Fix update of validity cache value, so that it reflects the correct state of any FormControlElement.

Source/core/html/HTMLTextAreaElement.h

Index: Source/core/html/HTMLTextAreaElement.h
diff --git a/Source/core/html/HTMLTextAreaElement.h b/Source/core/html/HTMLTextAreaElement.h
index b5f9ffb4b5f7320263d27d014696545f513cf86c..17e646a2bba4d34b9fdef5a5de525d20dae2466c 100644
--- a/Source/core/html/HTMLTextAreaElement.h
+++ b/Source/core/html/HTMLTextAreaElement.h
@@ -74,7 +74,6 @@ private:
     void handleBeforeTextInsertedEvent(BeforeTextInsertedEvent*) const;
     static String sanitizeUserInputValue(const String&, unsigned maxLength);
     void updateValue() const;
-    virtual void setInnerEditorValue(const String&) OVERRIDE;
     void setNonDirtyValue(const String&);
     void setValueCommon(const String&, TextFieldEventBehavior, SelectionOption = NotChangeSelection);
 
@@ -128,7 +127,6 @@ private:
     WrapMethod m_wrap;
     mutable String m_value;
     mutable bool m_isDirty;
-    bool m_valueIsUpToDate;

[spartha, 2014/09/02 14:09:55]

The flag is set in subTreeChanged() and the value was updated in the same call
stack (in the call to valid()). So there was no lazy update happening as such.
This flag has been replaced by a direct call to updateValue(). This change was
needed as we do a lazy update of validity.   

[tkent, 2014/09/03 02:16:35]

Don't change this.  The change will cause performance regression.

[spartha, 2014/09/03 06:14:47]

I have not come across m_valueIsUpToDate being set to false anywhere else, hence
I don't see it doing a lazy update. The call to setNeedsValidtyCheck() in
subTreeChanged() would update the value anyway. Have I missed any other
alternate flow that would benefit from this setup? 

[tkent, 2014/09/03 07:56:43]

- If willValidate() is false, updateValue() shouldn't be called immediately.
- By the lazy evaluation of validity, the number of updateValue() calls would be
reduced. We'll have performance benefit.

     String m_suggestedValue;
 };