Make custom element name checks faster and fewer

third_party/WebKit/Source/core/dom/custom/CustomElement.h

 // Copyright 2016 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #ifndef CustomElement_h
 #define CustomElement_h
 
 #include "core/CoreExport.h"
 #include "core/HTMLNames.h"
 #include "core/dom/Element.h"
+#include "platform/text/Character.h"
 #include "wtf/Allocator.h"
 #include "wtf/text/AtomicString.h"
 
 namespace blink {
 
 class Document;
 class Element;
 class HTMLElement;
 class QualifiedName;
 class CustomElementDefinition;
 class CustomElementReaction;
 class CustomElementRegistry;
 
 class CORE_EXPORT CustomElement {
     STATIC_ONLY(CustomElement);
 public:
     // Retrieves the CustomElementRegistry for Element, if any. This
     // may be a different object for a given element over its lifetime
     // as it moves between documents.
     static CustomElementRegistry* registry(const Element&);
     static CustomElementRegistry* registry(const Document&);
 
     static CustomElementDefinition* definitionForElement(const Element*);
 
-    static bool isValidName(const AtomicString& name);
+    static bool isValidName(const AtomicString& name)
+    {
+        // This quickly rejects all common built-in element names.
+        if (name.find('-', 1) == kNotFound)
+            return false;
+
+        // Note: All valid names start with an ASCII character.

[esprehn, 2016/09/02 00:53:20]

for v1 we could just lookup the element in the registry hash table too since we
don't need to do upgrades for things outside the document?

return registry.contains(name);

why do we care if the element is a custom element but isn't registered?

+        if (!Character::isPotentialCustomElementNameStartChar(name[0]))
+            return false;
+
+        if (name.is8Bit()) {
+            for (size_t i = 1; i < name.length(); ++i) {
+                if (!Character::isPotentialCustomElementName8BitChar(name[i]))

[esprehn, 2016/09/02 00:53:20]

this is checking is8Bit and isNull() repeatedly inside the loop in the
operator[] accessor of AtomicString.

const LChar* characters = name.characters8();

then use characters[i] in the loop

[esprehn, 2016/09/02 00:55:31]

This is also doing bounds checks, so you can save 3 branches per character here.

https://cs.chromium.org/chromium/src/third_party/WebKit/Source/wtf/text/WTFSt...

https://cs.chromium.org/chromium/src/third_party/WebKit/Source/wtf/text/Strin...

+                    return false;
+            }
+        } else {
+            for (size_t i = 1; i < name.length(); ) {
+                UChar32 ch;
+                U16_NEXT(name.characters16(), i, name.length(), ch);

[esprehn, 2016/09/02 00:53:20]

ditto, you're checking for null repeatedly inside the loop, before the loop do:

const UChar* characters = name.characters16();

+                if (!Character::isPotentialCustomElementNameChar(ch))
+                    return false;
+            }
+        }
+
+        return !isHyphenatedSpecElementName(name);
+    }
 
     static bool shouldCreateCustomElement(const AtomicString& localName);
     static bool shouldCreateCustomElement(const QualifiedName&);
 
     static HTMLElement* createCustomElementSync(Document&, const AtomicString& localName, ExceptionState&);
     static HTMLElement* createCustomElementSync(Document&, const QualifiedName&, ExceptionState&);
     static HTMLElement* createCustomElementSync(Document&, const QualifiedName&);
     static HTMLElement* createCustomElementAsync(Document&, const QualifiedName&);
 
     static HTMLElement* createFailedElement(Document&, const QualifiedName&);
 
     static void enqueue(Element*, CustomElementReaction*);
     static void enqueueConnectedCallback(Element*);
     static void enqueueDisconnectedCallback(Element*);
     static void enqueueAdoptedCallback(Element*);
     static void enqueueAttributeChangedCallback(Element*, const QualifiedName&,
         const AtomicString& oldValue, const AtomicString& newValue);
 
     static void tryToUpgrade(Element*);
 
 private:
+    // Some existing specs have element names with hyphens in them,
+    // like font-face in SVG. The custom elements spec explicitly
+    // disallows these as custom element names.
+    // https://html.spec.whatwg.org/#valid-custom-element-name
+    static bool isHyphenatedSpecElementName(const AtomicString&);
     static HTMLElement* createUndefinedElement(Document&, const QualifiedName&);
 };
 
 } // namespace blink
 
 #endif // CustomElement_h