From 1a0bb51069881e331d5c6489d2f1b8df4d2845a1 Mon Sep 17 00:00:00 2001
From: "lrn@chromium.org"
 <lrn@chromium.org@ce2b1a6d-e550-0410-aec6-3dcde31c8c00>
Date: Mon, 26 Apr 2010 15:10:42 +0000
Subject: [PATCH] Fix bug in word-boundary-lookahead followed by end-of-input
 assertion.

Review URL: http://codereview.chromium.org/1712013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4504 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
---
 src/jsregexp.cc                      | 19 +++++------
 src/regexp-macro-assembler-tracer.cc |  4 +--
 test/mjsunit/regexp.js               | 48 ++++++++++++++++++++++++++++
 3 files changed, 60 insertions(+), 11 deletions(-)

diff --git a/src/jsregexp.cc b/src/jsregexp.cc
index 0ae8bd34df..9a1f1f114c 100644
--- a/src/jsregexp.cc
+++ b/src/jsregexp.cc
@@ -4872,17 +4872,18 @@ void Analysis::VisitAssertion(AssertionNode* that) {
 
     SetRelation word_relation =
         CharacterRange::WordCharacterRelation(following_chars);
-    if (word_relation.ContainedIn()) {
-      // Following character is definitely a word character.
-      type = (type == AssertionNode::AT_BOUNDARY) ?
-                  AssertionNode::AFTER_NONWORD_CHARACTER :
-                  AssertionNode::AFTER_WORD_CHARACTER;
-      that->set_type(type);
-    } else if (word_relation.Disjoint()) {
+    if (word_relation.Disjoint()) {
+      // Includes the case where following_chars is empty (e.g., end-of-input).
       // Following character is definitely *not* a word character.
       type = (type == AssertionNode::AT_BOUNDARY) ?
-                  AssertionNode::AFTER_WORD_CHARACTER :
-                  AssertionNode::AFTER_NONWORD_CHARACTER;
+                 AssertionNode::AFTER_WORD_CHARACTER :
+                 AssertionNode::AFTER_NONWORD_CHARACTER;
+      that->set_type(type);
+    } else if (word_relation.ContainedIn()) {
+      // Following character is definitely a word character.
+      type = (type == AssertionNode::AT_BOUNDARY) ?
+                 AssertionNode::AFTER_NONWORD_CHARACTER :
+                 AssertionNode::AFTER_WORD_CHARACTER;
       that->set_type(type);
     }
   }
diff --git a/src/regexp-macro-assembler-tracer.cc b/src/regexp-macro-assembler-tracer.cc
index c5c2919c3f..522042262f 100644
--- a/src/regexp-macro-assembler-tracer.cc
+++ b/src/regexp-macro-assembler-tracer.cc
@@ -37,8 +37,8 @@ RegExpMacroAssemblerTracer::RegExpMacroAssemblerTracer(
     RegExpMacroAssembler* assembler) :
   assembler_(assembler) {
   unsigned int type = assembler->Implementation();
-  ASSERT(type < 3);
-  const char* impl_names[3] = {"IA32", "ARM", "Bytecode"};
+  ASSERT(type < 4);
+  const char* impl_names[3] = {"IA32", "ARM", "X64", "Bytecode"};
   PrintF("RegExpMacroAssembler%s();\n", impl_names[type]);
 }
 
diff --git a/test/mjsunit/regexp.js b/test/mjsunit/regexp.js
index c8dcc6fe89..a8891969f4 100644
--- a/test/mjsunit/regexp.js
+++ b/test/mjsunit/regexp.js
@@ -436,3 +436,51 @@ assertTrue(re.multiline);
 assertEquals(0, re.lastIndex);
 assertEquals(37, re.someOtherProperty);
 assertEquals(37, re[42]);
+
+// Test boundary-checks.
+function assertRegExpTest(re, input, test) { 
+  assertEquals(test, re.test(input), "test:" + re + ":" + input);
+}
+
+assertRegExpTest(/b\b/, "b", true);
+assertRegExpTest(/b\b$/, "b", true);
+assertRegExpTest(/\bb/, "b", true);
+assertRegExpTest(/^\bb/, "b", true);
+assertRegExpTest(/,\b/, ",", false);
+assertRegExpTest(/,\b$/, ",", false);
+assertRegExpTest(/\b,/, ",", false);
+assertRegExpTest(/^\b,/, ",", false);
+
+assertRegExpTest(/b\B/, "b", false);
+assertRegExpTest(/b\B$/, "b", false);
+assertRegExpTest(/\Bb/, "b", false);
+assertRegExpTest(/^\Bb/, "b", false);
+assertRegExpTest(/,\B/, ",", true);
+assertRegExpTest(/,\B$/, ",", true);
+assertRegExpTest(/\B,/, ",", true);
+assertRegExpTest(/^\B,/, ",", true);
+
+assertRegExpTest(/b\b/, "b,", true);
+assertRegExpTest(/b\b/, "ba", false);
+assertRegExpTest(/b\B/, "b,", false);
+assertRegExpTest(/b\B/, "ba", true);
+
+assertRegExpTest(/b\Bb/, "bb", true);
+assertRegExpTest(/b\bb/, "bb", false);
+
+assertRegExpTest(/b\b[,b]/, "bb", false);
+assertRegExpTest(/b\B[,b]/, "bb", true);
+assertRegExpTest(/b\b[,b]/, "b,", true);
+assertRegExpTest(/b\B[,b]/, "b,", false);
+
+assertRegExpTest(/[,b]\bb/, "bb", false);
+assertRegExpTest(/[,b]\Bb/, "bb", true);
+assertRegExpTest(/[,b]\bb/, ",b", true);
+assertRegExpTest(/[,b]\Bb/, ",b", false);
+
+assertRegExpTest(/[,b]\b[,b]/, "bb", false);
+assertRegExpTest(/[,b]\B[,b]/, "bb", true);
+assertRegExpTest(/[,b]\b[,b]/, ",b", true);
+assertRegExpTest(/[,b]\B[,b]/, ",b", false);
+assertRegExpTest(/[,b]\b[,b]/, "b,", true);
+assertRegExpTest(/[,b]\B[,b]/, "b,", false);