Bug 202097 - AudioTrackPrivateMediaStreamCocoa does not need to manipulate the audio unit in play/pause methods
Summary: AudioTrackPrivateMediaStreamCocoa does not need to manipulate the audio unit ...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebRTC (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: youenn fablet
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2019-09-22 23:38 PDT by youenn fablet
Modified: 2019-09-24 01:42 PDT (History)
11 users (show)

See Also:

Attachments
Patch (4.80 KB, patch)
2019-09-22 23:41 PDT, youenn fablet 		no flags Details | Formatted Diff | Diff
Patch (5.50 KB, patch)
2019-09-23 07:23 PDT, youenn fablet 		no flags Details | Formatted Diff | Diff
Patch (5.79 KB, patch)
2019-09-23 23:28 PDT, youenn fablet 		no flags Details | Formatted Diff | Diff
Show Obsolete (2) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description youenn fablet 2019-09-22 23:38:01 PDT 
AudioTrackPrivateMediaStreamCocoa does not need to manipulate the audio unit in play/pause methods
Comment 1 youenn fablet 2019-09-22 23:41:01 PDT 
Created attachment 379357 [details]
Patch
Comment 2 youenn fablet 2019-09-23 02:52:08 PDT 
<rdar://problem/51548144>
Comment 3 Eric Carlson 2019-09-23 06:06:27 PDT 
Comment on attachment 379357 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=379357&action=review

> Source/WebCore/platform/mediastream/mac/AudioTrackPrivateMediaStreamCocoa.h:89
> +    bool m_isTransitioningToPausing { false };
> +    bool m_isTransitioningToPlaying { false };

Couldn't we replace these with an enum? In a case where both are set simultaneously, wouldn't we always want the last one set to always "win"?
Comment 4 youenn fablet 2019-09-23 07:23:54 PDT 
Created attachment 379369 [details]
Patch
Comment 5 youenn fablet 2019-09-23 07:24:24 PDT 
> Couldn't we replace these with an enum? In a case where both are set
> simultaneously, wouldn't we always want the last one set to always "win"?

I tried to come up with a new approach without any transition values.
Comment 6 Eric Carlson 2019-09-23 07:44:43 PDT 
Comment on attachment 379369 [details]
Patch

Nice approach! 

r=me once the bots are happy.
Comment 7 youenn fablet 2019-09-23 23:28:08 PDT 
Created attachment 379433 [details]
Patch
Comment 8 youenn fablet 2019-09-24 00:55:58 PDT 
(In reply to youenn fablet from comment #7)
> Created attachment 379433 [details]
> Patch

Added some logging in case of failing to start the audio unit.
Comment 9 WebKit Commit Bot 2019-09-24 01:42:10 PDT 
Comment on attachment 379433 [details]
Patch

Clearing flags on attachment: 379433

Committed r250290: <https://trac.webkit.org/changeset/250290>
Comment 10 WebKit Commit Bot 2019-09-24 01:42:11 PDT 
All reviewed patches have been landed.  Closing bug.