Subj : Re: Supressing errors?
To   : Brian Genisio <BrianGenisio@yahoo.com>
From : Brendan Eich <brendan@meer.net>
Date : Mon Feb 02 2004 01:31 pm

Brian Genisio wrote:

> Sure :)  IE v6.0 XP sp2
>
> ///////////////////////////////////////////////////////
> var myText = document.createTextNode("THIS IS A TEST");
> alert(myText.substringData(window, 4));
> ///////////////////////////////////////////////////////
>
> The DOMTextData method of substringData takes two integers.  If I send 
> in the object of DOMWindow, it defaults to the integer of 0.  The 
> example given will print "THIS".


This is not really a JS engine issue.  What DOM spec, if any, governs 
the behavior shown above?  Why should window convert to a string 
starting with "THIS"?  Why should substringData silently convert that to 
0?  I'm going to go out on a limb and guess that there is no open 
standard here.

> Of course, when I implement the interface for this method, and do a 
> JS_ValueToInt32, it fails.  I can use the default of 0 in this case. 
> When that the conversion fails, I want to supress the error, because 
> IE does not show an error either.


You don't really want to use a JS API intended for numeric input, then.  
You should probably check yourself for a non-numeric string literal.  
Isn't the literal in fact supposed to be a non-negative integer 
literal?  That's easy to check for in your own code.

> Like I mentioned earlier, I can do other types of checks, and I 
> already override the error handler, so I have solutions for this, but 
> I guess I was wondering if there was an API way to surpress an error 
> type... Or, it is possible that I am doing something wrong still :)


Is reverse-engineering the IE DOM wrong?  Some would say so ;-).  Don't 
encourage non-standard quirks if you can help it.  What's all this for, 
anyway?

/be

.