Heuristics for grabber-continental.com.out (select-one)

chrome/browser/autofill/address_field.cc

 // Copyright (c) 2011 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.
 
 #include "chrome/browser/autofill/address_field.h"
 
 #include <stddef.h>
 
 #include "base/logging.h"
 #include "base/memory/scoped_ptr.h"
@@ skipping 177 matching lines @@
   if (is_ecml) {
     pattern = GetEcmlPattern(kEcmlShipToAddress1, kEcmlBillToAddress1, '|');
     if (!ParseField(scanner, pattern, &address_field->address1_))
       return false;
   } else {
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_LINE_1_RE);
     string16 label_pattern =
         l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_LINE_1_LABEL_RE);
 
     if (!ParseField(scanner, pattern, &address_field->address1_) &&
-        !ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL,
+        !ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL | MATCH_TEXT,
                              &address_field->address1_)) {
       return false;
     }
   }
 
   // Optionally parse more address lines, which may have empty labels.
   // Some pages have 3 address lines (eg SharperImageModifyAccount.html)
   // Some pages even have 4 address lines (e.g. uk/ShoesDirect2.html)!
   if (is_ecml) {
     pattern = GetEcmlPattern(kEcmlShipToAddress2, kEcmlBillToAddress2, '|');
     if (!ParseEmptyLabel(scanner, &address_field->address2_))
       ParseField(scanner, pattern, &address_field->address2_);
   } else {
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_LINE_2_RE);
     string16 label_pattern =
         l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_LINE_1_LABEL_RE);
     if (!ParseEmptyLabel(scanner, &address_field->address2_) &&
         !ParseField(scanner, pattern, &address_field->address2_)) {
-      ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL,
+      ParseFieldSpecifics(scanner, label_pattern, MATCH_LABEL | MATCH_TEXT,
                           &address_field->address2_);
     }
   }
 
   // Try for a third line, which we will promptly discard.
   if (address_field->address2_ != NULL) {
     if (is_ecml) {
       pattern = GetEcmlPattern(kEcmlShipToAddress3, kEcmlBillToAddress3, '|');
       ParseField(scanner, pattern, NULL);
     } else {
@@ skipping 15 matching lines @@
   // Note: ECML standard uses 2 letter country code (ISO 3166)
   if (address_field->country_ && !address_field->country_->IsEmpty())
     return false;
 
   string16 pattern;
   if (is_ecml)
     pattern = GetEcmlPattern(kEcmlShipToCountry, kEcmlBillToCountry, '|');
   else
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_COUNTRY_RE);
 
-  return ParseField(scanner, pattern, &address_field->country_);
+  return ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT | MATCH_SELECT,
+                             &address_field->country_);
 }
 
 // static
 bool AddressField::ParseZipCode(AutofillScanner* scanner,
                                 bool is_ecml,
                                 AddressField* address_field) {
   // Parse a zip code.  On some UK pages (e.g. The China Shop2.html) this
   // is called a "post code".
   //
   // HACK: Just for the MapQuest driving directions page we match the
@@ skipping 63 matching lines @@
                               AddressField* address_field) {
   if (address_field->state_)
     return false;
 
   string16 pattern;
   if (is_ecml)
     pattern = GetEcmlPattern(kEcmlShipToStateProv, kEcmlBillToStateProv, '|');
   else
     pattern = l10n_util::GetStringUTF16(IDS_AUTOFILL_STATE_RE);
 
-  return ParseField(scanner, pattern, &address_field->state_);
+  return ParseFieldSpecifics(scanner, pattern, MATCH_DEFAULT | MATCH_SELECT,
+                             &address_field->state_);
 }
 
 AddressType AddressField::AddressTypeFromText(const string16 &text) {
   if (text.find(l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_TYPE_SAME_AS_RE))
           != string16::npos ||
       text.find(l10n_util::GetStringUTF16(IDS_AUTOFILL_ADDRESS_TYPE_USE_MY_RE))
           != string16::npos)
     // This text could be a checkbox label such as "same as my billing
     // address" or "use my shipping address".
     // ++ It would help if we generally skipped all text that appears
@@ skipping 15 matching lines @@
     return kBillingAddress;
 
   if (bill == string16::npos && ship != string16::npos)
     return kShippingAddress;
 
   if (bill > ship)
     return kBillingAddress;
 
   return kShippingAddress;
 }