Added support of calc() for SVGLength

third_party/WebKit/Source/core/svg/SVGLengthTearOff.cpp

 /*
  * Copyright (C) 2014 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 86 matching lines @@
         return LengthTypePC;
     default:
         return LengthTypeUnknown;
     }
 }
 
 } // namespace
 
 bool SVGLengthTearOff::hasExposedLengthUnit()
 {
+    if (target()->isCalculated())
+        return false;
+
     CSSPrimitiveValue::UnitType unit = target()->typeWithCalcResolved();
     return isValidLengthUnit(unit)
         || unit == CSSPrimitiveValue::UnitType::Unknown
         || unit == CSSPrimitiveValue::UnitType::UserUnits;
 }
 
 SVGLengthType SVGLengthTearOff::unitType()
 {
     return hasExposedLengthUnit() ? toSVGLengthType(target()->typeWithCalcResolved()) : LengthTypeUnknown;
 }
@@ skipping 20 matching lines @@
         exceptionState.throwDOMException(NoModificationAllowedError, "The attribute is read-only.");
         return;
     }
 
     if (target()->isRelative() && !canResolveRelativeUnits(contextElement())) {
         exceptionState.throwDOMException(NotSupportedError, "Could not resolve relative length.");
         return;
     }
 
     SVGLengthContext lengthContext(contextElement());
-    target()->setValue(value, lengthContext);
+    if (target()->isCalculated())
+        target()->setValueAsNumber(value);

[fs, 2016/07/16 21:13:02]

Per the other CL (and current spec text) this would be the only case necessary?

[Shanmuga Pandi, 2016/07/18 13:35:33]

As per the spec, It should always "Set the SVGLength's value to a <number> whose
value is value.".
Since my previous patch is reverted, I added just isCalculated() check to handle
calc cases. 
But eventually, it should be as per spec. I will investigate why that patch is
reverted, if I get the testCase.

[fs, 2016/07/18 13:56:48]

Fair enough.

+    else
+        target()->setValue(value, lengthContext);
+
     commitChange();
 }
 
 float SVGLengthTearOff::valueInSpecifiedUnits()
 {
+    if (target()->isCalculated())
+        return 0;
+
     return target()->valueInSpecifiedUnits();
 }
 
 void SVGLengthTearOff::setValueInSpecifiedUnits(float value, ExceptionState& exceptionState)
 {
     if (isImmutable()) {
         exceptionState.throwDOMException(NoModificationAllowedError, "The attribute is read-only.");
         return;
     }
-    target()->setValueInSpecifiedUnits(value);
+
+    if (target()->isCalculated())
+        target()->setValueAsNumber(value);
+    else
+        target()->setValueInSpecifiedUnits(value);
+
     commitChange();
 }
 
 String SVGLengthTearOff::valueAsString()
 {
     // TODO(shanmuga.m@samsung.com): Not all <length> properties have 0 (with no unit) as the default (lacuna) value, Need to return default value instead of 0
     return hasExposedLengthUnit() ? target()->valueAsString() : String::number(0);
 }
 
 void SVGLengthTearOff::setValueAsString(const String& str, ExceptionState& exceptionState)
@@ skipping 62 matching lines @@
     : SVGPropertyTearOff<SVGLength>(target, contextElement, propertyIsAnimVal, attributeName)
 {
 }
 
 DEFINE_TRACE_WRAPPERS(SVGLengthTearOff)
 {
     visitor->traceWrappers(contextElement());
 }
 
 } // namespace blink