Merge pull request #404 from hamishknight/ban-empty-cc

Author: hamishknight

Sources/_RegexParser/Regex/Parse/Parse.swift

@@ -465,16 +465,6 @@ extension Parser {
     var members: Array<Member> = []
     try parseCCCMembers(into: &members)
 
-    // If we didn't parse any semantic members, we can eat a ']' character, as
-    // PCRE, Oniguruma, and ICU forbid empty character classes, and assume an
-    // initial ']' is literal.
-    if members.none(\.isSemantic) {
-      if let loc = source.tryEatWithLoc("]") {
-        members.append(.atom(.init(.char("]"), loc)))
-        try parseCCCMembers(into: &members)
-      }
-    }
-
     // If we have a binary set operator, parse it and the next members. Note
     // that this means we left associate for a chain of operators.
     // TODO: We may want to diagnose and require users to disambiguate, at least

Tests/RegexTests/MatchTests.swift

@@ -1282,7 +1282,6 @@ extension RegexTests {
     firstMatchTest(#"(?xx)[ \t]+"#, input: " \t \t", match: "\t")
 
     firstMatchTest("(?xx)[ a && ab ]+", input: " aaba ", match: "aa")
-    firstMatchTest("(?xx)[ ] a ]+", input: " a]]a ] ", match: "a]]a")
   }
 
   func testASCIIClasses() {

Tests/RegexTests/ParseTests.swift

@@ -528,23 +528,7 @@ extension RegexTests {
     ))
 
     parseTest("[-]", charClass("-"))
-
-    // Empty character classes are forbidden, therefore these are character
-    // classes containing literal ']'.
-    parseTest("[]]", charClass("]"))
-    parseTest("[]a]", charClass("]", "a"))
-    parseTest("(?x)[ ]]", concat(
-      changeMatchingOptions(matchingOptions(adding: .extended)),
-      charClass("]")
-    ))
-    parseTest("(?x)[ ]  ]", concat(
-      changeMatchingOptions(matchingOptions(adding: .extended)),
-      charClass("]")
-    ))
-    parseTest("(?x)[ ] a ]", concat(
-      changeMatchingOptions(matchingOptions(adding: .extended)),
-      charClass("]", "a")
-    ))
+    parseTest(#"[\]]"#, charClass("]"))
 
     // These are metacharacters in certain contexts, but normal characters
     // otherwise.
@@ -2497,10 +2481,15 @@ extension RegexTests {
 
     diagnosticTest("[a", .expected("]"))
 
-    // The first ']' of a custom character class is literal, so these are
-    // missing the closing bracket.
-    diagnosticTest("[]", .expected("]"))
-    diagnosticTest("(?x)[  ]", .expected("]"))
+    // Character classes may not be empty.
+    diagnosticTest("[]", .expectedCustomCharacterClassMembers)
+    diagnosticTest("[]]", .expectedCustomCharacterClassMembers)
+    diagnosticTest("[]a]", .expectedCustomCharacterClassMembers)
+    diagnosticTest("(?x)[  ]", .expectedCustomCharacterClassMembers)
+    diagnosticTest("(?x)[ ]  ]", .expectedCustomCharacterClassMembers)
+    diagnosticTest("(?x)[ ] a ]", .expectedCustomCharacterClassMembers)
+    diagnosticTest("(?xx)[ ] a ]+", .expectedCustomCharacterClassMembers)
+    diagnosticTest("(?x)[ ]]", .expectedCustomCharacterClassMembers)
 
     diagnosticTest("[&&]", .expectedCustomCharacterClassMembers)
     diagnosticTest("[a&&]", .expectedCustomCharacterClassMembers)