Use 'originIdentifier' in getUserMedia() steps (issue #339) #349
Conversation
| to have permission status "granted", resulting in a set of | ||
| provided media.</p> | ||
| <p>For the origin identified by <var>originIdentifier</var>, | ||
| <a>request permission</a> for use of the devices, while |
There was a problem hiding this comment.
I don't see how we're expected to verify anything about a "request permission" algorithm that doesn't exist yet.
There was a problem hiding this comment.
There was a problem hiding this comment.
Thanks, it's not in the TR which is what we link to, and we still have a disclaimer in our doc saying it doesn't exist. The link then seems like it should be _request a permission, not _request permission.
Also, maybe everyone else is aware, but this surprised me, that the other spec seems to follow a slightly different process for consensus than at least I was used to. Specifically, the editor of the other spec says he doesn't "claim that the current text represents any sort of consensus."
I'm surprised we're adding such dependencies when we're supposed to be way past last call.
There was a problem hiding this comment.
Is there a way in this spec now for getUserMedia to succeed without granting persistent permission? What does _request a permission_ return in that case?
See #350 for my take down of the request permission algorithm.
|
@jan-ivar, my interpretation is that your concerns are mainly with text that's not added by this PR (that you've also created a new issue for). And that shouldn't block us from merging this PR. Is that correct? Edit: Note that first commit is fixing indentation (mix of spaces and tabs), and the actual change is in the second commit. |
|
@adam-be Sounds like the algorithm referenced should be request _a_ permission, though I see no evidence that it takes an The level of meaninglessness seems neither increased nor diminished by this PR, so I have no problem with it merging. |
Fix for issue #339
First commit is purely editorial so please review the second.