Annotation of parser3/src/main/untaint.C, revision 1.115.2.12.2.3
1.7 paf 1: /** @file
1.8 paf 2: Parser: String class part: untaint mechanizm.
3:
1.115.2.6 paf 4: Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com)
1.89 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.103 paf 6: */
1.8 paf 7:
1.115.2.12.2. (paf 8:): static const char* IDENT_UNTAINT_C="$Date: 2003/03/19 16:28:09 $";
9:):
1.1 paf 10:
11: #include "pa_string.h"
12: #include "pa_hash.h"
13: #include "pa_exception.h"
1.13 paf 14: #include "pa_table.h"
1.32 paf 15: #include "pa_globals.h"
1.58 parser 16: #include "pa_dictionary.h"
1.66 parser 17: #include "pa_common.h"
1.85 paf 18: #include "pa_charset.h"
1.115.2.4 paf 19: #include "pa_request_charsets.h"
1.1 paf 20:
1.115.2.12.2. (paf 21:): //#define PA_SQL
22:):
23:): #ifdef PA_SQL
24:): #include "pa_sql_connection.h"
25:): #endif
26:):
1.18 paf 27: #define escape(action) \
1.1 paf 28: { \
1.115.2.6 paf 29: const char* src=fragment->ptr; \
1.115.2.1 paf 30: for(int size=fragment->size; size--; src++) \
1.18 paf 31: action \
1.1 paf 32: }
1.13 paf 33: #define _default default: *dest++=*src; break
34: #define encode(need_encode_func, prefix) \
35: if(need_encode_func(*src)) { \
1.115.2.6 paf 36: static const char* hex="0123456789ABCDEF"; \
1.9 paf 37: char chunk[3]={prefix}; \
1.13 paf 38: chunk[1]=hex[((unsigned char)*src)/0x10]; \
39: chunk[2]=hex[((unsigned char)*src)%0x10]; \
1.60 parser 40: memcpy(dest, chunk, 3); dest+=3; \
1.5 paf 41: } else \
1.13 paf 42: *dest++=*src; \
1.5 paf 43: break
1.18 paf 44: #define to_char(c) *dest++=c
45: #define to_string(b, bsize) \
1.60 parser 46: memcpy(dest, b, bsize); \
1.18 paf 47: dest+=bsize; \
1.4 paf 48:
1.9 paf 49: inline bool need_file_encode(unsigned char c){
1.108 paf 50: // russian letters and space ENABLED
51: // encoding only these...
52: return strchr(
53: "*?'\"<>|"
54: #ifndef WIN32
55: ":\\~"
1.31 paf 56: #endif
1.108 paf 57: , c)!=0;
1.9 paf 58: }
1.5 paf 59: inline bool need_uri_encode(unsigned char c){
1.13 paf 60: if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z'))
1.4 paf 61: return false;
62:
1.5 paf 63: return !strchr("_-./", c);
64: }
1.36 paf 65: inline bool need_http_header_encode(unsigned char c){
1.18 paf 66: if(strchr(" , :", c))
1.5 paf 67: return false;
68:
69: return need_uri_encode(c);
1.4 paf 70: }
1.1 paf 71:
1.56 parser 72: //
73:
1.115.2.6 paf 74: static const char* String_Untaint_lang_name[]={
1.56 parser 75: "U", ///< zero value handy for hash lookup @see untaint_lang_name2enum
76: "C", ///< clean
77: "T", ///< tainted, untaint language as assigned later
78: // untaint languages. assigned by ^untaint[lang]{...}
79: "P",
80: /**<
81: leave language built into string being appended.
82: just a flag, that value not stored
83: */
84: "A", ///< leave all characters intact
1.68 parser 85: "F", ///< file specification
86: "H", ///< ext in HTTP response header
1.56 parser 87: "M", ///< text in mail header
88: "URI", ///< text in uri
89: "T", ///< ^table:set body
90: "SQL", ///< ^table:sql body
91: "JS", ///< JavaScript code
1.68 parser 92: "XML", ///< ^dom:set xml
1.82 paf 93: "HTML" ///< HTML code (for editing)
1.56 parser 94: };
95:
96:
1.1 paf 97: // String
98:
1.41 paf 99: /*
100:
101: HTTP-header = field-name ":" [ field-value ] CRLF
102:
103: field-name = token
104: field-value = *( field-content | LWS )
105:
106: field-content = <the OCTETs making up the field-value
107: and consisting of either *TEXT or combinations
108: of token, tspecials, and quoted-string>
109:
110:
111:
112: word = token | quoted-string
113:
114: token = 1*<any CHAR except CTLs or tspecials>
115:
116:
117:
118: tspecials = "(" | ")" | "<" | ">" | "@"
119: | "," | ";" | ":" | "\" | <">
120: | "/" | "[" | "]" | "?" | "="
121: | "{" | "}" | SP | HT
122:
123: SP = <US-ASCII SP, space (32)>
124: HT = <US-ASCII HT, horizontal-tab (9)>
125:
126: LWS = [CRLF] 1*( SP | HT )
127: TEXT = <any OCTET except CTLs,
128: but including LWS>
129:
130: quoted-pair = "\" CHAR
131:
132: if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
133: */
1.115.2.6 paf 134: inline bool need_quote_http_header(const char* ptr, size_t size) {
1.41 paf 135: for(; size--; ptr++)
1.42 paf 136: if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
1.41 paf 137: return true;
138: return false;
139: }
140:
1.115.2.12.2. (paf 141:):
1.91 paf 142: /**
1.115.2.1 paf 143: appends to other String,
1.91 paf 144: marking all tainted pieces of it with @a lang.
1.92 paf 145: or marking ALL pieces of it with a @a lang when @a forced to,
146: and propagating OPTIMIZE language bit.
1.91 paf 147: */
1.115.2.12.2. (paf 148:): String& String::append_to(String& dest, Untaint_lang lang, bool forced) const {
149:): if(!*this)
1.115.2.1 paf 150: return dest;
151:
1.115.2.12.2. (paf 152:): // first: letters
153:): dest.body=CORD_cat(dest.body, body);
1.91 paf 154:
1.115.2.12.2. (paf 155:): // next: fragment infos
156:):
157:): if(lang==UL_PASS_APPENDED) // without language-chage string?
158:): dest.fragments.append(fragments, 0, fragments.count());
159:): else if(forced) //forcing passed lang?
160:): dest.fragments+=Fragment(lang, size());
161:): else if(lang&UL_OPTIMIZE_BIT) {
162:): Array_iterator<ArrayFragment::element_type> i(fragments);
163:): while(i.has_next()) {
164:): const Fragment& fragment=i.next();
165:): // main idea here:
166:): // tainted piece would get OPTIMIZED bit from 'lang'
167:): // clean piece would be marked OPTIMIZED manually
168:): // pieces with determined languages [not tainted|clean] would retain theirs langs
169:): dest.fragments+=Fragment(static_cast<Untaint_lang>(
170:): fragment.lang==UL_TAINTED?lang:(
171:): fragment.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag
172:): fragment.lang)),
173:): fragment.size);
174:): }
175:): } else { // The core idea: tainted pieces got marked with context's lang
176:): Array_iterator<ArrayFragment::element_type> i(fragments);
177:): while(i.has_next()) {
178:): const Fragment& fragment=i.next();
179:): dest.fragments+=Fragment(
180:): fragment.lang==UL_TAINTED?lang:fragment.lang,
181:): fragment.size);
182:): }
183:): }
1.115.2.1 paf 184:
185: return dest;
1.77 paf 186: }
187:
1.75 paf 188: size_t String::cstr_bufsize(Untaint_lang lang,
189: SQL_Connection *connection,
1.115.2.3 paf 190: const Request_charsets *charsets) const {
1.77 paf 191: size_t dest=1; // for terminating 0
1.115.2.12.2. (paf 192:): size_t fragment_start=0;
193:): size_t fragment_end;
194:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_start=fragment_end) {
195:): const Fragment& fragment=i.next();
196:): fragment_end=fragment_start+fragment.size;
197:):
198:): Untaint_lang to_lang=lang==UL_UNSPECIFIED?fragment.lang:lang;
1.77 paf 199:
200: switch(to_lang & ~UL_OPTIMIZE_BIT) {
201: case UL_CLEAN:
202: case UL_TAINTED:
203: case UL_AS_IS:
204: // clean piece
205:
206: // tainted piece, but undefined untaint language
207: // for VString.as_double of tainted values
208: // for ^process{body} evaluation
209:
210: // tainted, untaint language: as-is
1.115.2.12.2. (paf 211:): dest+=fragment.size;
1.77 paf 212: break;
213: case UL_FILE_SPEC:
214: // tainted, untaint language: file [name]
1.115.2.12.2. (paf 215:): dest+=fragment.size*3/* worst: Z->%XX */;
1.77 paf 216: break;
217: case UL_URI:
218: // tainted, untaint language: uri
1.115.2.12.2. (paf 219:): dest+=fragment.size*6*3/* worst utf8 x worst Z->%XX */;
1.77 paf 220: break;
221: case UL_HTTP_HEADER:
222: // tainted, untaint language: http-field-content-text
1.115.2.12.2. (paf 223:): dest+=fragment.size*3/* worst: Z->%XX */;
1.77 paf 224: break;
225: case UL_MAIL_HEADER:
226: // tainted, untaint language: mail-header
1.115.2.4 paf 227: if(charsets) {
1.111 paf 228: int parts_count=1;
1.115.2.12.2. (paf 229:): size_t size=fragment.size;
230:): const char* prev=CORD_substr(body, fragment_start, size);
1.115.2.6 paf 231: while(const char* comma_at=(const char*)memchr(prev, ',', size)) {
1.111 paf 232: parts_count++;
233: size-=comma_at-prev;
234: prev=comma_at+1;
235: }
1.77 paf 236: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
1.87 paf 237: dest+=
1.115.2.12.2. (paf 238:): fragment.size*3+
239:): parts_count*(charsets->mail().name().size()+MAX_STRING/* worst: =?charset?Q?=%XX?= */);
1.87 paf 240: } else
1.115.2.12.2. (paf 241:): dest+=fragment.size;
1.77 paf 242: break;
243: case UL_TABLE:
244: // tainted, untaint language: table
1.115.2.12.2. (paf 245:): dest+=fragment.size;
1.77 paf 246: break;
1.115.2.12.2. (paf 247:): #ifdef PA_SQL
1.77 paf 248: case UL_SQL:
249: // tainted, untaint language: sql
250: if(connection)
1.115.2.12.2. (paf 251:): dest+=connection->quote(0,
252:): CORD_substr(body, fragment_start, fragment.size),
253:): fragment.size);
1.77 paf 254: break;
1.115.2.12.2. (paf 255:): #endif
1.77 paf 256: case UL_JS:
257: escape(switch(*src) {
258: case '"': case '\'': case '\n': case '\\': case '\xFF':
259: dest+=2; break;
260: default:
261: dest++; break;
262: });
263: break;
264: case UL_XML:
265: escape(switch(*src) {
266: case '&': case '>': case '<': case '"': case '\'':
267: dest+= 6; break;
268: default:
269: dest++; break;
270: });
271: break;
272: case UL_HTML:
273: escape(switch(*src) {
274: case '&':
275: case '>':
276: case '<':
277: case '"':
278: dest+=6; break;
279: default:
280: dest++; break;
281: });
282: break;
1.75 paf 283: }
1.115.2.12.2. (paf 284:): }
1.75 paf 285: return dest;
1.51 parser 286: }
287:
1.109 paf 288: /** http://www.ietf.org/rfc/rfc2047.txt
289: RFC
290: (3) As a replacement for a 'word' entity within a 'phrase', for example,
291: one that precedes an address in a From, To, or Cc header. The ABNF
292: definition for 'phrase' from RFC 822 thus becomes:
293:
294: phrase = 1*( encoded-word / word )
295:
296: In this case the set of characters that may be used in a "Q"-encoded
297: 'encoded-word' is restricted to: <upper and lower case ASCII
298: letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_"
299: (underscore, ASCII 95.)>. An 'encoded-word' that appears within a
300: 'phrase' MUST be separated from any adjacent 'word', 'text' or
301: 'special' by 'linear-white-space'.
302: ...
303: (2) The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) may be
304: represented as "_" (underscore, ASCII 95.). (This character may
305: not pass through some internetwork mail gateways, but its use
306: will greatly enhance readability of "Q" encoded data with mail
307: readers that do not support this encoding.) Note that the "_"
308: always represents hexadecimal 20, even if the SPACE character
309: occupies a different code position in the character set in use.
310:
311: paf: obviously,
312: without "=", or one could not differ "=E0" and "russian letter a"
313: and without "_", or in would mean 0x20
314: */
315: static bool mail_header_char_valid_within_Qencoded(char c) {
316: return c>='A' && c<='Z'
317: || c>='a' && c<='Z'
318: || c>='0' && c<='9'
319: || strchr("!*+-/", c);
320: }
1.43 paf 321: char *String::store_to(char *dest, Untaint_lang lang,
322: SQL_Connection *connection,
1.115.2.3 paf 323: const Request_charsets *charsets) const {
1.75 paf 324: // WARNING:
325: // before any changes check cstr_bufsize first!!!
1.44 paf 326: bool whitespace=true;
1.115.2.1 paf 327: // STRING_FOREACH_FRAGMENT(
328: element_type *last = felements+fused; element_type *fragment = felements; for(; fragment<last; fragment++) {
1.111 paf 329:
1.115.2.12.2. (paf 330:): Untaint_lang to_lang=lang==UL_UNSPECIFIED?fragment->lang:lang;
1.77 paf 331:
332: char *start=dest;
333:
334: switch(to_lang & ~UL_OPTIMIZE_BIT) {
335: case UL_CLEAN:
336: case UL_TAINTED:
337: case UL_AS_IS:
338: // clean piece
339:
340: // tainted piece, but undefined untaint language
341: // for VString.as_double of tainted values
342: // for ^process{body} evaluation
343:
344: // tainted, untaint language: as-is
1.115.2.1 paf 345: memcpy(dest, fragment->ptr, fragment->size);
346: dest+=fragment->size;
1.77 paf 347: break;
348: case UL_FILE_SPEC:
349: // tainted, untaint language: file [name]
1.83 paf 350: escape(
1.113 paf 351: // Macintosh has problems with small Russian letter 'r'
1.115.2.4 paf 352: if( *src=='\xF0' && charsets && *charsets->source().name()=="windows-1251" ) {
1.113 paf 353: // fixing that letter for most common charset
1.115.2.2 paf 354: to_char('p');
1.114 paf 355: } else // fallback to default
356: encode(need_file_encode, '_');
1.83 paf 357: );
1.77 paf 358: break;
359: case UL_URI:
360: // tainted, untaint language: uri
1.85 paf 361: const void *client_ptr;
362: size_t client_size;
1.115.2.2 paf 363: if(charsets)
1.115.2.4 paf 364: Charset::transcode(charsets->pool(),
365: charsets->source(), fragment->ptr, fragment->size,
366: charsets->client(), client_ptr, client_size);
1.115.2.2 paf 367: else {
368: client_ptr=fragment->ptr;
369: client_size=fragment->size;
370: }
371:
1.85 paf 372: {
1.115.2.6 paf 373: const char* src=(const char* )client_ptr;
1.85 paf 374: for(int size=client_size; size--; src++)
1.108 paf 375: encode(need_uri_encode, '%');
1.85 paf 376: }
1.77 paf 377: break;
378: case UL_HTTP_HEADER:
379: // tainted, untaint language: http-field-content-text
1.108 paf 380: escape(
381: encode(need_uri_encode, '%');
382: );
1.77 paf 383: break;
384: case UL_MAIL_HEADER:
385: // tainted, untaint language: mail-header
1.105 paf 386: // http://www.ietf.org/rfc/rfc2047.txt
1.115.2.4 paf 387: if(charsets) {
1.115.2.12 paf 388: const void *mail_ptr=fragment->ptr;
389: size_t mail_size=fragment->size;
1.115.2.12.2. (paf 390:): const char* charset_name=charsets->mail().name()->cstr();
1.87 paf 391:
1.77 paf 392: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
1.115.2.6 paf 393: const char* src=(const char* )mail_ptr;
1.77 paf 394: bool to_quoted_printable=false;
1.105 paf 395:
1.111 paf 396: bool email=false;
1.115.2.6 paf 397: for(const char* end=src+mail_size; src<end; src++) {
1.111 paf 398: //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
399: if(to_quoted_printable && (*src==',' || *src=='<')) {
400: email=*src=='<';
1.105 paf 401: dest+=sprintf(dest, "?=");
1.107 paf 402: to_quoted_printable=false;
1.105 paf 403: }
1.111 paf 404: if(!email && (
1.109 paf 405: !to_quoted_printable && (*src & 0x80) // starting quote-printable-encoding on first 8bit char
406: || to_quoted_printable && !mail_header_char_valid_within_Qencoded(*src)
1.105 paf 407: )) {
1.77 paf 408: if(!to_quoted_printable) {
1.115.2.7 paf 409: dest+=sprintf(dest, "=?%s?Q?", charset_name.get());
1.77 paf 410: to_quoted_printable=true;
411: }
1.105 paf 412: //RFC Upper case should be used for hexadecimal digits "A" through "F"
1.109 paf 413: if(*src == 0x20) // RFC The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE)
414: *dest++='_'; // RFC may be represented as "_"
415: else
416: dest+=sprintf(dest, "=%02X", *src & 0xFF);
1.105 paf 417: } else
1.109 paf 418: *dest++=*src;
1.111 paf 419: if(*src=='>')
420: email=false;
1.44 paf 421: }
1.112 paf 422: if(to_quoted_printable) // close
1.77 paf 423: dest+=sprintf(dest, "?=");
1.87 paf 424:
1.77 paf 425: } else {
1.115.2.1 paf 426: memcpy(dest, fragment->ptr, fragment->size);
427: dest+=fragment->size;
1.1 paf 428: }
1.77 paf 429: break;
430: case UL_TABLE:
431: // tainted, untaint language: table
432: escape(switch(*src) {
433: case '\t': to_char(' '); break;
434: case '\n': to_char(' '); break;
435: _default;
436: });
437: break;
438: case UL_SQL:
439: // tainted, untaint language: sql
440: if(connection)
1.115.2.1 paf 441: dest+=connection->quote(dest, fragment->ptr, fragment->size);
1.77 paf 442: else
1.115.2.12.2. (paf 443:): throw Exception(0,
444:): 0,
1.77 paf 445: "untaint in SQL language failed - no connection specified");
446: break;
447: case UL_JS:
448: escape(switch(*src) {
449: case '"': to_string("\\\"", 2); break;
450: case '\'': to_string("\\'", 2); break;
451: case '\n': to_string("\\n", 2); break;
452: case '\\': to_string("\\\\", 2); break;
453: case '\xFF': to_string("\\\xFF", 2); break;
454: _default;
455: });
456: break;
457: case UL_XML:
458: escape(switch(*src) {
459: case '&': to_string("&", 5); break;
460: case '>': to_string(">", 4); break;
461: case '<': to_string("<", 4); break;
462: case '"': to_string(""", 6); break;
463: case '\'': to_string("'", 6); break;
464: _default;
465: });
466: break;
467: case UL_HTML:
468: escape(switch(*src) {
469: case '&': to_string("&", 5); break;
470: case '>': to_string(">", 4); break;
471: case '<': to_string("<", 4); break;
472: case '"': to_string(""", 6); break;
473: _default;
474: });
475: break;
476: default:
1.115.2.12.2. (paf 477:): throw Exception(0,
478:): 0,
1.81 paf 479: "unknown untaint language #%d",
1.110 paf 480: static_cast<int>(to_lang)); // should never
1.77 paf 481: break; // never
1.76 paf 482: }
1.55 parser 483:
1.77 paf 484: if(to_lang & UL_OPTIMIZE_BIT) {
485: // optimizing whitespace
486: char *stop=dest; dest=start;
487: for(char *src=start; src<stop; src++)
488: switch(*src) {
489: // of all consequent white space chars leaving only first one
1.80 paf 490: case ' ': case '\r': case '\n': case '\t':
1.77 paf 491: if(!whitespace) {
492: *dest++=*src;
493: whitespace=true;
494: }
495: break;
496: default:
497: whitespace=false;
498: *dest++=*src;
499: break;
500: };
501: } else // piece without optimization
502: whitespace=false;
1.115.2.1 paf 503: // );
504: }
505:
1.76 paf 506: return dest;
507: }
508:
1.115.2.12.2. (paf 509:): char *String::cstr_debug_origins() const {
1.81 paf 510: //_asm int 3;
1.115.2.12.2. (paf 511:): char *result=new char[size()+fused*MAX_STRING*2];
1.76 paf 512: char *dest=result;
513:
1.115.2.1 paf 514: STRING_FOREACH_FRAGMENT(
1.96 paf 515: IFNDEF_NO_STRING_ORIGIN(
1.115.2.1 paf 516: if(fragment->origin.file)
1.96 paf 517: dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT,
1.115.2.1 paf 518: fragment->origin.file,
519: 1+fragment->origin.line);
1.96 paf 520: else
521: dest+=sprintf(dest, "<unknown>");
522: );
1.115.2.12.2. (paf 523:): Untaint_lang show_lang=fragment->lang & ~UL_OPTIMIZE_BIT;
1.96 paf 524: if(show_lang>=sizeof(String_Untaint_lang_name)/sizeof(String_Untaint_lang_name[0]))
1.115.2.12.2. (paf 525:): throw Exception(0,
526:): 0,
1.96 paf 527: "unknown untaint language #%d",
528: static_cast<int>(show_lang)); // sould never
529:
530: dest+=sprintf(dest, "#%s%s: ",
531: String_Untaint_lang_name[show_lang],
1.115.2.1 paf 532: fragment->lang & UL_OPTIMIZE_BIT?".O":"");
1.96 paf 533: char *dest_after_origins=dest;
1.76 paf 534:
1.115.2.1 paf 535: memcpy(dest, fragment->ptr, fragment->size);
536: dest+=fragment->size;
1.76 paf 537:
1.96 paf 538: remove_crlf(dest_after_origins, dest);
539: to_char('\n');
540: );
1.64 parser 541:
1.76 paf 542: *dest=0;
543: return result;
1.1 paf 544: }
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to untaint.C CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / main