From 3521a00dbbe2f79b80f917bb8fe5e90870e3f497 Mon Sep 17 00:00:00 2001
From: Sam Harwell <sam@tunnelvisionlabs.com>
Date: Mon, 14 Jan 2013 18:02:17 -0600
Subject: [PATCH] Add error 146: non-fragment lexer rule can match the empty
 string

---
 .../antlr/v4/analysis/AnalysisPipeline.java   | 29 +++++++++++++++++--
 tool/src/org/antlr/v4/tool/ErrorType.java     |  1 +
 .../org/antlr/v4/test/TestSymbolIssues.java   | 16 ++++++++++
 3 files changed, 43 insertions(+), 3 deletions(-)

diff --git a/tool/src/org/antlr/v4/analysis/AnalysisPipeline.java b/tool/src/org/antlr/v4/analysis/AnalysisPipeline.java
index b24f7c94a..37472b499 100644
--- a/tool/src/org/antlr/v4/analysis/AnalysisPipeline.java
+++ b/tool/src/org/antlr/v4/analysis/AnalysisPipeline.java
@@ -31,10 +31,14 @@
 package org.antlr.v4.analysis;
 
 import org.antlr.v4.misc.Utils;
+import org.antlr.v4.runtime.Token;
 import org.antlr.v4.runtime.atn.DecisionState;
 import org.antlr.v4.runtime.atn.LL1Analyzer;
 import org.antlr.v4.runtime.misc.IntervalSet;
+import org.antlr.v4.tool.ErrorType;
 import org.antlr.v4.tool.Grammar;
+import org.antlr.v4.tool.Rule;
+import org.antlr.v4.tool.ast.GrammarAST;
 
 import java.util.ArrayList;
 import java.util.Arrays;
@@ -52,11 +56,30 @@ public class AnalysisPipeline {
 		lr.check();
 		if ( !lr.listOfRecursiveCycles.isEmpty() ) return; // bail out
 
-		// BUILD DFA FOR EACH DECISION
-		if ( !g.isLexer() ) processParser();
+		if (g.isLexer()) {
+			processLexer();
+		} else {
+			// BUILD DFA FOR EACH DECISION
+			processParser();
+		}
 	}
 
-	void processParser() {
+	protected void processLexer() {
+		// make sure all non-fragment lexer rules must match at least one symbol
+		for (Rule rule : g.rules.values()) {
+			if (rule.isFragment()) {
+				continue;
+			}
+
+			LL1Analyzer analyzer = new LL1Analyzer(g.atn);
+			IntervalSet look = analyzer.LOOK(g.atn.ruleToStartState[rule.index], null);
+			if (look.contains(Token.EPSILON)) {
+				g.tool.errMgr.grammarError(ErrorType.EPSILON_TOKEN, g.fileName, ((GrammarAST)rule.ast.getChild(0)).getToken(), rule.name);
+			}
+		}
+	}
+
+	protected void processParser() {
 		g.decisionLOOK = new ArrayList<IntervalSet[]>(g.atn.getNumberOfDecisions()+1);
 		for (DecisionState s : g.atn.decisionToState) {
             g.tool.log("LL1", "\nDECISION "+s.decision+" in rule "+g.getRule(s.ruleIndex).name);
diff --git a/tool/src/org/antlr/v4/tool/ErrorType.java b/tool/src/org/antlr/v4/tool/ErrorType.java
index 13654332e..542be506b 100644
--- a/tool/src/org/antlr/v4/tool/ErrorType.java
+++ b/tool/src/org/antlr/v4/tool/ErrorType.java
@@ -128,6 +128,7 @@ public enum ErrorType {
 	LOCAL_CONFLICTS_WITH_RETVAL(143, "local '<arg>' conflicts with return value with same name", ErrorSeverity.ERROR),
 	INVALID_LITERAL_IN_LEXER_SET(144, "multi-character literals are not allowed in lexer sets: <arg>", ErrorSeverity.ERROR),
 	MODE_WITHOUT_RULES(145, "lexer mode '<arg>' must contain at least one non-fragment rule", ErrorSeverity.ERROR),
+	EPSILON_TOKEN(146, "non-fragment lexer rule '<arg>' can match the empty string", 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/TestSymbolIssues.java b/tool/test/org/antlr/v4/test/TestSymbolIssues.java
index 8c6367bda..d4abe1a72 100644
--- a/tool/test/org/antlr/v4/test/TestSymbolIssues.java
+++ b/tool/test/org/antlr/v4/test/TestSymbolIssues.java
@@ -152,4 +152,20 @@ public class TestSymbolIssues extends BaseTest {
 
 		testErrors(test, false);
 	}
+
+	@Test public void testEmptyLexerRuleDetection() throws Exception {
+		String[] test = {
+			"lexer grammar L;\n" +
+			"A : 'a';\n" +
+			"WS : [ \t]* -> skip;\n" +
+			"mode X;\n" +
+			"  B : C;\n" +
+			"  fragment C : A | (A C)?;",
+
+			"error(" + ErrorType.EPSILON_TOKEN.code + "): L.g4:3:0: non-fragment lexer rule 'WS' can match the empty string\n" +
+			"error(" + ErrorType.EPSILON_TOKEN.code + "): L.g4:5:2: non-fragment lexer rule 'B' can match the empty string\n"
+		};
+
+		testErrors(test, false);
+	}
 }