--- parser3/src/main/untaint.C	2008/02/14 18:35:01	1.139
+++ parser3/src/main/untaint.C	2009/07/06 08:45:35	1.148
@@ -5,7 +5,7 @@
 	Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
 */
 
-static const char * const IDENT_UNTAINT_C="$Date: 2008/02/14 18:35:01 $";
+static const char * const IDENT_UNTAINT_C="$Date: 2009/07/06 08:45:35 $";
 
 
 #include "pa_string.h"
@@ -53,7 +53,7 @@ extern "C" { // author forgot to do that
     }
 
 
-#define escape(action) \
+#define escape_fragment(action) \
 	for(; fragment_length--; CORD_next(info->pos)) { \
 		char c=CORD_pos_fetch(info->pos); \
 		action \
@@ -74,26 +74,29 @@ inline bool need_file_encode(unsigned ch
 	// russian letters and space ENABLED
 	// encoding only these...
 	return strchr(
-			  "*?'\"<>|"
+			"*?'\"<>|"
 #ifndef WIN32
-			  ":\\"
+			":\\"
 #endif
-			  , c)!=0;
+			, c)!=0;
 }
+
 inline bool need_uri_encode(unsigned char c){
-    if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) 
+	if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z'))
 		return false;
 
-    return !strchr("_-./", c);
+	return !strchr("_-./", c);
 }
+
 inline bool need_http_header_encode(unsigned char c){
-    if(strchr(" , :", c))
+	if(strchr(" , :", c))
 		return false;
 
 	return need_uri_encode(c);
 }
+
 inline bool need_regex_escape(unsigned char c){
-	return strchr("\\^$.[]|()?*+{}", c)!=0;
+	return strchr("\\^$.[]|()?*+{}-", c)!=0;
 }
 
 // String
@@ -182,21 +185,37 @@ int append_fragment_nonoptimizing(char a
 	or marking ALL pieces of it with a @a lang when @a forced to,
 	and propagating OPTIMIZE language bit.
 */
-String& String::append_to(String& dest, Language lang, bool forced) const {
+String& String::append_to(String& dest, Language ilang, bool forced) const {
 	if(is_empty())
 		return dest;
 
 	// first: fragment infos
 	
-	if(lang==L_PASS_APPENDED) // without language-change?
-		dest.langs.append(dest.body, body.length(), langs);
+	if(ilang==L_PASS_APPENDED) // without language-change?
+		dest.langs.appendHelper(dest.body, langs, body);
 	else if(forced) //forcing passed lang?
-		dest.langs.append(dest.body, lang, length());
-	else { 
-		Append_fragment_info info={lang, &dest.langs, dest.body.length()};
-	    langs.for_each(body, lang&L_OPTIMIZE_BIT?
-			append_fragment_optimizing
-			:append_fragment_nonoptimizing, &info);
+		dest.langs.appendHelper(dest.body, ilang, body);
+	else {
+		if(langs.opt.is_not_just_lang){
+			Append_fragment_info info={ilang, &dest.langs, dest.body.length()};
+			langs.for_each(body, ilang&L_OPTIMIZE_BIT?
+				append_fragment_optimizing
+				:append_fragment_nonoptimizing, &info);
+		} else {
+			Language lang=langs.opt.lang;
+			// see append_fragment_* for explanation
+			if(ilang&L_OPTIMIZE_BIT){
+				dest.langs.appendHelper(dest.body,
+					lang==String::L_TAINTED?
+						ilang
+						:lang==String::L_CLEAN?
+							(String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT)
+							:lang,
+					body);
+			} else {
+				dest.langs.appendHelper(dest.body, lang==String::L_TAINTED ? ilang:lang, body);
+			}
+		}
 	}
 
 	// next: letters
@@ -298,14 +317,11 @@ struct Cstr_to_string_body_block_info {
 	const char* exception;
 };
 #endif
-int cstr_to_string_body_block(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info) {
-	const String::Language fragment_lang=(String::Language)(unsigned char)alang;
+int cstr_to_string_body_block(String::Language to_lang, size_t fragment_length, Cstr_to_string_body_block_info* info) {
 	bool& whitespace=info->whitespace;
 	size_t fragment_end=info->fragment_begin+fragment_length;
-	//fprintf(stderr, "%d, %d =%s=\n", fragment_lang, fragment_length, info->body->cstr());
-
+	//fprintf(stderr, "%d, %d =%s=\n", to_lang, fragment_length, info->body->cstr());
 	
-	String::Language to_lang=info->lang==String::L_UNSPECIFIED?fragment_lang:info->lang;
 	bool optimize=(to_lang & String::L_OPTIMIZE_BIT)!=0;
 	if(!optimize)
 		whitespace=false;
@@ -325,14 +341,26 @@ int cstr_to_string_body_block(char alang
 		break;
 	case String::L_FILE_SPEC:
 		// tainted, untaint language: file [name]
-		escape(
-			// Macintosh has problems with small Russian letter 'r'
-			if( c=='\xF0' && info->charsets && info->charsets->source().NAME()=="WINDOWS-1251" ) {
-				// fixing that letter for most common charset
-				to_char('p');
-			} else // fallback to default
-				encode(need_file_encode, '_', c); 
-		);
+		{
+			bool is1251=(info->charsets && info->charsets->source().NAME()=="WINDOWS-1251");
+			escape_fragment(
+				// Macintosh has problems with small Russian letter 'r'
+				if( is1251 && c=='\xF0' ) {
+					// fixing that letter for most common charset
+					to_char('p');
+				} else // fallback to default
+					encode(need_file_encode, '_', c); 
+			);
+		}
+		break;
+	case String::L_FILE_POST:
+		{
+			escape_fragment(switch(c) {
+				case '\0': to_string("\\0");  break;
+				case '\\': to_string("\\\\"); break;
+				default: _default; break;
+			});
+		}
 		break;
 	case String::L_URI:
 		// tainted, untaint language: uri
@@ -354,7 +382,7 @@ int cstr_to_string_body_block(char alang
 	case String::L_HTTP_HEADER:
 		// tainted, untaint language: http-field-content-text
 		// the same as L_URI BUT not transcoded into $response:charset before encoding
-		escape(
+		escape_fragment(
 			encode(need_uri_encode, '%', c);
 		);
 		break;
@@ -421,20 +449,22 @@ int cstr_to_string_body_block(char alang
 		}
 		break;
 	case String::L_JS:
-		escape(switch(c) {
+		escape_fragment(switch(c) {
+			case '\n': to_string("\\n");  break;
 			case '"': to_string("\\\"");  break;
 			case '\'': to_string("\\'");  break;
-			case '\n': to_string("\\n");  break;
 			case '\\': to_string("\\\\");  break;
 			case '\xFF': to_string("\\\xFF");  break;
+			case '\r': to_string("\\r");  break;
 			default: _default; break;
 		});
 		break;
 	case String::L_XML:
 		// [2]    Char    ::=    #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF] 
-		escape(switch(c) {
-			case '\x9': 
-			case '\xA': 
+		escape_fragment(switch(c) {
+			case '\x20':
+			case '\x9':
+			case '\xA':
 			case '\xD': // this is usually removed on input
 				_default; 
 				break;
@@ -463,7 +493,7 @@ int cstr_to_string_body_block(char alang
 		});
 		break;
 	case String::L_HTML:
-		escape(switch(c) {
+		escape_fragment(switch(c) {
 			case '&': to_string("&amp;");  break;
 			case '>': to_string("&gt;");  break;
 			case '<': to_string("&lt;");  break;
@@ -473,12 +503,26 @@ int cstr_to_string_body_block(char alang
 		break;
 	case String::L_REGEX:
 		// tainted, untaint language: regex
-		escape(
+		escape_fragment(
 			if(need_regex_escape(c))
 				to_char('\\')
 			_default;
 		);
 		break;
+	case String::L_HTTP_COOKIE:
+		// tainted, untaint language: cookie (3.3.0 and higher: %uXXXX in UTF-8)
+		{
+			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);
+
+		}
+		break;
 	default:
 		SAPI::abort("unknown untaint language #%d", 
 			static_cast<int>(to_lang)); // should never
@@ -491,9 +535,11 @@ int cstr_to_string_body_block(char alang
 }
 
 
-String::Body String::cstr_to_string_body(Language lang, 
-		  SQL_Connection* connection,
-		  const Request_charsets *charsets) const {
+int cstr_to_string_body_block_default(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info){
+	return cstr_to_string_body_block(info->lang==String::L_UNSPECIFIED ? (String::Language)(unsigned char)alang : info->lang, fragment_length, info);
+}
+
+String::Body String::cstr_to_string_body(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
 	if(is_empty())
 		return String::Body();
 
@@ -511,12 +557,81 @@ String::Body String::cstr_to_string_body
 	info.exception=0;
 	info.whitespace=true;
 
-	langs.for_each(body, cstr_to_string_body_block, &info);
-	if(info.exception){
+	langs.for_each(body, cstr_to_string_body_block_default, &info);
+	if(info.exception)
+		throw Exception(0,
+			0,
+			info.exception);
+
+	return String::Body(CORD_ec_to_cord(info.result));
+}
+
+String::Body String::cstr_to_string_body_taint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
+	if(is_empty())
+		return String::Body();
+
+	Cstr_to_string_body_block_info info;
+	// input
+	info.lang=lang;
+	info.connection=connection;
+	info.charsets=charsets;
+	info.body=&body;
+	// output
+	CORD_ec_init(info.result);
+	// private
+	body.set_pos(info.pos, 0);
+	info.fragment_begin=0;
+	info.exception=0;
+	info.whitespace=true;
+
+	cstr_to_string_body_block(lang, length(), &info);
+
+	if(info.exception)
+		throw Exception(0,
+			0,
+			info.exception);
+
+	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){
+	const String::Language lang=(String::Language)(unsigned char)alang;
+	// see append_fragment_* for explanation
+	if(info->lang&String::L_OPTIMIZE_BIT)
+		return cstr_to_string_body_block(
+			lang==String::L_TAINTED?
+				info->lang
+				:lang==String::L_CLEAN?
+					(String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT)
+					:lang,
+			fragment_length, info);
+	else
+		return cstr_to_string_body_block(lang==String::L_TAINTED ? info->lang:lang, fragment_length, info);
+}
+
+String::Body String::cstr_to_string_body_untaint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
+	if(is_empty())
+		return String::Body();
+
+	Cstr_to_string_body_block_info info;
+	// input
+	info.lang=lang;
+	info.connection=connection;
+	info.charsets=charsets;
+	info.body=&body;
+	// output
+	CORD_ec_init(info.result);
+	// private
+	body.set_pos(info.pos, 0);
+	info.fragment_begin=0;
+	info.exception=0;
+	info.whitespace=true;
+
+	langs.for_each(body, cstr_to_string_body_block_untaint, &info);
+	if(info.exception)
 		throw Exception(0,
 			0,
 			info.exception);
-	}
 
 	return String::Body(CORD_ec_to_cord(info.result));
 }