From 2f2fd4585e66d3b99a4ba5906bca84350930bd23 Mon Sep 17 00:00:00 2001
From: Sam Harwell <sam@tunnelvisionlabs.com>
Date: Sun, 26 May 2013 13:43:56 -0500
Subject: [PATCH] Improved error message for unterminated string literals
 (fixes #243)

---
 tool/src/org/antlr/v4/parse/ANTLRLexer.g         | 15 +++++++++++----
 tool/src/org/antlr/v4/tool/ErrorType.java        | 11 +++++++++++
 .../org/antlr/v4/test/TestToolSyntaxErrors.java  | 16 ++++++++++++++++
 3 files changed, 38 insertions(+), 4 deletions(-)

diff --git a/tool/src/org/antlr/v4/parse/ANTLRLexer.g b/tool/src/org/antlr/v4/parse/ANTLRLexer.g
index 25690fc70..fc46ca517 100644
--- a/tool/src/org/antlr/v4/parse/ANTLRLexer.g
+++ b/tool/src/org/antlr/v4/parse/ANTLRLexer.g
@@ -618,10 +618,17 @@ SRC : 'src' WSCHARS+ file=ACTION_STRING_LITERAL WSCHARS+ line=INT
 // may contain unicode escape sequences of the form \uxxxx, where x
 // is a valid hexadecimal number (as per Java basically).
 STRING_LITERAL
-@init {
-   int len = 0;
-}
-    :  '\'' ( ( ESC_SEQ | ~('\\'|'\''|'\r'|'\n') ) {len++;} )* '\''
+    :  '\'' ( ( ESC_SEQ | ~('\\'|'\''|'\r'|'\n') ) )*
+       (    '\''
+       |    // Unterminated string literal
+            {
+            Token t = new CommonToken(input, state.type, state.channel, state.tokenStartCharIndex, getCharIndex()-1);
+            t.setLine(state.tokenStartLine);
+            t.setText(state.text);
+            t.setCharPositionInLine(state.tokenStartCharPositionInLine);
+            grammarError(ErrorType.UNTERMINATED_STRING_LITERAL, t);
+            }
+       )
     ;
 
 // A valid hex digit specification
diff --git a/tool/src/org/antlr/v4/tool/ErrorType.java b/tool/src/org/antlr/v4/tool/ErrorType.java
index 605b83c31..b8b0cee3e 100644
--- a/tool/src/org/antlr/v4/tool/ErrorType.java
+++ b/tool/src/org/antlr/v4/tool/ErrorType.java
@@ -207,6 +207,17 @@ public enum ErrorType {
 	 * </pre>
 	 */
 	UNWANTED_LEXER_COMMAND_ARGUMENT(151, "lexer command '<arg>' does not take any arguments", ErrorSeverity.ERROR),
+	/**
+	 * The parser contains an unterminated string literal.
+	 * <p/>
+	 * The following rule produces this error.
+	 *
+	 * <pre>
+	 * x : 'x'; // ok
+	 * y : 'y;  // error 152
+	 * </pre>
+	 */
+	UNTERMINATED_STRING_LITERAL(152, "unterminated string literal", ErrorSeverity.ERROR),
 
 	// Backward incompatibility errors
 	V3_TREE_GRAMMAR(200, "tree grammars are not supported in ANTLR 4", ErrorSeverity.ERROR),
diff --git a/tool/test/org/antlr/v4/test/TestToolSyntaxErrors.java b/tool/test/org/antlr/v4/test/TestToolSyntaxErrors.java
index a4b42d744..13b589b5d 100644
--- a/tool/test/org/antlr/v4/test/TestToolSyntaxErrors.java
+++ b/tool/test/org/antlr/v4/test/TestToolSyntaxErrors.java
@@ -172,6 +172,22 @@ public class TestToolSyntaxErrors extends BaseTest {
 		super.testErrors(pair, true);
 	}
 
+	/**
+	 * This is a regression test for antlr/antlr4#243
+	 * "Generate a good message for unterminated strings"
+	 * https://github.com/antlr/antlr4/issues/243
+	 */
+	@Test public void testUnterminatedStringLiteral() {
+		String[] pair = new String[] {
+			"grammar A;\n" +
+			"a : 'x\n" +
+			"  ;\n",
+
+			"error(" + ErrorType.UNTERMINATED_STRING_LITERAL.code + "): A.g4:2:4: unterminated string literal\n"
+		};
+		super.testErrors(pair, true);
+	}
+
 	/**
 	 * This is a regression test for antlr/antlr4#262
 	 * "Parser Rule Name Starting With an Underscore"