--- parser3/src/main/untaint.C 2003/03/25 12:10:09 1.115.2.12.2.14 +++ parser3/src/main/untaint.C 2003/04/07 11:58:10 1.115.2.12.2.20 @@ -5,7 +5,7 @@ Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char* IDENT_UNTAINT_C="$Date: 2003/03/25 12:10:09 $"; +static const char* IDENT_UNTAINT_C="$Date: 2003/04/07 11:58:10 $"; #include "pa_string.h" @@ -17,21 +17,18 @@ static const char* IDENT_UNTAINT_C="$Dat #include "pa_common.h" #include "pa_charset.h" #include "pa_request_charsets.h" +#include "pa_sapi.h" extern "C" { // author forgot to do that #include "ec.h" } -//#define PA_SQL +#define PA_SQL #ifdef PA_SQL #include "pa_sql_connection.h" #endif -// forwards - -void CORD_ec_flush_buf_optimize_whitespace(CORD_ec x); - // defines @@ -67,14 +64,14 @@ void CORD_ec_flush_buf_optimize_whitespa action \ } #define _default default: CORD_ec_append(result, c); break -#define encode(need_encode_func, prefix) \ +#define encode(need_encode_func, prefix, otherwise) \ if(need_encode_func(c)) { \ static const char* hex="0123456789ABCDEF"; \ CORD_ec_append(result, prefix); \ CORD_ec_append(result, hex[((unsigned char)c)/0x10]); \ CORD_ec_append(result, hex[((unsigned char)c)%0x10]); \ } else \ - CORD_ec_append(result, c); + CORD_ec_append(result, otherwise); #define to_char(c) CORD_ec_append(result, c) #define to_string(s) CORD_ec_append_cord(result, s) @@ -182,7 +179,7 @@ String& String::append_to(String& dest, return dest; // first: letters - dest.body.append(body); + dest.body< i(fragments); while(i.has_next()) { - const Fragment& fragment=i.next(); + const Fragment fragment=i.next(); // main idea here: // tainted piece would get OPTIMIZED bit from 'lang' // clean piece would be marked OPTIMIZED manually @@ -207,7 +204,7 @@ String& String::append_to(String& dest, } else { // The core idea: tainted pieces got marked with context's lang Array_iterator i(fragments); while(i.has_next()) { - const Fragment& fragment=i.next(); + const Fragment fragment=i.next(); dest.fragments+=Fragment( fragment.lang==L_TAINTED?lang:fragment.lang, fragment.length); @@ -250,6 +247,15 @@ inline bool mail_header_char_valid_withi || c>='0' && c<='9' || strchr("!*+-/", c); } +/** + RFC + Upper case should be used for hexadecimal digits "A" through "F" + The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) + may be represented as "_" +*/ +inline bool mail_header_nonspace_char(char c) { + return c != 0x20; +} inline void ec_append(CORD_ec& result, bool& optimize, bool& whitespace, CORD_pos& pos, size_t size) { while(size--) { CORD_ec_append(result, CORD_pos_fetch(pos)); @@ -267,7 +273,7 @@ StringBody String::cstr_to_string_body(L size_t fragment_begin=0; size_t fragment_end; for(Array_iterator i(fragments); i.has_next(); fragment_begin=fragment_end) { - const Fragment& fragment=i.next(); + const Fragment fragment=i.next(); size_t fragment_length=fragment.length; fragment_end=fragment_begin+fragment_length; //fprintf(stderr, "%d, %d\n", fragment.lang, fragment.length); @@ -276,7 +282,7 @@ StringBody String::cstr_to_string_body(L bool optimize=(to_lang & L_OPTIMIZE_BIT)!=0; if(!optimize) whitespace=false; - + switch(to_lang & ~L_OPTIMIZE_BIT) { case L_CLEAN: case L_TAINTED: @@ -298,58 +304,59 @@ StringBody String::cstr_to_string_body(L // fixing that letter for most common charset to_char('p'); } else // fallback to default - encode(need_file_encode, '_'); + encode(need_file_encode, '_', c); ); break; case L_URI: // tainted, untaint language: uri { - size_t source_size; - const char *source_ptr= - body.mid(fragment_begin, source_size=fragment_length).cstr(); + const char *fragment_str=body.mid(fragment_begin, fragment_length).cstr(); // skip source [we use recoded version] - CORD_pos_advance(pos, source_size); + CORD_pos_advance(pos, fragment_length); const void *client_ptr; size_t client_size; if(charsets) Charset::transcode( - charsets->source(), source_ptr, source_size, + charsets->source(), fragment_str, fragment_length, charsets->client(), client_ptr, client_size); else { - client_ptr=source_ptr; - client_size=source_size; + client_ptr=fragment_str; + client_size=fragment_length; } char c; for(const char* src=(const char* )client_ptr; c=*src++; ) - encode(need_uri_encode, '%'); + encode(need_uri_encode, '%', c); } break; case L_HTTP_HEADER: // tainted, untaint language: http-field-content-text escape( - encode(need_uri_encode, '%'); + encode(need_uri_encode, '%', c); ); break; -#ifdef LATER case L_MAIL_HEADER: // tainted, untaint language: mail-header // http://www.ietf.org/rfc/rfc2047.txt if(charsets) { - const void *mail_ptr=fragment.ptr; - size_t mail_size=fragment_length; - const char* charset_name=charsets->mail().name()->cstr(); + size_t mail_size; + const char *mail_ptr= + body.mid(fragment_begin, mail_size=fragment_length).cstr(); + // skip source [we use recoded version] + CORD_pos_advance(pos, mail_size); + + const char* charset_name=charsets->mail().name().cstr(); // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= - const char* src=(const char* )mail_ptr; bool to_quoted_printable=false; bool email=false; - for(const char* end=src+mail_size; src') email=false; } if(to_quoted_printable) // close - dest+=sprintf(dest, "?="); + to_string("?="); - } else { - memcpy(dest, fragment.ptr, fragment_length); - dest+=fragment_length; - } + } else + ec_append(result, optimize, whitespace, pos, fragment_length); break; -#endif case L_TABLE: // tainted, untaint language: table escape(switch(c) { @@ -391,13 +393,12 @@ StringBody String::cstr_to_string_body(L case L_SQL: // tainted, untaint language: sql if(connection) { - dest+=connection->quote(0, - CORD_substr(body, fragment_begin, fragment_length), - fragment_length); -...^^^ + const char *fragment_str=body.mid(fragment_begin, fragment_length).cstr(); + // skip source [we use recoded version] + CORD_pos_advance(pos, fragment_length); - dest+=connection->quote(dest, fragment.ptr, fragment_length); - else + to_string(connection->quote(fragment_str, fragment_length)); + } else throw Exception(0, 0, "untaint in SQL language failed - no connection specified"); @@ -433,10 +434,8 @@ StringBody String::cstr_to_string_body(L }); break; default: - throw Exception(0, - 0, - "unknown untaint language #%d", - static_cast(to_lang)); // should never + SAPI::die("unknown untaint language #%d", + static_cast(to_lang)); // should never break; // never } }