From cad14436abff83a5025a661349b97c4f90f91a01 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Sylvain=20Hall=C3=A9?= <shalle@acm.org>
Date: Wed, 5 Aug 2015 11:57:29 -0400
Subject: [PATCH] Fixed parsing of epsilon character in an input file. It turns
 out there are two different Unicode symbols for lowercase epsilon.

---
 .../src/ca/uqac/lif/bullwinkle/BnfParser.java |  2 +-
 .../src/ca/uqac/lif/bullwinkle/BnfRule.java   |  3 +-
 .../ca/uqac/lif/bullwinkle/GrammarTests.java  | 40 ++++++++++++++++++-
 build.xml                                     |  1 +
 config.xml                                    |  2 +-
 5 files changed, 43 insertions(+), 5 deletions(-)

diff --git a/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfParser.java b/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfParser.java
index 11c11a5..76e701f 100644
--- a/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfParser.java
+++ b/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfParser.java
@@ -307,7 +307,7 @@ public void setStartRule(final NonTerminalToken token)
 				if (!alt_it.hasNext())
 				{
 					// We succeeded in parsing the complete string: done
-					if (level > 0 || n_input.length() == 0)
+					if (level > 0 || (level == 0 && n_input.toString().trim().length() == 0))
 					{
 						break;
 					}
diff --git a/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfRule.java b/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfRule.java
index f621f42..1566479 100644
--- a/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfRule.java
+++ b/Source/Parsing/src/ca/uqac/lif/bullwinkle/BnfRule.java
@@ -96,8 +96,9 @@ public static BnfRule parseRule(String input) throws BnfRule.InvalidRuleExceptio
             Token to_add = new NonTerminalToken(trimmed_word);
             alternative_to_add.add(to_add);
           }
-          else if (trimmed_word.compareTo("ε") == 0)
+          else if (trimmed_word.compareTo("\uCEB5") == 0 || trimmed_word.compareTo("\u03B5") == 0)
           {
+          	// There are two "lowercase epsilon" code points in Unicode; check for both
             Token to_add = new EpsilonTerminalToken();
             alternative_to_add.add(to_add);
           }
diff --git a/Source/ParsingTest/src/ca/uqac/lif/bullwinkle/GrammarTests.java b/Source/ParsingTest/src/ca/uqac/lif/bullwinkle/GrammarTests.java
index ec55f58..d7a0654 100644
--- a/Source/ParsingTest/src/ca/uqac/lif/bullwinkle/GrammarTests.java
+++ b/Source/ParsingTest/src/ca/uqac/lif/bullwinkle/GrammarTests.java
@@ -77,7 +77,7 @@ public void parseGrammar2a()
   public void parseGrammar2b()
   {
     String expression = "SELECT a FROM (SELECT b FROM t)";
-    ParseNode node = parseIt("data/Grammar-1.bnf", "<S>", expression, true);
+    ParseNode node = parseIt("data/Grammar-1.bnf", "<S>", expression, false);
     int size = node.getSize();
     int expected_size = 19;
     if (size != expected_size)
@@ -120,6 +120,13 @@ public void parseGrammar4()
     assertEquals("Incorrect parse tree size for expression " + expression, 5, node.getSize());
   }
   
+  @Test
+  public void parseGrammar5()
+  {
+    String expression = "a WHERE";
+    parseItNot("data/Grammar-10.bnf", "<S>", expression, false);
+  }
+  
   @Test
   public void parseGrammarLtlFo1()
   {
@@ -137,7 +144,7 @@ public void parseGrammarLtlFo1()
   public void parseGrammarWithEpsilon1()
   {
     String expression = "hello hello";
-    ParseNode node = parseIt("data/Grammar-3.bnf", "<S>", expression, false);
+    ParseNode node = parseIt("data/Grammar-3.bnf", "<S>", expression, true);
     int size = node.getSize();
     int expected_size = 6;
     if (size != expected_size)
@@ -218,6 +225,13 @@ private ParseNode parseIt(String grammar_filename, String start_symbol, String e
     return node;
   }
   
+  private void parseItNot(String grammar_filename, String start_symbol, String expression, boolean debug_mode)
+  {
+    BnfParser parser = readGrammar(grammar_filename, start_symbol, debug_mode);
+    shouldNotParse(expression, parser);
+  }
+
+  
   private ParseNode shouldParseAndNotNull(final String expression, final BnfParser parser)
   {
     ParseNode node = shouldParse(expression, parser);
@@ -248,6 +262,28 @@ private ParseNode shouldParse(final String expression, final BnfParser parser)
     }
     return node;
   }
+  
+  /**
+   * Attempts to parse an expression with the given parser, and
+   * expects the parsing not to raise an exception or to return null.
+   * @param expression The expression to parse
+   * @param parser The parser to use
+   */
+  private void shouldNotParse(final String expression, final BnfParser parser)
+  {
+    ParseNode node = null;
+    try
+    {
+      node = parser.parse(expression);      
+    }
+    catch (BnfParser.ParseException e)
+    {
+    }
+    if (node != null)
+    {
+    	fail("The parsing of " + expression + " should have failed");
+    }
+  }
 
   public static BnfParser readGrammar(final String filename, final String start_rule, boolean debug_mode)
   {
diff --git a/build.xml b/build.xml
index e637b84..011669d 100644
--- a/build.xml
+++ b/build.xml
@@ -158,6 +158,7 @@
     <mkdir dir="${build.test.bindir}"/>
     <javac
       target="1.6" source="1.6"
+      bootclasspath="${java6.boot.classpath}"
       srcdir="${build.test.srcdir}"
       destdir="${build.test.bindir}"
       includeantruntime="false">
diff --git a/config.xml b/config.xml
index 5369505..255ecb8 100644
--- a/config.xml
+++ b/config.xml
@@ -26,7 +26,7 @@
   <name>Bullwinkle</name>
   
   <!-- The project's version number -->
-  <version>1.1.3</version>
+  <version>1.1.4</version>
   
   <!-- The project's author. Currently this only
   shows up in the footer of the Javadoc documentation. -->