Re: [mediacapture-main] Reinstate strong language on permission ending when tracks stop, lost by editorial mistake

Reading the above, I'm concluding that we (media capture) need to have
 a "request permission" algorithm (which we are currently referencing 
from the permissions spec) that checks stored permissions, and on 
having the return be "prompt", prompts for permission, and returens 
"granted" or "denied", but does NOT (necessarily) cause "check for 
permission" to change its result.

That is:
Request permission (5.2)
   -> check permission according to 5.1 (returns "prompt") (A)
   -> prompt and return what the user wanted this time (B)

(time passes)

Request permission
    -> check permission (C)
    
At point (C), unless new information has been gathered, "check 
permission" should return the result that is consistent with point 
(A). What's returned at point (B) is irrelevant (unless the UA has 
concluded that the user has given permission for future access).

That is, "this algorithm" from the part Jan-Ivar quoted refers to 
section 5.1 "Check permission", which is invoked by "request 
permission". It does not refer to "request permission".

This seems like it needs clarification in the permissions spec, since 
Jan-Ivar and I read it differently.
(Or I might be reading it wrong, in which case it needs changing.)

-- 
GitHub Notification of comment by alvestrand
Please view or discuss this issue at 
https://github.com/w3c/mediacapture-main/issues/387#issuecomment-245507814
 using your GitHub account

Received on Thursday, 8 September 2016 06:41:37 UTC