Port and integrate the irregexp engine from V8

runtime/lib/regexp.cc

Index: runtime/lib/regexp.cc
diff --git a/runtime/lib/regexp.cc b/runtime/lib/regexp.cc
index 964a48850fd09cfc35171ccd0ea0f6ffb49a3207..e6286b28606985a250b635249ea4ac7a93f39319 100644
--- a/runtime/lib/regexp.cc
+++ b/runtime/lib/regexp.cc
@@ -4,6 +4,7 @@
 
 #include "platform/assert.h"
 #include "vm/bootstrap_natives.h"
+#include "vm/compiler.h"
 #include "vm/exceptions.h"
 #include "vm/native_entry.h"
 #include "vm/object.h"
@@ -13,6 +14,9 @@
 
 namespace dart {
 
+DECLARE_FLAG(bool, trace_irregexp);
+
+
 DEFINE_NATIVE_ENTRY(JSSyntaxRegExp_factory, 4) {
   ASSERT(TypeArguments::CheckedHandle(arguments->NativeArgAt(0)).IsNull());
   GET_NON_NULL_NATIVE_ARGUMENT(String, pattern, arguments->NativeArgAt(1));
@@ -22,7 +26,24 @@ DEFINE_NATIVE_ENTRY(JSSyntaxRegExp_factory, 4) {
       Instance, handle_case_sensitive, arguments->NativeArgAt(3));
   bool ignore_case = handle_case_sensitive.raw() != Bool::True().raw();
   bool multi_line = handle_multi_line.raw() == Bool::True().raw();
+
+#if defined(USE_JSCRE)
   return Jscre::Compile(pattern, multi_line, ignore_case);
+#else
+  // Parse the pattern once in order to throw any format exceptions within
+  // the factory constructor. It is parsed again upon compilation.
+  RegExpCompileData compileData;
+  if (!RegExpParser::ParseRegExp(pattern, multi_line, &compileData)) {
+    // Parsing failures throw an exception.
+    UNREACHABLE();
+  }
+
+  // Create a JSRegExp object containing only the initial parameters.
+  JSRegExp& regexp = JSRegExp::Handle(
+        isolate, RegExpEngine::New(isolate, pattern, multi_line, ignore_case));
+
+  return regexp.raw();
+#endif
 }
 
 
@@ -70,27 +91,26 @@ DEFINE_NATIVE_ENTRY(JSSyntaxRegExp_ExecuteMatch, 3) {
   GET_NON_NULL_NATIVE_ARGUMENT(String, str, arguments->NativeArgAt(1));
   GET_NON_NULL_NATIVE_ARGUMENT(Smi, start_index, arguments->NativeArgAt(2));
 
-  // The irregexp parser runs alongside the jscre parser while the port is
-  // still in progress. When done, it will replace jscre completely.
-  const String& pattern = String::Handle(regexp.pattern());
-  RegExpCompileData compileData;
-  if (RegExpParser::ParseRegExp(pattern, regexp.is_multi_line(),
-                                &compileData)) {
-    RegExpEngine::Compile(&compileData,
-                          regexp.is_ignore_case(),
-                          regexp.is_global(),  // A Dart regexp is always global
-                          regexp.is_multi_line(),
-                          pattern,
-                          str,
-                          str.IsOneByteString());
+#if defined(USE_JSCRE)
+  return Jscre::Execute(regexp, str, start_index.Value());
+#else
+  // This function is intrinsified. See Intrinsifier::JSRegExp_ExecuteMatch.
+  const intptr_t cid = str.GetClassId();
+
+  // Retrieve the cached function.
+  Function& fn = Function::Handle(regexp.function(cid));

[Florian Schneider, 2014/10/01 17:04:15]

const Function&

[jgruber1, 2014/10/03 18:59:53]

Done.

+  ASSERT(!fn.IsNull());
+
+  if (fn.CurrentCode() == StubCode::LazyCompile_entry()->code()) {

[Florian Schneider, 2014/10/01 17:04:14]

In the instrinc assembly code you check

(sample_subject == null)

here. It should be the equivalent (fn is compiled on first invocation) - but for
consistency I'd make it the same (i.e. use the same condition here.

[jgruber1, 2014/10/03 18:59:53]

Done.

+    // Set the sample subject now that we have it.
+    regexp.set_sample_subject(cid, str);
 #ifdef DEBUG

[Vyacheslav Egorov (Google), 2014/10/01 20:13:21]

Empty #ifdef ?

[jgruber1, 2014/10/03 18:59:53]

Done.

-    RegExpEngine::DotPrint(pattern.ToCString(),
-                           compileData.node,
-                           regexp.is_ignore_case());
 #endif
   }
 
-  return Jscre::Execute(regexp, str, start_index.Value());
+  // And finally call the generated code.
+  return IRRegExpMacroAssembler::Execute(fn, str, start_index, isolate);
+#endif
 }
 
 }  // namespace dart