--- parser3/src/main/untaint.C 2002/02/08 08:32:34 1.90 +++ parser3/src/main/untaint.C 2003/03/20 11:11:56 1.115.2.12.2.5 @@ -1,60 +1,59 @@ /** @file Parser: String class part: untaint mechanizm. - Copyright(c) 2001, 2002 ArtLebedev Group (http://www.artlebedev.com) + Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) - - $Id: untaint.C,v 1.90 2002/02/08 08:32:34 paf Exp $ */ -#include "pa_pool.h" +static const char* IDENT_UNTAINT_C="$Date: 2003/03/20 11:11:56 $"; + + #include "pa_string.h" #include "pa_hash.h" #include "pa_exception.h" #include "pa_table.h" #include "pa_globals.h" -#include "pa_sql_connection.h" #include "pa_dictionary.h" #include "pa_common.h" #include "pa_charset.h" +#include "pa_request_charsets.h" + +extern "C" { // author forgot to do that +#include "ec.h" +} + +//#define PA_SQL + +#ifdef PA_SQL +#include "pa_sql_connection.h" +#endif #define escape(action) \ - { \ - const char *src=row->item.ptr; \ - for(int size=row->item.size; size--; src++) \ - action \ + for(; fragment_size--; CORD_next(pos)) { \ + char c=CORD_pos_fetch(pos); \ + action \ } -#define _default default: *dest++=*src; break +#define _default default: CORD_ec_append(result, c); break #define encode(need_encode_func, prefix) \ - if(need_encode_func(*src)) { \ - static const char *hex="0123456789ABCDEF"; \ - char chunk[3]={prefix}; \ - chunk[1]=hex[((unsigned char)*src)/0x10]; \ - chunk[2]=hex[((unsigned char)*src)%0x10]; \ - memcpy(dest, chunk, 3); dest+=3; \ - } else \ - *dest++=*src; \ - break -#define to_char(c) *dest++=c -#define to_string(b, bsize) \ - memcpy(dest, b, bsize); \ - dest+=bsize; \ + 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); +#define to_char(c) CORD_ec_append(result, c) +#define to_string(s) CORD_ec_append_cord(result, s) inline bool need_file_encode(unsigned char c){ - // theoretical problem with, for instance, "_2B" and "." fragments, - // they would yield the same - // because need_file_encode('_')=false - // but we need to delete such files somehow, getting names from ^index - - if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) - return false; - - return !strchr( - " _./()-" -#ifdef WIN32 - ":\\~" + // russian letters and space ENABLED + // encoding only these... + return strchr( + "*?'\"<>|" +#ifndef WIN32 + ":\\~" #endif - , c); + , c)!=0; } inline bool need_uri_encode(unsigned char c){ if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) @@ -71,7 +70,7 @@ inline bool need_http_header_encode(unsi // -static const char * String_Untaint_lang_name[]={ +static const char* String_Untaint_lang_name[]={ "U", ///< zero value handy for hash lookup @see untaint_lang_name2enum "C", ///< clean "T", ///< tainted, untaint language as assigned later @@ -131,150 +130,115 @@ quoted-pair = "\" CHAR if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr)) */ -inline bool need_quote_http_header(const char *ptr, size_t size) { +inline bool need_quote_http_header(const char* ptr, size_t size) { for(; size--; ptr++) if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr)) return true; return false; } -String& String::append(const String& src, uchar lang, bool forced) { - // manually unrolled code to avoid do{if(const)} constructs - if(forced) - STRING_SRC_FOREACH_ROW( - APPEND(row->item.ptr, row->item.size, - lang, //forcing passed lang - row->item.origin.file, row->item.origin.line); - ) - else if(lang==UL_PASS_APPENDED) - STRING_SRC_FOREACH_ROW( - APPEND(row->item.ptr, row->item.size, - row->item.lang, // passing item's lang - row->item.origin.file, row->item.origin.line); - ) - else if(lang&UL_OPTIMIZE_BIT) // main idea here - // tainted piece would get OPTIMIZED bit from 'lang' - // clean piece would be marked OPTIMIZED manually - // pieces with determined languages [not tainted|clean] would retain theirs langs - STRING_SRC_FOREACH_ROW( - APPEND(row->item.ptr, row->item.size, - row->item.lang==UL_TAINTED?lang:( - row->item.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag - row->item.lang - ), - row->item.origin.file, row->item.origin.line); - ) - else - STRING_SRC_FOREACH_ROW( - APPEND(row->item.ptr, row->item.size, - row->item.lang==UL_TAINTED?lang:row->item.lang, - row->item.origin.file, row->item.origin.line); - ); -break2: - return *this; -} -size_t String::cstr_bufsize(Untaint_lang lang, - SQL_Connection *connection, - Charset *buf_charset) const { - size_t dest=1; // for terminating 0 - STRING_FOREACH_ROW( - uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; +/** + appends to other String, + marking all tainted pieces of it with @a lang. + 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, Untaint_lang lang, bool forced) const { + if(!*this) + return dest; - switch(to_lang & ~UL_OPTIMIZE_BIT) { - case UL_CLEAN: - case UL_TAINTED: - case UL_AS_IS: - // clean piece + // first: letters + dest.body=CORD_cat(dest.body, body); - // tainted piece, but undefined untaint language - // for VString.as_double of tainted values - // for ^process{body} evaluation - - // tainted, untaint language: as-is - dest+=row->item.size; - break; - case UL_FILE_SPEC: - // tainted, untaint language: file [name] - dest+=row->item.size*3/* worst: Z->%XX */; - break; - case UL_URI: - // tainted, untaint language: uri - dest+=row->item.size*6*3/* worst utf8 x worst Z->%XX */; - break; - case UL_HTTP_HEADER: - // tainted, untaint language: http-field-content-text - dest+=row->item.size*3/* worst: Z->%XX */; - break; - case UL_MAIL_HEADER: - // tainted, untaint language: mail-header - if(buf_charset) { - // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= - dest+= - row->item.size*3+ - buf_charset->name().size()+MAX_STRING/* worst: =?charset?Q?=%XX?= */; - } else - dest+=row->item.size; - break; - case UL_TABLE: - // tainted, untaint language: table - dest+=row->item.size; - break; - case UL_SQL: - // tainted, untaint language: sql - if(connection) - dest+=connection->quote(0, row->item.ptr, row->item.size); - break; - case UL_JS: - escape(switch(*src) { - case '"': case '\'': case '\n': case '\\': case '\xFF': - dest+=2; break; - default: - dest++; break; - }); - break; - case UL_XML: - escape(switch(*src) { - case '&': case '>': case '<': case '"': case '\'': - dest+= 6; break; - default: - dest++; break; - }); - break; - case UL_HTML: - escape(switch(*src) { - case '&': - case '>': - case '<': - case '"': - dest+=6; break; - default: - dest++; break; - }); - break; + // next: fragment infos + + if(lang==UL_PASS_APPENDED) // without language-chage string? + dest.fragments.append(fragments, 0, fragments.count()); + else if(forced) //forcing passed lang? + dest.fragments+=Fragment(lang, size()); + else if(lang&UL_OPTIMIZE_BIT) { + Array_iterator i(fragments); + while(i.has_next()) { + const Fragment& fragment=i.next(); + // main idea here: + // tainted piece would get OPTIMIZED bit from 'lang' + // clean piece would be marked OPTIMIZED manually + // pieces with determined languages [not tainted|clean] would retain theirs langs + dest.fragments+=Fragment(static_cast( + fragment.lang==UL_TAINTED?lang:( + fragment.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag + fragment.lang)), + fragment.size); } - ); -break2: + } 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(); + dest.fragments+=Fragment( + fragment.lang==UL_TAINTED?lang:fragment.lang, + fragment.size); + } + } + return dest; } -char *String::store_to(char *dest, Untaint_lang lang, - SQL_Connection *connection, - Charset *store_to_charset) const { - // WARNING: - // before any changes check cstr_bufsize first!!! +/** http://www.ietf.org/rfc/rfc2047.txt +RFC +(3) As a replacement for a 'word' entity within a 'phrase', for example, + one that precedes an address in a From, To, or Cc header. The ABNF + definition for 'phrase' from RFC 822 thus becomes: + + phrase = 1*( encoded-word / word ) + + In this case the set of characters that may be used in a "Q"-encoded + 'encoded-word' is restricted to: . An 'encoded-word' that appears within a + 'phrase' MUST be separated from any adjacent 'word', 'text' or + 'special' by 'linear-white-space'. +... + (2) The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) may be + represented as "_" (underscore, ASCII 95.). (This character may + not pass through some internetwork mail gateways, but its use + will greatly enhance readability of "Q" encoded data with mail + readers that do not support this encoding.) Note that the "_" + always represents hexadecimal 20, even if the SPACE character + occupies a different code position in the character set in use. + + paf: obviously, + without "=", or one could not differ "=E0" and "russian letter a" + 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); +} +inline void ec_append(CORD_ec result, CORD_pos& pos, size_t size) { + while(size--) { + CORD_ec_append(result, CORD_pos_fetch(pos)); + CORD_next(pos); + } +} +CORD String::cstr_to_cord(Untaint_lang lang, + SQL_Connection *connection, + const Request_charsets *charsets) const { + CORD_ec result; CORD_ec_init(result); + CORD_pos pos; CORD_set_pos(pos, body, 0); + bool whitespace=true; - // expanded STRING_FOREACH_ROW here for debugging purposes - const Chunk *chunk=&head; \ - do { \ - const Chunk::Row *row=chunk->rows; \ - for(uint i=0; icount; i++, row++) { \ - if(row==append_here) \ - goto break2; \ - \ - uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; - char *start=dest; + size_t fragment_start=0; + size_t fragment_end; + for(Array_iterator i(fragments); i.has_next(); fragment_start=fragment_end) { + const Fragment& fragment=i.next(); + size_t fragment_size=fragment.size; + fragment_end=fragment_start+fragment_size; + + Untaint_lang to_lang=lang==UL_UNSPECIFIED?fragment.lang:lang; switch(to_lang & ~UL_OPTIMIZE_BIT) { case UL_CLEAN: @@ -287,193 +251,179 @@ char *String::store_to(char *dest, Untai // for ^process{body} evaluation // tainted, untaint language: as-is - memcpy(dest, row->item.ptr, row->item.size); - dest+=row->item.size; + ec_append(result, pos, fragment_size); break; case UL_FILE_SPEC: // tainted, untaint language: file [name] escape( - encode(need_file_encode, '_'); + // Macintosh has problems with small Russian letter 'r' + if( c=='\xF0' && charsets && charsets->source().name()=="windows-1251" ) { + // fixing that letter for most common charset + to_char('p'); + } else // fallback to default + encode(need_file_encode, '_'); ); break; +#ifdef LATER case UL_URI: // tainted, untaint language: uri - const void *client_ptr; - size_t client_size; - Charset::transcode(pool(), - pool().get_source_charset(), row->item.ptr, row->item.size, - pool().get_client_charset(), client_ptr, client_size); { - const char *src=(const char *)client_ptr; - for(int size=client_size; size--; src++) - switch(*src) { - case ' ': to_char('+'); break; - default: encode(need_uri_encode, '%'); - }; + const char *substr= + CORD_to_const_char_star( + CORD_substr(body, fragment_start, fragment_size)); + const void *client_ptr; + size_t client_size; + if(charsets) + Charset::transcode( + charsets->source(), substr, fragment_size, + charsets->client(), client_ptr, client_size); + else { + client_ptr=substr; + client_size=fragment_size; + } + + char c; + for(const char* src=(const char* )client_ptr; c=*src++; ) + encode(need_uri_encode, '%'); } break; +#endif case UL_HTTP_HEADER: // tainted, untaint language: http-field-content-text - escape(switch(*src) { - case ' ': to_char('+'); break; - default: encode(need_uri_encode, '%'); - }); + escape( + encode(need_uri_encode, '%'); + ); break; +#ifdef LATER case UL_MAIL_HEADER: // tainted, untaint language: mail-header - if(store_to_charset) { - const void *mail_ptr; - size_t mail_size; - Charset::transcode(pool(), - pool().get_source_charset(), row->item.ptr, row->item.size, - *store_to_charset, mail_ptr, mail_size); + // http://www.ietf.org/rfc/rfc2047.txt + if(charsets) { + const void *mail_ptr=fragment.ptr; + size_t mail_size=fragment_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; + const char* src=(const char* )mail_ptr; bool to_quoted_printable=false; - for(int size=mail_size; size--; src++) { - if(*src & 0x80) { + + bool email=false; + for(const char* end=src+mail_size; srcname().cstr()); + dest+=sprintf(dest, "=?%s?Q?", charset_name.get()); to_quoted_printable=true; } - dest+=sprintf(dest, "=%02X", *src & 0xFF); - } else { - *dest++=*src; - } + //RFC Upper case should be used for hexadecimal digits "A" through "F" + if(c == 0x20) // RFC The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) + *dest++='_'; // RFC may be represented as "_" + else + dest+=sprintf(dest, "=%02X", c & 0xFF); + } else + *dest++=c; + if(c=='>') + email=false; } if(to_quoted_printable) // close dest+=sprintf(dest, "?="); } else { - memcpy(dest, row->item.ptr, row->item.size); - dest+=row->item.size; + memcpy(dest, fragment.ptr, fragment_size); + dest+=fragment_size; } break; +#endif case UL_TABLE: // tainted, untaint language: table - escape(switch(*src) { + escape(switch(c) { case '\t': to_char(' '); break; case '\n': to_char(' '); break; _default; }); break; +#ifdef PA_SQL case UL_SQL: // tainted, untaint language: sql - if(connection) - dest+=connection->quote(dest, row->item.ptr, row->item.size); + if(connection) { + dest+=connection->quote(0, + CORD_substr(body, fragment_start, fragment_size), + fragment_size); +...^^^ + + dest+=connection->quote(dest, fragment.ptr, fragment_size); else - throw Exception(0, 0, - this, + throw Exception(0, + 0, "untaint in SQL language failed - no connection specified"); break; +#endif case UL_JS: - escape(switch(*src) { - case '"': to_string("\\\"", 2); break; - case '\'': to_string("\\'", 2); break; - case '\n': to_string("\\n", 2); break; - case '\\': to_string("\\\\", 2); break; - case '\xFF': to_string("\\\xFF", 2); break; + escape(switch(c) { + case '"': to_string("\\\""); break; + case '\'': to_string("\\'"); break; + case '\n': to_string("\\n"); break; + case '\\': to_string("\\\\"); break; + case '\xFF': to_string("\\\xFF"); break; _default; }); break; case UL_XML: - escape(switch(*src) { - case '&': to_string("&", 5); break; - case '>': to_string(">", 4); break; - case '<': to_string("<", 4); break; - case '"': to_string(""", 6); break; - case '\'': to_string("'", 6); break; + escape(switch(c) { + case '&': to_string("&"); break; + case '>': to_string(">"); break; + case '<': to_string("<"); break; + case '"': to_string("""); break; + case '\'': to_string("'"); break; _default; }); break; case UL_HTML: - escape(switch(*src) { - case '&': to_string("&", 5); break; - case '>': to_string(">", 4); break; - case '<': to_string("<", 4); break; - case '"': to_string(""", 6); break; + escape(switch(c) { + case '&': to_string("&"); break; + case '>': to_string(">"); break; + case '<': to_string("<"); break; + case '"': to_string("""); break; _default; }); break; default: - throw Exception(0, 0, - this, + throw Exception(0, + 0, "unknown untaint language #%d", - static_cast(row->item.lang)); // sould never + static_cast(to_lang)); // should never break; // never } - +#ifdef LATER if(to_lang & UL_OPTIMIZE_BIT) { // optimizing whitespace char *stop=dest; dest=start; - for(char *src=start; srclink; \ - } while(chunk); \ - -break2: - return dest; -} - -char *String::cstr_debug_origins() const { - //_asm int 3; - char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2); - char *dest=result; - - const Chunk *chunk=&head; - do { - const Chunk::Row *row=chunk->rows; - for(uint i=0; icount; i++, row++) { - if(row==append_here) - goto break2; - -#ifndef NO_STRING_ORIGIN - if(row->item.origin.file) - dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT, - row->item.origin.file, - 1+row->item.origin.line); - else - dest+=sprintf(dest, ""); #endif - uchar show_lang=row->item.lang & ~UL_OPTIMIZE_BIT; - if(show_lang>=sizeof(String_Untaint_lang_name)/sizeof(String_Untaint_lang_name[0])) - throw Exception(0, 0, - this, - "unknown untaint language #%d", - static_cast(show_lang)); // sould never - - dest+=sprintf(dest, "#%s%s: ", - String_Untaint_lang_name[show_lang], - row->item.lang & UL_OPTIMIZE_BIT?".O":""); - char *dest_after_origins=dest; - - memcpy(dest, row->item.ptr, row->item.size); - dest+=row->item.size; - - remove_crlf(dest_after_origins, dest); - to_char('\n'); - } - chunk=row->link; - } while(chunk); + } -break2: - *dest=0; - return result; + return CORD_ec_to_cord(result); }