--- parser3/src/main/untaint.C 2002/02/08 08:31:32 1.89 +++ parser3/src/main/untaint.C 2003/01/29 11:06:44 1.115.2.3 @@ -1,12 +1,12 @@ /** @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 1.89 2002/02/08 08:31:32 paf Exp $ */ +static const char* IDENT_UNTAINT_C="$Date: 2003/01/29 11:06:44 $"; + #include "pa_pool.h" #include "pa_string.h" #include "pa_hash.h" @@ -20,8 +20,8 @@ #define escape(action) \ { \ - const char *src=row->item.ptr; \ - for(int size=row->item.size; size--; src++) \ + const char *src=fragment->ptr; \ + for(int size=fragment->size; size--; src++) \ action \ } #define _default default: *dest++=*src; break @@ -41,20 +41,14 @@ dest+=bsize; \ 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')) @@ -138,48 +132,66 @@ inline bool need_quote_http_header(const return false; } -String& String::append(const String& src, uchar lang, bool forced) { +//#include "pa_sapi.h" +/** + 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, uchar lang, bool forced) const { + if(is_empty()) + return dest; + + // without language-chage, small[fitting dest tail] string? + if(lang==UL_PASS_APPENDED && fused<=(dest.fallocated-dest.fused)) { + // felements[fused...]=src.felements[0..fused], fused+=src.fused + memcpy(&dest.felements[dest.fused], felements, sizeof(felements[0])*fused); + dest.fused+=fused; + return dest; + } + // manually unrolled code to avoid do{if(const)} constructs if(forced) - STRING_SRC_FOREACH_ROW( - APPEND(row->item.ptr, row->item.size, + STRING_FOREACH_FRAGMENT( + dest.APPEND(fragment->ptr, fragment->size, lang, //forcing passed lang - row->item.origin.file, row->item.origin.line); + fragment->origin.file, fragment->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); + STRING_FOREACH_FRAGMENT( + dest.APPEND(fragment->ptr, fragment->size, + fragment->lang, // passing item's lang + fragment->origin.file, fragment->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 + STRING_FOREACH_FRAGMENT( + dest.APPEND(fragment->ptr, fragment->size, + fragment->lang==UL_TAINTED?lang:( + fragment->lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag + fragment->lang ), - row->item.origin.file, row->item.origin.line); + fragment->origin.file, fragment->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); + STRING_FOREACH_FRAGMENT( + dest.APPEND(fragment->ptr, fragment->size, + fragment->lang==UL_TAINTED?lang:fragment->lang, + fragment->origin.file, fragment->origin.line); ); -break2: - return *this; + + return dest; } size_t String::cstr_bufsize(Untaint_lang lang, SQL_Connection *connection, - Charset *buf_charset) const { + const Request_charsets *charsets) const { size_t dest=1; // for terminating 0 - STRING_FOREACH_ROW( - uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; + STRING_FOREACH_FRAGMENT( + uchar to_lang=lang==UL_UNSPECIFIED?fragment->lang:lang; switch(to_lang & ~UL_OPTIMIZE_BIT) { case UL_CLEAN: @@ -192,38 +204,46 @@ size_t String::cstr_bufsize(Untaint_lang // for ^process{body} evaluation // tainted, untaint language: as-is - dest+=row->item.size; + dest+=fragment->size; break; case UL_FILE_SPEC: // tainted, untaint language: file [name] - dest+=row->item.size*3/* worst: Z->%XX */; + dest+=fragment->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 */; + dest+=fragment->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 */; + dest+=fragment->size*3/* worst: Z->%XX */; break; case UL_MAIL_HEADER: // tainted, untaint language: mail-header - if(buf_charset) { + if(charsets && charsets->mail) { + int parts_count=1; + const char *prev=fragment->ptr; + size_t size=fragment->size; + while(const char *comma_at=(const char*)memchr(prev, ',', size)) { + parts_count++; + size-=comma_at-prev; + prev=comma_at+1; + } // 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?= */; + fragment->size*3+ + parts_count*(charsets->mail->name()->size()+MAX_STRING/* worst: =?charset?Q?=%XX?= */); } else - dest+=row->item.size; + dest+=fragment->size; break; case UL_TABLE: // tainted, untaint language: table - dest+=row->item.size; + dest+=fragment->size; break; case UL_SQL: // tainted, untaint language: sql if(connection) - dest+=connection->quote(0, row->item.ptr, row->item.size); + dest+=connection->quote(0, fragment->ptr, fragment->size); break; case UL_JS: escape(switch(*src) { @@ -254,25 +274,52 @@ size_t String::cstr_bufsize(Untaint_lang break; } ); -break2: return dest; } +/** 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 +*/ +static bool mail_header_char_valid_within_Qencoded(char c) { + return c>='A' && c<='Z' + || c>='a' && c<='Z' + || c>='0' && c<='9' + || strchr("!*+-/", c); +} char *String::store_to(char *dest, Untaint_lang lang, SQL_Connection *connection, - Charset *store_to_charset) const { + const Request_charsets *charsets) const { // WARNING: // before any changes check cstr_bufsize first!!! 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; +// STRING_FOREACH_FRAGMENT( + element_type *last = felements+fused; element_type *fragment = felements; for(; fragmentlang:lang; char *start=dest; @@ -287,67 +334,91 @@ 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; + memcpy(dest, fragment->ptr, fragment->size); + dest+=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( *src=='\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; 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); + if(charsets) + Charset::transcode(*charsets->pool, + *charsets->source, fragment->ptr, fragment->size, + *charsets->client, client_ptr, client_size); + else { + client_ptr=fragment->ptr; + client_size=fragment->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, '%'); - }; + encode(need_uri_encode, '%'); } break; 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; case UL_MAIL_HEADER: // tainted, untaint language: mail-header - if(store_to_charset) { + // http://www.ietf.org/rfc/rfc2047.txt + if(charsets && charsets->mail) { 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); + Charset::transcode(*charsets->pool, + *charsets->source, fragment->ptr, fragment->size, + *charsets->mail, mail_ptr, mail_size); // 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; - 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); 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(*src == 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", *src & 0xFF); + } else + *dest++=*src; + if(*src=='>') + 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; case UL_TABLE: @@ -361,10 +432,10 @@ char *String::store_to(char *dest, Untai case UL_SQL: // tainted, untaint language: sql if(connection) - dest+=connection->quote(dest, row->item.ptr, row->item.size); + dest+=connection->quote(dest, fragment->ptr, fragment->size); else - throw Exception(0, 0, - this, + throw Exception(0, + ConstStringPtr(this), "untaint in SQL language failed - no connection specified"); break; case UL_JS: @@ -397,10 +468,10 @@ char *String::store_to(char *dest, Untai }); break; default: - throw Exception(0, 0, - this, + throw Exception(0, + ConstStringPtr(this), "unknown untaint language #%d", - static_cast(row->item.lang)); // sould never + static_cast(to_lang)); // should never break; // never } @@ -423,57 +494,45 @@ char *String::store_to(char *dest, Untai }; } else // piece without optimization whitespace=false; +// ); + } - } \ - chunk=row->link; \ - } while(chunk); \ - -break2: return dest; } -char *String::cstr_debug_origins() const { +char *String::cstr_debug_origins(Pool& pool) const { //_asm int 3; - char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2); + char *result=new(pool) char[size()+fused*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; + STRING_FOREACH_FRAGMENT( +IFNDEF_NO_STRING_ORIGIN( + if(fragment->origin.file) + dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT, + fragment->origin.file, + 1+fragment->origin.line); + else + dest+=sprintf(dest, ""); +); + uchar show_lang=fragment->lang & ~UL_OPTIMIZE_BIT; + if(show_lang>=sizeof(String_Untaint_lang_name)/sizeof(String_Untaint_lang_name[0])) + throw Exception(0, + ConstStringPtr(this), + "unknown untaint language #%d", + static_cast(show_lang)); // sould never - memcpy(dest, row->item.ptr, row->item.size); - dest+=row->item.size; + dest+=sprintf(dest, "#%s%s: ", + String_Untaint_lang_name[show_lang], + fragment->lang & UL_OPTIMIZE_BIT?".O":""); + char *dest_after_origins=dest; - remove_crlf(dest_after_origins, dest); - to_char('\n'); - } - chunk=row->link; - } while(chunk); + memcpy(dest, fragment->ptr, fragment->size); + dest+=fragment->size; + + remove_crlf(dest_after_origins, dest); + to_char('\n'); + ); -break2: *dest=0; return result; }