Autofill should use unicode/regex.h instead of WebKit::WebRegularExpression

Switches Autofill over to use ICU regex facilities.

BUG=84199, 84158
TEST=unit tests.

Review URL: http://codereview.chromium.org/7071048

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@87126 0039d316-1c4b-4281-b951-d872f2087c98

6 files changed, 58 insertions, 58 deletions

diff --git a/chrome/browser/autofill/autofill_scanner.cc b/chrome/browser/autofill/autofill_scanner.cc
index 1fd8a7e..6526fcc 100644
--- a/chrome/browser/autofill/autofill_scanner.cc
+++ b/chrome/browser/autofill/autofill_scanner.cc
@@ -6,6 +6,7 @@
 
 #include "base/logging.h"
 #include "chrome/browser/autofill/autofill_field.h"
+#include "unicode/regex.h"
 
 AutofillScanner::AutofillScanner(
     const std::vector<const AutofillField*>& fields)
@@ -43,3 +44,21 @@ void AutofillScanner::Rewind() {
 void AutofillScanner::SaveCursor() {
   saved_cursors_.push_back(cursor_);
 }
+
+namespace autofill {
+
+bool MatchString(const string16& input, const string16& pattern) {
+  UErrorCode status = U_ZERO_ERROR;
+  icu::UnicodeString icu_pattern(pattern.data(), pattern.length());
+  icu::UnicodeString icu_input(input.data(), input.length());
+  icu::RegexMatcher matcher(icu_pattern, icu_input,
+                            UREGEX_CASE_INSENSITIVE, status);
+  DCHECK(U_SUCCESS(status));
+
+  UBool match = matcher.find(0, status);
+  DCHECK(U_SUCCESS(status));
+  return !!match;
+}
+
+}  // namespace autofill
+
diff --git a/chrome/browser/autofill/autofill_scanner.h b/chrome/browser/autofill/autofill_scanner.h
index 8c582dd..7eca0ce 100644
--- a/chrome/browser/autofill/autofill_scanner.h
+++ b/chrome/browser/autofill/autofill_scanner.h
@@ -9,6 +9,7 @@
 #include <vector>
 
 #include "base/basictypes.h"
+#include "base/string16.h"
 
 class AutofillField;
 
@@ -48,4 +49,13 @@ class AutofillScanner {
   DISALLOW_COPY_AND_ASSIGN(AutofillScanner);
 };
 
+// Parsing utilities.
+namespace autofill {
+
+// Case-insensitive regular expression matching.  Returns true if |pattern| is
+// found in |input|.
+bool MatchString(const string16& input, const string16& pattern);
+
+}  // namespace autofill
+
 #endif  // CHROME_BROWSER_AUTOFILL_AUTOFILL_SCANNER_H_
diff --git a/chrome/browser/autofill/credit_card.cc b/chrome/browser/autofill/credit_card.cc
index d6ea407..4e487da 100644
--- a/chrome/browser/autofill/credit_card.cc
+++ b/chrome/browser/autofill/credit_card.cc
@@ -14,13 +14,11 @@
 #include "base/string_split.h"
 #include "base/string_util.h"
 #include "base/utf_string_conversions.h"
+#include "chrome/browser/autofill/autofill_scanner.h"
 #include "chrome/browser/autofill/autofill_type.h"
 #include "chrome/browser/autofill/field_types.h"
 #include "chrome/common/guid.h"
 #include "grit/generated_resources.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebRegularExpression.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebString.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebTextCaseSensitivity.h"
 #include "ui/base/l10n/l10n_util.h"
 
 namespace {
@@ -322,22 +320,20 @@ const string16 CreditCard::Label() const {
 
 void CreditCard::SetInfoForMonthInputType(const string16& value) {
   // Check if |text| is "yyyy-mm" format first, and check normal month format.
-  WebKit::WebRegularExpression re(WebKit::WebString("^[0-9]{4}\\-[0-9]{1,2}$"),
-                                  WebKit::WebTextCaseInsensitive);
-  bool match = re.match(WebKit::WebString(StringToLowerASCII(value))) != -1;
-  if (match) {
-    std::vector<string16> year_month;
-    base::SplitString(value, L'-', &year_month);
-    DCHECK_EQ((int)year_month.size(), 2);
-    int num = 0;
-    bool converted = false;
-    converted = base::StringToInt(year_month[0], &num);
-    DCHECK(converted);
-    SetExpirationYear(num);
-    converted = base::StringToInt(year_month[1], &num);
-    DCHECK(converted);
-    SetExpirationMonth(num);
-  }
+  if (!autofill::MatchString(value, UTF8ToUTF16("^[0-9]{4}-[0-9]{1,2}$")))
+    return;
+
+  std::vector<string16> year_month;
+  base::SplitString(value, L'-', &year_month);
+  DCHECK_EQ((int)year_month.size(), 2);
+  int num = 0;
+  bool converted = false;
+  converted = base::StringToInt(year_month[0], &num);
+  DCHECK(converted);
+  SetExpirationYear(num);
+  converted = base::StringToInt(year_month[1], &num);
+  DCHECK(converted);
+  SetExpirationMonth(num);
 }
 
 string16 CreditCard::ObfuscatedNumber() const {
diff --git a/chrome/browser/autofill/form_field.cc b/chrome/browser/autofill/form_field.cc
index 5cd6ae6..01f2ee1 100644
--- a/chrome/browser/autofill/form_field.cc
+++ b/chrome/browser/autofill/form_field.cc
@@ -24,9 +24,6 @@
 #include "chrome/browser/autofill/name_field.h"
 #include "chrome/browser/autofill/phone_field.h"
 #include "grit/autofill_resources.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebRegularExpression.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebString.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebTextCaseSensitivity.h"
 #include "ui/base/l10n/l10n_util.h"
 
 namespace {
@@ -34,26 +31,6 @@ namespace {
 using autofill::GetEcmlPattern;
 using autofill::HasECMLField;
 
-bool MatchName(const AutofillField* field, const string16& pattern) {
-  // TODO(jhawkins): Remove StringToLowerASCII.  WebRegularExpression needs to
-  // be fixed to take WebTextCaseInsensitive into account.
-  WebKit::WebRegularExpression re(WebKit::WebString(pattern),
-                                  WebKit::WebTextCaseInsensitive);
-  bool match = re.match(
-      WebKit::WebString(StringToLowerASCII(field->name))) != -1;
-  return match;
-}
-
-bool MatchLabel(const AutofillField* field, const string16& pattern) {
-  // TODO(jhawkins): Remove StringToLowerASCII.  WebRegularExpression needs to
-  // be fixed to take WebTextCaseInsensitive into account.
-  WebKit::WebRegularExpression re(WebKit::WebString(pattern),
-                                  WebKit::WebTextCaseInsensitive);
-  bool match = re.match(
-      WebKit::WebString(StringToLowerASCII(field->label))) != -1;
-  return match;
-}
-
 bool IsTextField(const string16& type) {
   return type == ASCIIToUTF16("text");
 }
@@ -173,11 +150,15 @@ bool FormField::MatchAndAdvance(AutofillScanner* scanner,
 bool FormField::Match(const AutofillField* field,
                       const string16& pattern,
                       int match_type) {
-  if ((match_type & FormField::MATCH_LABEL) && MatchLabel(field, pattern))
-      return true;
+  if ((match_type & FormField::MATCH_LABEL) &&
+      autofill::MatchString(field->label, pattern)) {
+    return true;
+  }
 
-  if ((match_type & FormField::MATCH_NAME) && MatchName(field, pattern))
-      return true;
+  if ((match_type & FormField::MATCH_NAME) &&
+      autofill::MatchString(field->name, pattern)) {
+    return true;
+  }
 
   return false;
 }
diff --git a/chrome/browser/autofill/form_structure_unittest.cc b/chrome/browser/autofill/form_structure_unittest.cc
index 983f8e8..7f80582 100644
--- a/chrome/browser/autofill/form_structure_unittest.cc
+++ b/chrome/browser/autofill/form_structure_unittest.cc
@@ -833,7 +833,7 @@ TEST(FormStructureTest, ThreeAddressLinesExpedia) {
   form_structure->DetermineHeuristicTypes();
   EXPECT_TRUE(form_structure->IsAutofillable(true));
   ASSERT_EQ(4U, form_structure->field_count());
-  ASSERT_EQ(3U, form_structure->autofill_count());
+  EXPECT_EQ(3U, form_structure->autofill_count());
 
   // Address Line 1.
   EXPECT_EQ(ADDRESS_HOME_LINE1, form_structure->field(0)->heuristic_type());
@@ -1125,7 +1125,7 @@ TEST(FormStructureTest, HeuristicsInfernoCC) {
 
   // Expect the correct number of fields.
   ASSERT_EQ(5U, form_structure->field_count());
-  ASSERT_EQ(5U, form_structure->autofill_count());
+  EXPECT_EQ(5U, form_structure->autofill_count());
 
   // Name on Card.
   EXPECT_EQ(CREDIT_CARD_NAME, form_structure->field(0)->heuristic_type());
diff --git a/chrome/browser/autofill/personal_data_manager.cc b/chrome/browser/autofill/personal_data_manager.cc
index 03eb64e..d3d78ad 100644
--- a/chrome/browser/autofill/personal_data_manager.cc
+++ b/chrome/browser/autofill/personal_data_manager.cc
@@ -13,6 +13,7 @@
 #include "chrome/browser/autofill/autofill-inl.h"
 #include "chrome/browser/autofill/autofill_field.h"
 #include "chrome/browser/autofill/autofill_metrics.h"
+#include "chrome/browser/autofill/autofill_scanner.h"
 #include "chrome/browser/autofill/form_structure.h"
 #include "chrome/browser/autofill/phone_number.h"
 #include "chrome/browser/autofill/phone_number_i18n.h"
@@ -23,8 +24,6 @@
 #include "chrome/browser/webdata/web_data_service.h"
 #include "chrome/common/pref_names.h"
 #include "content/browser/browser_thread.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebRegularExpression.h"
-#include "third_party/WebKit/Source/WebKit/chromium/public/WebString.h"
 
 namespace {
 
@@ -76,20 +75,15 @@ T* address_of(T& v) {
 bool IsValidEmail(const string16& value) {
   // This regex is more permissive than the official rfc2822 spec on the
   // subject, but it does reject obvious non-email addresses.
-  const string16 kEmailPattern =
-      ASCIIToUTF16("^[^@]+@[^@]+\\.[a-z]{2,6}$");
-  WebKit::WebRegularExpression re(WebKit::WebString(kEmailPattern),
-                                  WebKit::WebTextCaseInsensitive);
-  return re.match(WebKit::WebString(StringToLowerASCII(value))) != -1;
+  const string16 kEmailPattern = ASCIIToUTF16("^[^@]+@[^@]+\\.[a-z]{2,6}$");
+  return autofill::MatchString(value, kEmailPattern);
 }
 
 // Valid for US zip codes only.
 bool IsValidZip(const string16& value) {
   // Basic US zip code matching.
   const string16 kZipPattern = ASCIIToUTF16("^\\d{5}(-\\d{4})?$");
-  WebKit::WebRegularExpression re(WebKit::WebString(kZipPattern),
-                                  WebKit::WebTextCaseInsensitive);
-  return re.match(WebKit::WebString(StringToLowerASCII(value))) != -1;
+  return autofill::MatchString(value, kZipPattern);
 }
 
 // Returns true if minimum requirements for import of a given |profile| have