Change CachingWordShapeIterator to delimit CJK characters

third_party/WebKit/Source/platform/fonts/shaping/CachingWordShaperTest.cpp

 // Copyright (c) 2014 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 "config.h"
 #include "platform/fonts/shaping/CachingWordShaper.h"
 
 #include "platform/fonts/FontCache.h"
 #include "platform/fonts/GlyphBuffer.h"
 #include "platform/fonts/shaping/CachingWordShapeIterator.h"
@@ skipping 162 matching lines @@
     ASSERT_GT(shaper.width(&font, textRun, nullptr, &glyphBounds), 0);
 
     GlyphBuffer glyphBuffer;
     shaper.fillGlyphBuffer(&font, textRun, fallbackFonts, &glyphBuffer, 0, 8);
 
     FloatPoint point;
     int height = 16;
     shaper.selectionRect(&font, textRun, point, height, 0, 8);
 }
 
+TEST_F(CachingWordShaperTest, SegmentCJKByCharacter)
+{
+    const UChar str[] = {
+        0x56FD, 0x56FD, // CJK Unified Ideograph
+        'a', 'b',
+        0x56FD, // CJK Unified Ideograph
+        'x', 'y', 'z',
+        0x3042, // HIRAGANA LETTER A
+        0x56FD, // CJK Unified Ideograph
+        0x0
+    };
+    TextRun textRun(str, 10);
+
+    RefPtr<ShapeResult> wordResult;
+    CachingWordShapeIterator iterator(cache.get(), textRun, &font);
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(2u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(3u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_FALSE(iterator.next(&wordResult));
+}
+
+TEST_F(CachingWordShaperTest, SegmentCJKAndCommon)
+{
+    const UChar str[] = {
+        'a', 'b',
+        0xFF08, // FULLWIDTH LEFT PARENTHESIS (script=common)
+        0x56FD, // CJK Unified Ideograph
+        0x56FD, // CJK Unified Ideograph
+        0x56FD, // CJK Unified Ideograph
+        0x3002, // IDEOGRAPHIC FULL STOP (script=common)
+        0x0
+    };
+    TextRun textRun(str, 7);
+
+    RefPtr<ShapeResult> wordResult;
+    CachingWordShapeIterator iterator(cache.get(), textRun, &font);
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(2u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(2u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(2u, wordResult->numCharacters());
+
+    ASSERT_FALSE(iterator.next(&wordResult));
+}
+
+TEST_F(CachingWordShaperTest, SegmentCJKAndInherit)
+{
+    const UChar str[] = {
+        0x304B, // HIRAGANA LETTER KA
+        0x304B, // HIRAGANA LETTER KA
+        0x3009, // COMBINING KATAKANA-HIRAGANA VOICED SOUND MARK
+        0x304B, // HIRAGANA LETTER KA
+        0x0
+    };
+    TextRun textRun(str, 4);
+
+    RefPtr<ShapeResult> wordResult;
+    CachingWordShapeIterator iterator(cache.get(), textRun, &font);
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(2u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_FALSE(iterator.next(&wordResult));
+}
+
+TEST_F(CachingWordShaperTest, SegmentCJKAndNonCJKCommon)
+{
+    const UChar str[] = {
+        0x56FD, // CJK Unified Ideograph
+        ' ',
+        0x0
+    };
+    TextRun textRun(str, 2);
+
+    RefPtr<ShapeResult> wordResult;
+    CachingWordShapeIterator iterator(cache.get(), textRun, &font);
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_FALSE(iterator.next(&wordResult));
+}
+
+TEST_F(CachingWordShaperTest, SegmentCJKCommonAndNonCJK)

[drott, 2015/12/21 14:57:41]

Could you add one test with only COMMON, no actual script values? IIUC we would
expect to use the full length of the run then.

[kojii, 2015/12/21 15:34:53]

Done.

+{
+    const UChar str[] = {
+        0xFF08, // FULLWIDTH LEFT PARENTHESIS (script=common)
+        'a', 'b',
+        0x0
+    };
+    TextRun textRun(str, 3);
+
+    RefPtr<ShapeResult> wordResult;
+    CachingWordShapeIterator iterator(cache.get(), textRun, &font);
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(1u, wordResult->numCharacters());
+
+    ASSERT_TRUE(iterator.next(&wordResult));
+    EXPECT_EQ(2u, wordResult->numCharacters());
+
+    ASSERT_FALSE(iterator.next(&wordResult));
+}
+
 TEST_F(CachingWordShaperTest, TextOrientationFallbackShouldNotInFallbackList)
 {
     const UChar str[] = {
         'A', // code point for verticalRightOrientationFontData()
         // Ideally we'd like to test uprightOrientationFontData() too
         // using code point such as U+3042, but it'd fallback to system
         // fonts as the glyph is missing.
         0x0
     };
     TextRun textRun(str, 1);
@@ skipping 23 matching lines @@
     float periodsAndSpacesWidth = shaper.width(&font, periodsAndSpaces, nullptr, &periodsAndSpacesGlyphBounds);
 
     // The total width of periods and spaces should be longer than the width of periods alone.
     ASSERT_GT(periodsAndSpacesWidth, periodsWidth);
 
     // The glyph bounds of periods and spaces should be longer than the glyph bounds of periods alone.
     ASSERT_GT(periodsAndSpacesGlyphBounds.width(), periodsGlyphBounds.width());
 }
 
 } // namespace blink