Eric Rescorla wrote on 11/30/12 7:38 PM:

>> > {picture: true, audio: true}
>> > {picture: true, video: true}
>> > {video: true, audio: true, picture: true}
>>
>> Also why does `mozGetUserMedia(1, function() {}, function() {}' throw
>> instead of calling the onErrorCallback as given by the spec? I haven't
>> tested yet how Chrome handles that but if we decide to take another
>> route as given by the spec we will indeed cause trouble by developers
>> who want to use this feature.
> 
> Actually, the WG isn't completely clear on this point. There is some
> feeling that simple programming errors (like the one you illustrate
> above) should just throw exceptions.

Ok, that makes sense but it doesn't answer the first half of my
question. So if your answer also belongs to the example objects above,
where can I find updated specs or actual discussions? Is there an email
list I should subscribe to read that stuff? When do we expect that the
spec gets updated? What is the turnaround for it?

IMO it makes the code kinda hard to maintain as long as `{picture: true,
video: true}` throws an exception. You will have a mixture of
understandable asynchronous code, but also a mess with try/catch
statements around.

>> Also what is the reason why the onErrorCallback is mandatory in our
>> implementation while the spec says it's optional?
> 
> We decided to make it mandatory at the most recent TPAC. The
> spec is out of date.

Just out of interest, have there been taken any notes which are
available for the public? It looks like I miss a lot of information
before I can really qualify the webrtc code by writing mochitests for
gUM and pc.

>> I would love if we can document anything we implement differently. If we
>> don't do this it will be hard for everyone to get started.
> 
> Why don't you record these issues somewhere and then when things
> stabilize, we can figure out which ones are really variants and need
> real documentation.

I see problems once we really open it up for developers in Firefox 20
builds. Fundamental things like the above mentioned issue should be
clear. Otherwise a lot of code would have to be updated/changed.

I can start such a list by asking all those questions step by step, but
I think that you and Randell have the most knowledge about differences
in our code to the spec which is currently available. That also includes
the above mentioned changes which haven't made it into the specs yet.

-- 
Henrik
_______________________________________________
dev-media mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-media

Reply via email to