summaryrefslogtreecommitdiffstats
path: root/base/i18n/break_iterator_unittest.cc
diff options
context:
space:
mode:
Diffstat (limited to 'base/i18n/break_iterator_unittest.cc')
-rw-r--r--base/i18n/break_iterator_unittest.cc117
1 files changed, 117 insertions, 0 deletions
diff --git a/base/i18n/break_iterator_unittest.cc b/base/i18n/break_iterator_unittest.cc
new file mode 100644
index 0000000..8add918
--- /dev/null
+++ b/base/i18n/break_iterator_unittest.cc
@@ -0,0 +1,117 @@
+// 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.
+
+#include "base/i18n/break_iterator.h"
+
+#include "base/string_piece.h"
+#include "base/string_util.h"
+#include "base/utf_string_conversions.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+TEST(BreakIteratorTest, BreakWord) {
+ string16 space(UTF8ToUTF16(" "));
+
+ string16 str(UTF8ToUTF16(" foo bar! \npouet boom"));
+ base::BreakIterator iter(&str, base::BreakIterator::BREAK_WORD);
+ ASSERT_TRUE(iter.Init());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(space, iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("foo"), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(space, iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("bar"), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("!"), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(space, iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("\n"), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("pouet"), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(space, iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("boom"), iter.GetString());
+ EXPECT_FALSE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+}
+
+TEST(BreakIteratorTest, BreakSpace) {
+ string16 str(UTF8ToUTF16(" foo bar! \npouet boom"));
+ base::BreakIterator iter(&str, base::BreakIterator::BREAK_SPACE);
+ ASSERT_TRUE(iter.Init());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16(" "), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("foo "), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("bar! \n"), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("pouet "), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("boom"), iter.GetString());
+ EXPECT_FALSE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+}
+
+TEST(BreakIteratorTest, BreakWide16) {
+ // "Παγκόσμιος Ιστός"
+ const string16 str(WideToUTF16(
+ L"\x03a0\x03b1\x03b3\x03ba\x03cc\x03c3\x03bc\x03b9"
+ L"\x03bf\x03c2\x0020\x0399\x03c3\x03c4\x03cc\x03c2"));
+ const string16 word1(str.substr(0, 10));
+ const string16 word2(str.substr(11, 5));
+ base::BreakIterator iter(&str, base::BreakIterator::BREAK_WORD);
+ ASSERT_TRUE(iter.Init());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(word1, iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16(" "), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(word2, iter.GetString());
+ EXPECT_FALSE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+}
+
+TEST(BreakIteratorTest, BreakWide32) {
+ // U+1D49C MATHEMATICAL SCRIPT CAPITAL A
+ const char* very_wide_char = "\xF0\x9D\x92\x9C";
+ const string16 str(
+ UTF8ToUTF16(StringPrintf("%s a", very_wide_char)));
+ const string16 very_wide_word(str.substr(0, 2));
+
+ base::BreakIterator iter(&str, base::BreakIterator::BREAK_WORD);
+ ASSERT_TRUE(iter.Init());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(very_wide_word, iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16(" "), iter.GetString());
+ EXPECT_TRUE(iter.Advance());
+ EXPECT_TRUE(iter.IsWord());
+ EXPECT_EQ(UTF8ToUTF16("a"), iter.GetString());
+ EXPECT_FALSE(iter.Advance());
+ EXPECT_FALSE(iter.IsWord());
+}