Optimize KURL protocols

third_party/WebKit/Source/platform/weborigin/KURL.cpp

Index: third_party/WebKit/Source/platform/weborigin/KURL.cpp
diff --git a/third_party/WebKit/Source/platform/weborigin/KURL.cpp b/third_party/WebKit/Source/platform/weborigin/KURL.cpp
index 48edd0ef58e8855386d8a68fd54b5f9909915ab7..d6b59269cc37aba2d11450f8b004a54df21236e6 100644
--- a/third_party/WebKit/Source/platform/weborigin/KURL.cpp
+++ b/third_party/WebKit/Source/platform/weborigin/KURL.cpp
@@ -34,6 +34,7 @@
 #include "wtf/StdLibExtras.h"
 #include "wtf/text/CString.h"
 #include "wtf/text/StringHash.h"
+#include "wtf/text/StringStatics.h"
 #include "wtf/text/StringUTF8Adaptor.h"
 #include "wtf/text/TextEncoding.h"
 #include <algorithm>
@@ -48,6 +49,7 @@ static const int invalidPortNumber = 0xFFFF;
 
 static void assertProtocolIsGood(const char* protocol) {
 #if ENABLE(ASSERT)
+  DCHECK_NE(protocol, "");
   const char* p = protocol;
   while (*p) {
     ASSERT(*p > ' ' && *p < 0x7F && !(*p >= 'A' && *p <= 'Z'));
@@ -241,7 +243,7 @@ KURL::KURL(const AtomicString& canonicalString,
       m_protocolIsInHTTPFamily(false),
       m_parsed(parsed),
       m_string(canonicalString) {
-  initProtocolIsInHTTPFamily();
+  initProtocolMetadata();
   initInnerURL();
 }
 
@@ -253,6 +255,7 @@ KURL::KURL(WTF::HashTableDeletedValueType)
 KURL::KURL(const KURL& other)
     : m_isValid(other.m_isValid),
       m_protocolIsInHTTPFamily(other.m_protocolIsInHTTPFamily),
+      m_protocol(other.m_protocol),
       m_parsed(other.m_parsed),
       m_string(other.m_string) {
   if (other.m_innerURL.get())
@@ -264,6 +267,7 @@ KURL::~KURL() {}
 KURL& KURL::operator=(const KURL& other) {
   m_isValid = other.m_isValid;
   m_protocolIsInHTTPFamily = other.m_protocolIsInHTTPFamily;
+  m_protocol = other.m_protocol;
   m_parsed = other.m_parsed;
   m_string = other.m_string;
   if (other.m_innerURL)
@@ -277,6 +281,7 @@ KURL KURL::copy() const {
   KURL result;
   result.m_isValid = m_isValid;
   result.m_protocolIsInHTTPFamily = m_protocolIsInHTTPFamily;
+  result.m_protocol = m_protocol.isolatedCopy();
   result.m_parsed = m_parsed;
   result.m_string = m_string.isolatedCopy();
   if (m_innerURL)
@@ -336,7 +341,7 @@ String KURL::lastPathComponent() const {
 }
 
 String KURL::protocol() const {
-  return componentString(m_parsed.scheme);
+  return m_protocol;
 }
 
 String KURL::host() const {
@@ -365,6 +370,9 @@ unsigned short KURL::port() const {
   return static_cast<unsigned short>(port);
 }
 
+// TODO(csharrison): Migrate pass() and user() to return a StringView. Most
+// consumers just need to know if the string is empty.
+
 String KURL::pass() const {
   // Bug: https://bugs.webkit.org/show_bug.cgi?id=21015 this function returns
   // a null string when the password is empty, which we duplicate here.
@@ -433,6 +441,7 @@ bool KURL::setProtocol(const String& protocol) {
   replacements.SetScheme(charactersOrEmpty(newProtocolUTF8),
                          url::Component(0, newProtocolUTF8.length()));
   replaceComponents(replacements);
+  initProtocolMetadata();
 
   // isValid could be false but we still return true here. This is because
   // WebCore or JS scripts can build up a URL by setting individual
@@ -767,7 +776,7 @@ void KURL::init(const KURL& base,
     m_string = AtomicString::fromUTF8(output.data(), output.length());
   }
 
-  initProtocolIsInHTTPFamily();
+  initProtocolMetadata();
   initInnerURL();
   DCHECK_EQ(protocol(), protocol().lower());
 }
@@ -786,50 +795,24 @@ void KURL::initInnerURL() {
     m_innerURL.reset();
 }
 
-template <typename CHAR>
-bool internalProtocolIs(const url::Component& scheme,
-                        const CHAR* spec,
-                        const char* protocol) {
-  const CHAR* begin = spec + scheme.begin;
-  const CHAR* end = begin + scheme.len;
-
-  while (begin != end && *protocol) {
-    ASSERT(toASCIILower(*protocol) == *protocol);
-    if (toASCIILower(*begin++) != *protocol++)
-      return false;
-  }
-
-  // Both strings are equal (ignoring case) if and only if all of the characters
-  // were equal, and the end of both has been reached.
-  return begin == end && !*protocol;
-}
-
-template <typename CHAR>
-bool checkIfProtocolIsInHTTPFamily(const url::Component& scheme,
-                                   const CHAR* spec) {
-  if (scheme.len == 4)
-    return internalProtocolIs(scheme, spec, "http");
-  if (scheme.len == 5)
-    return internalProtocolIs(scheme, spec, "https");
-  if (scheme.len == 7)
-    return internalProtocolIs(scheme, spec, "http-so");

[esprehn, 2016/12/12 22:18:46]

This was all case insensitive before.

[Charlie Harrison, 2016/12/12 22:26:06]

We should always have a canonical, lowercase scheme here, so the case
insensitivity was not necessary. We could probably DCHECK this fact though.

[Charlie Harrison, 2016/12/13 16:03:22]

Actually, we have a DCHECK_EQ(protocol(), protocol().lower()) in KURL::init from
a previous patch which removed a lot of random protocol.lower() in a bunch of
places in the codebase. Let me move it to initProtocolMetadata which will catch
more cases.

-  if (scheme.len == 8)
-    return internalProtocolIs(scheme, spec, "https-so");
-  return false;
-}
-
-void KURL::initProtocolIsInHTTPFamily() {
+void KURL::initProtocolMetadata() {
   if (!m_isValid) {
     m_protocolIsInHTTPFamily = false;
     return;
   }
 
   ASSERT(!m_string.isNull());
-  m_protocolIsInHTTPFamily =
-      m_string.is8Bit() ? checkIfProtocolIsInHTTPFamily(m_parsed.scheme,
-                                                        m_string.characters8())
-                        : checkIfProtocolIsInHTTPFamily(
-                              m_parsed.scheme, m_string.characters16());
+  StringView protocol = componentStringView(m_parsed.scheme);
+  m_protocolIsInHTTPFamily = true;
+  if (protocol == WTF::httpsAtom) {
+    m_protocol = WTF::httpsAtom;
+  } else if (protocol == WTF::httpAtom) {
+    m_protocol = WTF::httpAtom;
+  } else {
+    m_protocol = AtomicString(protocol.toString());
+    m_protocolIsInHTTPFamily =
+        m_protocol == "http-so" || m_protocol == "https-so";

[esprehn, 2016/12/12 22:18:46]

What strange protocols are these?

[Charlie Harrison, 2016/12/12 22:26:06]

suborigin protocols:
https://w3c.github.io/webappsec-suborigins/#representation

I think these are rare so I didn't add them to the atomic string table.

+  }
 }
 
 bool KURL::protocolIs(const char* protocol) const {
@@ -840,15 +823,7 @@ bool KURL::protocolIs(const char* protocol) const {
   // instead.
   // FIXME: Chromium code needs to be fixed for this assert to be enabled.
   // ASSERT(strcmp(protocol, "javascript"));
-
-  if (m_string.isNull() || m_parsed.scheme.len <= 0)
-    return *protocol == '\0';
-
-  return m_string.is8Bit()
-             ? internalProtocolIs(m_parsed.scheme, m_string.characters8(),
-                                  protocol)
-             : internalProtocolIs(m_parsed.scheme, m_string.characters16(),
-                                  protocol);
+  return m_protocol == protocol;

[esprehn, 2016/12/12 22:18:46]

internalProtocolIs was case insensitive, are you sure this is right?

[Charlie Harrison, 2016/12/12 22:26:06]

assertProtocolIsGood checks that the input protocol wasn't uppercase ASCII.

 }
 
 String KURL::stringForInvalidComponent() const {
@@ -857,7 +832,7 @@ String KURL::stringForInvalidComponent() const {
   return emptyString();
 }
 
-String KURL::componentString(const url::Component& component) const {
+StringView KURL::componentStringView(const url::Component& component) const {
   if (!m_isValid || component.len <= 0)
     return stringForInvalidComponent();
   // begin and len are in terms of bytes which do not match
@@ -868,7 +843,14 @@ String KURL::componentString(const url::Component& component) const {
   // byte) will be longer than what's needed by 'mid'. However, mid
   // truncates len to avoid go past the end of a string so that we can
   // get away without doing anything here.
-  return getString().substring(component.begin, component.len);
+
+  int maxLength = getString().length() - component.begin;
+  return StringView(getString(), component.begin,
+                    component.len > maxLength ? maxLength : component.len);
+}
+
+String KURL::componentString(const url::Component& component) const {
+  return componentStringView(component).toString();
 }
 
 template <typename CHAR>