|
|
| version 1.51, 2001/05/19 19:52:38 | version 1.149, 2009/07/07 05:48:05 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: String class part: untaint mechanizm. | Parser: String class part: untaint mechanizm. |
| Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) | Copyright(c) 2001-2005 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | |
| */ | |
| Author: Alexander Petrosyan <paf@design.ru>(http://design.ru/paf) | static const char * const IDENT_UNTAINT_C="$Date$"; |
| $Id$ | |
| */ | |
| #include "pa_pool.h" | |
| #include "pa_string.h" | #include "pa_string.h" |
| #include "pa_hash.h" | #include "pa_hash.h" |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_table.h" | #include "pa_table.h" |
| #include "pa_globals.h" | #include "pa_globals.h" |
| #include "pa_dictionary.h" | |
| #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" | |
| } | |
| #include "pa_sql_connection.h" | #include "pa_sql_connection.h" |
| #define escape(action) \ | // defines |
| { \ | |
| const char *src=row->item.ptr; \ | #undef CORD_ec_append |
| for(int size=row->item.size; size--; src++) \ | // redefining to intercept flushes and implement whitespace optimization |
| action \ | // of all consequent white space chars leaving only first one |
| #define CORD_ec_append(x, c) \ | |
| { \ | |
| bool skip=false; \ | |
| if(optimize) switch(c) { \ | |
| case ' ': case '\n': case '\t': \ | |
| if(whitespace) \ | |
| skip=true; /*skipping subsequent*/ \ | |
| else \ | |
| whitespace=true; \ | |
| break; \ | |
| default: \ | |
| whitespace=false; \ | |
| break; \ | |
| } \ | |
| if(!skip) { \ | |
| if ((x)[0].ec_bufptr == (x)[0].ec_buf + CORD_BUFSZ) { \ | |
| CORD_ec_flush_buf(x); \ | |
| } \ | |
| *((x)[0].ec_bufptr)++ = (c); \ | |
| } \ | |
| } | |
| #define escape_fragment(action) \ | |
| for(; fragment_length--; CORD_next(info->pos)) { \ | |
| char c=CORD_pos_fetch(info->pos); \ | |
| action \ | |
| } | } |
| #define _default default: *dest++=*src; break | #define _default CORD_ec_append(info->result, c) |
| #define encode(need_encode_func, prefix) \ | #define encode(need_encode_func, prefix, otherwise) \ |
| default: \ | if(need_encode_func(c)) { \ |
| if(need_encode_func(*src)) { \ | static const char* hex="0123456789ABCDEF"; \ |
| static const char *hex="0123456789ABCDEF"; \ | CORD_ec_append(info->result, prefix); \ |
| char chunk[3]={prefix}; \ | CORD_ec_append(info->result, hex[((unsigned char)c)/0x10]); \ |
| chunk[1]=hex[((unsigned char)*src)/0x10]; \ | CORD_ec_append(info->result, hex[((unsigned char)c)%0x10]); \ |
| chunk[2]=hex[((unsigned char)*src)%0x10]; \ | } else \ |
| strncpy(dest, chunk, 3); dest+=3; \ | CORD_ec_append(info->result, otherwise); |
| } else \ | #define to_char(c) { CORD_ec_append(info->result, c); whitespace=false; } |
| *dest++=*src; \ | #define to_string(s) { CORD_ec_append_cord(info->result, s); whitespace=false; } |
| break | |
| #define to_char(c) *dest++=c | |
| #define to_string(b, bsize) \ | |
| strncpy(dest, b, bsize); \ | |
| dest+=bsize; \ | |
| inline bool need_file_encode(unsigned char c){ | inline bool need_file_encode(unsigned char c){ |
| if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) | // russian letters and space ENABLED |
| return false; | // encoding only these... |
| return strchr( | |
| return !strchr( | "*?'\"<>|" |
| #ifdef WIN32 | #ifndef WIN32 |
| ":\\~" | ":\\" |
| #endif | #endif |
| "./()_-", c); | , c)!=0; |
| } | } |
| inline bool need_uri_encode(unsigned char c){ | inline bool need_uri_encode(unsigned char c){ |
| if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) | if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) |
| return false; | return false; |
| return !strchr("_-./", c); | return !strchr("_-./", c); |
| } | } |
| inline bool need_http_header_encode(unsigned char c){ | inline bool need_http_header_encode(unsigned char c){ |
| if(strchr(" , :", c)) | if(strchr(" , :", c)) |
| return false; | return false; |
| return need_uri_encode(c); | return need_uri_encode(c); |
| } | } |
| // String | inline bool need_regex_escape(unsigned char c){ |
| return strchr("\\^$.[]|()?*+{}-", c)!=0; | |
| static bool typo_present(Array::Item *value, const void *info) { | |
| Array *row=static_cast<Array *>(value); | |
| const char *src=static_cast<const char *>(info); | |
| int partial; | |
| row->get_string(0)->cmp(partial, src); | |
| return | |
| partial==0 || // full match | |
| partial==1; // typo left column starts 'src' | |
| } | } |
| /* | // String |
| /* | |
| HTTP-header = field-name ":" [ field-value ] CRLF | HTTP-header = field-name ":" [ field-value ] CRLF |
| field-name = token | field-name = token |
| Line 110 quoted-pair = "\" CHAR | Line 135 quoted-pair = "\" CHAR |
| if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr)) | 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++) | for(; size--; ptr++) |
| if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr)) | if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr)) |
| return true; | return true; |
| return false; | return false; |
| } | } |
| /// @todo maybe additional check "are all pieces are clean?" would be profitable? | #ifndef DOXYGEN |
| size_t String::cstr_bufsize(Untaint_lang lang) const { | struct Append_fragment_info { |
| return (lang==UL_AS_IS?size():size()*UNTAINT_TIMES_BIGGER) +1; | String::Language lang; |
| } | String::Languages* dest_languages; |
| size_t dest_body_plan_length; | |
| /// @todo fix theoretical \n mem overrun in TYPO replacements | }; |
| char *String::store_to(char *dest, Untaint_lang lang, | #endif |
| SQL_Connection *connection, | int append_fragment_optimizing(char alang, size_t asize, Append_fragment_info* info) { |
| const char *charset) const { | const String::Language lang=(String::Language)(unsigned char)alang; |
| // $MAIN:html-typo table | // main idea here: |
| Table *user_typo_table=static_cast<Table *>(pool().tag()); | // tainted piece would get OPTIMIZED bit from 'lang' |
| Table *typo_table=user_typo_table?user_typo_table:default_typo_table; | // clean piece would be marked OPTIMIZED manually |
| // pieces with determined languages [not tainted|clean] would retain theirs langs | |
| bool whitespace=true; | info->dest_languages->append(info->dest_body_plan_length, |
| const Chunk *chunk=&head; | lang==String::L_TAINTED? |
| do { | info->lang |
| const Chunk::Row *row=chunk->rows; | :lang==String::L_CLEAN? |
| for(size_t i=0; i<chunk->count; i++, row++) { | (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT) // ORing with OPTIMIZED flag |
| if(row==append_here) | :lang, |
| goto break2; | asize); |
| info->dest_body_plan_length+=asize; | |
| // WARNING: | |
| // string can grow only UNTAINT_TIMES_BIGGER | return 0; // 0=continue |
| switch(lang==UL_UNSPECIFIED?row->item.lang:lang) { | } |
| case UL_CLEAN: | int append_fragment_nonoptimizing(char alang, size_t asize, Append_fragment_info* info) { |
| // clean piece | const String::Language lang=(String::Language)(unsigned char)alang; |
| { // optimizing whitespace | // The core idea: tainted pieces got marked with context's lang |
| const char *src=row->item.ptr; | info->dest_languages->append(info->dest_body_plan_length, |
| for(int size=row->item.size; size--; src++) | lang==String::L_TAINTED? |
| switch(*src) { | info->lang |
| case ' ': case '\n': case '\r': case '\t': | :lang, |
| if(!whitespace) { | asize); |
| *dest++=*src; | info->dest_body_plan_length+=asize; |
| whitespace=true; | |
| } | return 0; // 0=continue |
| break; | } |
| default: | |
| whitespace=false; | /** |
| *dest++=*src; | appends to other String, |
| break; | 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. |
| break; | */ |
| case UL_TAINTED: | String& String::append_to(String& dest, Language ilang, bool forced) const { |
| // tainted piece, but undefined untaint language | if(is_empty()) |
| // for VString.as_double of tainted values | return dest; |
| // for ^process{body} evaluation | |
| case UL_AS_IS: | // first: fragment infos |
| // tainted, untaint language: as-is | |
| memcpy(dest, row->item.ptr, row->item.size); | if(ilang==L_PASS_APPENDED) // without language-change? |
| dest+=row->item.size; | dest.langs.appendHelper(dest.body, langs, body); |
| break; | else if(forced) //forcing passed lang? |
| case UL_FILE_NAME: | dest.langs.appendHelper(dest.body, ilang, body); |
| // tainted, untaint language: file [name] | else { |
| escape(switch(*src) { | if(langs.opt.is_not_just_lang){ |
| case ' ': to_char('_'); break; | Append_fragment_info info={ilang, &dest.langs, dest.body.length()}; |
| encode(need_file_encode, '+'); | langs.for_each(body, ilang&L_OPTIMIZE_BIT? |
| }); | append_fragment_optimizing |
| break; | :append_fragment_nonoptimizing, &info); |
| case UL_URI: | } else { |
| // tainted, untaint language: uri | Language lang=langs.opt.lang; |
| escape(switch(*src) { | // see append_fragment_* for explanation |
| case ' ': to_char('+'); break; | if(ilang&L_OPTIMIZE_BIT){ |
| encode(need_uri_encode, '%'); | dest.langs.appendHelper(dest.body, |
| }); | lang==String::L_TAINTED? |
| break; | ilang |
| case UL_HTTP_HEADER: | :lang==String::L_CLEAN? |
| // tainted, untaint language: http-header | (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT) |
| if(need_quote_http_header(row->item.ptr, row->item.size)) { | :lang, |
| *dest++='\"'; | body); |
| escape(switch(*src) { | } else { |
| case '\"': to_string("\\\"", 2); break; | dest.langs.appendHelper(dest.body, lang==String::L_TAINTED ? ilang:lang, body); |
| _default; | } |
| }); | } |
| *dest++='\"'; | } |
| } else { | |
| memcpy(dest, row->item.ptr, row->item.size); | // next: letters |
| dest+=row->item.size; | dest.body<<body; |
| ASSERT_STRING_INVARIANT(dest); | |
| 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: <upper and lower case ASCII | |
| letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_" | |
| (underscore, ASCII 95.)>. 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 bool addr_spec_soon(const char *src) { | |
| for(char c; (c=*src); src++) | |
| if(c=='<') | |
| return true; | |
| else if(!(c==' ' || c=='\t')) | |
| return false; | |
| return false; | |
| } | |
| /** | |
| 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)); | |
| CORD_next(pos); | |
| } | |
| } | |
| inline void pa_CORD_pos_advance(CORD_pos pos, size_t n) { | |
| while(true) { | |
| long avail=CORD_pos_chars_left(pos); | |
| if(avail<=0) { | |
| CORD_next(pos); | |
| if(!--n) | |
| break; | |
| } else if((size_t)avail<n) { | |
| CORD_pos_advance(pos, avail); | |
| n-=avail; | |
| } else { // avail>=n | |
| CORD_pos_advance(pos, n); | |
| break; | |
| } | |
| } | |
| } | |
| #ifndef DOXYGEN | |
| struct Cstr_to_string_body_block_info { | |
| // input | |
| String::Language lang; | |
| SQL_Connection* connection; | |
| const Request_charsets* charsets; | |
| const String::Body* body; | |
| // output | |
| CORD_ec result; | |
| // private | |
| CORD_pos pos; | |
| size_t fragment_begin; | |
| bool whitespace; | |
| const char* exception; | |
| }; | |
| #endif | |
| int cstr_to_string_body_block(String::Language to_lang, size_t fragment_length, Cstr_to_string_body_block_info* info) { | |
| bool& whitespace=info->whitespace; | |
| size_t fragment_end=info->fragment_begin+fragment_length; | |
| //fprintf(stderr, "%d, %d =%s=\n", to_lang, fragment_length, info->body->cstr()); | |
| bool optimize=(to_lang & String::L_OPTIMIZE_BIT)!=0; | |
| if(!optimize) | |
| whitespace=false; | |
| switch(to_lang & ~String::L_OPTIMIZE_BIT) { | |
| case String::L_CLEAN: | |
| case String::L_TAINTED: | |
| case String::L_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 | |
| ec_append(info->result, optimize, whitespace, info->pos, fragment_length); | |
| break; | |
| case String::L_FILE_SPEC: | |
| // tainted, untaint language: file [name] | |
| { | |
| bool is1251=(info->charsets && info->charsets->source().NAME()=="WINDOWS-1251"); | |
| escape_fragment( | |
| // Macintosh has problems with small Russian letter 'r' | |
| if( is1251 && c=='\xF0' ) { | |
| // fixing that letter for most common charset | |
| to_char('p'); | |
| } else // fallback to default | |
| encode(need_file_encode, '_', c); | |
| ); | |
| } | |
| break; | |
| case String::L_FILE_POST: | |
| { | |
| escape_fragment(switch(c) { | |
| case '\0': to_string("\\0"); break; | |
| case '\\': to_string("\\\\"); break; | |
| default: _default; break; | |
| }); | |
| } | |
| break; | |
| case String::L_URI: | |
| // tainted, untaint language: uri | |
| { | |
| const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr(); | |
| // skip source [we use recoded version] | |
| pa_CORD_pos_advance(info->pos, fragment_length); | |
| String::C output(fragment_str, fragment_length); | |
| if(info->charsets) | |
| output=Charset::transcode(output, | |
| info->charsets->source(), | |
| info->charsets->client()); | |
| char c; | |
| for(const char* src=output.str; (c=*src++); ) | |
| encode(need_uri_encode, '%', c); | |
| } | |
| break; | |
| case String::L_HTTP_HEADER: | |
| // tainted, untaint language: http-field-content-text | |
| // the same as L_URI BUT not transcoded into $response:charset before encoding | |
| escape_fragment( | |
| encode(need_uri_encode, '%', c); | |
| ); | |
| break; | |
| case String::L_MAIL_HEADER: | |
| // tainted, untaint language: mail-header | |
| // http://www.ietf.org/rfc/rfc2047.txt | |
| if(info->charsets) { | |
| size_t mail_size; | |
| const char *mail_ptr= | |
| info->body->mid(info->fragment_begin, mail_size=fragment_length).cstr(); | |
| // skip source [we use recoded version] | |
| pa_CORD_pos_advance(info->pos, mail_size); | |
| const char* charset_name=info->charsets->mail().NAME().cstr(); | |
| // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= | |
| bool to_quoted_printable=false; | |
| bool email=false; | |
| uchar c; | |
| for(const char* src=mail_ptr; (c=(uchar)*src++); ) { | |
| if(c=='\r' || c=='\n') | |
| c=' '; | |
| if(to_quoted_printable && (c==',' || c == '"' || addr_spec_soon(src-1/*position to 'c'*/))) { | |
| email=c=='<'; | |
| to_string("?="); | |
| to_quoted_printable=false; | |
| } | } |
| break; | //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'. |
| case UL_MAIL_HEADER: | if(!email && ( |
| // tainted, untaint language: mail-header | !to_quoted_printable && (c & 0x80) // starting quote-printable-encoding on first 8bit char |
| if(charset) { | || to_quoted_printable && !mail_header_char_valid_within_Qencoded(c) |
| // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= | )) { |
| const char *src=row->item.ptr; | if(!to_quoted_printable) { |
| bool to_base_64=false; | to_string("=?"); |
| for(int size=row->item.size; size--; src++) { | to_string(charset_name); |
| if(*src & 0x80) { | to_string("?Q?"); |
| if(!to_base_64) { | to_quoted_printable=true; |
| dest+=sprintf(dest, "=?%.15s?Q?", charset); | |
| to_base_64=true; | |
| } | |
| dest+=sprintf(dest, "=%02X", *src & 0xFF); | |
| } else { | |
| *dest++=*src; | |
| } | |
| } | } |
| if(to_base_64) // close | encode(mail_header_nonspace_char, '=', '_'); |
| dest+=sprintf(dest, "?="); | } else |
| to_char(c); | |
| if(c=='>') | |
| email=false; | |
| } | |
| if(to_quoted_printable) // close | |
| to_string("?="); | |
| } else | |
| ec_append(info->result, optimize, whitespace, info->pos, fragment_length); | |
| break; | |
| case String::L_SQL: | |
| // tainted, untaint language: sql | |
| if(info->connection) { | |
| const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr(); | |
| // skip source [we use recoded version] | |
| pa_CORD_pos_advance(info->pos, fragment_length); | |
| to_string(info->connection->quote(fragment_str, fragment_length)); | |
| } else { | |
| info->exception="untaint in SQL language failed - no connection specified"; | |
| info->fragment_begin=fragment_end; | |
| return 1; // stop processing. can't throw exception here | |
| } | |
| break; | |
| case String::L_JS: | |
| escape_fragment(switch(c) { | |
| case '\n': to_string("\\n"); break; | |
| case '"': to_string("\\\""); break; | |
| case '\'': to_string("\\'"); break; | |
| case '\\': to_string("\\\\"); break; | |
| case '\xFF': to_string("\\\xFF"); break; | |
| case '\r': to_string("\\r"); break; | |
| default: _default; break; | |
| }); | |
| break; | |
| case String::L_XML: | |
| // [2] Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF] | |
| escape_fragment(switch(c) { | |
| case '\x20': | |
| case '\x9': | |
| case '\xA': | |
| case '\xD': // this is usually removed on input | |
| _default; | |
| break; | |
| case '&': to_string("&"); break; | |
| case '>': to_string(">"); break; | |
| case '<': to_string("<"); break; | |
| case '"': to_string("""); break; | |
| case '\'': to_string("'"); break; | |
| default: | |
| if(((unsigned char)c)<0x20) { | |
| // fixing it, so that libxml would not result | |
| // in fatal error parsing text | |
| // though it really violates standard. | |
| // to indicate there were an error | |
| // replace bad char not to it's code, | |
| // which we can do, | |
| // but rather to '!' to show that input were actually | |
| // invalid. | |
| // life: shows that MSIE can somehow garble form values | |
| // so that they contain these chars. | |
| to_char('!'); | |
| } else { | } else { |
| memcpy(dest, row->item.ptr, row->item.size); | _default; |
| dest+=row->item.size; | |
| } | } |
| break; | break; |
| case UL_TABLE: | }); |
| // tainted, untaint language: table | break; |
| escape(switch(*src) { | case String::L_HTML: |
| case '\t': to_char(' '); break; | escape_fragment(switch(c) { |
| case '\n': to_char(' '); break; | case '&': to_string("&"); break; |
| _default; | case '>': to_string(">"); break; |
| }); | case '<': to_string("<"); break; |
| break; | case '"': to_string("""); break; |
| case UL_SQL: | default: _default; break; |
| // tainted, untaint language: sql | }); |
| if(connection) | break; |
| dest+=connection->quote(dest, row->item.ptr, row->item.size); | case String::L_REGEX: |
| else | // tainted, untaint language: regex |
| THROW(0, 0, | escape_fragment( |
| this, | if(need_regex_escape(c)) |
| "untaint in SQL language failed - no connection specified"); | to_char('\\') |
| break; | _default; |
| case UL_JS: | ); |
| escape(switch(*src) { | break; |
| case '"': to_string("\\\"", 2); break; | case String::L_HTTP_COOKIE: |
| case '\'': to_string("\\'", 2); break; | // tainted, untaint language: cookie (3.3.0 and higher: %uXXXX in UTF-8) |
| case '\n': to_string("\\n", 2); break; | { |
| case '\\': to_string("\\\\", 2); break; | const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr(); |
| case '\xFF': to_string("\\\xFF", 2); break; | // skip source [we use recoded version] |
| _default; | pa_CORD_pos_advance(info->pos, fragment_length); |
| }); | String::C output(fragment_str, fragment_length); |
| break; | |
| case UL_HTML: | output=Charset::escape(output, info->charsets->source()); |
| escape(switch(*src) { | //throw Exception(0, 0, output); |
| case '&': to_string("&", 5); break; | to_string(output); |
| case '>': to_string(">", 4); break; | |
| case '<': to_string("<", 4); break; | |
| case '"': to_string(""", 6); break; | |
| //TODO: XSLT case '\'': to_string("'", 6); break; | |
| _default; | |
| }); | |
| break; | |
| case UL_USER_HTML: { | |
| // tainted, untaint language: html-typo | |
| char *html_for_typo= | |
| (char *)malloc(row->item.size*2/* '\n' -> '\' 'n' */+1); | |
| // note: | |
| // there still is a possibility that user | |
| // would not replace \n as she supposed to | |
| // and rather replace \ and n into huge strings | |
| // thus causing memory overrun | |
| // this can be dealed by allocating *2 memory, but that's too expensive | |
| size_t html_for_typo_size; | |
| { // local dest | |
| char *dest=html_for_typo; | |
| escape(switch(*src) { | |
| // convinient name for typo match "\n" | |
| case '\r': | |
| if(typo_table) { | |
| *dest++='\\'; *dest++='n'; // \r -> \n | |
| if(src[1]=='\n') { // \r\n -> remove \n | |
| size--; src++; | |
| } | |
| } | |
| break; | |
| case '\n': | |
| if(typo_table) | |
| to_string("\\n", 2); | |
| break; | |
| //TODO: XSLT case '\'': to_string("'", 6); break; | |
| _default; | |
| }); | |
| *dest=0; | |
| html_for_typo_size=dest-html_for_typo; | |
| } | |
| // typo table replacements | |
| const char *src=html_for_typo; | |
| do { | |
| // there is a row where first column starts 'src' | |
| if(Table::Item *item=typo_table->first_that(typo_present, src)) { | |
| // get a=>b values | |
| const String& a=*static_cast<Array *>(item)->get_string(0); | |
| const String& b=*static_cast<Array *>(item)->get_string(1); | |
| // empty 'a' | 'b' checks | |
| if(a.size()==0 || b.size()==0) { | |
| pool().set_tag(default_typo_table); // avoid recursion | |
| THROW(0, 0, | |
| typo_table->origin_string(), | |
| "typo table column elements must not be empty"); | |
| } | |
| // overflow check: | |
| // b allowed to be max UNTAINT_TIMES_BIGGER then a | |
| if(b.size()>UNTAINT_TIMES_BIGGER*a.size()) { | |
| pool().set_tag(default_typo_table); // avoid recursion | |
| THROW(0, 0, | |
| &b, | |
| "is %g times longer then '%s', " | |
| "while maximum, handled by Parser, is %d", | |
| ((double)b.size())/a.size(), | |
| a.cstr(), | |
| UNTAINT_TIMES_BIGGER); | |
| } | |
| // skip 'a' in 'src' | |
| src+=a.size(); | |
| // write 'b' to 'dest' | |
| b.store_to(dest); | |
| dest+=b.size(); | |
| } else | |
| *dest++=*src++; | |
| } while(*src); | |
| break; | |
| } | |
| default: | |
| THROW(0, 0, | |
| this, | |
| "unknown untaint language #%d of %d piece", | |
| static_cast<int>(row->item.lang), | |
| i); // never | |
| break; // never | |
| } | |
| if((lang==UL_UNSPECIFIED?row->item.lang:lang)!=UL_CLEAN) | |
| whitespace=false; | |
| } | } |
| chunk=row->link; | break; |
| } while(chunk); | default: |
| break2: | SAPI::abort("unknown untaint language #%d", |
| return dest; | static_cast<int>(to_lang)); // should never |
| break; // never | |
| } | |
| info->fragment_begin=fragment_end; | |
| return 0; // 0=continue | |
| } | |
| String::Body String::cstr_to_string_body_taint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const { | |
| if(is_empty()) | |
| return String::Body(); | |
| Cstr_to_string_body_block_info info; | |
| // input | |
| info.lang=lang; | |
| info.connection=connection; | |
| info.charsets=charsets; | |
| info.body=&body; | |
| // output | |
| CORD_ec_init(info.result); | |
| // private | |
| body.set_pos(info.pos, 0); | |
| info.fragment_begin=0; | |
| info.exception=0; | |
| info.whitespace=true; | |
| cstr_to_string_body_block(lang, length(), &info); | |
| if(info.exception) | |
| throw Exception(0, | |
| 0, | |
| info.exception); | |
| return String::Body(CORD_ec_to_cord(info.result)); | |
| } | |
| int cstr_to_string_body_block_untaint(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info){ | |
| const String::Language lang=(String::Language)(unsigned char)alang; | |
| // see append_fragment_* for explanation | |
| if(info->lang&String::L_OPTIMIZE_BIT) | |
| return cstr_to_string_body_block( | |
| lang==String::L_TAINTED? | |
| info->lang | |
| :lang==String::L_CLEAN? | |
| (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT) | |
| :lang, | |
| fragment_length, info); | |
| else | |
| return cstr_to_string_body_block(lang==String::L_TAINTED ? info->lang:lang, fragment_length, info); | |
| } | |
| String::Body String::cstr_to_string_body_untaint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const { | |
| if(is_empty()) | |
| return String::Body(); | |
| Cstr_to_string_body_block_info info; | |
| // input | |
| info.lang=lang; | |
| info.connection=connection; | |
| info.charsets=charsets; | |
| info.body=&body; | |
| // output | |
| CORD_ec_init(info.result); | |
| // private | |
| body.set_pos(info.pos, 0); | |
| info.fragment_begin=0; | |
| info.exception=0; | |
| info.whitespace=true; | |
| langs.for_each(body, cstr_to_string_body_block_untaint, &info); | |
| if(info.exception) | |
| throw Exception(0, | |
| 0, | |
| info.exception); | |
| return String::Body(CORD_ec_to_cord(info.result)); | |
| } | } |