|
|
| version 1.3, 2001/03/18 13:22:07 | version 1.79, 2001/11/21 08:33:56 |
|---|---|
| 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) | Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) |
| Author: Alexander Petrosyan <paf@design.ru>(http://paf.design.ru) | |
| $Id$ | $Id$ |
| */ | */ |
| #include <string.h> | |
| #include "pa_pool.h" | #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_sql_connection.h" | |
| #include "pa_dictionary.h" | |
| #include "pa_common.h" | |
| #define escape(cases) \ | #define escape(action) \ |
| { \ | { \ |
| const char *ptr=row->item.ptr; \ | const char *src=row->item.ptr; \ |
| for (int size=row->item.size; size--; ptr++) \ | for(int size=row->item.size; size--; src++) \ |
| switch(*ptr) { \ | action \ |
| cases \ | |
| default: *copy_here++=*ptr; break; \ | |
| } \ | |
| } | } |
| #define escape_value(a, c) case a: *copy_here++=c; break; | #define _default default: *dest++=*src; break |
| #define escape_subst(a, b, bsize) \ | #define encode(need_encode_func, prefix) \ |
| case a: \ | default: \ |
| { \ | if(need_encode_func(*src)) { \ |
| strncpy(copy_here, b, bsize); \ | static const char *hex="0123456789ABCDEF"; \ |
| copy_here+=bsize; \ | char chunk[3]={prefix}; \ |
| } \ | chunk[1]=hex[((unsigned char)*src)/0x10]; \ |
| break; | 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; \ | |
| inline bool need_file_encode(unsigned char c){ | |
| if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) | |
| return false; | |
| return !strchr( | |
| #ifdef WIN32 | |
| ":\\~" | |
| #endif | |
| "./()_-", c); | |
| } | |
| 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) | |
| "UHTML", ///< HTML code with USER chars | |
| }; | |
| // String | // String |
| char *String::cstr() const { | /* |
| char *result=(char *)malloc(size()*UNTAINT_TIMES_BIGGER+1); | |
| HTTP-header = field-name ":" [ field-value ] CRLF | |
| field-name = token | |
| field-value = *( field-content | LWS ) | |
| field-content = <the OCTETs making up the field-value | |
| and consisting of either *TEXT or combinations | |
| of token, tspecials, and quoted-string> | |
| word = token | quoted-string | |
| token = 1*<any CHAR except CTLs or tspecials> | |
| tspecials = "(" | ")" | "<" | ">" | "@" | |
| | "," | ";" | ":" | "\" | <"> | |
| | "/" | "[" | "]" | "?" | "=" | |
| | "{" | "}" | SP | HT | |
| SP = <US-ASCII SP, space (32)> | |
| HT = <US-ASCII HT, horizontal-tab (9)> | |
| LWS = [CRLF] 1*( SP | HT ) | |
| TEXT = <any OCTET except CTLs, | |
| but including LWS> | |
| quoted-pair = "\" CHAR | |
| if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr)) | |
| */ | |
| inline bool need_quote_http_header(const char *ptr, size_t size) { | |
| for(; size--; ptr++) | |
| if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr)) | |
| return true; | |
| return false; | |
| } | |
| char *copy_here=result; | String& String::append(const String& src, uchar lang, bool forced) { |
| // manually unrolled code to avoid do{if(const)} constructs | |
| if(forced) | |
| STRING_SRC_FOREACH_ROW( | |
| APPEND(row->item.ptr, row->item.size, | |
| lang, //forcing passed lang | |
| row->item.origin.file, row->item.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); | |
| ) | |
| 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 | |
| ), | |
| row->item.origin.file, row->item.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); | |
| ); | |
| break2: | |
| return *this; | |
| } | |
| size_t String::cstr_bufsize(Untaint_lang lang, | |
| SQL_Connection *connection, | |
| const char *charset) const { | |
| size_t dest=1; // for terminating 0 | |
| STRING_FOREACH_ROW( | |
| uchar to_lang=lang==UL_UNSPECIFIED?row->item.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+=row->item.size; | |
| break; | |
| case UL_FILE_SPEC: | |
| // tainted, untaint language: file [name] | |
| dest+=row->item.size*3/* worst: Z->%XX */; | |
| break; | |
| case UL_URI: | |
| // tainted, untaint language: uri | |
| dest+=row->item.size*3/* worst: Z->%XX */; | |
| break; | |
| case UL_HTTP_HEADER: | |
| // tainted, untaint language: http-field-content-text | |
| dest+=row->item.size*3/* worst: Z->%XX */; | |
| break; | |
| case UL_MAIL_HEADER: | |
| // tainted, untaint language: mail-header | |
| if(charset) { | |
| // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= | |
| dest+=row->item.size*3+MAX_STRING/* worst: =?charset?Q?=%XX?= */; | |
| } else { | |
| dest+=row->item.size; | |
| } | |
| break; | |
| case UL_TABLE: | |
| // tainted, untaint language: table | |
| dest+=row->item.size; | |
| break; | |
| case UL_SQL: | |
| // tainted, untaint language: sql | |
| if(connection) | |
| dest+=connection->quote(0, row->item.ptr, row->item.size); | |
| break; | |
| 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; | |
| } | |
| ); | |
| break2: | |
| return dest; | |
| } | |
| char *String::store_to(char *dest, Untaint_lang lang, | |
| SQL_Connection *connection, | |
| const char *charset) 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; i<chunk->count; i++, row++) { \ | |
| if(row==append_here) \ | |
| goto break2; \ | |
| \ | |
| uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; | |
| char *start=dest; | |
| 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 | |
| memcpy(dest, row->item.ptr, row->item.size); | |
| dest+=row->item.size; | |
| break; | |
| case UL_FILE_SPEC: | |
| // tainted, untaint language: file [name] | |
| escape(switch(*src) { | |
| case ' ': to_char('_'); break; | |
| encode(need_file_encode, '+'); | |
| }); | |
| break; | |
| case UL_URI: | |
| // tainted, untaint language: uri | |
| escape(switch(*src) { | |
| case ' ': to_char('+'); break; | |
| encode(need_uri_encode, '%'); | |
| }); | |
| break; | |
| case UL_HTTP_HEADER: | |
| // tainted, untaint language: http-field-content-text | |
| escape(switch(*src) { | |
| case ' ': to_char('+'); break; | |
| encode(need_uri_encode, '%'); | |
| }); | |
| break; | |
| case UL_MAIL_HEADER: | |
| // tainted, untaint language: mail-header | |
| if(charset) { | |
| // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= | |
| const char *src=row->item.ptr; | |
| bool to_quoted_printable=false; | |
| for(int size=row->item.size; size--; src++) { | |
| if(*src & 0x80) { | |
| if(!to_quoted_printable) { | |
| dest+=sprintf(dest, "=?%.15s?Q?", charset); | |
| to_quoted_printable=true; | |
| } | |
| dest+=sprintf(dest, "=%02X", *src & 0xFF); | |
| } else { | |
| *dest++=*src; | |
| } | |
| } | |
| if(to_quoted_printable) // close | |
| dest+=sprintf(dest, "?="); | |
| } else { | |
| memcpy(dest, row->item.ptr, row->item.size); | |
| dest+=row->item.size; | |
| } | |
| break; | |
| case UL_TABLE: | |
| // tainted, untaint language: table | |
| escape(switch(*src) { | |
| case '\t': to_char(' '); break; | |
| case '\n': to_char(' '); break; | |
| _default; | |
| }); | |
| break; | |
| case UL_SQL: | |
| // tainted, untaint language: sql | |
| if(connection) | |
| dest+=connection->quote(dest, row->item.ptr, row->item.size); | |
| else | |
| throw Exception(0, 0, | |
| this, | |
| "untaint in SQL language failed - no connection specified"); | |
| break; | |
| 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; | |
| _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; | |
| _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; | |
| _default; | |
| }); | |
| break; | |
| default: | |
| throw Exception(0, 0, | |
| this, | |
| "unknown untaint language #%d of %d piece", | |
| static_cast<int>(row->item.lang), | |
| i); // never | |
| break; // never | |
| } | |
| if(to_lang & UL_OPTIMIZE_BIT) { | |
| // optimizing whitespace | |
| char *stop=dest; dest=start; | |
| for(char *src=start; src<stop; src++) | |
| switch(*src) { | |
| // of all consequent white space chars leaving only first one | |
| case ' ': case '\n': case '\t': | |
| if(!whitespace) { | |
| *dest++=*src; | |
| whitespace=true; | |
| } | |
| break; | |
| default: | |
| whitespace=false; | |
| *dest++=*src; | |
| break; | |
| }; | |
| } else // piece without optimization | |
| whitespace=false; | |
| } \ | |
| chunk=row->link; \ | |
| } while(chunk); \ | |
| break2: | |
| return dest; | |
| } | |
| char *String::cstr_debug_origins() const { | |
| char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2); | |
| char *dest=result; | |
| const Chunk *chunk=&head; | const Chunk *chunk=&head; |
| // TODO: оптимизировать whitespaces для всех, кроме 'html' | |
| do { | do { |
| const Chunk::Row *row=chunk->rows; | const Chunk::Row *row=chunk->rows; |
| for(int i=0; i<chunk->count; i++) { | for(uint i=0; i<chunk->count; i++, row++) { |
| if(row==append_here) | if(row==append_here) |
| goto break2; | goto break2; |
| // WARNING: | #ifndef NO_STRING_ORIGIN |
| // string can grow only UNTAINT_TIMES_BIGGER | if(row->item.origin.file) |
| switch(row->item.lang) { | dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT, |
| case NO: | row->item.origin.file, |
| // clean piece | 1+row->item.origin.line); |
| case YES: | else |
| // tainted piece, but undefined untaint language | dest+=sprintf(dest, "<unknown>"); |
| // for VString.get_double of tainted values | #endif |
| // for ^process{body} evaluation | dest+=sprintf(dest, "#%s: ", |
| case AS_IS: | String_Untaint_lang_name[row->item.lang]); |
| // tainted, untaint language: as-is | char *dest_after_origins=dest; |
| memcpy(copy_here, row->item.ptr, row->item.size); | |
| copy_here+=row->item.size; | memcpy(dest, row->item.ptr, row->item.size); |
| break; | dest+=row->item.size; |
| case TABLE: | |
| escape( | remove_crlf(dest_after_origins, dest); |
| escape_value('\t', ' ') | to_char('\n'); |
| escape_value('\n', ' ') | |
| ); | |
| break; | |
| case SQL: | |
| // tainted, untaint language: sql | |
| // TODO: зависимость от sql сервера | |
| memset(copy_here, '?', row->item.size); | |
| copy_here+=row->item.size; | |
| break; | |
| case JS: | |
| escape( | |
| escape_subst('"', "\\\"", 2) | |
| escape_subst('\'', "\\'", 2) | |
| escape_subst('\n', "\\n", 2) | |
| escape_subst('\r', "\\r", 2) | |
| escape_subst('\\', "\\\\", 2) | |
| escape_subst('я', "\\я", 2) | |
| ); | |
| break; | |
| case HTML: | |
| escape( | |
| escape_subst('&', "&", 5) // BEFORE consequent relpaces yelding '&' | |
| escape_subst('>', ">", 4) | |
| escape_subst('<', "<",4) | |
| escape_subst('"', """,6) | |
| escape_value('\t', ' ') | |
| //TODO: XSLT escape_subst('\'', "'", 6) | |
| ); | |
| break; | |
| case HTML_TYPO: | |
| // tainted, untaint language: html-typo | |
| escape( | |
| escape_subst('&', "&", 5) // BEFORE consequent relpaces yelding '&' | |
| escape_subst('>', ">", 4) | |
| escape_subst('<', "<",4) | |
| escape_subst('"', """,6) | |
| escape_value('\t', ' ') | |
| //TODO: $MAIN:html-type table replace, max length(b)==UNTAINT_TIMES_BIGGER*length(a) | |
| ); | |
| break; | |
| default: | |
| THROW(0,0, | |
| this, | |
| "unknown untaint language #%d of %d piece", | |
| static_cast<int>(row->item.lang), | |
| i); | |
| } | |
| row++; | |
| } | } |
| chunk=row->link; | chunk=row->link; |
| } while(chunk); | } while(chunk); |
| break2: | break2: |
| *copy_here=0; | *dest=0; |
| return result; | return result; |
| } | } |