Changeset 199497 in webkit

Timestamp:

Apr 13, 2016 9:16:13 AM (8 years ago)

Author:

fred.wang@free.fr

Message:

Fix two coding mistakes in MathMLInlineContainerElement::childrenChanged
​https://bugs.webkit.org/show_bug.cgi?id=156538

Patch by Frederic Wang <​fwang@igalia.com> on 2016-04-13
Reviewed by Darin Adler.

We fix the call to updateOperatorProperties inside MathMLInlineContainerElement::childrenChanged
for the <math> and <msqrt> tags.

The <math> tag is already a RenderMathMLRow so the hasTagName(mathTag)
conditional is never executed. The tag does not create any anonymous
wrapper so we do not need a special case for it anyway.

The <msqrt> tag is not a RenderMathMLRow (yet). However, the anonymous
wrapper behaving as a RenderMathMLRow is actually the last child, not
the first one.

No new tests, this is already covered by mathml/presentation/mo-form-dynamic.html
Note that for some reason the coding error for <msqrt> only shows up
after the refactoring of bug 152244.

• mathml/MathMLInlineContainerElement.cpp:

(WebCore::MathMLInlineContainerElement::childrenChanged): Fix the two mistakes and add some FIXME comments.

Location:

trunk/Source/WebCore

Files:

• ChangeLog (modified) (1 diff)
• mathml/MathMLInlineContainerElement.cpp (modified) (1 diff)

trunk/Source/WebCore/ChangeLog

@@ +1 @@
+2016-04-13  Frederic Wang  <fwang@igalia.com>
+
+        Fix two coding mistakes in MathMLInlineContainerElement::childrenChanged
+        https://bugs.webkit.org/show_bug.cgi?id=156538
+
+        Reviewed by Darin Adler.
+
+        We fix the call to updateOperatorProperties inside MathMLInlineContainerElement::childrenChanged
+        for the <math> and <msqrt> tags.
+
+        The <math> tag is already a RenderMathMLRow so the hasTagName(mathTag)
+        conditional is never executed. The tag does not create any anonymous
+        wrapper so we do not need a special case for it anyway.
+
+        The <msqrt> tag is not a RenderMathMLRow (yet). However, the anonymous
+        wrapper behaving as a RenderMathMLRow is actually the last child, not
+        the first one.
+
+        No new tests, this is already covered by mathml/presentation/mo-form-dynamic.html
+        Note that for some reason the coding error for <msqrt> only shows up
+        after the refactoring of bug 152244.
+
+        * mathml/MathMLInlineContainerElement.cpp:
+        (WebCore::MathMLInlineContainerElement::childrenChanged): Fix the two mistakes and add some FIXME comments.
+
 2016-04-12  Chris Dumez  <cdumez@apple.com>
 

trunk/Source/WebCore/mathml/MathMLInlineContainerElement.cpp

@@ -59 +59 @@
 {
     if (renderer()) {
+        // FIXME: Parsing of operator properties should be done in the element classes rather than in the renderer classes.
+        // See http://webkit.org/b/156537
         if (is<RenderMathMLRow>(*renderer()))
             downcast<RenderMathMLRow>(*renderer()).updateOperatorProperties();
-        else if (hasTagName(mathTag) || hasTagName(msqrtTag)) {
-            auto* childRenderer = renderer()->firstChild();
+        else if (hasTagName(msqrtTag)) {
+            // Update operator properties for the base wrapper.
+            // FIXME: This won't be necessary when RenderMathMLSquareRoot derives from RenderMathMLRow and does not use anonymous wrappers.
+            // See http://webkit.org/b/153987
+            auto* childRenderer = renderer()->lastChild();
             if (is<RenderMathMLRow>(childRenderer))
                 downcast<RenderMathMLRow>(*childRenderer).updateOperatorProperties();