patch: correctness of <script> tag attributes

[John Sullivan]

On Oct 14, 2004, at 12:01 PM, David Faure wrote:


> On Friday 24 September 2004 04:19, John Sullivan wrote:
>> We discovered a web page that hangs in an infinite JavaScript loop 
>> that
>> doesn't hang on other browsers because the <script> tag had a
>> deliberately bad value for its language attribute of
>> "disabled_JavaScript". Our test for valid values of the language and
>> type attributes of the <script> tag was much more permissive than 
>> other
>> browsers, so we accepted "disabled_JavaScript" as a valid value.
>>
>> Here's the patch, and a test page that shows a variety of values of
>> "language" and "type" that can be compared against different browsers.
>
> Thanks a lot for the patch and analysis. Merged, tested and committed.

One followup to this: I neglected to include "application/x-javascript" 
in the list of legal types. Mozilla accepts this, though WinIE doesn't; 
I've added it to the list. After some further investigation, I did not 
find any other "application/xxx" types that should be included.

>
>
> (This DomStringImpl::value() method must be new in safari, I had to use
> DOMString(a).string())

I don't know about this detail; maybe someone else does?

John