Annotation of parser3/src/main/untaint.C, revision 1.115.2.12.2.12
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. 2(paf 8:3): static const char* IDENT_UNTAINT_C="$Date: 2003/03/24 14:53:58 $";
(paf 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:): extern "C" { // author forgot to do that
22:): #include "ec.h"
23:): }
24:):
25:): //#define PA_SQL
26:):
27:): #ifdef PA_SQL
28:): #include "pa_sql_connection.h"
29:): #endif
30:):
1.18 paf 31: #define escape(action) \
1.115.2.12.2. (paf 32:): for(; fragment_length--; CORD_next(pos)) { \
33:): char c=CORD_pos_fetch(pos); \
34:): action \
1.1 paf 35: }
1.115.2.12.2. (paf 36:): #define _default default: CORD_ec_append(result, c); break
1.13 paf 37: #define encode(need_encode_func, prefix) \
1.115.2.12.2. (paf 38:): if(need_encode_func(c)) { \
39:): static const char* hex="0123456789ABCDEF"; \
40:): CORD_ec_append(result, prefix); \
41:): CORD_ec_append(result, hex[((unsigned char)c)/0x10]); \
42:): CORD_ec_append(result, hex[((unsigned char)c)%0x10]); \
43:): } else \
44:): CORD_ec_append(result, c);
45:): #define to_char(c) CORD_ec_append(result, c)
46:): #define to_string(s) CORD_ec_append_cord(result, s)
1.4 paf 47:
1.9 paf 48: inline bool need_file_encode(unsigned char c){
1.108 paf 49: // russian letters and space ENABLED
50: // encoding only these...
51: return strchr(
52: "*?'\"<>|"
53: #ifndef WIN32
54: ":\\~"
1.31 paf 55: #endif
1.108 paf 56: , c)!=0;
1.9 paf 57: }
1.5 paf 58: inline bool need_uri_encode(unsigned char c){
1.13 paf 59: if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z'))
1.4 paf 60: return false;
61:
1.5 paf 62: return !strchr("_-./", c);
63: }
1.36 paf 64: inline bool need_http_header_encode(unsigned char c){
1.18 paf 65: if(strchr(" , :", c))
1.5 paf 66: return false;
67:
68: return need_uri_encode(c);
1.4 paf 69: }
1.1 paf 70:
1.56 parser 71: //
72:
1.115.2.6 paf 73: static const char* String_Untaint_lang_name[]={
1.56 parser 74: "U", ///< zero value handy for hash lookup @see untaint_lang_name2enum
75: "C", ///< clean
76: "T", ///< tainted, untaint language as assigned later
77: // untaint languages. assigned by ^untaint[lang]{...}
78: "P",
79: /**<
80: leave language built into string being appended.
81: just a flag, that value not stored
82: */
83: "A", ///< leave all characters intact
1.68 parser 84: "F", ///< file specification
85: "H", ///< ext in HTTP response header
1.56 parser 86: "M", ///< text in mail header
87: "URI", ///< text in uri
88: "T", ///< ^table:set body
89: "SQL", ///< ^table:sql body
90: "JS", ///< JavaScript code
1.68 parser 91: "XML", ///< ^dom:set xml
1.82 paf 92: "HTML" ///< HTML code (for editing)
1.56 parser 93: };
94:
95:
1.1 paf 96: // String
97:
1.41 paf 98: /*
99:
100: HTTP-header = field-name ":" [ field-value ] CRLF
101:
102: field-name = token
103: field-value = *( field-content | LWS )
104:
105: field-content = <the OCTETs making up the field-value
106: and consisting of either *TEXT or combinations
107: of token, tspecials, and quoted-string>
108:
109:
110:
111: word = token | quoted-string
112:
113: token = 1*<any CHAR except CTLs or tspecials>
114:
115:
116:
117: tspecials = "(" | ")" | "<" | ">" | "@"
118: | "," | ";" | ":" | "\" | <">
119: | "/" | "[" | "]" | "?" | "="
120: | "{" | "}" | SP | HT
121:
122: SP = <US-ASCII SP, space (32)>
123: HT = <US-ASCII HT, horizontal-tab (9)>
124:
125: LWS = [CRLF] 1*( SP | HT )
126: TEXT = <any OCTET except CTLs,
127: but including LWS>
128:
129: quoted-pair = "\" CHAR
130:
131: if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
132: */
1.115.2.6 paf 133: inline bool need_quote_http_header(const char* ptr, size_t size) {
1.41 paf 134: for(; size--; ptr++)
1.42 paf 135: if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
1.41 paf 136: return true;
137: return false;
138: }
139:
1.115.2.12.2. (paf 140:):
1.91 paf 141: /**
1.115.2.1 paf 142: appends to other String,
1.91 paf 143: marking all tainted pieces of it with @a lang.
1.92 paf 144: or marking ALL pieces of it with a @a lang when @a forced to,
145: and propagating OPTIMIZE language bit.
1.91 paf 146: */
1.115.2.12.2. (paf 147:): String& String::append_to(String& dest, Language lang, bool forced) const {
0(paf 148:3): if(is_empty())
1.115.2.1 paf 149: return dest;
150:
1.115.2.12.2. (paf 151:): // first: letters
0(paf 152:3): dest.body.append(body);
1.91 paf 153:
1.115.2.12.2. (paf 154:): // next: fragment infos
155:):
156:): if(lang==L_PASS_APPENDED) // without language-chage string?
157:): dest.fragments.append(fragments);
158:): else if(forced) //forcing passed lang?
159:): dest.fragments+=Fragment(lang, length());
160:): else if(lang&L_OPTIMIZE_BIT) {
161:): Array_iterator<ArrayFragment::element_type> i(fragments);
162:): while(i.has_next()) {
163:): const Fragment& fragment=i.next();
164:): // main idea here:
165:): // tainted piece would get OPTIMIZED bit from 'lang'
166:): // clean piece would be marked OPTIMIZED manually
167:): // pieces with determined languages [not tainted|clean] would retain theirs langs
168:): dest.fragments+=Fragment(static_cast<Language>(
169:): fragment.lang==L_TAINTED?lang:(
170:): fragment.lang==L_CLEAN?L_CLEAN|L_OPTIMIZE_BIT: // ORing with OPTIMIZED flag
171:): fragment.lang)),
172:): fragment.length);
173:): }
174:): } else { // The core idea: tainted pieces got marked with context's lang
175:): Array_iterator<ArrayFragment::element_type> i(fragments);
176:): while(i.has_next()) {
177:): const Fragment& fragment=i.next();
178:): dest.fragments+=Fragment(
179:): fragment.lang==L_TAINTED?lang:fragment.lang,
180:): fragment.length);
181:): }
182:): }
1.115.2.1 paf 183:
184: return dest;
1.77 paf 185: }
186:
1.109 paf 187: /** http://www.ietf.org/rfc/rfc2047.txt
188: RFC
189: (3) As a replacement for a 'word' entity within a 'phrase', for example,
190: one that precedes an address in a From, To, or Cc header. The ABNF
191: definition for 'phrase' from RFC 822 thus becomes:
192:
193: phrase = 1*( encoded-word / word )
194:
195: In this case the set of characters that may be used in a "Q"-encoded
196: 'encoded-word' is restricted to: <upper and lower case ASCII
197: letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_"
198: (underscore, ASCII 95.)>. An 'encoded-word' that appears within a
199: 'phrase' MUST be separated from any adjacent 'word', 'text' or
200: 'special' by 'linear-white-space'.
201: ...
202: (2) The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) may be
203: represented as "_" (underscore, ASCII 95.). (This character may
204: not pass through some internetwork mail gateways, but its use
205: will greatly enhance readability of "Q" encoded data with mail
206: readers that do not support this encoding.) Note that the "_"
207: always represents hexadecimal 20, even if the SPACE character
208: occupies a different code position in the character set in use.
209:
210: paf: obviously,
211: without "=", or one could not differ "=E0" and "russian letter a"
212: and without "_", or in would mean 0x20
213: */
1.115.2.12.2. (paf 214:): inline bool mail_header_char_valid_within_Qencoded(char c) {
1.109 paf 215: return c>='A' && c<='Z'
216: || c>='a' && c<='Z'
217: || c>='0' && c<='9'
218: || strchr("!*+-/", c);
219: }
1.115.2.12.2. 1(paf 220:3): inline void ec_append(CORD_ec& result, CORD_pos& pos, size_t size) {
(paf 221:): while(size--) {
222:): CORD_ec_append(result, CORD_pos_fetch(pos));
223:): CORD_next(pos);
224:): }
225:): }
0(paf 226:3): StringBody String::cstr_to_string_body(Language lang,
(paf 227:): SQL_Connection* connection,
228:): const Request_charsets *charsets) const {
229:): CORD_ec result; CORD_ec_init(result);
0(paf 230:3): CORD_pos pos; body.set_pos(pos, 0);
(paf 231:):
1.44 paf 232: bool whitespace=true;
1.111 paf 233:
1.115.2.12.2. (paf 234:): size_t fragment_begin=0;
235:): size_t fragment_end;
236:): for(Array_iterator<Fragment> i(fragments); i.has_next(); fragment_begin=fragment_end) {
237:): const Fragment& fragment=i.next();
238:): size_t fragment_length=fragment.length;
239:): fragment_end=fragment_begin+fragment_length;
2(paf 240:3): //fprintf(stderr, "%d, %d\n", fragment.lang, fragment.length);
1.77 paf 241:
1.115.2.12.2. (paf 242:): Language to_lang=lang==L_UNSPECIFIED?fragment.lang:lang;
1.77 paf 243:
1.115.2.12.2. (paf 244:): switch(to_lang & ~L_OPTIMIZE_BIT) {
245:): case L_CLEAN:
246:): case L_TAINTED:
247:): case L_AS_IS:
1.77 paf 248: // clean piece
249:
250: // tainted piece, but undefined untaint language
251: // for VString.as_double of tainted values
252: // for ^process{body} evaluation
253:
254: // tainted, untaint language: as-is
1.115.2.12.2. (paf 255:): ec_append(result, pos, fragment_length);
1.77 paf 256: break;
1.115.2.12.2. (paf 257:): case L_FILE_SPEC:
1.77 paf 258: // tainted, untaint language: file [name]
1.83 paf 259: escape(
1.113 paf 260: // Macintosh has problems with small Russian letter 'r'
1.115.2.12.2. (paf 261:): if( c=='\xF0' && charsets && charsets->source().name()=="windows-1251" ) {
1.113 paf 262: // fixing that letter for most common charset
1.115.2.2 paf 263: to_char('p');
1.114 paf 264: } else // fallback to default
265: encode(need_file_encode, '_');
1.83 paf 266: );
1.77 paf 267: break;
1.115.2.12.2. (paf 268:): #ifdef LATER
269:): case L_URI:
1.77 paf 270: // tainted, untaint language: uri
1.85 paf 271: {
1.115.2.12.2. (paf 272:): const char *substr=
273:): CORD_to_const_char_star(
274:): CORD_substr(body, fragment_begin, fragment_length));
275:): const void *client_ptr;
276:): size_t client_size;
277:): if(charsets)
278:): Charset::transcode(
279:): charsets->source(), substr, fragment_length,
280:): charsets->client(), client_ptr, client_size);
281:): else {
282:): client_ptr=substr;
283:): client_size=fragment_length;
284:): }
285:):
286:): char c;
287:): for(const char* src=(const char* )client_ptr; c=*src++; )
1.108 paf 288: encode(need_uri_encode, '%');
1.85 paf 289: }
1.77 paf 290: break;
1.115.2.12.2. (paf 291:): #endif
292:): case L_HTTP_HEADER:
1.77 paf 293: // tainted, untaint language: http-field-content-text
1.108 paf 294: escape(
295: encode(need_uri_encode, '%');
296: );
1.77 paf 297: break;
1.115.2.12.2. (paf 298:): #ifdef LATER
299:): case L_MAIL_HEADER:
1.77 paf 300: // tainted, untaint language: mail-header
1.105 paf 301: // http://www.ietf.org/rfc/rfc2047.txt
1.115.2.4 paf 302: if(charsets) {
1.115.2.12.2. (paf 303:): const void *mail_ptr=fragment.ptr;
304:): size_t mail_size=fragment_length;
305:): const char* charset_name=charsets->mail().name()->cstr();
1.87 paf 306:
1.77 paf 307: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
1.115.2.6 paf 308: const char* src=(const char* )mail_ptr;
1.77 paf 309: bool to_quoted_printable=false;
1.105 paf 310:
1.111 paf 311: bool email=false;
1.115.2.6 paf 312: for(const char* end=src+mail_size; src<end; src++) {
1.111 paf 313: //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
1.115.2.12.2. (paf 314:): if(to_quoted_printable && (c==',' || c=='<')) {
315:): email=c=='<';
1.105 paf 316: dest+=sprintf(dest, "?=");
1.107 paf 317: to_quoted_printable=false;
1.105 paf 318: }
1.111 paf 319: if(!email && (
1.115.2.12.2. (paf 320:): !to_quoted_printable && (c & 0x80) // starting quote-printable-encoding on first 8bit char
321:): || to_quoted_printable && !mail_header_char_valid_within_Qencoded(c)
1.105 paf 322: )) {
1.77 paf 323: if(!to_quoted_printable) {
1.115.2.7 paf 324: dest+=sprintf(dest, "=?%s?Q?", charset_name.get());
1.77 paf 325: to_quoted_printable=true;
326: }
1.105 paf 327: //RFC Upper case should be used for hexadecimal digits "A" through "F"
1.115.2.12.2. (paf 328:): if(c == 0x20) // RFC The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE)
1.109 paf 329: *dest++='_'; // RFC may be represented as "_"
330: else
1.115.2.12.2. (paf 331:): dest+=sprintf(dest, "=%02X", c & 0xFF);
1.105 paf 332: } else
1.115.2.12.2. (paf 333:): *dest++=c;
334:): if(c=='>')
1.111 paf 335: email=false;
1.44 paf 336: }
1.112 paf 337: if(to_quoted_printable) // close
1.77 paf 338: dest+=sprintf(dest, "?=");
1.87 paf 339:
1.77 paf 340: } else {
1.115.2.12.2. (paf 341:): memcpy(dest, fragment.ptr, fragment_length);
342:): dest+=fragment_length;
1.1 paf 343: }
1.77 paf 344: break;
1.115.2.12.2. (paf 345:): #endif
346:): case L_TABLE:
1.77 paf 347: // tainted, untaint language: table
1.115.2.12.2. (paf 348:): escape(switch(c) {
1.77 paf 349: case '\t': to_char(' '); break;
350: case '\n': to_char(' '); break;
351: _default;
352: });
353: break;
1.115.2.12.2. (paf 354:): #ifdef PA_SQL
355:): case L_SQL:
1.77 paf 356: // tainted, untaint language: sql
1.115.2.12.2. (paf 357:): if(connection) {
358:): dest+=connection->quote(0,
359:): CORD_substr(body, fragment_begin, fragment_length),
360:): fragment_length);
361:): ...^^^
362:):
363:): dest+=connection->quote(dest, fragment.ptr, fragment_length);
1.77 paf 364: else
1.115.2.12.2. (paf 365:): throw Exception(0,
366:): 0,
1.77 paf 367: "untaint in SQL language failed - no connection specified");
368: break;
1.115.2.12.2. (paf 369:): #endif
370:): case L_JS:
371:): escape(switch(c) {
372:): case '"': to_string("\\\""); break;
373:): case '\'': to_string("\\'"); break;
374:): case '\n': to_string("\\n"); break;
375:): case '\\': to_string("\\\\"); break;
376:): case '\xFF': to_string("\\\xFF"); break;
1.77 paf 377: _default;
378: });
379: break;
1.115.2.12.2. (paf 380:): case L_XML:
381:): escape(switch(c) {
382:): case '&': to_string("&"); break;
383:): case '>': to_string(">"); break;
384:): case '<': to_string("<"); break;
385:): case '"': to_string("""); break;
386:): case '\'': to_string("'"); break;
1.77 paf 387: _default;
388: });
389: break;
1.115.2.12.2. (paf 390:): case L_HTML:
391:): escape(switch(c) {
392:): case '&': to_string("&"); break;
393:): case '>': to_string(">"); break;
394:): case '<': to_string("<"); break;
395:): case '"': to_string("""); break;
1.77 paf 396: _default;
397: });
398: break;
399: default:
1.115.2.12.2. (paf 400:): throw Exception(0,
401:): 0,
1.81 paf 402: "unknown untaint language #%d",
1.110 paf 403: static_cast<int>(to_lang)); // should never
1.77 paf 404: break; // never
1.76 paf 405: }
1.115.2.12.2. (paf 406:): #ifdef LATER
407:): if(to_lang & L_OPTIMIZE_BIT) {
1.77 paf 408: // optimizing whitespace
409: char *stop=dest; dest=start;
1.115.2.12.2. (paf 410:): for(char c=start; src<stop; src++)
411:): switch(c) {
1.77 paf 412: // of all consequent white space chars leaving only first one
1.80 paf 413: case ' ': case '\r': case '\n': case '\t':
1.77 paf 414: if(!whitespace) {
1.115.2.12.2. (paf 415:): *dest++=c;
1.77 paf 416: whitespace=true;
417: }
418: break;
419: default:
420: whitespace=false;
1.115.2.12.2. (paf 421:): *dest++=c;
1.77 paf 422: break;
423: };
424: } else // piece without optimization
425: whitespace=false;
1.115.2.12.2. (paf 426:): #endif
1.115.2.1 paf 427: }
428:
1.115.2.12.2. 0(paf 429:3): return StringBody(CORD_ec_to_cord(result));
1.1 paf 430: }
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