Annotation of parser3/src/main/untaint.C, revision 1.115.2.12.2.8
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 15:05:59 $";
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, Untaint_lang lang, bool forced) const {
148:): if(!*this)
1.115.2.1 paf 149: return dest;
150:
1.115.2.12.2. (paf 151:): // first: letters
152:): dest.body=CORD_cat(dest.body, body);
1.91 paf 153:
1.115.2.12.2. (paf 154:): // next: fragment infos
155:):
156:): if(lang==UL_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&UL_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<Untaint_lang>(
169:): fragment.lang==UL_TAINTED?lang:(
170:): fragment.lang==UL_CLEAN?UL_CLEAN|UL_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==UL_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. (paf 220:): inline void ec_append(CORD_ec result, CORD_pos& pos, size_t size) {
221:): while(size--) {
222:): CORD_ec_append(result, CORD_pos_fetch(pos));
223:): CORD_next(pos);
224:): }
225:): }
226:): CORD String::cstr_to_cord(Untaint_lang lang,
227:): SQL_Connection *connection,
228:): const Request_charsets *charsets) const {
229:): CORD_ec result; CORD_ec_init(result);
230:): CORD_pos pos; CORD_set_pos(pos, body, 0);
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;
1.77 paf 240:
1.115.2.12.2. (paf 241:): Untaint_lang to_lang=lang==UL_UNSPECIFIED?fragment.lang:lang;
1.77 paf 242:
243: switch(to_lang & ~UL_OPTIMIZE_BIT) {
244: case UL_CLEAN:
245: case UL_TAINTED:
246: case UL_AS_IS:
247: // clean piece
248:
249: // tainted piece, but undefined untaint language
250: // for VString.as_double of tainted values
251: // for ^process{body} evaluation
252:
253: // tainted, untaint language: as-is
1.115.2.12.2. (paf 254:): ec_append(result, pos, fragment_length);
1.77 paf 255: break;
256: case UL_FILE_SPEC:
257: // tainted, untaint language: file [name]
1.83 paf 258: escape(
1.113 paf 259: // Macintosh has problems with small Russian letter 'r'
1.115.2.12.2. (paf 260:): if( c=='\xF0' && charsets && charsets->source().name()=="windows-1251" ) {
1.113 paf 261: // fixing that letter for most common charset
1.115.2.2 paf 262: to_char('p');
1.114 paf 263: } else // fallback to default
264: encode(need_file_encode, '_');
1.83 paf 265: );
1.77 paf 266: break;
1.115.2.12.2. (paf 267:): #ifdef LATER
1.77 paf 268: case UL_URI:
269: // tainted, untaint language: uri
1.85 paf 270: {
1.115.2.12.2. (paf 271:): const char *substr=
272:): CORD_to_const_char_star(
273:): CORD_substr(body, fragment_begin, fragment_length));
274:): const void *client_ptr;
275:): size_t client_size;
276:): if(charsets)
277:): Charset::transcode(
278:): charsets->source(), substr, fragment_length,
279:): charsets->client(), client_ptr, client_size);
280:): else {
281:): client_ptr=substr;
282:): client_size=fragment_length;
283:): }
284:):
285:): char c;
286:): for(const char* src=(const char* )client_ptr; c=*src++; )
1.108 paf 287: encode(need_uri_encode, '%');
1.85 paf 288: }
1.77 paf 289: break;
1.115.2.12.2. (paf 290:): #endif
1.77 paf 291: case UL_HTTP_HEADER:
292: // tainted, untaint language: http-field-content-text
1.108 paf 293: escape(
294: encode(need_uri_encode, '%');
295: );
1.77 paf 296: break;
1.115.2.12.2. (paf 297:): #ifdef LATER
1.77 paf 298: case UL_MAIL_HEADER:
299: // tainted, untaint language: mail-header
1.105 paf 300: // http://www.ietf.org/rfc/rfc2047.txt
1.115.2.4 paf 301: if(charsets) {
1.115.2.12.2. (paf 302:): const void *mail_ptr=fragment.ptr;
303:): size_t mail_size=fragment_length;
304:): const char* charset_name=charsets->mail().name()->cstr();
1.87 paf 305:
1.77 paf 306: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
1.115.2.6 paf 307: const char* src=(const char* )mail_ptr;
1.77 paf 308: bool to_quoted_printable=false;
1.105 paf 309:
1.111 paf 310: bool email=false;
1.115.2.6 paf 311: for(const char* end=src+mail_size; src<end; src++) {
1.111 paf 312: //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
1.115.2.12.2. (paf 313:): if(to_quoted_printable && (c==',' || c=='<')) {
314:): email=c=='<';
1.105 paf 315: dest+=sprintf(dest, "?=");
1.107 paf 316: to_quoted_printable=false;
1.105 paf 317: }
1.111 paf 318: if(!email && (
1.115.2.12.2. (paf 319:): !to_quoted_printable && (c & 0x80) // starting quote-printable-encoding on first 8bit char
320:): || to_quoted_printable && !mail_header_char_valid_within_Qencoded(c)
1.105 paf 321: )) {
1.77 paf 322: if(!to_quoted_printable) {
1.115.2.7 paf 323: dest+=sprintf(dest, "=?%s?Q?", charset_name.get());
1.77 paf 324: to_quoted_printable=true;
325: }
1.105 paf 326: //RFC Upper case should be used for hexadecimal digits "A" through "F"
1.115.2.12.2. (paf 327:): if(c == 0x20) // RFC The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE)
1.109 paf 328: *dest++='_'; // RFC may be represented as "_"
329: else
1.115.2.12.2. (paf 330:): dest+=sprintf(dest, "=%02X", c & 0xFF);
1.105 paf 331: } else
1.115.2.12.2. (paf 332:): *dest++=c;
333:): if(c=='>')
1.111 paf 334: email=false;
1.44 paf 335: }
1.112 paf 336: if(to_quoted_printable) // close
1.77 paf 337: dest+=sprintf(dest, "?=");
1.87 paf 338:
1.77 paf 339: } else {
1.115.2.12.2. (paf 340:): memcpy(dest, fragment.ptr, fragment_length);
341:): dest+=fragment_length;
1.1 paf 342: }
1.77 paf 343: break;
1.115.2.12.2. (paf 344:): #endif
1.77 paf 345: case UL_TABLE:
346: // tainted, untaint language: table
1.115.2.12.2. (paf 347:): escape(switch(c) {
1.77 paf 348: case '\t': to_char(' '); break;
349: case '\n': to_char(' '); break;
350: _default;
351: });
352: break;
1.115.2.12.2. (paf 353:): #ifdef PA_SQL
1.77 paf 354: case UL_SQL:
355: // tainted, untaint language: sql
1.115.2.12.2. (paf 356:): if(connection) {
357:): dest+=connection->quote(0,
358:): CORD_substr(body, fragment_begin, fragment_length),
359:): fragment_length);
360:): ...^^^
361:):
362:): dest+=connection->quote(dest, fragment.ptr, fragment_length);
1.77 paf 363: else
1.115.2.12.2. (paf 364:): throw Exception(0,
365:): 0,
1.77 paf 366: "untaint in SQL language failed - no connection specified");
367: break;
1.115.2.12.2. (paf 368:): #endif
1.77 paf 369: case UL_JS:
1.115.2.12.2. (paf 370:): escape(switch(c) {
371:): case '"': to_string("\\\""); break;
372:): case '\'': to_string("\\'"); break;
373:): case '\n': to_string("\\n"); break;
374:): case '\\': to_string("\\\\"); break;
375:): case '\xFF': to_string("\\\xFF"); break;
1.77 paf 376: _default;
377: });
378: break;
379: case UL_XML:
1.115.2.12.2. (paf 380:): escape(switch(c) {
381:): case '&': to_string("&"); break;
382:): case '>': to_string(">"); break;
383:): case '<': to_string("<"); break;
384:): case '"': to_string("""); break;
385:): case '\'': to_string("'"); break;
1.77 paf 386: _default;
387: });
388: break;
389: case UL_HTML:
1.115.2.12.2. (paf 390:): escape(switch(c) {
391:): case '&': to_string("&"); break;
392:): case '>': to_string(">"); break;
393:): case '<': to_string("<"); break;
394:): case '"': to_string("""); break;
1.77 paf 395: _default;
396: });
397: break;
398: default:
1.115.2.12.2. (paf 399:): throw Exception(0,
400:): 0,
1.81 paf 401: "unknown untaint language #%d",
1.110 paf 402: static_cast<int>(to_lang)); // should never
1.77 paf 403: break; // never
1.76 paf 404: }
1.115.2.12.2. (paf 405:): #ifdef LATER
1.77 paf 406: if(to_lang & UL_OPTIMIZE_BIT) {
407: // optimizing whitespace
408: char *stop=dest; dest=start;
1.115.2.12.2. (paf 409:): for(char c=start; src<stop; src++)
410:): switch(c) {
1.77 paf 411: // of all consequent white space chars leaving only first one
1.80 paf 412: case ' ': case '\r': case '\n': case '\t':
1.77 paf 413: if(!whitespace) {
1.115.2.12.2. (paf 414:): *dest++=c;
1.77 paf 415: whitespace=true;
416: }
417: break;
418: default:
419: whitespace=false;
1.115.2.12.2. (paf 420:): *dest++=c;
1.77 paf 421: break;
422: };
423: } else // piece without optimization
424: whitespace=false;
1.115.2.12.2. (paf 425:): #endif
1.115.2.1 paf 426: }
427:
1.115.2.12.2. (paf 428:): return CORD_ec_to_cord(result);
1.1 paf 429: }
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