Subj : Re: spidermonkey: where global vars live
To   : Brian Barnes <ggadwa@voyager.net>
From : Brendan Eich <brendan@meer.net>
Date : Mon May 12 2003 12:07 pm

Brian Barnes wrote:

> I thought I knew it all, but little pieces of information keep 
> escaping me no matter how hard Brendan tries to pound them in :)
>
> Take this script:
>
> var x=5;
>
> Where does the global reside?  In the global object I initiated the 
> standard classes on and use in CallFunction?


Those are (or can be) two different objects (1) and (2), as I think you 
say below.

If you're calling a function defined by the script, you must be 
evaluating the script first, with JS_ExecuteScript, right?  The obj 
param you pass to that (2) is used to find the place where global 
variables are bound.

Do read the JSOPTION_VAROBJFIX comments in jsapi.h for a wrinkle.

>   Or in the "global object" I can get from the context with 
> js_getglobalobject?


That's the object you passed to JS_InitStandardClasses, if not to 
JS_SetGlobalObject.  So that's (1), above.

> What I'm trying to do is use the same compiled script on a 
> standard-class initialized global object


You don't need a standard-class-initialized global object when precompiling.

FYI, tangential point only: with recent versions of SpiderMonkey, when 
precompiling a function using any of the JS_Compile*Function* APIs, you 
don't need any global object at all -- you can pass NULL as the obj 
parameter.  I hope to relax the JS_Compile*Script* API rules so you can 
pass null for the obj param with those functions, too, but that will 
require some jsparse.c changes.

> over and over with a different set of global data (instead of using 20 
> different global objects each with the same script).
>
> What I got now works, it just seems a bit unecessary.


Not clear what you have now -- are you recompiling the script for each 
of those 20 global objects?  You don't need to do that.

/be

>
>
> [>] Brian
>
>

.