Fix emoji ZWJ and modifier sequence line breaking

third_party/WebKit/Source/platform/text/TextBreakIterator.cpp

Index: third_party/WebKit/Source/platform/text/TextBreakIterator.cpp
diff --git a/third_party/WebKit/Source/platform/text/TextBreakIterator.cpp b/third_party/WebKit/Source/platform/text/TextBreakIterator.cpp
index e1205bfd90dd2eecfc285b970d88b2f1dea0ecbf..8c84c5443e775dd4b161be44cbf75882223c9dad 100644
--- a/third_party/WebKit/Source/platform/text/TextBreakIterator.cpp
+++ b/third_party/WebKit/Source/platform/text/TextBreakIterator.cpp
@@ -22,6 +22,7 @@
 
 #include "platform/text/TextBreakIterator.h"
 
+#include "platform/fonts/Character.h"
 #include "wtf/ASCIICType.h"
 #include "wtf/StdLibExtras.h"
 #include "wtf/text/CharacterNames.h"
@@ -221,6 +222,56 @@ inline bool needsLineBreakIterator(UChar ch)
     return ch > asciiLineBreakTableLastChar && ch != noBreakSpaceCharacter;
 }
 
+// Customization for ICU line breaking behavior. This allows us to reject ICU
+// line break suggestions which would split an emoji sequence.
+// FIXME crbug.com/593260: Remove this customization once ICU implements this
+// natively.
+static bool isBreakValid(const UChar* buf, size_t length, size_t breakPos)
+{
+    UChar32 codepoint;
+    size_t prevOffset = breakPos;
+    U16_PREV(buf, 0, prevOffset, codepoint);
+    uint32_t nextCodepoint;
+    size_t nextOffset = breakPos;
+    U16_NEXT(buf, nextOffset, length, nextCodepoint);
+
+    // Possible Emoji ZWJ sequence
+    if (codepoint == zeroWidthJoinerCharacter) {
+        if (nextCodepoint == 0x2764 // HEAVY BLACK HEART
+            || nextCodepoint == 0x1F466 // BOY
+            || nextCodepoint == 0x1F467 // GIRL
+            || nextCodepoint == 0x1F468 // MAN
+            || nextCodepoint == 0x1F469 // WOMAN
+            || nextCodepoint == 0x1F48B // KISS MARK
+            || nextCodepoint == 0x1F5E8) // LEFT SPEECH BUBBLE
+        {
+            return false;
+        }
+    }
+
+    // Possible emoji modifier sequence
+    // Proposed Rule LB30b from http://www.unicode.org/L2/L2016/16011r3-break-prop-emoji.pdf
+    // EB x EM
+    if (Character::isModifier(nextCodepoint)) {
+        if (codepoint == variationSelector16Character && prevOffset > 0) {
+            // Skip over emoji variation selector.
+            U16_PREV(buf, 0, prevOffset, codepoint);
+        }
+        if (Character::isEmojiModifierBase(codepoint)) {
+            return false;
+        }
+    }
+    return true;
+}
+
+// Trivial implementation to match possible template paramters in
+// nextBreakablePosition. There are no emoji sequences in 8bit strings, so we
+// accept all break opportunities.
+static bool isBreakValid(const LChar*, size_t, size_t)
+{
+    return true;
+}
+
 template<typename CharacterType, LineBreakType lineBreakType>
 static inline int nextBreakablePosition(LazyLineBreakIterator& lazyBreakIterator, const CharacterType* str, unsigned length, int pos)
 {
@@ -260,8 +311,9 @@ static inline int nextBreakablePosition(LazyLineBreakIterator& lazyBreakIterator
                     }
                 }
             }
-            if (i == nextBreak && !isBreakableSpace(lastCh))
+            if (i == nextBreak && !isBreakableSpace(lastCh) && isBreakValid(str, length, i)) {
                 return i;
+            }
         }
 
         lastLastCh = lastCh;
@@ -307,7 +359,7 @@ static inline int nextBreakablePositionKeepAllInternal(LazyLineBreakIterator& la
                     }
                 }
             }
-            if (i == nextBreak && !isBreakableSpace(lastCh))
+            if (i == nextBreak && !isBreakableSpace(lastCh) && isBreakValid(str, length, i))
                 return i;
         }