Another step towards bug 156537. This class will handle all the code specific to (anonymous) mfenced operators. In particular many members of the RenderMathMLOperator class are going to be removed while fixing bug 156537 but they may still need to stored and exposed in some way by mfenced operators. This work applies after bug 160241 and the new class will first be used to expose the m_textContent and update it.
Created attachment 284694 [details] Patch (applies after bug 160241)
Created attachment 284850 [details] Patch
Created attachment 284855 [details] Patch
Comment on attachment 284855 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=284855&action=review > Source/WebCore/rendering/mathml/RenderMathMLFenced.cpp:87 > +RenderPtr<RenderMathMLFencedOperator> RenderMathMLFenced::createMathMLOperator(const String& operatorString, MathMLOperatorDictionary::Form form, MathMLOperatorDictionary::Flag flag) Probably not a change to be don in this patch, but I wonder whether we should remove the RenderPtr and use std:unique instead.
Comment on attachment 284855 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=284855&action=review > Source/WebCore/rendering/mathml/RenderMathMLFencedOperator.cpp:39 > + m_operatorForm = form; > + m_operatorFlags = flags; Can we pass these to a base class constructor instead of using assignment?
(In reply to comment #4) > Probably not a change to be don in this patch, but I wonder whether we > should remove the RenderPtr and use std:unique instead. Yes, I opened 160396. (In reply to comment #5) > Comment on attachment 284855 [details] > Can we pass these to a base class constructor instead of using assignment? OK. The plan is to move these members to the derived class later, but for now I'll keep initialization in the base class.
Committed r203973: <http://trac.webkit.org/changeset/203973>