Two bespoke test cases from ICU4X

unicode-org · Jan 18, 2024 · c395de9 · c395de9
1 parent 77e3ce6
commit c395de9
Showing 1 changed file with 7 additions and 0 deletions.
diff --git a/unicodetools/src/main/java/org/unicode/text/UCD/GenerateData.java b/unicodetools/src/main/java/org/unicode/text/UCD/GenerateData.java
@@ -1051,6 +1051,13 @@ static final String comma(String s) {
         "\u0592\u05B7\u05BC\u05A5\u05B0\u05C0\u05C4\u05AD",
         "\u1100\uAC00\u11A8",
         "\u1100\uAC00\u11A8\u11A8",
+        // Some implementations have an edge case when a character whose
+        // decomposition contains multiple starters and ends with a non-starter
+        // is followed by a non-starter of lower CCC.
+        // See https://github.com/unicode-org/unicodetools/issues/656
+        // and https://github.com/unicode-org/icu4x/pull/4530.
+        "\u01C4\u0323",
+        "\u0DDD\u0334",
     };
     /*
     static final void backwardsCompat(String directory, String filename, int[] list) throws IOException {