Improve Autofill heuristics when detecting labels from previous elements.

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"
 #include "base/string16.h"
 #include "base/string_util.h"
 #include "base/utf_string_conversions.h"
 #include "chrome/browser/autofill/autofill_ecml.h"
 #include "chrome/browser/autofill/autofill_field.h"
 #include "chrome/browser/autofill/autofill_scanner.h"
 #include "grit/autofill_resources.h"
 #include "ui/base/l10n/l10n_util.h"
 
 using autofill::GetEcmlPattern;
 
 FormField* AddressField::Parse(AutofillScanner* scanner, bool is_ecml) {
   if (scanner->IsEnd())
     return NULL;
 
   scoped_ptr<AddressField> address_field(new AddressField);
-  const AutofillField* initial_field = scanner->Cursor();
-  scanner->SaveCursor();
+  const AutofillField* const initial_field = scanner->Cursor();
+  size_t saved_cursor = scanner->SaveCursor();
 
   string16 attention_ignored =
       l10n_util::GetStringUTF16(IDS_AUTOFILL_ATTENTION_IGNORED_RE);
   string16 region_ignored =
       l10n_util::GetStringUTF16(IDS_AUTOFILL_REGION_IGNORED_RE);
 
   // Allow address fields to appear in any order.
+  size_t begin_trailing_non_labeled_fields = 0;
+  bool has_trailing_non_labeled_fields = false;
   while (!scanner->IsEnd()) {
+    const size_t cursor = scanner->SaveCursor();
     if (ParseAddressLines(scanner, is_ecml, address_field.get()) ||
         ParseCity(scanner, is_ecml, address_field.get()) ||
         ParseState(scanner, is_ecml, address_field.get()) ||
         ParseZipCode(scanner, is_ecml, address_field.get()) ||
         ParseCountry(scanner, is_ecml, address_field.get()) ||
         ParseCompany(scanner, is_ecml, address_field.get())) {
+      has_trailing_non_labeled_fields = false;
       continue;
     } else if (ParseField(scanner, attention_ignored, NULL) ||
                ParseField(scanner, region_ignored, NULL)) {
       // We ignore the following:
       // * Attention.
       // * Province/Region/Other.
       continue;
     } else if (scanner->Cursor() != initial_field &&
                ParseEmptyLabel(scanner, NULL)) {
       // Ignore non-labeled fields within an address; the page
       // MapQuest Driving Directions North America.html contains such a field.
       // We only ignore such fields after we've parsed at least one other field;
       // otherwise we'd effectively parse address fields before other field
       // types after any non-labeled fields, and we want email address fields to
       // have precedence since some pages contain fields labeled
       // "Email address".
+      if (!has_trailing_non_labeled_fields) {
+        has_trailing_non_labeled_fields = true;
+        begin_trailing_non_labeled_fields = cursor;
+      }
+
       continue;
     } else {
       // No field found.
       break;
     }
   }
 
   // If we have identified any address fields in this field then it should be
   // added to the list of fields.
   if (address_field->company_ != NULL ||
       address_field->address1_ != NULL || address_field->address2_ != NULL ||
       address_field->city_ != NULL || address_field->state_ != NULL ||
       address_field->zip_ != NULL || address_field->zip4_ ||
       address_field->country_ != NULL) {
+    // Don't slurp non-labeled fields at the end into the address.
+    if (has_trailing_non_labeled_fields)
+      scanner->RewindTo(begin_trailing_non_labeled_fields);
+
     address_field->type_ = address_field->FindType();
     return address_field.release();
   }
 
-  scanner->Rewind();
+  scanner->RewindTo(saved_cursor);
   return NULL;
 }
 
 AddressType AddressField::FindType() const {
   // First look at the field name, which itself will sometimes contain
   // "bill" or "ship".  We could check for the ECML type prefixes
   // here, but there's no need to since ECML's prefixes Ecom_BillTo
   // and Ecom_ShipTo contain "bill" and "ship" anyway.
   if (company_) {
     string16 name = StringToLowerASCII(company_->name);
@@ skipping 301 matching lines @@
     return kBillingAddress;
 
   if (bill == string16::npos && ship != string16::npos)
     return kShippingAddress;
 
   if (bill > ship)
     return kBillingAddress;
 
   return kShippingAddress;
 }