Annotation of parser3/src/main/untaint.C, revision 1.115.2.12.2.4
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: 2003/03/20 08:11:15 $";
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:): #include "ec.h"
22:):
23:): //#define PA_SQL
24:):
25:): #ifdef PA_SQL
26:): #include "pa_sql_connection.h"
27:): #endif
28:):
1.18 paf 29: #define escape(action) \
1.115.2.12.2. (paf 30:): for(; fragment_size--; CORD_next(pos)) { \
31:): char c=CORD_pos_fetch(pos); \
32:): action \
1.1 paf 33: }
1.115.2.12.2. (paf 34:): #define _default default: CORD_ec_append(result, c); break
1.13 paf 35: #define encode(need_encode_func, prefix) \
1.115.2.12.2. (paf 36:): if(need_encode_func(c)) { \
37:): static const char* hex="0123456789ABCDEF"; \
38:): CORD_ec_append(result, prefix); \
39:): CORD_ec_append(result, hex[((unsigned char)c)/0x10]); \
40:): CORD_ec_append(result, hex[((unsigned char)c)%0x10]); \
41:): } else \
42:): CORD_ec_append(result, c);
43:): #define to_char(c) CORD_ec_append(result, c)
44:): #define to_string(s) CORD_ec_append_cord(result, s)
1.4 paf 45:
1.9 paf 46: inline bool need_file_encode(unsigned char c){
1.108 paf 47: // russian letters and space ENABLED
48: // encoding only these...
49: return strchr(
50: "*?'\"<>|"
51: #ifndef WIN32
52: ":\\~"
1.31 paf 53: #endif
1.108 paf 54: , c)!=0;
1.9 paf 55: }
1.5 paf 56: inline bool need_uri_encode(unsigned char c){
1.13 paf 57: if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z'))
1.4 paf 58: return false;
59:
1.5 paf 60: return !strchr("_-./", c);
61: }
1.36 paf 62: inline bool need_http_header_encode(unsigned char c){
1.18 paf 63: if(strchr(" , :", c))
1.5 paf 64: return false;
65:
66: return need_uri_encode(c);
1.4 paf 67: }
1.1 paf 68:
1.56 parser 69: //
70:
1.115.2.6 paf 71: static const char* String_Untaint_lang_name[]={
1.56 parser 72: "U", ///< zero value handy for hash lookup @see untaint_lang_name2enum
73: "C", ///< clean
74: "T", ///< tainted, untaint language as assigned later
75: // untaint languages. assigned by ^untaint[lang]{...}
76: "P",
77: /**<
78: leave language built into string being appended.
79: just a flag, that value not stored
80: */
81: "A", ///< leave all characters intact
1.68 parser 82: "F", ///< file specification
83: "H", ///< ext in HTTP response header
1.56 parser 84: "M", ///< text in mail header
85: "URI", ///< text in uri
86: "T", ///< ^table:set body
87: "SQL", ///< ^table:sql body
88: "JS", ///< JavaScript code
1.68 parser 89: "XML", ///< ^dom:set xml
1.82 paf 90: "HTML" ///< HTML code (for editing)
1.56 parser 91: };
92:
93:
1.1 paf 94: // String
95:
1.41 paf 96: /*
97:
98: HTTP-header = field-name ":" [ field-value ] CRLF
99:
100: field-name = token
101: field-value = *( field-content | LWS )
102:
103: field-content = <the OCTETs making up the field-value
104: and consisting of either *TEXT or combinations
105: of token, tspecials, and quoted-string>
106:
107:
108:
109: word = token | quoted-string
110:
111: token = 1*<any CHAR except CTLs or tspecials>
112:
113:
114:
115: tspecials = "(" | ")" | "<" | ">" | "@"
116: | "," | ";" | ":" | "\" | <">
117: | "/" | "[" | "]" | "?" | "="
118: | "{" | "}" | SP | HT
119:
120: SP = <US-ASCII SP, space (32)>
121: HT = <US-ASCII HT, horizontal-tab (9)>
122:
123: LWS = [CRLF] 1*( SP | HT )
124: TEXT = <any OCTET except CTLs,
125: but including LWS>
126:
127: quoted-pair = "\" CHAR
128:
129: if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
130: */
1.115.2.6 paf 131: inline bool need_quote_http_header(const char* ptr, size_t size) {
1.41 paf 132: for(; size--; ptr++)
1.42 paf 133: if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
1.41 paf 134: return true;
135: return false;
136: }
137:
1.115.2.12.2. (paf 138:):
1.91 paf 139: /**
1.115.2.1 paf 140: appends to other String,
1.91 paf 141: marking all tainted pieces of it with @a lang.
1.92 paf 142: or marking ALL pieces of it with a @a lang when @a forced to,
143: and propagating OPTIMIZE language bit.
1.91 paf 144: */
1.115.2.12.2. (paf 145:): String& String::append_to(String& dest, Untaint_lang lang, bool forced) const {
146:): if(!*this)
1.115.2.1 paf 147: return dest;
148:
1.115.2.12.2. (paf 149:): // first: letters
150:): dest.body=CORD_cat(dest.body, body);
1.91 paf 151:
1.115.2.12.2. (paf 152:): // next: fragment infos
153:):
154:): if(lang==UL_PASS_APPENDED) // without language-chage string?
155:): dest.fragments.append(fragments, 0, fragments.count());
156:): else if(forced) //forcing passed lang?
157:): dest.fragments+=Fragment(lang, size());
158:): else if(lang&UL_OPTIMIZE_BIT) {
159:): Array_iterator<ArrayFragment::element_type> i(fragments);
160:): while(i.has_next()) {
161:): const Fragment& fragment=i.next();
162:): // main idea here:
163:): // tainted piece would get OPTIMIZED bit from 'lang'
164:): // clean piece would be marked OPTIMIZED manually
165:): // pieces with determined languages [not tainted|clean] would retain theirs langs
166:): dest.fragments+=Fragment(static_cast<Untaint_lang>(
167:): fragment.lang==UL_TAINTED?lang:(
168:): fragment.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag
169:): fragment.lang)),
170:): fragment.size);
171:): }
172:): } else { // The core idea: tainted pieces got marked with context's lang
173:): Array_iterator<ArrayFragment::element_type> i(fragments);
174:): while(i.has_next()) {
175:): const Fragment& fragment=i.next();
176:): dest.fragments+=Fragment(
177:): fragment.lang==UL_TAINTED?lang:fragment.lang,
178:): fragment.size);
179:): }
180:): }
1.115.2.1 paf 181:
182: return dest;
1.77 paf 183: }
184:
1.109 paf 185: /** http://www.ietf.org/rfc/rfc2047.txt
186: RFC
187: (3) As a replacement for a 'word' entity within a 'phrase', for example,
188: one that precedes an address in a From, To, or Cc header. The ABNF
189: definition for 'phrase' from RFC 822 thus becomes:
190:
191: phrase = 1*( encoded-word / word )
192:
193: In this case the set of characters that may be used in a "Q"-encoded
194: 'encoded-word' is restricted to: <upper and lower case ASCII
195: letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_"
196: (underscore, ASCII 95.)>. An 'encoded-word' that appears within a
197: 'phrase' MUST be separated from any adjacent 'word', 'text' or
198: 'special' by 'linear-white-space'.
199: ...
200: (2) The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) may be
201: represented as "_" (underscore, ASCII 95.). (This character may
202: not pass through some internetwork mail gateways, but its use
203: will greatly enhance readability of "Q" encoded data with mail
204: readers that do not support this encoding.) Note that the "_"
205: always represents hexadecimal 20, even if the SPACE character
206: occupies a different code position in the character set in use.
207:
208: paf: obviously,
209: without "=", or one could not differ "=E0" and "russian letter a"
210: and without "_", or in would mean 0x20
211: */
1.115.2.12.2. (paf 212:): inline bool mail_header_char_valid_within_Qencoded(char c) {
1.109 paf 213: return c>='A' && c<='Z'
214: || c>='a' && c<='Z'
215: || c>='0' && c<='9'
216: || strchr("!*+-/", c);
217: }
1.115.2.12.2. (paf 218:): inline void ec_append(CORD_ec result, CORD_pos& pos, size_t size) {
219:): while(size--) {
220:): CORD_ec_append(result, CORD_pos_fetch(pos));
221:): CORD_next(pos);
222:): }
223:): }
224:): CORD String::cstr_to_cord(Untaint_lang lang,
225:): SQL_Connection *connection,
226:): const Request_charsets *charsets) const {
227:): CORD_ec result; CORD_ec_init(result);
228:): CORD_pos pos; CORD_set_pos(pos, body, 0);
229:):
1.44 paf 230: bool whitespace=true;
1.111 paf 231:
1.115.2.12.2. (paf 232:): size_t fragment_start=0;
233:): size_t fragment_end;
234:): for(Array_iterator<Fragment> i(fragments); i.has_next(); fragment_start=fragment_end) {
235:): const Fragment& fragment=i.next();
236:): size_t fragment_size=fragment.size;
237:): fragment_end=fragment_start+fragment_size;
1.77 paf 238:
1.115.2.12.2. (paf 239:): Untaint_lang to_lang=lang==UL_UNSPECIFIED?fragment.lang:lang;
1.77 paf 240:
241: switch(to_lang & ~UL_OPTIMIZE_BIT) {
242: case UL_CLEAN:
243: case UL_TAINTED:
244: case UL_AS_IS:
245: // clean piece
246:
247: // tainted piece, but undefined untaint language
248: // for VString.as_double of tainted values
249: // for ^process{body} evaluation
250:
251: // tainted, untaint language: as-is
1.115.2.12.2. (paf 252:): ec_append(result, pos, fragment_size);
1.77 paf 253: break;
254: case UL_FILE_SPEC:
255: // tainted, untaint language: file [name]
1.83 paf 256: escape(
1.113 paf 257: // Macintosh has problems with small Russian letter 'r'
1.115.2.12.2. (paf 258:): if( c=='\xF0' && charsets && charsets->source().name()=="windows-1251" ) {
1.113 paf 259: // fixing that letter for most common charset
1.115.2.2 paf 260: to_char('p');
1.114 paf 261: } else // fallback to default
262: encode(need_file_encode, '_');
1.83 paf 263: );
1.77 paf 264: break;
265: case UL_URI:
266: // tainted, untaint language: uri
1.85 paf 267: {
1.115.2.12.2. (paf 268:): const char *substr=
269:): CORD_to_const_char_star(
270:): CORD_substr(body, fragment_start, fragment_size));
271:): const void *client_ptr;
272:): size_t client_size;
273:): if(charsets)
274:): Charset::transcode(
275:): charsets->source(), substr, fragment_size,
276:): charsets->client(), client_ptr, client_size);
277:): else {
278:): client_ptr=substr;
279:): client_size=fragment_size;
280:): }
281:):
282:): char c;
283:): for(const char* src=(const char* )client_ptr; c=*src++; )
1.108 paf 284: encode(need_uri_encode, '%');
1.85 paf 285: }
1.77 paf 286: break;
287: case UL_HTTP_HEADER:
288: // tainted, untaint language: http-field-content-text
1.108 paf 289: escape(
290: encode(need_uri_encode, '%');
291: );
1.77 paf 292: break;
1.115.2.12.2. (paf 293:): #ifdef LATER
1.77 paf 294: case UL_MAIL_HEADER:
295: // tainted, untaint language: mail-header
1.105 paf 296: // http://www.ietf.org/rfc/rfc2047.txt
1.115.2.4 paf 297: if(charsets) {
1.115.2.12.2. (paf 298:): const void *mail_ptr=fragment.ptr;
299:): size_t mail_size=fragment_size;
300:): const char* charset_name=charsets->mail().name()->cstr();
1.87 paf 301:
1.77 paf 302: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
1.115.2.6 paf 303: const char* src=(const char* )mail_ptr;
1.77 paf 304: bool to_quoted_printable=false;
1.105 paf 305:
1.111 paf 306: bool email=false;
1.115.2.6 paf 307: for(const char* end=src+mail_size; src<end; src++) {
1.111 paf 308: //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
1.115.2.12.2. (paf 309:): if(to_quoted_printable && (c==',' || c=='<')) {
310:): email=c=='<';
1.105 paf 311: dest+=sprintf(dest, "?=");
1.107 paf 312: to_quoted_printable=false;
1.105 paf 313: }
1.111 paf 314: if(!email && (
1.115.2.12.2. (paf 315:): !to_quoted_printable && (c & 0x80) // starting quote-printable-encoding on first 8bit char
316:): || to_quoted_printable && !mail_header_char_valid_within_Qencoded(c)
1.105 paf 317: )) {
1.77 paf 318: if(!to_quoted_printable) {
1.115.2.7 paf 319: dest+=sprintf(dest, "=?%s?Q?", charset_name.get());
1.77 paf 320: to_quoted_printable=true;
321: }
1.105 paf 322: //RFC Upper case should be used for hexadecimal digits "A" through "F"
1.115.2.12.2. (paf 323:): if(c == 0x20) // RFC The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE)
1.109 paf 324: *dest++='_'; // RFC may be represented as "_"
325: else
1.115.2.12.2. (paf 326:): dest+=sprintf(dest, "=%02X", c & 0xFF);
1.105 paf 327: } else
1.115.2.12.2. (paf 328:): *dest++=c;
329:): if(c=='>')
1.111 paf 330: email=false;
1.44 paf 331: }
1.112 paf 332: if(to_quoted_printable) // close
1.77 paf 333: dest+=sprintf(dest, "?=");
1.87 paf 334:
1.77 paf 335: } else {
1.115.2.12.2. (paf 336:): memcpy(dest, fragment.ptr, fragment_size);
337:): dest+=fragment_size;
1.1 paf 338: }
1.77 paf 339: break;
1.115.2.12.2. (paf 340:): #endif
1.77 paf 341: case UL_TABLE:
342: // tainted, untaint language: table
1.115.2.12.2. (paf 343:): escape(switch(c) {
1.77 paf 344: case '\t': to_char(' '); break;
345: case '\n': to_char(' '); break;
346: _default;
347: });
348: break;
1.115.2.12.2. (paf 349:): #ifdef PA_SQL
1.77 paf 350: case UL_SQL:
351: // tainted, untaint language: sql
1.115.2.12.2. (paf 352:): if(connection) {
353:): dest+=connection->quote(0,
354:): CORD_substr(body, fragment_start, fragment_size),
355:): fragment_size);
356:): ...^^^
357:):
358:): dest+=connection->quote(dest, fragment.ptr, fragment_size);
1.77 paf 359: else
1.115.2.12.2. (paf 360:): throw Exception(0,
361:): 0,
1.77 paf 362: "untaint in SQL language failed - no connection specified");
363: break;
1.115.2.12.2. (paf 364:): #endif
1.77 paf 365: case UL_JS:
1.115.2.12.2. (paf 366:): escape(switch(c) {
367:): case '"': to_string("\\\""); break;
368:): case '\'': to_string("\\'"); break;
369:): case '\n': to_string("\\n"); break;
370:): case '\\': to_string("\\\\"); break;
371:): case '\xFF': to_string("\\\xFF"); break;
1.77 paf 372: _default;
373: });
374: break;
375: case UL_XML:
1.115.2.12.2. (paf 376:): escape(switch(c) {
377:): case '&': to_string("&"); break;
378:): case '>': to_string(">"); break;
379:): case '<': to_string("<"); break;
380:): case '"': to_string("""); break;
381:): case '\'': to_string("'"); break;
1.77 paf 382: _default;
383: });
384: break;
385: case UL_HTML:
1.115.2.12.2. (paf 386:): escape(switch(c) {
387:): case '&': to_string("&"); break;
388:): case '>': to_string(">"); break;
389:): case '<': to_string("<"); break;
390:): case '"': to_string("""); break;
1.77 paf 391: _default;
392: });
393: break;
394: default:
1.115.2.12.2. (paf 395:): throw Exception(0,
396:): 0,
1.81 paf 397: "unknown untaint language #%d",
1.110 paf 398: static_cast<int>(to_lang)); // should never
1.77 paf 399: break; // never
1.76 paf 400: }
1.115.2.12.2. (paf 401:): #ifdef LATER
1.77 paf 402: if(to_lang & UL_OPTIMIZE_BIT) {
403: // optimizing whitespace
404: char *stop=dest; dest=start;
1.115.2.12.2. (paf 405:): for(char c=start; src<stop; src++)
406:): switch(c) {
1.77 paf 407: // of all consequent white space chars leaving only first one
1.80 paf 408: case ' ': case '\r': case '\n': case '\t':
1.77 paf 409: if(!whitespace) {
1.115.2.12.2. (paf 410:): *dest++=c;
1.77 paf 411: whitespace=true;
412: }
413: break;
414: default:
415: whitespace=false;
1.115.2.12.2. (paf 416:): *dest++=c;
1.77 paf 417: break;
418: };
419: } else // piece without optimization
420: whitespace=false;
1.115.2.12.2. (paf 421:): #endif
1.115.2.1 paf 422: }
423:
1.115.2.12.2. (paf 424:): return CORD_ec_to_cord(result);
1.1 paf 425: }
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