--- parser3/src/main/pa_string.C	2003/09/26 06:53:27	1.178
+++ parser3/src/main/pa_string.C	2003/09/30 05:20:16	1.183
@@ -5,7 +5,7 @@
 	Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
 */
 
-static const char* IDENT_STRING_C="$Date: 2003/09/26 06:53:27 $";
+static const char* IDENT_STRING_C="$Date: 2003/09/30 05:20:16 $";
 
 #include "pcre.h"
 
@@ -167,11 +167,18 @@ String& String::mid(size_t substr_begin,
 }
 
 size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
-	size_t substr_begin=body.pos(substr, this_offset);
-	if(substr_begin==CORD_NOT_FOUND || !langs.check_lang(lang, substr_begin, substr.length()))
-		return STRING_NOT_FOUND;
+	size_t substr_length=substr.length();
+	while(true) {
+		size_t substr_begin=body.pos(substr, this_offset);
+		
+		if(substr_begin==CORD_NOT_FOUND)
+			return STRING_NOT_FOUND;
+
+		if(langs.check_lang(lang, substr_begin, substr_length))
+			return substr_begin;
 
-	return substr_begin;
+		this_offset=substr_begin+substr_length;
+	}
 }
 
 size_t String::pos(const String& substr, 
@@ -365,37 +372,49 @@ String& String::change_case(Charset& sou
 	if(is_empty())
 		return result;
 
-	const unsigned char *tables=source_charset.pcre_tables;
-
-	const unsigned char *a;
-	const unsigned char *b;
-	switch(kind) {
-	case CC_UPPER:
-		a=tables+lcc_offset;
-		b=tables+fcc_offset;
-		break;
-	case CC_LOWER:
-		a=tables+lcc_offset;
-		b=0;
-		break;
-	default:
-		throw Exception(0, 
-			this, 
-			"unknown change case kind #%d", 
-				static_cast<int>(kind)); // never
-		a=b=0; // calm, compiler
-		break; // never
-	}	
-
 	char* new_cstr=cstrm();
 	char *dest=new_cstr;
-	unsigned char index;
-	for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
-		unsigned char c=a[index];
-		if(b)
-			c=b[c];
+	if(source_charset.isUTF8()) {
+		switch(kind) {
+		case CC_UPPER:
+			change_case_UTF8((const XMLByte*)new_cstr, (XMLByte*)new_cstr, UTF8CaseToUpper);
+			break;
+		case CC_LOWER:
+			change_case_UTF8((const XMLByte*)new_cstr, (XMLByte*)new_cstr, UTF8CaseToLower);
+			break;
+		default:
+			assert(!"unknown change case kind");
+			break; // never
+		}	
+		
+	} else {
+		const unsigned char *tables=source_charset.pcre_tables;
+
+		const unsigned char *a;
+		const unsigned char *b;
+		switch(kind) {
+		case CC_UPPER:
+			a=tables+lcc_offset;
+			b=tables+fcc_offset;
+			break;
+		case CC_LOWER:
+			a=tables+lcc_offset;
+			b=0;
+			break;
+		default:
+			assert(!"unknown change case kind");
+			a=b=0; // calm, compiler
+			break; // never
+		}	
+
+		unsigned char index;
+		for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
+			unsigned char c=a[index];
+			if(b)
+				c=b[c];
 
-		*dest++=(char)c;
+			*dest++=(char)c;
+		}
 	}
 	result.langs=langs;
 	result.body=new_cstr;
@@ -413,7 +432,7 @@ const String& String::replace(const Dict
 		if(Table::element_type row=dict.first_that_begins(current)) {
 			// prematch
 			if(size_t prematch_length=current-prematch_begin) {
-				result.langs.append(result.body, langs, prematch_begin-old_cstr, current-old_cstr);
+				result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length);
 				result.body.append_strdup_know_length(prematch_begin, prematch_length);
 			}
 
@@ -433,7 +452,7 @@ const String& String::replace(const Dict
 
 	// postmatch
 	if(size_t postmatch_length=current-prematch_begin) {
-		result.langs.append(result.body, langs, prematch_begin-old_cstr, current-old_cstr);
+		result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length);
 		result.body.append_strdup_know_length(prematch_begin, postmatch_length);
 	}
 
@@ -496,6 +515,10 @@ int String::as_int() const {
 	return result;
 }
 
+static int serialize_body_char(char c, char** cur) {
+	*((*cur)++)=c;
+	return 0; // 0=continue
+};
 static int serialize_body_piece(const char* s, char** cur) {
 	size_t length=strlen(s);
 	memcpy(*cur, s, length);  *cur+=length;
@@ -526,8 +549,9 @@ String::Cm String::serialize(size_t prol
 	// 3: lang info
 	langs.for_each(body, serialize_lang_piece, &cur);
 	// 4: letters
-	body.for_each(serialize_body_piece, &cur);
-	// 5: zero terminator already there put by new(PointerFreeGC)
+	body.for_each(serialize_body_char, serialize_body_piece, &cur);
+	// 5: zero terminator
+	*cur=0;
 
 	return result;
 }