Detecting form locale

chrome/browser/autofill/autofill_locale_model.h

+// Copyright (c) 2010 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.
+
+#ifndef CHROME_BROWSER_AUTOFILL_AUTOFILL_LOCALE_MODEL_H_
+#define CHROME_BROWSER_AUTOFILL_AUTOFILL_LOCALE_MODEL_H_
+#pragma once
+
+#include <string>
+#include <vector>
+
+#include "base/basictypes.h"
+#include "base/hash_tables.h"
+
+class FormStructure;
+class GURL;
+class TabContents;
+
+// This class is responsible for collating and making sense of the various
+// locale indicators for auto-fillable forms.
+class AutoFillLocaleModel {
+ public:
+  explicit AutoFillLocaleModel(const TabContents* tab_contents);
+  virtual ~AutoFillLocaleModel();
+
+  // Updates the |form|'s locale using heuristics that combine three signals:
+  // * the author-specified locale, which we get from the DOM;
+  // * the automatically detected language; and
+  // * the page url's registry (e.g. ".co.uk").
+  void UpdateLocale(FormStructure* form) const;
+
+  // Returns whether |language_tag| is a known language subtag.
+  static bool IsValidLanguageTag(const std::string& language_tag);
+
+  // Returns whether |region_tag| is a known region subtag.
+  static bool IsValidRegionTag(const std::string& region_tag);
+
+  bool tab_language_determined() const {
+    return tab_language_determined_;
+  }
+  void set_tab_language_determined(bool language_determined) {
+    tab_language_determined_ = language_determined;
+  }
+
+ protected:
+  // Region and language data for known domain registries (roughly, top level
+  // domains).
+  struct AutoFillRegistryInfo {
+    std::string region;
+    std::vector<std::string> languages;
+  };
+  typedef base::hash_map<std::string, AutoFillRegistryInfo> AutoFillRegistryMap;
+
+  // For tests.
+  const AutoFillRegistryMap* registries() { return &registries_; }
+
+ private:
+  std::string RegionFromRegistry(const std::string& registry,
+                                 std::string* guessed_language) const;
+
+  // Hash table mapping registries to their language and region data.
+  AutoFillRegistryMap registries_;
+
+  // The |TabContents| hosting this language model.  Weak reference.
+  const TabContents* tab_contents_;
+
+  // Has the tab language been determined?
+  bool tab_language_determined_;
+
+  DISALLOW_COPY_AND_ASSIGN(AutoFillLocaleModel);
+};
+
+#endif  // CHROME_BROWSER_AUTOFILL_AUTOFILL_LOCALE_MODEL_H_