Subj : Re: SM: JS_ExecuteScript and rval
To   : Franky Braem <fa225992@skynet.be>
From : Brendan Eich <brendan@meer.net>
Date : Sun Apr 06 2003 03:45 pm

Franky Braem wrote:
> Hi,
> 
> I read the following in the API-docs
> 
> JS_ExecuteScript executes a previously compiled script, script. On
> successful completion, rval is a
> pointer to a variable that holds the value from the last executed expression
> statement processed in
> the script.
> 
> What happens when the script only contains a function? I returns undefined.
> Wouldn't it be nice when the function object is returned?
> 
> I want to use this to implement a better way for event handling in wxJS. eg.
> 
> frame.onClose = wxScript("close.js");


This seems wrong: the value returned by wxScript is not the result of 
evaluating the script in close.js, correct?  It's a script object that 
can be executed later, whose code is compiled from the source in 
close.js, right?

You are removing one level of reference here.  I think that makes for 
problems where users explicitly want a script object, to defer multiple 
executions and amortize compilation overhead.  What if there are several 
functions in close.js?  Why can't I refer to each of them?  Why must I 
break each handler into its own source file?

It would be better to require users to say what they mean:

frame.onClose = wxScript("close.js").onCloseHandler;

You would have to extend wxScript so it could look up functions declared 
in the compiled script.

/be

.