--- parser3/src/main/untaint.C 2002/02/20 10:40:08 1.92 +++ parser3/src/main/untaint.C 2003/03/17 14:48:06 1.115.4.1 @@ -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 1.92 2002/02/20 10:40:08 paf Exp $ */ +static const char* IDENT_UNTAINT_C="$Date: 2003/03/17 14:48:06 $"; + #include "pa_pool.h" #include "pa_string.h" #include "pa_hash.h" @@ -18,13 +18,12 @@ #include "pa_common.h" #include "pa_charset.h" -#define DEBUG_STRING_APPENDS_VS_EXPANDS +//#define DEBUG_STRING_APPENDS_VS_EXPANDS #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS ulong string_string_shortcut_economy=0; #endif - #define escape(action) \ { \ const char *src=row->item.ptr; \ @@ -48,20 +47,14 @@ ulong string_string_shortcut_economy=0; 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')) @@ -156,21 +149,66 @@ inline bool need_quote_http_header(const algorithm: if no language-change specified and src not yet appended to some other string[last_chunk!=0] shrinking dest last_chunk[preparing it for linking], - shrinking src last_chunk[preparing it to be linked, consequent dest.appends would go there], - linking[dest.last_chunk = src.head] + ///shrinking src last_chunk[preparing it to be linked, consequent dest.appends would go there], + linking[dest.last_chunk = src.head.chunk] if some language-change specified or src already appended to some other string[last_chunk==0] cloning pieces. */ String& String::append(const String& src, uchar lang, bool forced) { - if(lang==UL_PASS_APPENDED && src.last_chunk) { + if(!last_chunk) // growth stopped [we're appended as string to somebody] + throw Exception(0, + this, + "string growth stopped (append string)"); + + if(src.is_empty()) + return *this; + + // without language-chage, not-appended-before, big[not fitting our tail] string? + if(lang==UL_PASS_APPENDED + && src.last_chunk + && (uint(&last_chunk->rows[last_chunk->count]-append_here) < src.used_rows())) { #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS string_string_shortcut_economy+=src.used_rows()*sizeof(String::Chunk::Row); #endif -// if(!head.rows[head.count].link && + // using fact: + // src.head.chunk.count initally equeals this.head.chunk.count and shrinks-only, + // so can't be more than this.head.chunk.count, + // which means that we know that + // src.head.chunk would fit into this.head.chunk + if(is_empty()) { // our head.chunk is empty + // they have more than head.chunk? we need all head.chunk : we need only filled-part of head.chunk + Chunk *src_head_link=src.head.chunk.rows[src.head.chunk.count].link; + size_t head_count=src_head_link?src.head.chunk.count:(src.append_here-src.head.chunk.rows); + // "your head.chunk is my head.chunk" + memcpy(head.chunk.rows, src.head.chunk.rows, sizeof(Chunk::Row)*(head_count)); + if(src_head_link) { + // "your body is my body" + head.chunk.rows[head.chunk.count=head_count].link=src_head_link; + // "your last_chunk is mine now" + last_chunk=src.last_chunk; + // "your append_here is mine now" + append_here=src.append_here; + } else { + // "your last_chunk is mine now" + last_chunk=&head.chunk; + // "your append_here is recalc-mine now" + append_here=head.chunk.rows+head_count; + } + } else { // our head.chunk contains something + // "chopping off my tail-reserve" + last_chunk->count=append_here-last_chunk->rows; + // "you is my tail" + append_here->link=&src.head.chunk; + // "your last_chunk is mine now" + last_chunk=src.last_chunk; + // "your append_here is mine now" + append_here=src.append_here; + } - src.last_chunk=0; // stop growing - //return; + // stop-growing mark + src.last_chunk=0; + return *this; } // manually unrolled code to avoid do{if(const)} constructs @@ -204,7 +242,10 @@ String& String::append(const String& src row->item.lang==UL_TAINTED?lang:row->item.lang, row->item.origin.file, row->item.origin.line); ); -break2: +/* + for(Chunk::Row *row=last_chunk->rows; rowlink==(void*)0xcdcdcdcd) + _asm int 3;*/ return *this; } @@ -243,10 +284,18 @@ size_t String::cstr_bufsize(Untaint_lang case UL_MAIL_HEADER: // tainted, untaint language: mail-header if(buf_charset) { + int parts_count=1; + const char *prev=row->item.ptr; + size_t size=row->item.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?= */; + parts_count*(buf_charset->name().size()+MAX_STRING/* worst: =?charset?Q?=%XX?= */); } else dest+=row->item.size; break; @@ -288,24 +337,62 @@ 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 { + Charset *store_to_charset, + const char *store_to_charset_name) 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; \ - \ +//debug + const Chunk *chunk=&(*this).head.chunk; \ + const Chunk::Row *row=chunk->rows; \ + uint countdown=chunk->count; \ + while(row!=(*this).append_here) { \ + if(countdown==0) { \ + chunk=row->link; \ + row=chunk->rows; \ + countdown=chunk->count; \ + }; +//*/ +// STRING_FOREACH_ROW( + uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; char *start=dest; @@ -327,7 +414,12 @@ char *String::store_to(char *dest, Untai 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' && pool().get_source_charset().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: @@ -340,22 +432,19 @@ char *String::store_to(char *dest, Untai { 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(store_to_charset && store_to_charset_name) { const void *mail_ptr; size_t mail_size; Charset::transcode(pool(), @@ -365,16 +454,32 @@ char *String::store_to(char *dest, Untai // 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?", store_to_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, "?="); @@ -397,7 +502,7 @@ char *String::store_to(char *dest, Untai if(connection) dest+=connection->quote(dest, row->item.ptr, row->item.size); else - throw Exception(0, 0, + throw Exception(0, this, "untaint in SQL language failed - no connection specified"); break; @@ -431,10 +536,10 @@ char *String::store_to(char *dest, Untai }); break; default: - throw Exception(0, 0, + throw Exception(0, this, "unknown untaint language #%d", - static_cast(row->item.lang)); // sould never + static_cast(to_lang)); // should never break; // never } @@ -456,58 +561,49 @@ char *String::store_to(char *dest, Untai break; }; } else // piece without optimization +//debug whitespace=false; - - } \ - chunk=row->link; \ - } while(chunk); \ - -break2: + row++; countdown--; \ + } +/* + ); +*/ return dest; } char *String::cstr_debug_origins() const { //_asm int 3; - char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2); + char *result=(char *)malloc_atomic(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; + STRING_FOREACH_ROW( +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, ""); +); + 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, + 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], + row->item.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, row->item.ptr, row->item.size); + dest+=row->item.size; + + remove_crlf(dest_after_origins, dest); + to_char('\n'); + ); -break2: *dest=0; return result; }