Merge pull request #20599 from milseman/emoji_tag

[Character] Permit tagged emoji Character literals

Author: milseman

include/swift/Basic/Unicode.h

@@ -54,17 +54,6 @@ extern const uint16_t ExtendedGraphemeClusterNoBoundaryRulesMatrix[];
 /// point.
 GraphemeClusterBreakProperty getGraphemeClusterBreakProperty(uint32_t C);
 
-/// Returns true if there is always an extended grapheme cluster boundary
-/// after a code point with a given property value.  Use only for optimization,
-/// to skip calculating Grapheme_Cluster_Break property for the second code
-/// point.
-static inline bool
-isExtendedGraphemeClusterBoundaryAfter(GraphemeClusterBreakProperty GCB1) {
-  auto RuleRow =
-      ExtendedGraphemeClusterNoBoundaryRulesMatrix[static_cast<unsigned>(GCB1)];
-  return RuleRow == 0;
-}
-
 /// Determine if there is an extended grapheme cluster boundary between code
 /// points with given Grapheme_Cluster_Break property values.
 static inline bool

lib/Basic/Unicode.cpp

@@ -27,7 +27,22 @@ using namespace swift;
 // break between them. That is, whether we're overriding the behavior of the
 // hard coded Unicode 8 rules surrounding ZWJ and emoji modifiers.
 static inline bool graphemeBreakOverride(llvm::UTF32 lhs, llvm::UTF32 rhs) {
-  return lhs == 0x200D || (rhs >= 0x1F3FB && rhs <= 0x1F3FF);
+  // Assume ZWJ sequences produce new emoji
+  if (lhs == 0x200D) {
+    return true;
+  }
+
+  // Permit continuing regional indicators
+  if (rhs >= 0x1F3FB && rhs <= 0x1F3FF) {
+    return true;
+  }
+
+  // Permit emoji tag sequences
+  if (rhs >= 0xE0020 && rhs <= 0xE007F) {
+    return true;
+  }
+
+  return false;
 }
 
 StringRef swift::unicode::extractFirstExtendedGraphemeCluster(StringRef S) {
@@ -52,9 +67,6 @@ StringRef swift::unicode::extractFirstExtendedGraphemeCluster(StringRef S) {
 
   GraphemeClusterBreakProperty GCBForC0 = getGraphemeClusterBreakProperty(C[0]);
   while (true) {
-    if (isExtendedGraphemeClusterBoundaryAfter(GCBForC0))
-      return S.slice(0, SourceNext - SourceStart);
-
     size_t C1Offset = SourceNext - SourceStart;
     ConvertUTF8toUTF32(&SourceNext, SourceStart + S.size(), &TargetStart, C + 2,
                        llvm::lenientConversion);

test/Parse/enum.swift

@@ -197,6 +197,7 @@ enum RawTypeWithCharacterValues_Correct : Character {
   case First = "😅" // ok
   case Second = "👩‍👩‍👧‍👦" // ok
   case Third = "👋🏽" // ok
+  case Fourth = "\u{1F3F4}\u{E0067}\u{E0062}\u{E0065}\u{E006E}\u{E0067}\u{E007F}" // ok
 }
 
 enum RawTypeWithCharacterValues_Error1 : Character { // expected-error {{'RawTypeWithCharacterValues_Error1' declares raw type 'Character', but does not conform to RawRepresentable and conformance could not be synthesized}}