Allow most metacharacters in character classes, e.g. allow "[+-]"

Allows most metacharacters, e.g. "*+()[", to appear unescaped within a
character class.  This, along with hyphens at the end, allows "[+-]" to
match a plus or a minus.

src/lalr/RegexParser.cpp

@@ -241,7 +241,7 @@ bool RegexParser::match_item()
         syntax_tree_->item_xdigit();
         return true;
     }
-    else if ( match_character() )
+    else if ( match_character_in_character_class() )
     {
         int character = escape( lexeme_begin_, lexeme_end_ );
         if ( match_end_of_range() )
@@ -325,7 +325,7 @@ bool RegexParser::match_negative_item()
         syntax_tree_->negative_item_xdigit();
         return true;
     }
-    else if ( match_character() )
+    else if ( match_character_in_character_class() )
     {
         int character = escape( lexeme_begin_, lexeme_end_ );
         if ( match_end_of_range() )
@@ -344,6 +344,17 @@ bool RegexParser::match_negative_item()
 
 bool RegexParser::match_character()
 {
+    return match_character( "|*+?[]()-" );
+}
+
+bool RegexParser::match_character_in_character_class()
+{
+    return match_character( "^]-" );
+}
+
+bool RegexParser::match_character( const char* metacharacters )
+{
+    LALR_ASSERT( metacharacters );
     const char* position = position_;
     if ( position != end_ )
     {
@@ -375,7 +386,7 @@ bool RegexParser::match_character()
             position_ = position;
             return true;
         }
-        else if ( !strchr("|*+?[]()-", *position) )
+        else if ( !strchr(metacharacters, *position) )
         {
             ++position;
             lexeme_begin_ = position_;
@@ -390,7 +401,7 @@ bool RegexParser::match_character()
 bool RegexParser::match_end_of_range()
 {
     const char* position = position_;
-    if ( match("-") && match_character() )
+    if ( match("-") && match_character_in_character_class() )
     {
         return true;
     }

src/lalr/RegexParser.hpp

@@ -32,6 +32,9 @@ private:
     bool match_item();
     bool match_negative_item();
     bool match_character();
+    bool match_character_in_character_class();
+    bool match_character( const char* metacharacters );
+    bool match_bracket_expression_character();
     bool match_end_of_range();
     bool match_identifier();
     bool match( const char* lexeme );

src/lalr/lalr_test/TestRegularExpressions.cpp

@@ -1378,4 +1378,29 @@ SUITE( RegularExpressions )
         CHECK_EQUAL( "NOT BETWEEN", lexer.lexeme() );
         CHECK( lexer.symbol() == &not_between );
     }
+
+
+    TEST( CharacterClassHyphens )
+    {
+        void* plus_or_minus;
+        RegexCompiler compiler;
+        compiler.compile( "[+-]", &plus_or_minus );
+        Lexer<const char*> lexer( compiler.state_machine() );
+        
+        const char* regex = "+";
+        lexer.reset( regex, regex + strlen(regex) );
+        lexer.advance();
+        CHECK( lexer.symbol() == &plus_or_minus );
+        CHECK( lexer.lexeme() == "+" );
+        lexer.advance();
+        CHECK( lexer.symbol() == nullptr );
+
+        regex = "-";
+        lexer.reset( regex, regex + strlen(regex) );
+        lexer.advance();
+        CHECK( lexer.symbol() == &plus_or_minus );
+        CHECK( lexer.lexeme() == "-" );
+        lexer.advance();
+        CHECK( lexer.symbol() == nullptr );
+    }
 }