--- parser3/src/main/untaint.C	2010/10/28 21:49:46	1.158
+++ parser3/src/main/untaint.C	2015/10/08 18:29:16	1.168
@@ -1,11 +1,11 @@
 /** @file
 	Parser: String class part: untaint mechanizm.
 
-	Copyright(c) 2001-2005 ArtLebedev Group (http://www.artlebedev.com)
+	Copyright (c) 2001-2012 Art. Lebedev Studio (http://www.artlebedev.com)
 	Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
 */
 
-static const char * const IDENT_UNTAINT_C="$Date: 2010/10/28 21:49:46 $";
+volatile const char * IDENT_UNTAINT_C="$Id: untaint.C,v 1.168 2015/10/08 18:29:16 moko Exp $";
 
 
 #include "pa_string.h"
@@ -52,29 +52,34 @@ extern "C" { // author forgot to do that
 	} \
     }
 
-
 #define escape_fragment(action) \
 	for(; fragment_length--; CORD_next(info->pos)) { \
 		char c=CORD_pos_fetch(info->pos); \
 		action \
 	}
-#define _default CORD_ec_append(info->result, c)
+
 #define encode(need_encode_func, prefix, otherwise)  \
 	if(need_encode_func(c)) { \
-		static const char* hex="0123456789ABCDEF"; \
 		CORD_ec_append(info->result, prefix); \
-		CORD_ec_append(info->result, hex[((unsigned char)c)/0x10]); \
-		CORD_ec_append(info->result, hex[((unsigned char)c)%0x10]); \
+		to_hex(c); \
 	} else \
 		CORD_ec_append(info->result, otherwise);
+
+#define to_hex(c) \
+	{ \
+		CORD_ec_append(info->result, hex_digits[((unsigned char)c) >> 4]); \
+		CORD_ec_append(info->result, hex_digits[((unsigned char)c) & 0x0F]); \
+	}
+
 #define to_char(c)  { CORD_ec_append(info->result, c); whitespace=false; }
-#define to_string(s)  { CORD_ec_append_cord(info->result, s); whitespace=false; }
+#define to_string(s)  { CORD_ec_append_cord(info->result, (CORD)(s)); whitespace=false; }
+#define _default CORD_ec_append(info->result, c)
 
 inline bool need_file_encode(unsigned char c){
 	// russian letters and space ENABLED
 	// encoding only these...
 	return strchr(
-			"*?'\"<>|"
+			"*?\"<>|"
 #ifndef WIN32
 			":\\"
 #endif
@@ -82,10 +87,7 @@ inline bool need_file_encode(unsigned ch
 }
 
 inline bool need_uri_encode(unsigned char c){
-	if((c>='0') && (c<='9') || (c>='A') && (c<='Z') || (c>='a') && (c<='z'))
-		return false;
-
-	return !strchr("_-./*", c);
+	return !(pa_isalnum(c) || strchr("_-./*", c));
 }
 
 inline bool need_regex_escape(unsigned char c){
@@ -251,10 +253,7 @@ RFC
 		and without "_", or in would mean 0x20
 */
 inline bool mail_header_char_valid_within_Qencoded(char c) {
-	return c>='A' && c<='Z'
-		|| c>='a' && c<='Z'
-		|| c>='0' && c<='9'
-		|| strchr("!*+-/", c);
+	return (pa_isalnum((unsigned char)c) || strchr("!*+-/", c));
 }
 inline bool addr_spec_soon(const char *src) {
 	for(char c; (c=*src); src++)
@@ -366,9 +365,11 @@ int cstr_to_string_body_block(String::La
 		break;
 	case String::L_HTTP_HEADER:
 		// tainted, untaint language: http-field-content-text
-		escape_fragment(
-			encode(need_uri_encode, '%', c);
-		);
+		escape_fragment(switch(c) {
+			case '\n': 
+			case '\r': to_string(" ");  break;
+			default: _default; break;
+		});
 		break;
 	case String::L_MAIL_HEADER:
 		// tainted, untaint language: mail-header
@@ -397,8 +398,8 @@ int cstr_to_string_body_block(String::La
 				}
 				//RFC   + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
 				if(!email && (
-					!to_quoted_printable && (c & 0x80)  // starting quote-printable-encoding on first 8bit char
-					|| to_quoted_printable && !mail_header_char_valid_within_Qencoded(c)
+					( !to_quoted_printable && (c & 0x80) )  // starting quote-printable-encoding on first 8bit char
+					|| ( to_quoted_printable && !mail_header_char_valid_within_Qencoded(c) )
 					)) {
 					if(!to_quoted_printable) {
 						to_string("=?");
@@ -497,9 +498,9 @@ int cstr_to_string_body_block(String::La
 		// tainted, untaint language: json <http://json.org/>
 		// escape '"' '\' '/' '\n' '\t' '\r' '\b' '\f' chars and escape chars as \uXXXX if output charset != UTF-8
 		{
-			if(info->charsets->client().isUTF8()){
+			if(info->charsets==NULL || info->charsets->client().isUTF8()){
 				// escaping to \uXXXX is not needed
-				escape_fragment(switch(c) {
+				escape_fragment(switch((unsigned char)c) {
 					case '\n': to_string("\\n");  break;
 					case '"' : to_string("\\\""); break;
 					case '\\': to_string("\\\\"); break;
@@ -508,7 +509,33 @@ int cstr_to_string_body_block(String::La
 					case '\r': to_string("\\r");  break;
 					case '\b': to_string("\\b");  break;
 					case '\f': to_string("\\f");  break;
-					default  : _default; break;
+					case 0xE2: // \u2028 and \u2029 (line/paragraph separators), check bug #1023
+						if(info->charsets && info->charsets->source().isUTF8() && fragment_length>=2){
+							CORD_next(info->pos);
+							char c1=CORD_pos_fetch(info->pos);
+							CORD_next(info->pos);
+							char c2=CORD_pos_fetch(info->pos);
+							if((unsigned char)c1 == 0x80 && ((unsigned char)c2 >= 0xA8 && (unsigned char)c2 <= 0xAF)){
+								to_string("\\u20");
+								to_hex(((unsigned char)c2-0x80));
+							} else {
+								CORD_ec_append(info->result, c);
+								CORD_ec_append(info->result, c1);
+								CORD_ec_append(info->result, c2);
+							}
+							fragment_length-=2;
+						} else {
+							_default;
+						}
+						break;
+					default: 
+						if((unsigned char)c < 0x20){
+							to_string("\\u00");
+							to_hex(c);
+						} else {
+							_default;
+						}
+						break;
 				});
 			} else {
 				const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
@@ -517,23 +544,22 @@ int cstr_to_string_body_block(String::La
 				String::C output(fragment_str, fragment_length);
 
 				output=Charset::escape_JSON(output, info->charsets->source());
-				to_string(output);
+				to_string(output.str);
 			}
 		}
 		break;
 	case String::L_HTTP_COOKIE:
 		// tainted, untaint language: cookie (3.3.0 and higher: %uXXXX in UTF-8)
-		{
+		if(info->charsets) {
 			const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
 			// skip source [we use recoded version]
 			pa_CORD_pos_advance(info->pos, fragment_length);
 			String::C output(fragment_str, fragment_length);
 			
 			output=Charset::escape(output, info->charsets->source());
-			//throw Exception(0, 0, output);
-			to_string(output);
-
-		}
+			to_string(output.str);
+		} else
+			ec_append(info->result, optimize, whitespace, info->pos, fragment_length);
 		break;
 	case String::L_PARSER_CODE:
 		// for auto-untaint in process
@@ -580,7 +606,7 @@ String::Body String::cstr_to_string_body
 			0,
 			info.exception);
 
-	return String::Body(CORD_ec_to_cord(info.result), info.fragment_begin);
+	return String::Body(CORD_ec_to_cord(info.result));
 }
 
 int cstr_to_string_body_block_untaint(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info){
@@ -623,7 +649,7 @@ String::Body String::cstr_to_string_body
 			0,
 			info.exception);
 
-	return String::Body(CORD_ec_to_cord(info.result), info.fragment_begin);
+	return String::Body(CORD_ec_to_cord(info.result));
 }
 
 const char* String::untaint_and_transcode_cstr(Language lang, const Request_charsets *charsets) const {