Updates SpellcheckWordIterator::GetNextWord to return an enum.

chrome/renderer/spellchecker/spellcheck_worditerator_unittest.cc

 // Copyright (c) 2012 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 <string>
 #include <vector>
 
 #include "base/format_macros.h"
 #include "base/i18n/break_iterator.h"
 #include "base/strings/string_split.h"
@@ skipping 13 matching lines @@
 };
 
 base::string16 GetRulesForLanguage(const std::string& language) {
   SpellcheckCharAttribute attribute;
   attribute.SetDefaultLanguage(language);
   return attribute.GetRuleSet(true);
 }
 
 }  // namespace
 
-// Tests whether or not our SpellcheckWordIterator can extract only words used
-// by the specified language from a multi-language text.
+// Tests whether or not our SpellcheckWordIterator can extract words used by the
+// specified language from a multi-language text.
 TEST(SpellcheckWordIteratorTest, SplitWord) {
   // An input text. This text includes words of several languages. (Some words
   // are not separated with whitespace characters.) Our SpellcheckWordIterator
-  // should extract only the words used by the specified language from this text
-  // and normalize them so our spell-checker can check their spellings.
+  // should extract the words used by the specified language from this text and
+  // normalize them so our spell-checker can check their spellings. If
+  // characters are found that are not from the specified language we skip them.

[please use gerrit instead, 2015/08/13 00:13:28]

"we" can be confusing in code. Specify the subject. I think it's "the test" that
skips the characters.

[Julius, 2015/08/13 01:32:04]

Done.

   const wchar_t kTestText[] =
       // Graphic characters
       L"!@#$%^&*()"
       // Latin (including a contraction character and a ligature).
       L"hello:hello a\xFB03x"
       // Greek
       L"\x03B3\x03B5\x03B9\x03AC\x0020\x03C3\x03BF\x03C5"
       // Cyrillic
       L"\x0437\x0434\x0440\x0430\x0432\x0441\x0442\x0432"
       L"\x0443\x0439\x0442\x0435"
@@ skipping 83 matching lines @@
                                     kTestCases[i].allow_contraction));
     EXPECT_TRUE(iterator.SetText(input.c_str(), input.length()));
 
     std::vector<base::string16> expected_words = base::SplitString(
         base::WideToUTF16(kTestCases[i].expected_words),
         base::string16(1, ' '), base::TRIM_WHITESPACE, base::SPLIT_WANT_ALL);
 
     base::string16 actual_word;
     int actual_start, actual_end;
     size_t index = 0;
-    while (iterator.GetNextWord(&actual_word, &actual_start, &actual_end)) {
+    for (SpellcheckWordIterator::WordIteratorStatus status =
+             iterator.GetNextWord(&actual_word, &actual_start, &actual_end);
+         status != SpellcheckWordIterator::IS_END_OF_TEXT;
+         status =
+             iterator.GetNextWord(&actual_word, &actual_start, &actual_end)) {
+      if (status == SpellcheckWordIterator::WordIteratorStatus::IS_SKIPPABLE)
+        continue;
+
       EXPECT_TRUE(index < expected_words.size());
       if (index < expected_words.size())
         EXPECT_EQ(expected_words[index], actual_word);
       ++index;
     }
   }
 }
 
 // Tests whether our SpellcheckWordIterator extracts an empty word without
 // getting stuck in an infinite loop when inputting a Khmer text. (This is a
 // regression test for Issue 46278.)
 TEST(SpellcheckWordIteratorTest, RuleSetConsistency) {
   SpellcheckCharAttribute attributes;
   attributes.SetDefaultLanguage("en-US");
 
   const wchar_t kTestText[] = L"\x1791\x17c1\x002e";
   base::string16 input(base::WideToUTF16(kTestText));
 
   SpellcheckWordIterator iterator;
   EXPECT_TRUE(iterator.Initialize(&attributes, true));
   EXPECT_TRUE(iterator.SetText(input.c_str(), input.length()));
 
   // When SpellcheckWordIterator uses an inconsistent ICU ruleset, the following
-  // iterator.GetNextWord() call gets stuck in an infinite loop. Therefore, this
+  // iterator.GetNextWord() calls get stuck in an infinite loop. Therefore, this
   // test succeeds if this call returns without timeouts.
   base::string16 actual_word;
   int actual_start, actual_end;
-  EXPECT_FALSE(iterator.GetNextWord(&actual_word, &actual_start, &actual_end));
+  SpellcheckWordIterator::WordIteratorStatus status;
+  for (status = iterator.GetNextWord(&actual_word, &actual_start, &actual_end);
+       status == SpellcheckWordIterator::IS_SKIPPABLE;
+       status =
+           iterator.GetNextWord(&actual_word, &actual_start, &actual_end)) {
+    continue;
+  }
+
+  EXPECT_EQ(SpellcheckWordIterator::WordIteratorStatus::IS_END_OF_TEXT, status);
   EXPECT_EQ(0, actual_start);
   EXPECT_EQ(0, actual_end);
 }
 
 // Vertify our SpellcheckWordIterator can treat ASCII numbers as word characters
 // on LTR languages. On the other hand, it should not treat ASCII numbers as
 // word characters on RTL languages because they change the text direction from
 // RTL to LTR.
 TEST(SpellcheckWordIteratorTest, TreatNumbersAsWordCharacters) {
   // A set of a language, a dummy word, and a text direction used in this test.
@@ skipping 40 matching lines @@
     SpellcheckCharAttribute attributes;
     attributes.SetDefaultLanguage(kTestCases[i].language);
 
     base::string16 input_word(base::WideToUTF16(kTestCases[i].text));
     SpellcheckWordIterator iterator;
     EXPECT_TRUE(iterator.Initialize(&attributes, true));
     EXPECT_TRUE(iterator.SetText(input_word.c_str(), input_word.length()));
 
     base::string16 actual_word;
     int actual_start, actual_end;
-    EXPECT_TRUE(iterator.GetNextWord(&actual_word, &actual_start, &actual_end));
+    SpellcheckWordIterator::WordIteratorStatus status;
+    for (status =
+             iterator.GetNextWord(&actual_word, &actual_start, &actual_end);
+         status == SpellcheckWordIterator::IS_SKIPPABLE;
+         status =
+             iterator.GetNextWord(&actual_word, &actual_start, &actual_end)) {
+      continue;
+    }
+
+    EXPECT_EQ(SpellcheckWordIterator::WordIteratorStatus::IS_WORD, status);
     if (kTestCases[i].left_to_right)
       EXPECT_EQ(input_word, actual_word);
     else
       EXPECT_NE(input_word, actual_word);
   }
 }
 
 // Vertify SpellcheckWordIterator treats typographical apostrophe as a part of
 // the word.
 TEST(SpellcheckWordIteratorTest, TypographicalApostropheIsPartOfWord) {
@@ skipping 27 matching lines @@
     SpellcheckCharAttribute attributes;
     attributes.SetDefaultLanguage(kTestCases[i].language);
 
     base::string16 input_word(base::WideToUTF16(kTestCases[i].word));
     SpellcheckWordIterator iterator;
     EXPECT_TRUE(iterator.Initialize(&attributes, true));
     EXPECT_TRUE(iterator.SetText(input_word.c_str(), input_word.length()));
 
     base::string16 actual_word;
     int actual_start, actual_end;
-    EXPECT_TRUE(iterator.GetNextWord(&actual_word, &actual_start, &actual_end));
+    SpellcheckWordIterator::WordIteratorStatus status;
+    for (status =
+             iterator.GetNextWord(&actual_word, &actual_start, &actual_end);
+         status == SpellcheckWordIterator::IS_SKIPPABLE;
+         iterator.GetNextWord(&actual_word, &actual_start, &actual_end)) {
+      continue;
+    }
+
+    EXPECT_EQ(SpellcheckWordIterator::WordIteratorStatus::IS_WORD, status);
     EXPECT_EQ(input_word, actual_word);
     EXPECT_EQ(0, actual_start);
     EXPECT_EQ(input_word.length(),
               static_cast<base::string16::size_type>(actual_end));
   }
 }
 
 TEST(SpellcheckWordIteratorTest, Initialization) {
   // Test initialization works when a default language is set.
   {
@@ skipping 167 matching lines @@
   EXPECT_EQ(base::UTF8ToUTF16("."), iter.GetString());
   EXPECT_EQ(iter.GetWordBreakStatus(), BreakIterator::IS_SKIPPABLE_WORD);
   EXPECT_TRUE(iter.Advance());
   EXPECT_EQ(base::UTF8ToUTF16(" "), iter.GetString());
   EXPECT_EQ(iter.GetWordBreakStatus(), BreakIterator::IS_SKIPPABLE_WORD);
   EXPECT_TRUE(iter.Advance());
   EXPECT_EQ(base::UTF8ToUTF16(","), iter.GetString());
   EXPECT_EQ(iter.GetWordBreakStatus(), BreakIterator::IS_SKIPPABLE_WORD);
   EXPECT_FALSE(iter.Advance());
 }