Do not allow line breaks before currency symbols

Implement the change proposed in UTC document L2/16-043R
(http://www.unicode.org/L2/L2016/16043r-line-break-pr-po.txt) to make
sure we do not break between letters and currency symbols.

Bug: 24959657
Change-Id: Ia29d0e5625f84870bd910d0c6e19036d17206704

2 files changed, 29 insertions, 0 deletions

diff --git a/libs/minikin/WordBreaker.cpp b/libs/minikin/WordBreaker.cpp
index 721c5bf..d420a6a 100644
--- a/libs/minikin/WordBreaker.cpp
+++ b/libs/minikin/WordBreaker.cpp
@@ -79,6 +79,18 @@ static bool isBreakValid(const uint16_t* buf, size_t bufEnd, size_t i) {
     uint32_t next_codepoint;
     size_t next_offset = i;
     U16_NEXT(buf, next_offset, bufEnd, next_codepoint);
+
+    // Proposed change to LB24 from http://www.unicode.org/L2/L2016/16043r-line-break-pr-po.txt
+    //(AL | HL) × (PR | PO)
+    int32_t lineBreak = u_getIntPropertyValue(codePoint, UCHAR_LINE_BREAK);
+    if (lineBreak == U_LB_ALPHABETIC || lineBreak == U_LB_HEBREW_LETTER) {
+        lineBreak = u_getIntPropertyValue(next_codepoint, UCHAR_LINE_BREAK);
+        if (lineBreak == U_LB_PREFIX_NUMERIC || lineBreak == U_LB_POSTFIX_NUMERIC) {
+            return false;
+        }
+    }
+
+    // Known emoji ZWJ sequences
     if (codePoint == CHAR_ZWJ) {
         // Possible emoji ZWJ sequence
         if (next_codepoint == 0x2764 ||       // HEAVY BLACK HEART
@@ -91,6 +103,7 @@ static bool isBreakValid(const uint16_t* buf, size_t bufEnd, size_t i) {
             return false;
         }
     }
+
     // Proposed Rule LB30b from http://www.unicode.org/L2/L2016/16011r3-break-prop-emoji.pdf
     // EB x EM
     if (isEmojiModifier(next_codepoint)) {
diff --git a/tests/WordBreakerTests.cpp b/tests/WordBreakerTests.cpp
index cb12722..480c57d 100644
--- a/tests/WordBreakerTests.cpp
+++ b/tests/WordBreakerTests.cpp
@@ -69,6 +69,22 @@ TEST_F(WordBreakerTest, softHyphen) {
     EXPECT_EQ(0, breaker.breakBadness());
 }
 
+TEST_F(WordBreakerTest, postfixAndPrefix) {
+    uint16_t buf[] = {'U', 'S', 0x00A2, ' ', 'J', 'P', 0x00A5}; // US¢ JP¥
+    WordBreaker breaker;
+    breaker.setLocale(icu::Locale::getEnglish());
+    breaker.setText(buf, NELEM(buf));
+    EXPECT_EQ(0, breaker.current());
+
+    EXPECT_EQ(4, breaker.next());  // after CENT SIGN
+    EXPECT_EQ(0, breaker.wordStart());  // "US¢"
+    EXPECT_EQ(3, breaker.wordEnd());
+
+    EXPECT_EQ((ssize_t)NELEM(buf), breaker.next());  // end of string
+    EXPECT_EQ(4, breaker.wordStart());  // "JP¥"
+    EXPECT_EQ((ssize_t)NELEM(buf), breaker.wordEnd());
+}
+
 TEST_F(WordBreakerTest, zwjEmojiSequences) {
     uint16_t buf[] = {
         // man + zwj + heart + zwj + man