the single execute.apply is faster. I can double check though...

Tyler

ps I'm referring to some code internal to Signal for dispatching.

[simon]

You might want to try a if-statement in your tests as switch-statement
is considered slower then if-statement according to these tests
http://jacksondunstan.com/articles/1007

si

[Robert Penner]

Jackson's article on switch vs. if-else vs. ternary was an eye-opener.
I thought the bytecode was basically the same. Switches were generally
10-15% slower in his tests, but the impact in the context of a Signal
dispatch would need to be tested.

Robert

[David,Linse]

----- Ursprüngliche Nachricht -----
Von: simon <stick...@gmail.com>
Gesendet: Dienstag, 1. Februar 2011 12:10
An: as3-signals <as3-s...@googlegroups.com>
Betreff: [as3-signals] Re: execute0

[Tyler Wright]

I think my tests were with a single listener - switch vs always function.apply. It makes sense that if there were many listeners, using a switch/if-else/ternary initially wouldn't hurt the overall performance gain of choosing a more optimized execute statement.

But I still look at how much less code there would be, longingly, of skipping the executes altogether. Yes Stray, I said 'less code' ... I just can't help it, I'm flawed, obsessive.

Tyler