The user pressing a button on a gamepad should cause gamepads to become visible to all NavigatorGamepads

[Brady Eidson]

The user pressing a button on a gamepad should cause gamepads to become visible to all NavigatorGamepads

[Brady Eidson]

Created attachment 233967 [details]
Patch v1

[Brady Eidson]

Ignore the can't-buildness that EWS sees - In my tree this patch is built on previous patches.  I will pass it through EWS before landing.

[Darin Adler]

Comment on attachment 233967 [details]
Patch v1

View in context: https://bugs.webkit.org/attachment.cgi?id=233967&action=review

> Source/WebCore/Modules/gamepad/GamepadManager.cpp:81
> +        const Vector<PlatformGamepad*>& platformGamepads = GamepadProvider::shared().platformGamepads();

Seems like a good place to use auto, although maybe Alexey wouldn’t agree.

> Source/WebCore/Modules/gamepad/GamepadManager.cpp:82
> +        for (unsigned i = 0; i < platformGamepads.size(); ++i) {

Normally it’s an anti-pattern to actually call size each time through. My own preference is to put it into a local variable named size, either inside the for or just before it:

    for (unsigned i = 0, size = platformGamepads.size(); i < size; ++i)

> Source/WebCore/platform/mac/HIDGamepadProvider.cpp:85
> +    , m_inputNotificationTimer(this, &HIDGamepadProvider::inputNotificationTimerFired)

Consider the new style Timer that uses a function/lambda instead of a pointer to member function.

> Source/WebCore/platform/mac/HIDGamepadProvider.cpp:215
> +    if (!m_inputNotificationTimer.isActive())
> +        m_inputNotificationTimer.startOneShot(InputNotificationDelay);

A why comment would be good here to explain why the isActive check. We want to delay from the time of the first input, not postpone when we get additional input, I assume.

[Brady Eidson]

http://trac.webkit.org/changeset/170553