Simply SVGSVGElement::getElementById method

Source/core/svg/SVGSVGElement.cpp

 /*
  * Copyright (C) 2004, 2005, 2006 Nikolas Zimmermann <zimmermann@kde.org>
  * Copyright (C) 2004, 2005, 2006, 2007, 2008, 2010 Rob Buis <buis@kde.org>
  * Copyright (C) 2007 Apple Inc. All rights reserved.
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Library General Public
  * License as published by the Free Software Foundation; either
  * version 2 of the License, or (at your option) any later version.
  *
@@ skipping 749 matching lines @@
     if (viewElement->hasAttribute(SVGNames::zoomAndPanAttr))
         view->setZoomAndPan(viewElement->zoomAndPan());
     else
         view->setZoomAndPan(zoomAndPan());
 }
 
 // getElementById on SVGSVGElement is restricted to only the child subtree defined by the <svg> element.
 // See http://www.w3.org/TR/SVG11/struct.html#InterfaceSVGSVGElement
 Element* SVGSVGElement::getElementById(const AtomicString& id) const
 {
-    Element* element = treeScope().getElementById(id);
-    if (element && element->isDescendantOf(this))
-        return element;
+    Vector<Element*> elements = treeScope().getAllElementsById(id);

[Inactive, 2014/02/24 20:12:42]

This may cause a performance regression in the non-duplicate id case, which is
the most common case. Your CL causes some unnecessary overhead in the single-id
case (Vector creation, some checks in getAllElementsById).

It would be safer to use TreeScope::containsMultipleElementsWithId() and use
getElementById() if it returns true (fast path). Then fallback to using
getAllElementsById() instead of traversing.

 
-    // Fall back to traversing our subtree. Duplicate ids are allowed, the first found will
-    // be returned.
-    for (Node* node = firstChild(); node; node = NodeTraversal::next(*node, this)) {
-        if (!node->isElementNode())
-            continue;
-
-        Element* element = toElement(node);
-        if (element->getIdAttribute() == id)
+    for (Vector<Element*>::iterator it = elements.begin(); it != elements.end(); ++it) {

[pdr., 2014/02/24 18:39:43]

I'm not sure how much it matters here, but I typically see the end() moved into
a local variable like so:

Vector<Element*>::iterator end = elements.end();
for (Vector<Element*>::iterator it = elements.begin(); it != end; ++it) {

I don't feel strongly about this though.

+        Element* element = *it;
+        if (element->isDescendantOf(this))
             return element;
     }
+
     return 0;
 }
 
 }