--- parser3/src/main/pa_string.C	2008/07/17 09:11:25	1.211
+++ parser3/src/main/pa_string.C	2008/08/21 15:58:12	1.217
@@ -5,9 +5,7 @@
 	Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
 */
 
-static const char * const IDENT_STRING_C="$Date: 2008/07/17 09:11:25 $";
-
-#include "pcre.h"
+static const char * const IDENT_STRING_C="$Date: 2008/08/21 15:58:12 $";
 
 #include "pa_string.h"
 #include "pa_exception.h"
@@ -15,6 +13,8 @@ static const char * const IDENT_STRING_C
 #include "pa_dictionary.h"
 #include "pa_charset.h"
 
+#include "pcre.h"
+
 const String String::Empty;
 
 int pa_atoi(const char* str, const String* problem_source) {
@@ -328,7 +328,7 @@ String& String::mid(Charset& charset, si
 		const XMLByte* srcPtr=(const XMLByte*)cstrm();
 		const XMLByte* srcEnd=srcPtr+body.length();
 
-		// convert from and substr_length from 'characters' to 'bytes'
+		// convert 'from' and 'substr_length' from 'characters' to 'bytes'
 		from=getUTF8BytePos(srcPtr, srcEnd, from);
 		substr_length=getUTF8BytePos(srcPtr+from, srcEnd, substr_length);
 		if(!substr_length)
@@ -367,13 +367,24 @@ size_t String::pos(const String& substr,
 size_t String::pos(Charset& charset, const String& substr, 
 				size_t this_offset, Language lang) const {
 
-	size_t result=pos(substr.body, this_offset, lang);
-	if(result==CORD_NOT_FOUND)
-		return STRING_NOT_FOUND;
-
+	size_t result;
 	if(charset.isUTF8()){
 		const XMLByte* srcPtr=(const XMLByte*)cstrm();
-		result=getUTF8CharPos(srcPtr, srcPtr+body.length(), result);
+		const XMLByte* srcEnd=srcPtr+body.length();
+
+		// convert 'this_offset' from 'characters' to 'bytes'
+		this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset);
+
+		result=pos(substr.body, this_offset, lang);
+		if(result==CORD_NOT_FOUND)
+			return STRING_NOT_FOUND;
+
+		// convert 'result' from 'bytes' to 'characters'
+		result=getUTF8CharPos(srcPtr, srcEnd, result);
+	} else {
+		result=pos(substr.body, this_offset, lang);
+		if(result==CORD_NOT_FOUND)
+			return STRING_NOT_FOUND;
 	}
 
 	return result;
@@ -431,30 +442,31 @@ enum Match_feature {
 
 static void regex_options(const String* options, int* result, int* match_features){
     struct Regex_option {
-		const char* keyL;
-		const char* keyU;
+		const char* key;
+		const char* keyAlt;
 		int clear;
 		int set;
 		int *result;
 		int flag;
     } regex_option[]={
 		{"i", "I", 0, PCRE_CASELESS, result, 0}, // a=A
-		{"s", "S", 0, PCRE_DOTALL, result, 0}, // \n\n$ [default]
-		{"x", "U", 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored
+		{"s", "S", 0, PCRE_DOTALL, result, 0}, // ^\n\n$ [default]
 		{"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result, 0}, // ^aaa\n$^bbb\n$
+		{"x", 0, 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored
+		{"U", 0, 0, PCRE_UNGREEDY, result, 0}, // ungreedy patterns (greedy by default)
 		{"g", "G", 0, 1, result+1, 0}, // many rows
 		{"'", 0, 0, 0, 0, MF_NEED_PRE_POST_MATCH},
-		{"n", "N", 0, 0, 0, MF_JUST_COUNT_MATCHES},
+		{"n", 0, 0, 0, 0, MF_JUST_COUNT_MATCHES},
 		{0, 0, 0, 0, 0, 0}
     };
 	result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
 	result[1]=0;
 
     if(options && !options->is_empty()) 
-		for(Regex_option *o=regex_option; o->keyL; o++) 
+		for(Regex_option *o=regex_option; o->key; o++) 
 			if(
-				options->pos(o->keyL)!=STRING_NOT_FOUND
-				|| (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)
+				options->pos(o->key)!=STRING_NOT_FOUND
+				|| (o->keyAlt && options->pos(o->keyAlt)!=STRING_NOT_FOUND)
 			){
 				if(o->flag){
 					(*match_features) |= o->flag;
@@ -484,6 +496,10 @@ Table* String::match(Charset& source_cha
 	bool need_pre_post_match=(match_features & MF_NEED_PRE_POST_MATCH) != 0;
 	bool just_count_matches=(match_features & MF_JUST_COUNT_MATCHES) != 0;
 	bool global=option_bits[1]!=0;
+
+	if(source_charset.isUTF8())
+		option_bits[0]=option_bits[0] | PCRE_UTF8;
+
 	pcre *code=pcre_compile(pattern, option_bits[0], 
 		&errptr, &erroffset,
 		source_charset.pcre_tables);
@@ -535,7 +551,7 @@ Table* String::match(Charset& source_cha
 			pcre_free(code);
 			throw Exception(0,
 				&regexp,
-				"regular expression execute error (%d)", 
+				print_pcre_exec_error_text(exec_substrings),
 					exec_substrings);
 		}
 
@@ -581,8 +597,9 @@ String& String::change_case(Charset& sou
 		return result;
 
 	char* new_cstr=cstrm();
-	size_t new_cstr_len=length();
+
 	if(source_charset.isUTF8()) {
+		size_t new_cstr_len=length();
 		switch(kind) {
 		case CC_UPPER:
 			change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
@@ -631,6 +648,13 @@ String& String::change_case(Charset& sou
 	return result;
 }
 
+const String& String::escape(Charset& source_charset) const {
+	if(is_empty())
+		return *this;
+
+	return Charset::escape(*this, source_charset);
+}
+
 const String& String::replace(const Dictionary& dict) const {
 	String& result=*new String();
 	const char* old_cstr=cstr();