Merge pull request #40607 from LakshanWeerasinghe/fix-#40499

Fix compiler crash for interpolations inside regex char sets
ballerina-platform · Jul 6, 2023 · 3775e22 · 3775e22
2 parents 531f246 + 93b4834
commit 3775e22
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 0 deletions.
diff --git a/.../ballerina-parser/src/main/java/io/ballerina/compiler/internal/parser/BallerinaLexer.java b/.../ballerina-parser/src/main/java/io/ballerina/compiler/internal/parser/BallerinaLexer.java
@@ -1520,6 +1520,11 @@ private STToken readRegExpTemplateToken() {
                     shouldProcessInterpolations = false;
                 }
                 while (!reader.isEOF()) {
+                    if (shouldProcessInterpolations && this.reader.peek() == LexerTerminals.BACKSLASH 
+                            && this.reader.peek(1) == LexerTerminals.OPEN_BRACKET) {
+                        // Escaped open brackets are not considered as the start of a no interpolation context.
+                        reader.advance();
+                    }
                     reader.advance();
                     nextChar = this.reader.peek();
                     switch (nextChar) {
@@ -1536,6 +1541,11 @@ private STToken readRegExpTemplateToken() {
                         case LexerTerminals.CLOSE_BRACKET:
                             shouldProcessInterpolations = true;
                             continue;
+                        case LexerTerminals.BACKSLASH:
+                            if (!shouldProcessInterpolations && this.reader.peek(1) == LexerTerminals.CLOSE_BRACKET) {
+                                reader.advance();
+                            }
+                            continue;    
                         default:
                             continue;
                     }

diff --git a/...er/ballerina-parser/src/main/java/io/ballerina/compiler/internal/parser/RegExpParser.java b/...er/ballerina-parser/src/main/java/io/ballerina/compiler/internal/parser/RegExpParser.java
@@ -427,6 +427,9 @@ private STNode parseCharSetAtom(STToken nextToken, STNode prevNode) {
                     return STNodeFactory.createToken(SyntaxKind.ESCAPED_MINUS_TOKEN, minusToken.leadingMinutiae(),
                             minusToken.trailingMinutiae());
                 }
+                if (token.kind == SyntaxKind.CLOSE_BRACKET_TOKEN) {
+                    this.tokenReader.startMode(ParserMode.RE_CHAR_CLASS);
+                }
                 return parseReEscape();
             default:
                 STNode consumedToken = consume();

diff --git a/tests/jballerina-unit-test/src/test/resources/test-src/types/regexp/regexp_value_test.bal b/tests/jballerina-unit-test/src/test/resources/test-src/types/regexp/regexp_value_test.bal
@@ -81,6 +81,15 @@ function testRegExpValueWithCharacterClass() {
 
     string:RegExp x7 = re `[\p{sc=Latin}\p{gc=Lu}\p{Lt}\tA\)]??`;
     assertEquality("[\\p{sc=Latin}\\p{gc=Lu}\\p{Lt}\\tA\\)]??", x7.toString());
+
+    string:RegExp x8 = re `abc[a-z\]A-Z${"abc"}]`;
+    assertEquality("abc[a-z\\]A-Z${\"abc\"}]", x8.toString());
+
+    string:RegExp x9 = re `abc\[a-z\]A-Z${"abc"}`;
+    assertEquality("abc\\[a-z\\]A-Zabc", x9.toString());
+
+    string:RegExp x10 = re `\[${"a"}\]`;
+    assertEquality("\\[a\\]", x10.toString());
 }
 
 function testRegExpValueWithCharacterClass2() {