Refactor code to avoid direct dependency upon ICU: phishing_term_feature_extractor

base/i18n/break_iterator.h

 // 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.
 
 #ifndef BASE_I18N_BREAK_ITERATOR_H_
 #define BASE_I18N_BREAK_ITERATOR_H_
 
 #include "base/basictypes.h"
 #include "base/i18n/base_i18n_export.h"
 #include "base/strings/string16.h"
@@ skipping 88 matching lines @@
   // have advanced to somewhere useful.
   string16 GetString() const;
 
   // Returns the value of pos() returned before Advance() was last called.
   size_t prev() const { return prev_; }
 
   // Returns the current break position within the string,
   // or BreakIterator::npos when done.
   size_t pos() const { return pos_; }
 
+  // Returns whether or not the iterator is currently at a valid position.
+  bool IsValid() const {
+    return pos_ != static_cast<size_t>(-1);
+  }
+
  private:
   // ICU iterator, avoiding ICU ubrk.h dependence.
   // This is actually an ICU UBreakiterator* type, which turns out to be
   // a typedef for a void* in the ICU headers. Using void* directly prevents
   // callers from needing access to the ICU public headers directory.
   void* iter_;
 
   // The string we're iterating over.
   const string16& string_;
 
   // The breaking style (word/space/newline).
   BreakType break_type_;
 
   // Previous and current iterator positions.
   size_t prev_, pos_;
 
   DISALLOW_COPY_AND_ASSIGN(BreakIterator);
 };
 
 }  // namespace i18n
 }  // namespace base
 
 #endif  // BASE_I18N_BREAK_ITERATOR_H_