Re: [widgets] WARP testing

On May 11, 2010, at 9:00 PM, Scott Wilson <scott.bradley.wilson@gmail.com 
 > wrote:

> On 11 May 2010, at 15:58, Marcos Caceres wrote:
>
>> On Tue, May 11, 2010 at 3:55 PM, Scott Wilson
>> <scott.bradley.wilson@gmail.com> wrote:
>>> Hi Marcos,
>>>
>>> I'll make a start on tests for the assertions about correctly  
>>> processing the element (6-13).
>
> I've checked in tests for ta-6 through ta-9.
>
>>> I'm not sure about assertion 5, however  - how would you tell if  
>>> it had been processed at an inappropriate point? What difference  
>>> would it make? And how could you test it?
>>>
>>> "Assertion ta-5: Secondly, a user agent MUST apply the rule for  
>>> processing an access element at the appropriate point in the  
>>> algorithm to process a configuration document: the appropriate  
>>> point is where the algorithm allows for processing 'any other type  
>>> of element' [[!WIDGETS]]." [1]
>>
>> Tests would be having an access element at the top, middle, and end  
>> of
>> the document and also making sure that it does not get processed when
>> nested inside another element.
>>
>> so 1:
>> <widget..>
>> <name/>
>> <access .../ >
>> </widget>
>>
>> 2:
>> <widget..>
>> <access ... />
>> <name/>
>> </widget>
>>
>> 3:
>> <widget..>
>> <name/>
>> <access ... >
>> <description/>
>> </widget>
>
> So presumably the access element is ignored in cases 2 & 3, even if  
> the element itself is valid? I'm not sure if that's such a great  
> idea for interoperability.

No, it is valid in all cases except 4.

>
> Unless of course it is processed, just after processing the P&C  
> elements, which has no discernible effect on the processed widget  
> and doesn't really make it testable.
>
> Maybe I'm just missing something here!
>

Yeah, you are over thinking it:) - all elements are processed as  
normal. My examples were just supposed to test that. The P&C "hook"  
does not really apply here (I defined it as an i18n extension point).

>> 4: (which would cause it to be ignored)
>> <widget..>
>> <name><access ... ></name>
>> </widget>
>
> OK, I'll make a test for that case.

So yeah, 4 is the only fail if access is granted.

>
>>
>>
>>
>>
>>> [1] http://dev.w3.org/2006/waf/widgets-access/test-suite/
>>>
>>> S
>>>
>>> On 4 May 2010, at 15:23, Marcos Caceres wrote:
>>>
>>>> On Thu, Apr 1, 2010 at 8:55 PM, Scott Wilson
>>>> <scott.bradley.wilson@gmail.com> wrote:
>>>>> Wookie has implemented WARP, so we can try out the tests as soon  
>>>>> as they are
>>>>> available.
>>>>
>>>> I'm seriously not getting the cycles to do this. Scott, any  
>>>> chance you
>>>> could help us out?
>>>>
>>>> According to [1], there is only around 15 assertions... that  
>>>> roughly
>>>> equates to 45 tests.
>>>>
>>>> Kind regards,
>>>> Marcos
>>>>
>>>> [1] http://dev.w3.org/2006/waf/widgets-access/test-suite/
>>>>
>>>> --
>>>> Marcos Caceres
>>>> Opera Software ASA, http://www.opera.com/
>>>> http://datadriven.com.au
>>>
>>>
>>>
>>
>>
>>
>> -- 
>> Marcos Caceres
>> Opera Software ASA, http://www.opera.com/
>> http://datadriven.com.au
>

Received on Tuesday, 11 May 2010 23:03:38 UTC