From: Mikko Rasa <tdb@tdb.fi>
Date: Fri, 10 Aug 2007 19:43:55 +0000 (+0000)
Subject: Fix a bug in Regex with brackets
X-Git-Tag: strings-1.0~9
X-Git-Url: http://git.tdb.fi/?a=commitdiff_plain;h=df1a66180655104af15f9bb6067b4eaa1421f5ff;p=libs%2Fcore.git

Fix a bug in Regex with brackets
Make the disassembly output of grep optional
---

diff --git a/grep.cpp b/grep.cpp
index 8c369a5..eaf1663 100644
--- a/grep.cpp
+++ b/grep.cpp
@@ -1,6 +1,7 @@
 /* $Id$ */
 #include <iostream>
 #include <string>
+#include <msp/core/getopt.h>
 #include "regex.h"
 
 using namespace std;
@@ -8,21 +9,27 @@ using namespace Msp;
 
 int main(int argc, char **argv)
 {
-	if(argc<2)
+	bool debug=false;
+	GetOpt getopt;
+	getopt.add_option('d', "debug", debug, GetOpt::NO_ARG);
+	getopt(argc, argv);
+
+	const list<string> &args=getopt.get_args();
+
+	if(args.empty())
 	{
 		cerr<<"Usage: "<<argv[0]<<" <regex>\n";
 		return 1;
 	}
 
 	Regex regex(argv[1]);
-	cout<<regex.disassemble();
+	if(debug)
+		cout<<regex.disassemble();
 	string line;
 	while(getline(cin, line))
 	{
 		if(RegMatch match=regex.match(line))
-		{
 			cout<<line<<'\n';
-		}
 	}
 
 	return 0;
diff --git a/source/regex.cpp b/source/regex.cpp
index ae3576c..320705e 100644
--- a/source/regex.cpp
+++ b/source/regex.cpp
@@ -11,8 +11,6 @@ Distributed under the LGPL
 
 using namespace std;
 
-#include <iostream>
-
 namespace {
 
 /**
@@ -309,14 +307,14 @@ Regex::Code Regex::parse_brackets(const string &str, string::const_iterator &ite
 	}
 
 	string::const_iterator end=iter;
-	for(; (end!=str.end() && *end!=']'); ++end);
+	for(; (end!=str.end() && (end==iter || *end!=']')); ++end);
 	if(end==str.end())
 		throw InvalidParameterValue("Unmatched '['");
 
 	uint8_t mask[32]={0};
 	unsigned type=0;
 	bool range=false;
-	unsigned char first, last;
+	unsigned char first=0, last=0;
 	for(string::const_iterator i=iter; i!=end; ++i)
 	{
 		unsigned char c=*i;
@@ -502,7 +500,7 @@ bool Regex::group_compare(const RegMatch::Group &g1, const RegMatch::Group &g2)
 {
 	if(!g1.match)
 		return false;
-	
+
 	// Any match is better than no match
 	if(!g2.match)
 		return true;
diff --git a/source/regex.h b/source/regex.h
index 31b66d2..3750025 100644
--- a/source/regex.h
+++ b/source/regex.h
@@ -54,7 +54,7 @@ the input string, respectively.  They do not consume input.
 
 The MATCH_CHAR instruction consumes the input character and matches it against
 a single character.  Since regexes often match sequences of printable character,
-a match for such a character may be encoded as the character itself.
+a match for a non-opcode character may be encoded as the character itself.
 
 The MATCH_RANGE instruction consumes the input character and matches it against
 an inclusive character range.