Bug 165139 – [MediaStream] Sync video preview layer and parent layer sizes

• Component: Media
• Status: RESOLVED FIXED

[Eric Carlson]

Resize video preview layer when its parent layer size changes.

[Eric Carlson]

Created attachment 295593 [details]
Proposed patch.

[WebKit Commit Bot]

Attachment 295593 [details] did not pass style-queue:


ERROR: Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:656:  The parameter name "object" adds no information, so it should be removed.  [readability/parameter_name] [5]
ERROR: Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:660:  The parameter name "object" adds no information, so it should be removed.  [readability/parameter_name] [5]
Total errors found: 2 in 2 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Daniel Bates]

Comment on attachment 295593 [details]
Proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=295593&action=review

> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:100
> +    AVVideoSourcePreview* _parent;

Nit: * is on the wrong side.

> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:101
> +    BOOL _haveObserver;

This is OK as-is. Would using the present tense, _hasObserver, read better?

> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:117
> +    PlatformLayer* platformLayer() const final { return m_previewBackgroundLayer.get(); }

It is an unwritten convention to that const functions should return pointers to const objects. When returning a non-const pointer the function should not be const. The motivation behind this idiom is that the object pointed by the non-const pointer can be mutated, which could have implications to the behavior of the class. Humans not well versed in the C++ concept of const typically consider const functions to convey that an instance of a class will not be modified by the call. Returning a non-const pointer circumvents this notion.

> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:640
> +        [_parent->platformLayer() addObserver:self forKeyPath:@"bounds" options:0 context:0];

We should pass nullptr instead of 0 for context.

> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:644
> +- (void)observeValueForKeyPath:keyPath ofObject:(id)object change:(NSDictionary *)change context:(void*)context

Nit: Missing a space after "void".

> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:656
> +        if ([keyPath isEqual:@"bounds"] && object == _parent->platformLayer())

We repeatedly hardcode the string literal "bounds" throughout this patch. I suggest that we extract it into a constant and reference this constant throughout the patch.

[Eric Carlson]

Comment on attachment 295593 [details]
Proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=295593&action=review

>> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:100
>> +    AVVideoSourcePreview* _parent;
> 
> Nit: * is on the wrong side.

OK

>> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:101
>> +    BOOL _haveObserver;
> 
> This is OK as-is. Would using the present tense, _hasObserver, read better?

Changed.

>> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:117
>> +    PlatformLayer* platformLayer() const final { return m_previewBackgroundLayer.get(); }
> 
> It is an unwritten convention to that const functions should return pointers to const objects. When returning a non-const pointer the function should not be const. The motivation behind this idiom is that the object pointed by the non-const pointer can be mutated, which could have implications to the behavior of the class. Humans not well versed in the C++ concept of const typically consider const functions to convey that an instance of a class will not be modified by the call. Returning a non-const pointer circumvents this notion.

Sounds like this should be codified in our style guidelines. 

This isn't new to this patch, and changing it will requires a change a number of other classes so I will address this in another patch.

>> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:640
>> +        [_parent->platformLayer() addObserver:self forKeyPath:@"bounds" options:0 context:0];
> 
> We should pass nullptr instead of 0 for context.

OK

>> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:644
>> +- (void)observeValueForKeyPath:keyPath ofObject:(id)object change:(NSDictionary *)change context:(void*)context
> 
> Nit: Missing a space after "void".

Fixed.

>> Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:656
>> +        if ([keyPath isEqual:@"bounds"] && object == _parent->platformLayer())
> 
> We repeatedly hardcode the string literal "bounds" throughout this patch. I suggest that we extract it into a constant and reference this constant throughout the patch.

Fixed.

[Eric Carlson]

Created attachment 295698 [details]
Updated patch.

[WebKit Commit Bot]

Attachment 295698 [details] did not pass style-queue:


ERROR: Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:647:  The parameter name "object" adds no information, so it should be removed.  [readability/parameter_name] [5]
ERROR: Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:651:  The parameter name "object" adds no information, so it should be removed.  [readability/parameter_name] [5]
Total errors found: 2 in 2 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Eric Carlson]

Created attachment 295714 [details]
YAUP.

[WebKit Commit Bot]

Attachment 295714 [details] did not pass style-queue:


ERROR: Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:647:  The parameter name "object" adds no information, so it should be removed.  [readability/parameter_name] [5]
ERROR: Source/WebCore/platform/mediastream/mac/AVVideoCaptureSource.mm:651:  The parameter name "object" adds no information, so it should be removed.  [readability/parameter_name] [5]
Total errors found: 2 in 2 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[WebKit Commit Bot]

Comment on attachment 295714 [details]
YAUP.

Clearing flags on attachment: 295714

Committed r209141: <http://trac.webkit.org/changeset/209141>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.