Re: Orientation event draft

On Thu, Apr 22, 2010 at 11:22 AM, Olli Pettay <Olli.Pettay@helsinki.fi> wrote:
> On 4/22/10 1:16 PM, Olli Pettay wrote:
>>
>> On 4/22/10 12:53 PM, Andrei Popescu wrote:
>>>
>>> On Thu, Apr 22, 2010 at 10:41 AM, Olli Pettay<Olli.Pettay@helsinki.fi>
>>> wrote:
>>>>
>>>> On 4/21/10 7:32 PM, Steve Block wrote:
>>>>>
>>>>>> "Implementations that are unable to provide all three angles must
>>>>>> set the values of the unknown angles to null" That is strange. The
>>>>>> type of the property is double. That kind of properties don't
>>>>>> usually give suddenly non-numeric values.
>>>>>
>>>>> There is precedent for this in the Geolocation spec.
>>>>
>>>>
>>>> Sounds like a bug in Geolocation draft, IMO
>>>> If the value isn't available, implementation could throw something
>>>> like NOT_AVAILABLE.
>>>>
>>>
>>> But then wouldn't you have to access all these properties from try /
>>> catch blocks? I don't think that's ideal and it's just more convenient
>>> to let them be null if they aren't available.
>>
>> And then all the callers need to check whether the returned value is
>> null. And if (thevalue) isn't enough, because the value can be 0.
>>
>> I don't see how it is really more convenient to make it null in some
>> cases. And making a double attribute null isn't something used in DOM
>> Core for example.
>> (Not all the things WebIDL specifies should be used, like PutForwards)
>>
>> -Olli
>>
>
> But I could still add that I don't care too much about this issue.
> It is just very ugly, and API wise rather inconsistent if some
> double attributes can suddenly become null.
>

Yeah, question of taste, I guess. I don't see it as ugly and I am not
convinced that changing the API to throw buys us anything.

Andrei

Received on Thursday, 22 April 2010 10:27:36 UTC