[webkit-dev] Isolated world tests

[Maciej Stachowiak]

On Jul 1, 2009, at 7:21 PM, Adam Barth wrote:

> On Wed, Jul 1, 2009 at 7:08 PM, Oliver Hunt<oliver at apple.com> wrote:
>> Given this apparently needs to exist in WebCore you will need to  
>> implement
>> this feature for the JSC bindings as well.  This will also make it  
>> possible
>> for it to be tested in the normal layout tests.
>
> Huh?  I don't understand the technical reason for this requirement.
> Can you explain why we should spend double the engineering effort (and
> introduce the corresponding instability into JSC ports) to build
> something that no one plans to use?
>
> Regardless of whether we implement the feature for JSC, we'll still
> need to enhance the layoutTestController to exercise the feature
> because the feature isn't exposed to normal web pages.  Isn't a better
> solution just to add the tests to the skipped list for non-supporting
> ports?

We generally wouldn't accept WebKit features that only work with V8,  
even if other ports may not immediately plan to use them. Similarly,  
we'd be unlikely to take a patch for the DOM that only works if you  
replace the style system. I haven't thought through whether this  
particular feature should be an exception.

In general, I think using regression tests for features that are not  
directly exposed to Web content, but implemented in WebCore/WebKit, is  
reasonable. For example we have tests that check that WebKit's  
delegate methods relating to load progress are dispatched in the  
correct order.

Regards,
Maciej