Bug 210658 – REGRESSION (r234105): [iOS] WKColorButton leaks a UIColor

• Component: WebKit2
• Status: RESOLVED FIXED

[David Kilzer (:ddkilzer)]

WKColorButton leaks a UIColor.

Found by clang static analyzer:

'WKColorButton' lacks a 'dealloc' instance method but must release '_color'

--

#pragma mark - WKColorButton

@interface WKColorButton : UIButton
@property (nonatomic, strong) UIColor *color;

+ (instancetype)colorButtonWithColor:(UIColor *)color;
@end

@implementation WKColorButton

+ (instancetype)colorButtonWithColor:(UIColor *)color
{
    WKColorButton *colorButton = [WKColorButton buttonWithType:UIButtonTypeCustom];
    colorButton.color = color;
    colorButton.backgroundColor = color;
    return colorButton;
}

@end

[Radar WebKit Bug Importer]

<rdar://problem/61938137>

[David Kilzer (:ddkilzer)]

Created attachment 396770 [details]
Patch v1

[Darin Adler]

Comment on attachment 396770 [details]
Patch v1

View in context: https://bugs.webkit.org/attachment.cgi?id=396770&action=review

> Source/WebCore/page/Chrome.cpp:428
>  #if PLATFORM(IOS_FAMILY)
>      return nullptr;
> -#endif
> +#else
>      notifyPopupOpeningObservers();
>      return m_client.createColorChooser(client, initialColor);
> +#endif

We need UNUSED_PARAM here.

I also think the "always-null color chooser" design strategy is a little messy, but that goes beyond this function

[David Kilzer (:ddkilzer)]

Created attachment 396803 [details]
Patch for landing

[EWS]

Committed r260290: <https://trac.webkit.org/changeset/260290>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 396803 [details].