Re: [Bug 22593] Capabilities are buggy

On 4/3/14 5:21 AM, Harald Alvestrand wrote:
> (continuing actual discussion on list, since it clutters up bugzilla)
>
> On 04/03/2014 11:14 AM, bugzilla@jessica.w3.org wrote:
>> https://www.w3.org/Bugs/Public/show_bug.cgi?id=22593
>>
>> --- Comment #4 from Dominique Hazael-Massieux <dom@w3.org> ---
>> Makes sense; so that leaves the question of the meaningfulness of having an
>> array of values for facingMode in capabilities.
>>
>> In particular, the example in 11.3 has:
>>     facingMode": ["user", "environment"]
>> which doesn't make sense to me; it may be that there is some (obsolete?)
>> ambiguity about whether capabilities are at the UA level (which the text in
>> 11.3 seems to imply) or at the source level.
> [...] Each camera can only have a single value for facingMode, but it makes
> perfect sense to say that multiple values for facingMode are acceptable.

If cameras only had single values then wouldn't getSettings() suffice? :-)

A camera that supports more than one facingMode is perhaps odd, but 
maybe it swivels? Why limit future cameras by limiting expression of 
limitations?

Presumably, getCapabilities() in the context of a MediaStreamTrack is 
there to support applyConstraints() so that one can use it with some 
probability of success.

If the UA factors out multi-functional devices into separate logical 
devices, then at some point there wont be much left to constrain.

.: Jan-Ivar :.

Received on Thursday, 3 April 2014 16:23:08 UTC