|
|
| version 1.2, 2001/03/12 21:54:20 | version 1.118.2.4, 2003/09/23 14:00:53 |
|---|---|
| Line 1 | Line 1 |
| /* | /** @file |
| Parser | Parser: String class part: untaint mechanizm. |
| Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) | |
| Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) | |
| $Id$ | Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | |
| */ | */ |
| #include <string.h> | static const char* IDENT_UNTAINT_C="$Date$"; |
| #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_globals.h" | |
| #include "pa_dictionary.h" | |
| #include "pa_common.h" | |
| #include "pa_charset.h" | |
| #include "pa_request_charsets.h" | |
| #include "pa_sapi.h" | |
| #define escape(cases) \ | extern "C" { // author forgot to do that |
| { \ | #include "ec.h" |
| const char *ptr=row->item.ptr; \ | } |
| int size=row->item.size; \ | |
| for (;*ptr;ptr++) \ | #define PA_SQL |
| switch(*ptr) { \ | |
| cases \ | #ifdef PA_SQL |
| default: *copy_here++=*ptr; break; \ | #include "pa_sql_connection.h" |
| } \ | #endif |
| } | |
| #define escape_value(a, c) case a: *copy_here++=c; break; | // defines |
| #define escape_subst(a, b, bsize) \ | |
| case a: \ | |
| { \ | #undef CORD_ec_append |
| strncpy(copy_here, b, bsize); \ | // redefining to intercept flushes and implement whitespace optimization |
| copy_here+=bsize; \ | // 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 '\r': 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); \ | |
| } \ | } \ |
| break; | *((x)[0].ec_bufptr)++ = (c); \ |
| } \ | |
| } | |
| #define escape(action) \ | |
| for(; fragment_length--; CORD_next(info->pos)) { \ | |
| char c=CORD_pos_fetch(info->pos); \ | |
| action \ | |
| } | |
| #define _default default: CORD_ec_append(info->result, c); break | |
| #define encode(need_encode_func, prefix, otherwise) \ | |
| if(need_encode_func(c)) { \ | |
| static const char* hex="0123456789ABCDEF"; \ | |
| CORD_ec_append(info->result, prefix); \ | |
| CORD_ec_append(info->result, hex[((unsigned char)c)/0x10]); \ | |
| CORD_ec_append(info->result, hex[((unsigned char)c)%0x10]); \ | |
| } else \ | |
| CORD_ec_append(info->result, otherwise); | |
| #define to_char(c) CORD_ec_append(info->result, c) | |
| #define to_string(s) CORD_ec_append_cord(info->result, s) | |
| inline bool need_file_encode(unsigned char c){ | |
| // russian letters and space ENABLED | |
| // encoding only these... | |
| return strchr( | |
| "*?'\"<>|" | |
| #ifndef WIN32 | |
| ":\\" | |
| #endif | |
| , c)!=0; | |
| } | |
| inline bool need_uri_encode(unsigned char c){ | |
| if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) | |
| return false; | |
| return !strchr("_-./", c); | |
| } | |
| inline bool need_http_header_encode(unsigned char c){ | |
| if(strchr(" , :", c)) | |
| return false; | |
| return need_uri_encode(c); | |
| } | |
| // | |
| 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 | |
| // untaint languages. assigned by ^untaint[lang]{...} | |
| "P", | |
| /**< | |
| leave language built into string being appended. | |
| just a flag, that value not stored | |
| */ | |
| "A", ///< leave all characters intact | |
| "F", ///< file specification | |
| "H", ///< ext in HTTP response header | |
| "M", ///< text in mail header | |
| "URI", ///< text in uri | |
| "T", ///< ^table:set body | |
| "SQL", ///< ^table:sql body | |
| "JS", ///< JavaScript code | |
| "XML", ///< ^dom:set xml | |
| "HTML" ///< HTML code (for editing) | |
| }; | |
| // String | // String |
| char *String::cstr() const { | /* |
| char *result=(char *)malloc(size()*UNTAINT_TIMES_BIGGER+1); | |
| char *copy_here=result; | HTTP-header = field-name ":" [ field-value ] CRLF |
| const Chunk *chunk=&head; | |
| // TODO: оптимизировать whitespaces для всех, кроме 'html' | field-name = token |
| do { | field-value = *( field-content | LWS ) |
| const Chunk::Row *row=chunk->rows; | |
| for(int i=0; i<chunk->count; i++) { | field-content = <the OCTETs making up the field-value |
| if(row==append_here) | and consisting of either *TEXT or combinations |
| goto break2; | of token, tspecials, and quoted-string> |
| // WARNING: | |
| // string can grow only UNTAINT_TIMES_BIGGER | |
| switch(row->item.lang) { | word = token | quoted-string |
| case NO: | |
| // clean piece | token = 1*<any CHAR except CTLs or tspecials> |
| case YES: | |
| // tainted piece, but undefined untaint language | |
| // for VString.get_double of tainted values | |
| // for ^process{body} evaluation | tspecials = "(" | ")" | "<" | ">" | "@" |
| case AS_IS: | | "," | ";" | ":" | "\" | <"> |
| // tainted, untaint language: as-is | | "/" | "[" | "]" | "?" | "=" |
| memcpy(copy_here, row->item.ptr, row->item.size); | | "{" | "}" | SP | HT |
| copy_here+=row->item.size; | |
| break; | SP = <US-ASCII SP, space (32)> |
| case TABLE: | HT = <US-ASCII HT, horizontal-tab (9)> |
| escape( | |
| escape_value('\t', ' ') | LWS = [CRLF] 1*( SP | HT ) |
| escape_value('\n', ' ') | TEXT = <any OCTET except CTLs, |
| ); | but including LWS> |
| break; | |
| case SQL: | quoted-pair = "\" CHAR |
| // tainted, untaint language: sql | |
| // TODO: зависимость от sql сервера | if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr)) |
| memset(copy_here, '?', row->item.size); | */ |
| copy_here+=row->item.size; | inline bool need_quote_http_header(const char* ptr, size_t size) { |
| break; | for(; size--; ptr++) |
| case JS: | if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr)) |
| escape( | return true; |
| escape_subst('"', "\\\"", 2) | return false; |
| escape_subst('\'', "\\'", 2) | } |
| escape_subst('\n', "\\n", 2) | |
| escape_subst('\r', "\\r", 2) | #ifndef DOXYGEN |
| escape_subst('\\', "\\\\", 2) | struct Append_fragment_info { |
| escape_subst('я', "\\я", 2) | String::Language lang; |
| ); | const StringBody dest_body; |
| break; | String::Languages* dest_languages; |
| case HTML: | size_t dest_body_plan_length; |
| escape( | }; |
| escape_subst('&', "&", 5) // BEFORE consequent relpaces yelding '&' | #endif |
| escape_subst('>', ">", 4) | int append_fragment_optimizing(char alang, size_t asize, Append_fragment_info* info) { |
| escape_subst('<', "<",4) | const String::Language lang=(String::Language)alang; |
| escape_subst('"', """,6) | // main idea here: |
| escape_value('\t', ' ') | // tainted piece would get OPTIMIZED bit from 'lang' |
| //TODO: XSLT escape_subst('\'', "'", 6) | // clean piece would be marked OPTIMIZED manually |
| ); | // pieces with determined languages [not tainted|clean] would retain theirs langs |
| break; | info->dest_languages->append(info->dest_body_plan_length, |
| case HTML_TYPO: | lang==String::Language::L_TAINTED? |
| // tainted, untaint language: html-typo | info->lang |
| escape( | :lang==String::Language::L_CLEAN? |
| escape_subst('&', "&", 5) // BEFORE consequent relpaces yelding '&' | (String::Language)(String::Language::L_CLEAN|String::Language::L_OPTIMIZE_BIT) // ORing with OPTIMIZED flag |
| escape_subst('>', ">", 4) | :lang, |
| escape_subst('<', "<",4) | asize); |
| escape_subst('"', """,6) | info->dest_body_plan_length+=asize; |
| escape_value('\t', ' ') | |
| //TODO: $MAIN:html-type table replace, max length(b)==UNTAINT_TIMES_BIGGER*length(a) | return 0; // 0=continue |
| ); | } |
| int append_fragment_nonoptimizing(char alang, size_t asize, Append_fragment_info* info) { | |
| const String::Language lang=(String::Language)alang; | |
| // The core idea: tainted pieces got marked with context's lang | |
| info->dest_languages->append(info->dest_body_plan_length, | |
| lang==String::Language::L_TAINTED? | |
| info->lang | |
| :lang, | |
| asize); | |
| info->dest_body_plan_length+=asize; | |
| return 0; // 0=continue | |
| } | |
| /** | |
| 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, Language lang, bool forced) const { | |
| if(is_empty()) | |
| return dest; | |
| // first: fragment infos | |
| if(lang==L_PASS_APPENDED) // without language-change? | |
| dest.langs.append(dest.body, body.length(), langs); | |
| else if(forced) //forcing passed lang? | |
| dest.langs.append(dest.body, lang, length()); | |
| else { | |
| Append_fragment_info info={lang, dest.body, &dest.langs, dest.body.length()}; | |
| langs.for_each(body, lang&L_OPTIMIZE_BIT? | |
| append_fragment_optimizing | |
| :append_fragment_nonoptimizing, &info); | |
| } | |
| // next: letters | |
| 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) { | |
| size_t avail=CORD_pos_chars_left(pos); | |
| if(avail==0) { | |
| CORD_next(pos); | |
| if(!--n) | |
| break; | break; |
| default: | } else if(avail<n) { |
| THROW(0,0, | CORD_pos_advance(pos, avail); |
| this, | n-=avail; |
| "unknown untaint language #%d of %d piece", | } else { // avail>=n |
| static_cast<int>(row->item.lang), | CORD_pos_advance(pos, n); |
| i); | break; |
| } | } |
| row++; | } |
| } | |
| #ifndef DOXYGEN | |
| struct Cstr_to_string_body_block_info { | |
| // input | |
| String::Language lang; | |
| SQL_Connection* connection; | |
| const Request_charsets* charsets; | |
| const StringBody* body; | |
| // output | |
| CORD_ec result; | |
| // private | |
| CORD_pos pos; | |
| size_t fragment_begin; | |
| bool whitespace; | |
| }; | |
| #endif | |
| int cstr_to_string_body_block(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info) { | |
| const String::Language fragment_lang=(String::Language)alang; | |
| bool& whitespace=info->whitespace; | |
| size_t fragment_end=info->fragment_begin+fragment_length; | |
| //fprintf(stderr, "%d, %d\n", fragment.lang, fragment.length); | |
| String::Language to_lang=info->lang==String::Language::L_UNSPECIFIED?fragment_lang:info->lang; | |
| bool optimize=(to_lang & String::Language::L_OPTIMIZE_BIT)!=0; | |
| if(!optimize) | |
| whitespace=false; | |
| switch(to_lang & ~String::Language::L_OPTIMIZE_BIT) { | |
| case String::Language::L_CLEAN: | |
| case String::Language::L_TAINTED: | |
| case String::Language::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::Language::L_FILE_SPEC: | |
| // tainted, untaint language: file [name] | |
| escape( | |
| // Macintosh has problems with small Russian letter 'r' | |
| if( c=='\xF0' && info->charsets && info->charsets->source().NAME()=="WINDOWS-1251" ) { | |
| // fixing that letter for most common charset | |
| to_char('p'); | |
| } else // fallback to default | |
| encode(need_file_encode, '_', c); | |
| ); | |
| break; | |
| case String::Language::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); | |
| } | } |
| chunk=row->link; | break; |
| } while(chunk); | case String::Language::L_HTTP_HEADER: |
| break2: | // tainted, untaint language: http-field-content-text |
| *copy_here=0; | escape( |
| return result; | encode(need_uri_encode, '%', c); |
| ); | |
| break; | |
| case String::Language::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++; ) { | |
| //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'. | |
| if(to_quoted_printable && (c==',' || addr_spec_soon(src) || c == '"')) { | |
| email=c=='<'; | |
| to_string("?="); | |
| to_quoted_printable=false; | |
| } | |
| if(!email && ( | |
| !to_quoted_printable && (c & 0x80) // starting quote-printable-encoding on first 8bit char | |
| || to_quoted_printable && !mail_header_char_valid_within_Qencoded(c) | |
| )) { | |
| if(!to_quoted_printable) { | |
| to_string("=?"); | |
| to_string(charset_name); | |
| to_string("?Q?"); | |
| to_quoted_printable=true; | |
| } | |
| encode(mail_header_nonspace_char, '=', '_'); | |
| } 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::Language::L_TABLE: | |
| // tainted, untaint language: table | |
| escape(switch(c) { | |
| case '\t': to_char(' '); break; | |
| case '\n': to_char(' '); break; | |
| _default; | |
| }); | |
| break; | |
| #ifdef PA_SQL | |
| case String::Language::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 | |
| throw Exception(0, | |
| 0, | |
| "untaint in SQL language failed - no connection specified"); | |
| break; | |
| #endif | |
| case String::Language::L_JS: | |
| 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 String::Language::L_XML: | |
| 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 String::Language::L_HTML: | |
| escape(switch(c) { | |
| case '&': to_string("&"); break; | |
| case '>': to_string(">"); break; | |
| case '<': to_string("<"); break; | |
| case '"': to_string("""); break; | |
| _default; | |
| }); | |
| break; | |
| default: | |
| SAPI::die("unknown untaint language #%d", | |
| static_cast<int>(to_lang)); // should never | |
| break; // never | |
| } | |
| info->fragment_begin=fragment_end; | |
| return 0; // 0=continue | |
| } | |
| StringBody String::cstr_to_string_body(Language lang, | |
| SQL_Connection* connection, | |
| const Request_charsets *charsets) const { | |
| 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.whitespace=true; | |
| if(!is_empty()) | |
| langs.for_each(body, cstr_to_string_body_block, &info); | |
| return StringBody(CORD_ec_to_cord(info.result)); | |
| } | } |