--- parser3/src/main/untaint.C 2003/03/20 08:11:15 1.115.2.12.2.3 +++ parser3/src/main/untaint.C 2003/03/20 14:19:10 1.115.2.12.2.6 @@ -5,7 +5,7 @@ Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char* IDENT_UNTAINT_C="$Date: 2003/03/20 08:11:15 $"; +static const char* IDENT_UNTAINT_C="$Date: 2003/03/20 14:19:10 $"; #include "pa_string.h" @@ -18,6 +18,10 @@ static const char* IDENT_UNTAINT_C="$Dat #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 @@ -25,26 +29,21 @@ static const char* IDENT_UNTAINT_C="$Dat #endif #define escape(action) \ - { \ - const char* src=fragment->ptr; \ - for(int size=fragment->size; size--; src++) \ - action \ + for(; fragment_length--; 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){ // russian letters and space ENABLED @@ -157,7 +156,7 @@ String& String::append_to(String& dest, 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()); + dest.fragments+=Fragment(lang, length()); else if(lang&UL_OPTIMIZE_BIT) { Array_iterator i(fragments); while(i.has_next()) { @@ -170,7 +169,7 @@ String& String::append_to(String& dest, fragment.lang==UL_TAINTED?lang:( fragment.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag fragment.lang)), - fragment.size); + fragment.length); } } else { // The core idea: tainted pieces got marked with context's lang Array_iterator i(fragments); @@ -178,113 +177,13 @@ String& String::append_to(String& dest, const Fragment& fragment=i.next(); dest.fragments+=Fragment( fragment.lang==UL_TAINTED?lang:fragment.lang, - fragment.size); + fragment.length); } } return dest; } -size_t String::cstr_bufsize(Untaint_lang lang, - SQL_Connection *connection, - const Request_charsets *charsets) const { - size_t dest=1; // for terminating 0 - 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(); - 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: - case UL_TAINTED: - case UL_AS_IS: - // clean piece - - // tainted piece, but undefined untaint language - // for VString.as_double of tainted values - // for ^process{body} evaluation - - // tainted, untaint language: as-is - dest+=fragment.size; - break; - case UL_FILE_SPEC: - // tainted, untaint language: file [name] - dest+=fragment.size*3/* worst: Z->%XX */; - break; - case UL_URI: - // tainted, untaint language: uri - dest+=fragment.size*6*3/* worst utf8 x worst Z->%XX */; - break; - case UL_HTTP_HEADER: - // tainted, untaint language: http-field-content-text - dest+=fragment.size*3/* worst: Z->%XX */; - break; - case UL_MAIL_HEADER: - // tainted, untaint language: mail-header - if(charsets) { - int parts_count=1; - size_t size=fragment.size; - const char* prev=CORD_substr(body, fragment_start, 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+= - fragment.size*3+ - parts_count*(charsets->mail().name().size()+MAX_STRING/* worst: =?charset?Q?=%XX?= */); - } else - dest+=fragment.size; - break; - case UL_TABLE: - // tainted, untaint language: table - dest+=fragment.size; - break; -#ifdef PA_SQL - case UL_SQL: - // tainted, untaint language: sql - if(connection) - dest+=connection->quote(0, - CORD_substr(body, fragment_start, fragment.size), - fragment.size); - break; -#endif - 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; - } - } - return dest; -} - /** http://www.ietf.org/rfc/rfc2047.txt RFC (3) As a replacement for a 'word' entity within a 'phrase', for example, @@ -312,24 +211,34 @@ RFC 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) { +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); } -char *String::store_to(char *dest, Untaint_lang lang, - SQL_Connection *connection, - const Request_charsets *charsets) const { - // WARNING: - // before any changes check cstr_bufsize first!!! +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; -// STRING_FOREACH_FRAGMENT( - element_type *last = felements+fused; element_type *fragment = felements; for(; fragmentlang:lang; + 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_length=fragment.length; + fragment_end=fragment_start+fragment_length; - char *start=dest; + Untaint_lang to_lang=lang==UL_UNSPECIFIED?fragment.lang:lang; switch(to_lang & ~UL_OPTIMIZE_BIT) { case UL_CLEAN: @@ -342,51 +251,56 @@ char *String::store_to(char *dest, Untai // for ^process{body} evaluation // tainted, untaint language: as-is - memcpy(dest, fragment->ptr, fragment->size); - dest+=fragment->size; + ec_append(result, pos, fragment_length); break; case UL_FILE_SPEC: // tainted, untaint language: file [name] escape( // Macintosh has problems with small Russian letter 'r' - if( *src=='\xF0' && charsets && *charsets->source().name()=="windows-1251" ) { + 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; - 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++) + const char *substr= + CORD_to_const_char_star( + CORD_substr(body, fragment_start, fragment_length)); + const void *client_ptr; + size_t client_size; + if(charsets) + Charset::transcode( + charsets->source(), substr, fragment_length, + charsets->client(), client_ptr, client_size); + else { + client_ptr=substr; + client_size=fragment_length; + } + + 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( encode(need_uri_encode, '%'); ); break; +#ifdef LATER case UL_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->size; + const void *mail_ptr=fragment.ptr; + size_t mail_size=fragment_length; const char* charset_name=charsets->mail().name()->cstr(); // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= @@ -396,80 +310,88 @@ char *String::store_to(char *dest, Untai bool email=false; for(const char* end=src+mail_size; src') + *dest++=c; + if(c=='>') email=false; } if(to_quoted_printable) // close dest+=sprintf(dest, "?="); } else { - memcpy(dest, fragment->ptr, fragment->size); - dest+=fragment->size; + memcpy(dest, fragment.ptr, fragment_length); + dest+=fragment_length; } 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, fragment->ptr, fragment->size); + if(connection) { + dest+=connection->quote(0, + CORD_substr(body, fragment_start, fragment_length), + fragment_length); +...^^^ + + dest+=connection->quote(dest, fragment.ptr, fragment_length); else 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; @@ -480,65 +402,28 @@ char *String::store_to(char *dest, Untai 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; srcorigin.file) - dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT, - fragment->origin.file, - 1+fragment->origin.line); - else - dest+=sprintf(dest, ""); -); - Untaint_lang show_lang=fragment->lang & ~UL_OPTIMIZE_BIT; - if(show_lang>=sizeof(String_Untaint_lang_name)/sizeof(String_Untaint_lang_name[0])) - throw Exception(0, - 0, - "unknown untaint language #%d", - static_cast(show_lang)); // sould never - - dest+=sprintf(dest, "#%s%s: ", - String_Untaint_lang_name[show_lang], - fragment->lang & UL_OPTIMIZE_BIT?".O":""); - char *dest_after_origins=dest; - - memcpy(dest, fragment->ptr, fragment->size); - dest+=fragment->size; - - remove_crlf(dest_after_origins, dest); - to_char('\n'); - ); - - *dest=0; - return result; + return CORD_ec_to_cord(result); }