Bug 233722 – DisplayCaptureSource doesn't initialize base class correctly

• Status: RESOLVED
• Resolution: FIXED
• Component: WebRTC

[Eric Carlson]

UserMediaCaptureManagerProxy doesn't pass the site hash salt the display capture factory, so RealtimeMediaSource::m_hashedID is empty.

[Eric Carlson]

Created attachment 445614 [details]
Patch

[youenn fablet]

Comment on attachment 445614 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=445614&action=review

> Source/WebCore/ChangeLog:8
> +        No change in functionality.

Is the goal to expose a deviceId in settings? If so, we should call setDeviceId in DisplayCaptureSourceCocoa::settings() and add a test.
If not, why do we care about having an empty hashed ID?

Looking at other browsers, we are doing like Firefox (no deviceId) but not like Chrome.
I guess the spec should be clear about that.

Also RealtimeMediaSource::hashedId is checking that m_hashedId is not empty if device is not screen/window.
I guess we could update this check (though other sources might still have empty hashedId, canvas or WebAudio for instance).

> Source/WebCore/platform/mediastream/RealtimeMediaSourceCenter.h:126
> +    void getDisplayMediaDevices(const MediaStreamRequest&, String&&, Vector<DeviceInfo>&, String&);

Would be worth changing to Expected<Vector<DeviceInfo>, String> getDisplayMediaDevices(..) at some point (probably not this patch though).

[Eric Carlson]

Created attachment 445777 [details]
Patch

[Eric Carlson]

(In reply to youenn fablet from comment #2)
> Comment on attachment 445614 [details]
> 
> Is the goal to expose a deviceId in settings? If so, we should call
> setDeviceId in DisplayCaptureSourceCocoa::settings() and add a test.
> If not, why do we care about having an empty hashed ID?
> 

Good point, I added a test.

> Looking at other browsers, we are doing like Firefox (no deviceId) but not
> like Chrome.
> I guess the spec should be clear about that.
> 
> Also RealtimeMediaSource::hashedId is checking that m_hashedId is not empty
> if device is not screen/window.
> I guess we could update this check (though other sources might still have
> empty hashedId, canvas or WebAudio for instance).
> 

I updated the check.

[EWS]

Committed r286491 (244830@main): <https://commits.webkit.org/244830@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 445777 [details].

[Radar WebKit Bug Importer]

<rdar://problem/86019589>