--- parser3/src/main/untaint.C	2001/04/09 11:04:12	1.37
+++ parser3/src/main/untaint.C	2001/04/23 08:52:24	1.43
@@ -5,7 +5,7 @@
 
 	Author: Alexander Petrosyan <paf@design.ru>(http://design.ru/paf)
 
-	$Id: untaint.C,v 1.37 2001/04/09 11:04:12 paf Exp $
+	$Id: untaint.C,v 1.43 2001/04/23 08:52:24 paf Exp $
 */
 
 #include "pa_config_includes.h"
@@ -49,7 +49,7 @@ inline bool need_file_encode(unsigned ch
 #ifdef WIN32
 		":\\~"
 #endif
-		"./", c);
+		"./()_-", c);
 }
 inline bool need_uri_encode(unsigned char c){
     if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) 
@@ -77,12 +77,55 @@ static bool typo_present(Array::Item *va
 		partial==1; // typo left column starts 'src'
 }
 
+/*
+
+HTTP-header    = field-name ":" [ field-value ] CRLF
+
+       field-name     = token
+       field-value    = *( field-content | LWS )
+
+       field-content  = <the OCTETs making up the field-value
+                        and consisting of either *TEXT or combinations
+                        of token, tspecials, and quoted-string>
+
+
+
+word           = token | quoted-string
+
+token          = 1*<any CHAR except CTLs or tspecials>
+
+
+
+tspecials      = "(" | ")" | "<" | ">" | "@"
+                      | "," | ";" | ":" | "\" | <">
+                      | "/" | "[" | "]" | "?" | "="
+                      | "{" | "}" | SP | HT
+
+SP             = <US-ASCII SP, space (32)>
+HT             = <US-ASCII HT, horizontal-tab (9)>
+
+LWS            = [CRLF] 1*( SP | HT )
+TEXT           = <any OCTET except CTLs,
+                        but including LWS>
+
+quoted-pair    = "\" CHAR
+
+  if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
+*/
+inline bool need_quote_http_header(const char *ptr, size_t size) {
+	for(; size--; ptr++)
+		if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
+			return true;
+	return false;
+}
+
 /**
 	@test optimize whitespaces for all but 'html'
 	@todo fix theoretical \n mem overrun in TYPO replacements
-	@test mail-header
 */
-char *String::store_to(char *dest, Untaint_lang lang, SQL_Connection *connection) const {
+char *String::store_to(char *dest, Untaint_lang lang, 
+					   SQL_Connection *connection,
+					   const char *charset) const {
 	// $MAIN:html-typo table
 	Table *user_typo_table=static_cast<Table *>(pool().tag());
 	Table *typo_table=user_typo_table?user_typo_table:default_typo_table;
@@ -112,7 +155,7 @@ char *String::store_to(char *dest, Untai
 				// tainted, untaint language: file [name]
 				escape(switch(*src) {
 					case ' ': to_char('_');  break;
-					encode(need_file_encode, '-');
+					encode(need_file_encode, '+');
 				});
 				break;
 			case UL_URI:
@@ -124,14 +167,42 @@ char *String::store_to(char *dest, Untai
 				break;
 			case UL_HTTP_HEADER:
 				// tainted, untaint language: http-header
-				escape(switch(*src) {
-					encode(need_http_header_encode, '%');
-				});
+				if(need_quote_http_header(row->item.ptr, row->item.size)) {
+					*dest++='\"';
+					escape(switch(*src) {
+						case '\"': to_string("\\\"", 2);  break;
+						_default;
+					});
+					*dest++='\"';
+				} else {
+					memcpy(dest, row->item.ptr, row->item.size); 
+					dest+=row->item.size;
+				}
 				break;
 			case UL_MAIL_HEADER:
 				// tainted, untaint language: mail-header
-				memcpy(dest, row->item.ptr, row->item.size); 
-				dest+=row->item.size;
+				{
+					// Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
+					const char *src=row->item.ptr; 
+					bool down=false;
+					for(int size=row->item.size; size--; src++) {
+						if(*src & 0x80) {
+							if(!down) {
+								dest+=sprintf(dest, "=?%.15s?Q?", charset);
+								down=true;
+							}
+							dest+=sprintf(dest, "=%02X", *src & 0xFF);						
+						} else {
+							if(down) {
+								down=false;
+								dest+=sprintf(dest, "?=");
+							}
+							*dest++=*src;						
+						}
+					}
+					if(down) // close unclosed
+						dest+=sprintf(dest, "?=");
+				}
 				break;
 			case UL_TABLE: 
 				// tainted, untaint language: table
@@ -245,7 +316,7 @@ char *String::store_to(char *dest, Untai
 					this, 
 					"unknown untaint language #%d of %d piece", 
 						static_cast<int>(row->item.lang), 
-						i);
+						i); // never
 			}
 		}
 		chunk=row->link;