Updates SpellcheckWordIterator::GetNextWord to return an enum.

chrome/renderer/spellchecker/spellcheck_worditerator.h

Index: chrome/renderer/spellchecker/spellcheck_worditerator.h
diff --git a/chrome/renderer/spellchecker/spellcheck_worditerator.h b/chrome/renderer/spellchecker/spellcheck_worditerator.h
index 6b6a2f4eac20bc1884c79a2ce1a99aa604aaa09f..728e2a13c1e70a8651904b7ad31c6fdec6e5963e 100644
--- a/chrome/renderer/spellchecker/spellcheck_worditerator.h
+++ b/chrome/renderer/spellchecker/spellcheck_worditerator.h
@@ -112,6 +112,17 @@ class SpellcheckCharAttribute {
 //
 class SpellcheckWordIterator {
  public:
+  enum WordIteratorStatus {
+    // The end of a sequence of text that the iterator recognizes as characters
+    // that can form a word.
+    IS_WORD,
+    // Non-word characters that the iterator can skip past, such as punctuation,
+    // whitespace, and characters from another character set.
+    IS_SKIPPABLE_CHAR,

[please use gerrit instead, 2015/08/11 22:54:10]

IS_SKIPPABLE

Putting "_CHAR" in there might confuse people into thinking that only a single
character was encountered. Although the implementation might behave this way,
let's not expose too much implementation details in the API for the next layer
up the stack.

[Julius, 2015/08/12 20:25:59]

Done.

+    // The end of the text that the iterator is going over.
+    IS_END_OF_TEXT
+  };
+
   SpellcheckWordIterator();
   ~SpellcheckWordIterator();
 
@@ -130,19 +141,20 @@ class SpellcheckWordIterator {
   // without calling Initialize().
   bool SetText(const base::char16* text, size_t length);
 
-  // Retrieves a word (or a contraction), stores its copy to 'word_string', and
-  // stores the position and the length for input word to 'word_start'. Since
-  // this function normalizes the output word, the length of 'word_string' may
-  // be different from the 'word_length'. Therefore, when we call functions that
-  // changes the input text, such as string16::replace(), we need to use
-  // 'word_start' and 'word_length' as listed in the following snippet.
-  //
-  //   while(iterator.GetNextWord(&word, &offset, &length))

[please use gerrit instead, 2015/08/11 22:54:10]

Keep this example in the comment please.

[Julius, 2015/08/12 20:25:59]

Done.

-  //     text.replace(offset, length, word);
-  //
-  bool GetNextWord(base::string16* word_string,
-                   int* word_start,
-                   int* word_length);
+  // Advances |iterator_| through |text_| and gets the current status of the
+  // word iterator within the |text|:
+  //  - Returns IS_WORD if the iterator just found the end of a sequence of word

[please use gerrit instead, 2015/08/11 22:54:10]

Put a newline before a bullet point to make the comment easier to read.

[Julius, 2015/08/12 20:25:59]

Done.

+  //    characters and it was able to normalize the sequence. This stores the
+  //    normalized string into |word_string| and stores the position and length
+  //    into |word_start| and |word_length| respectively. Keep in mind that
+  //    since this function normalizes the output word, the length of
+  //    |word_string| may be different from the |word_length|.
+  //  - Returns IS_SKIPPABLE_CHAR if the iterator just found a character that
+  //    the iterator can skip past such as punctuation, whitespace, and
+  //    characters from another character set.
+  //  - Returns IS_END_OF_TEXT if the iterator has reached the end of |text_|.
+  SpellcheckWordIterator::WordIteratorStatus
+  GetNextWord(base::string16* word_string, int* word_start, int* word_length);
 
   // Releases all the resources attached to this object.
   void Reset();