…ur when the defaultResponder was the pane itself.
…er must be overridden to return null, not the defaultResponder.
…e pane if a default responder doesn't handle the event.
… pane's responder chain handles an event, the pane would be asked to perform the event twice. Also removed unnecessary check for evt && evt.mouseHandler at the end of checking for a target. `mouseHandler` isn't defined anywhere in the framework and if it were, since it has the ability to override target, why wouldn't it be checked immediately?
…in SC.Pane. Improves the initialization speed of SC.Pane.
…lass's prototype is completely unnecessary and simply slows down the initialization process of SC.Pane. No code changes, just moved the split up code into pane.js.
… for some refactoring of methods and removal of unnecessary reopen calls.
…tate of the child view before calling detach on it. This removes excess console warnings about detaching a view in an invalid state.
… bleedthrough of adjacent sprites on retina displays (noticeable as pink in development mode).
…ule and pointer images at every control size. Previously, the capsule theme was only supported at SC.REGULAR_CONTROL_SIZE and the pointer theme was only supported at SC.REGULAR_CONTROL_SIZE or SC.HUGE_CONTROL_SIZE.
…ay change signature slightly
…ains a cache for String and Number GUIDs. Because the GUID for a String or a Number is essentially its own value, the caching process actually takes longer than it does to simply generate a GUID key from the given String or Number (see http://jsperf.com/cache-vs-manipulate). More importantly, this removes the memory overhead of maintaining the GUID cache, which was also unable to be cleaned. Also removed the completely unused _guidPrefixes, _numberGuids and _stringGuids objects.