Return to untaint.C CVS log

Up to [parser3project] / parser3 / src / main

2002->2003

/** @file
	Parser: String class part: untaint mechanizm.

	Copyright(c) 2001, 2003 ArtLebedev Group (http://www.artlebedev.com)
	Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
*/

static const char* IDENT_UNTAINT_C="$Date: 2003/01/21 15:51:15 $";

#include "pa_pool.h"
#include "pa_string.h"
#include "pa_hash.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"
#include "pa_charset.h"

//#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; \
		for(int size=row->item.size; size--; src++) \
			action \
	}
#define _default  default: *dest++=*src; 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; \

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

/*

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;
}

//#include "pa_sapi.h"
/** 
	appends 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.

    using architecture advantage: after string-to-string-append string never modified.
	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.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(!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

		// 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;
		}

		// stop-growing mark
		src.last_chunk=0;
		return *this;
	}

	// 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);
		);
/*
			for(Chunk::Row *row=last_chunk->rows; row<append_here; row++)
				if(row->link==(void*)0xcdcdcdcd)
					_asm int 3;*/
	return *this;
}

size_t String::cstr_bufsize(Untaint_lang lang,
							SQL_Connection *connection,
							Charset *buf_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*6*3/* worst utf8 x 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(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+
					parts_count*(buf_charset->name().size()+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;
		}
	);
	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
*/
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 char *store_to_charset_name) const {
	// WARNING:
	//	 before any changes check cstr_bufsize first!!!
	bool whitespace=true;
//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;

		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(
				// 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:
			// tainted, untaint language: uri
			const void *client_ptr;
			size_t client_size;
			Charset::transcode(pool(), 
				pool().get_source_charset(), row->item.ptr, row->item.size,
				pool().get_client_charset(), client_ptr, client_size);
			{
				const char *src=(const char *)client_ptr;
				for(int size=client_size; size--; src++) 
					encode(need_uri_encode, '%');
			}
			break;
		case UL_HTTP_HEADER:
			// tainted, untaint language: http-field-content-text
			escape(
				encode(need_uri_encode, '%');
			);
			break;
		case UL_MAIL_HEADER:
			// tainted, untaint language: mail-header
			// 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(), 
					pool().get_source_charset(), row->item.ptr, row->item.size,
					*store_to_charset, mail_ptr, mail_size);

				// 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;

				bool email=false;
				for(const char *end=src+mail_size; src<end; src++) {
					//RFC   + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
					if(to_quoted_printable && (*src==',' || *src=='<')) {
						email=*src=='<';
						dest+=sprintf(dest, "?=");
						to_quoted_printable=false;
					}
					if(!email && (
						!to_quoted_printable && (*src & 0x80)  // starting quote-printable-encoding on first 8bit char
						|| to_quoted_printable && !mail_header_char_valid_within_Qencoded(*src)
						)) {
						if(!to_quoted_printable) {
							dest+=sprintf(dest, "=?%s?Q?", store_to_charset_name);
							to_quoted_printable=true;
						}
						//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, "?=");
			
			} 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,
					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("&amp;", 5);  break;
				case '>': to_string("&gt;", 4);  break;
				case '<': to_string("&lt;", 4);  break;
				case '"': to_string("&quot;", 6);  break;
				case '\'': to_string("&apos;", 6);  break;
				_default;
			});
			break;
		case UL_HTML:
			escape(switch(*src) {
				case '&': to_string("&amp;", 5);  break;
				case '>': to_string("&gt;", 4);  break;
				case '<': to_string("&lt;", 4);  break;
				case '"': to_string("&quot;", 6);  break;
				_default;
			});
			break;
		default:
			throw Exception(0, 
				this, 
				"unknown untaint language #%d", 
					static_cast<int>(to_lang)); // should 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 '\r': case '\n': case '\t':
					if(!whitespace) {
						*dest++=*src;
						whitespace=true;
					}
					break;
				default:
					whitespace=false;
					*dest++=*src;
					break;
				};
		} else // piece without optimization
//debug
			whitespace=false;
		row++; countdown--; \
	} 
/*
	);
*/
	return dest;
}

char *String::cstr_debug_origins() const {
	//_asm int 3;
	char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2);
	char *dest=result;
	
	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, "<unknown>");
);
		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<int>(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;

		memcpy(dest, row->item.ptr, row->item.size); 
		dest+=row->item.size;

		remove_crlf(dest_after_origins, dest);
		to_char('\n');
	);

	*dest=0;
	return result;
}