Text incorporating the Permissions API's definition of permission handling #319
Conversation
| <code><a>MediaStream</a></code> object representing a media | ||
| stream.</p> | ||
| <p>Choose a device of each requested media type from | ||
| <var>finalSet</var>. |
There was a problem hiding this comment.
Alternate: Present the whole set to the user (somehow) and let them pick. Implementations should be allowed to do this.
|
@martinthomson we'd really like you to take a look at this to make sure it is compliant with FF UI. |
|
fwiw, the Travis error depends on getting fixed w3c/respec#597 I think |
| permission for the source in question, the given permission is revoked | ||
| is <dfn id="source-stopped">stopped</dfn>. If | ||
| the <a>permission</a> for the source is marked "transient", the | ||
| User Agent will <a href="#remove-permission">remove permission</a>, |
There was a problem hiding this comment.
The 'remove permission' link is broken. We could simply use <a>remove permission</a> to fix the link or <a href="#dfn-remove-permission">remove the permission</a> to make it more natural to read.
There was a problem hiding this comment.
removed this part. Not the best design.
|
This might be a follow up change.. In the description of the deviceId attribute [1] we talk about 'stored permission' and link to 'Best Practice 2: Stored Permissions'. Should that be expressed via the permissions terminology you introduce here? [1] https://w3c.github.io/mediacapture-main/#widl-MediaDeviceInfo-deviceId |
| permission</dfn> and <dfn>create a PermissionStatus</dfn> are | ||
| defined in [[!permissions]]. | ||
| <p class="note">"request permission" isn't defined yet. | ||
| https://github.com/w3c/permissions/issues/62 |
There was a problem hiding this comment.
Putting it as a link rather than a plain-text URL would be more reader-friendly
|
except for some minor nits and the corrections that @adam-be identified, LGTM |
alvestrand
force-pushed
the
hta-permissions
branch
from
2b42d87
to
2837e23
Compare
|
See if this works for people. In particular pinging @martinthomson to see if text is consistent with Firefox' behavior. |
| @@ -2731,6 +2757,17 @@ | |||
| and can thus be used as a fingerprinting surface.</p> | |||
| </li> | |||
| <li> | |||
| <p><a>Retrieve the permission state</a> for all | |||
| candidate devices that are not open in the current | |||
| browsing context. Remove from the set any device for | |||
There was a problem hiding this comment.
"open" seems to not be defined in mediacapture-main. Could we use something along the lines of "... for all candidate live sources not currently attached to at least one live MediaStreamTrack"?
Earlier in the document "live source" is used as a term for microphones and cameras OTOH in the introduction we use "local multimedia devices". Or we could reference MediaDevices, or define device.
Perhaps defining device is the best thing to do.
There was a problem hiding this comment.
The "life cycle and media flow" section defines the term "live" for MediaStreamTracks as "not ended". I'll use that.
|
I'm leaving the "best practice: stored permissions" for a later change. This might involve cross-spec surgery. |
alvestrand
force-pushed
the
hta-permissions
branch
from
7c17934
to
d03317d
Compare
In particular, make "temporary permissions" a thing that exists within the getusermedia spec, not in the permissions spec.
alvestrand
force-pushed
the
hta-permissions
branch
from
d03317d
to
bbdee24
Compare
|
PTAL. |
|
LGTM (I made a comment, but am fine merging regardless of if we do anything about it or not) |
|
LGTM |
|
I think we should stick to "live" MediaStreamTrack, but other than that, I think this is mergable. |
|
Found 2 more "active" in an unrelated section. Changed them to "live". |
|
PTAL. Adam, if happy, merge. |
|
LGTM - we should merge this one now IMO. |
This is text for review. ReSpec complains about one error still, but I can't see it.
Fixes #305