Annotation of parser3/src/main/untaint.C, revision 1.166
1.7 paf 1: /** @file
1.8 paf 2: Parser: String class part: untaint mechanizm.
3:
1.163 moko 4: Copyright (c) 2001-2012 Art. Lebedev Studio (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:
1.166 ! moko 8: volatile const char * IDENT_UNTAINT_C="$Id: untaint.C,v 1.165 2013/10/14 21:17:38 moko Exp $";
1.117 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"
1.117 paf 19: #include "pa_request_charsets.h"
20: #include "pa_sapi.h"
1.1 paf 21:
1.117 paf 22: extern "C" { // author forgot to do that
23: #include "ec.h"
24: }
1.91 paf 25:
1.117 paf 26: #include "pa_sql_connection.h"
1.91 paf 27:
1.117 paf 28: // defines
29:
30: #undef CORD_ec_append
31: // redefining to intercept flushes and implement whitespace optimization
32: // of all consequent white space chars leaving only first one
33: #define CORD_ec_append(x, c) \
34: { \
35: bool skip=false; \
36: if(optimize) switch(c) { \
1.131 paf 37: case ' ': case '\n': case '\t': \
1.117 paf 38: if(whitespace) \
39: skip=true; /*skipping subsequent*/ \
40: else \
41: whitespace=true; \
42: break; \
43: default: \
44: whitespace=false; \
45: break; \
46: } \
47: if(!skip) { \
48: if ((x)[0].ec_bufptr == (x)[0].ec_buf + CORD_BUFSZ) { \
49: CORD_ec_flush_buf(x); \
50: } \
51: *((x)[0].ec_bufptr)++ = (c); \
52: } \
53: }
54:
55:
1.140 misha 56: #define escape_fragment(action) \
1.119 paf 57: for(; fragment_length--; CORD_next(info->pos)) { \
58: char c=CORD_pos_fetch(info->pos); \
1.117 paf 59: action \
1.1 paf 60: }
1.133 paf 61: #define _default CORD_ec_append(info->result, c)
1.117 paf 62: #define encode(need_encode_func, prefix, otherwise) \
63: if(need_encode_func(c)) { \
1.119 paf 64: CORD_ec_append(info->result, prefix); \
1.165 moko 65: CORD_ec_append(info->result, hex_digits[((unsigned char)c) >> 4]); \
66: CORD_ec_append(info->result, hex_digits[((unsigned char)c) & 0x0F]); \
1.117 paf 67: } else \
1.119 paf 68: CORD_ec_append(info->result, otherwise);
1.126 paf 69: #define to_char(c) { CORD_ec_append(info->result, c); whitespace=false; }
70: #define to_string(s) { CORD_ec_append_cord(info->result, s); whitespace=false; }
1.4 paf 71:
1.9 paf 72: inline bool need_file_encode(unsigned char c){
1.108 paf 73: // russian letters and space ENABLED
74: // encoding only these...
75: return strchr(
1.164 misha 76: "*?\"<>|"
1.108 paf 77: #ifndef WIN32
1.143 misha 78: ":\\"
1.31 paf 79: #endif
1.143 misha 80: , c)!=0;
1.9 paf 81: }
1.143 misha 82:
1.5 paf 83: inline bool need_uri_encode(unsigned char c){
1.162 misha 84: return !(pa_isalnum(c) || strchr("_-./*", c));
1.4 paf 85: }
1.143 misha 86:
1.136 paf 87: inline bool need_regex_escape(unsigned char c){
1.142 misha 88: return strchr("\\^$.[]|()?*+{}-", c)!=0;
1.136 paf 89: }
1.1 paf 90:
1.150 misha 91: inline bool need_parser_code_escape(unsigned char c){
92: return strchr("^$;@()[]{}:#\"", c)!=0;
93: }
94:
1.1 paf 95: // String
96:
1.41 paf 97: /*
98: HTTP-header = field-name ":" [ field-value ] CRLF
99:
1.152 misha 100: field-name = token
101: field-value = *( field-content | LWS )
1.41 paf 102:
1.152 misha 103: field-content = <the OCTETs making up the field-value
1.41 paf 104: and consisting of either *TEXT or combinations
105: of token, tspecials, and quoted-string>
106:
107:
1.152 misha 108: token = 1*<any CHAR except CTLs or tspecials>
1.41 paf 109: word = token | quoted-string
1.152 misha 110: quoted-string = ( <"> *(qdtext | quoted-pair ) <"> )
111: qdtext = <any TEXT except <">>
112: quoted-pair = "\" CHAR
1.41 paf 113:
1.152 misha 114: OCTET = <any 8-bit sequence of data>
115: CHAR = <any US-ASCII character (octets 0 - 127)>
1.41 paf 116:
117: tspecials = "(" | ")" | "<" | ">" | "@"
1.152 misha 118: | "," | ";" | ":" | "\" | <">
119: | "/" | "[" | "]" | "?" | "="
120: | "{" | "}" | SP | HT
1.41 paf 121:
122: SP = <US-ASCII SP, space (32)>
123: HT = <US-ASCII HT, horizontal-tab (9)>
124:
125: LWS = [CRLF] 1*( SP | HT )
1.152 misha 126: TEXT = <any OCTET except CTLs, but including LWS>
127: CTL = <any US-ASCII control character (octets 0 - 31) and DEL (127)>
1.41 paf 128:
129:
130: if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
131: */
1.117 paf 132: inline bool need_quote_http_header(const char* ptr, size_t size) {
1.41 paf 133: for(; size--; ptr++)
1.42 paf 134: if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
1.41 paf 135: return true;
136: return false;
137: }
138:
1.119 paf 139: #ifndef DOXYGEN
140: struct Append_fragment_info {
141: String::Language lang;
142: String::Languages* dest_languages;
143: size_t dest_body_plan_length;
144: };
145: #endif
146: int append_fragment_optimizing(char alang, size_t asize, Append_fragment_info* info) {
147: const String::Language lang=(String::Language)(unsigned char)alang;
148: // main idea here:
149: // tainted piece would get OPTIMIZED bit from 'lang'
150: // clean piece would be marked OPTIMIZED manually
151: // pieces with determined languages [not tainted|clean] would retain theirs langs
152: info->dest_languages->append(info->dest_body_plan_length,
1.120 paf 153: lang==String::L_TAINTED?
1.119 paf 154: info->lang
1.120 paf 155: :lang==String::L_CLEAN?
156: (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT) // ORing with OPTIMIZED flag
1.119 paf 157: :lang,
158: asize);
159: info->dest_body_plan_length+=asize;
160:
161: return 0; // 0=continue
162: }
163: int append_fragment_nonoptimizing(char alang, size_t asize, Append_fragment_info* info) {
164: const String::Language lang=(String::Language)(unsigned char)alang;
165: // The core idea: tainted pieces got marked with context's lang
166: info->dest_languages->append(info->dest_body_plan_length,
1.120 paf 167: lang==String::L_TAINTED?
1.119 paf 168: info->lang
169: :lang,
170: asize);
171: info->dest_body_plan_length+=asize;
172:
173: return 0; // 0=continue
174: }
1.117 paf 175:
1.91 paf 176: /**
1.117 paf 177: appends to other String,
1.91 paf 178: marking all tainted pieces of it with @a lang.
1.92 paf 179: or marking ALL pieces of it with a @a lang when @a forced to,
180: and propagating OPTIMIZE language bit.
1.117 paf 181: */
1.147 misha 182: String& String::append_to(String& dest, Language ilang, bool forced) const {
1.117 paf 183: if(is_empty())
184: return dest;
1.91 paf 185:
1.119 paf 186: // first: fragment infos
1.117 paf 187:
1.147 misha 188: if(ilang==L_PASS_APPENDED) // without language-change?
189: dest.langs.appendHelper(dest.body, langs, body);
1.117 paf 190: else if(forced) //forcing passed lang?
1.147 misha 191: dest.langs.appendHelper(dest.body, ilang, body);
192: else {
193: if(langs.opt.is_not_just_lang){
194: Append_fragment_info info={ilang, &dest.langs, dest.body.length()};
195: langs.for_each(body, ilang&L_OPTIMIZE_BIT?
196: append_fragment_optimizing
197: :append_fragment_nonoptimizing, &info);
198: } else {
199: Language lang=langs.opt.lang;
200: // see append_fragment_* for explanation
201: if(ilang&L_OPTIMIZE_BIT){
202: dest.langs.appendHelper(dest.body,
203: lang==String::L_TAINTED?
204: ilang
205: :lang==String::L_CLEAN?
206: (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT)
207: :lang,
208: body);
209: } else {
210: dest.langs.appendHelper(dest.body, lang==String::L_TAINTED ? ilang:lang, body);
211: }
212: }
1.91 paf 213: }
214:
1.119 paf 215: // next: letters
216: dest.body<<body;
217:
1.117 paf 218: ASSERT_STRING_INVARIANT(dest);
1.75 paf 219: return dest;
1.51 parser 220: }
221:
1.109 paf 222: /** http://www.ietf.org/rfc/rfc2047.txt
223: RFC
224: (3) As a replacement for a 'word' entity within a 'phrase', for example,
225: one that precedes an address in a From, To, or Cc header. The ABNF
226: definition for 'phrase' from RFC 822 thus becomes:
227:
228: phrase = 1*( encoded-word / word )
229:
230: In this case the set of characters that may be used in a "Q"-encoded
231: 'encoded-word' is restricted to: <upper and lower case ASCII
232: letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_"
233: (underscore, ASCII 95.)>. An 'encoded-word' that appears within a
234: 'phrase' MUST be separated from any adjacent 'word', 'text' or
235: 'special' by 'linear-white-space'.
236: ...
237: (2) The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) may be
238: represented as "_" (underscore, ASCII 95.). (This character may
239: not pass through some internetwork mail gateways, but its use
240: will greatly enhance readability of "Q" encoded data with mail
241: readers that do not support this encoding.) Note that the "_"
242: always represents hexadecimal 20, even if the SPACE character
243: occupies a different code position in the character set in use.
244:
245: paf: obviously,
246: without "=", or one could not differ "=E0" and "russian letter a"
247: and without "_", or in would mean 0x20
248: */
1.117 paf 249: inline bool mail_header_char_valid_within_Qencoded(char c) {
1.162 misha 250: return (pa_isalnum((unsigned char)c) || strchr("!*+-/", c));
1.109 paf 251: }
1.118 paf 252: inline bool addr_spec_soon(const char *src) {
1.124 paf 253: for(char c; (c=*src); src++)
1.118 paf 254: if(c=='<')
255: return true;
256: else if(!(c==' ' || c=='\t'))
257: return false;
258: return false;
259: }
1.117 paf 260: /**
261: RFC
262: Upper case should be used for hexadecimal digits "A" through "F"
263: The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE)
264: may be represented as "_"
265: */
266: inline bool mail_header_nonspace_char(char c) {
267: return c != 0x20;
268: }
1.125 paf 269:
1.117 paf 270: inline void ec_append(CORD_ec& result, bool& optimize, bool& whitespace, CORD_pos pos, size_t size) {
271: while(size--) {
272: CORD_ec_append(result, CORD_pos_fetch(pos));
273: CORD_next(pos);
274: }
275: }
276: inline void pa_CORD_pos_advance(CORD_pos pos, size_t n) {
277: while(true) {
1.125 paf 278: long avail=CORD_pos_chars_left(pos);
279: if(avail<=0) {
1.117 paf 280: CORD_next(pos);
281: if(!--n)
282: break;
1.125 paf 283: } else if((size_t)avail<n) {
1.117 paf 284: CORD_pos_advance(pos, avail);
285: n-=avail;
286: } else { // avail>=n
287: CORD_pos_advance(pos, n);
288: break;
289: }
290: }
291: }
292:
1.119 paf 293: #ifndef DOXYGEN
294: struct Cstr_to_string_body_block_info {
295: // input
296: String::Language lang;
297: SQL_Connection* connection;
298: const Request_charsets* charsets;
299: const String::Body* body;
300:
301: // output
302: CORD_ec result;
303:
304: // private
305: CORD_pos pos;
306: size_t fragment_begin;
307: bool whitespace;
1.138 misha 308: const char* exception;
1.119 paf 309: };
310: #endif
1.151 misha 311:
312: // @todo: replace info->body->mid with something that uses info->pos
1.148 misha 313: int cstr_to_string_body_block(String::Language to_lang, size_t fragment_length, Cstr_to_string_body_block_info* info) {
1.119 paf 314: bool& whitespace=info->whitespace;
315: size_t fragment_end=info->fragment_begin+fragment_length;
1.148 misha 316: //fprintf(stderr, "%d, %d =%s=\n", to_lang, fragment_length, info->body->cstr());
1.119 paf 317:
1.120 paf 318: bool optimize=(to_lang & String::L_OPTIMIZE_BIT)!=0;
1.119 paf 319: if(!optimize)
320: whitespace=false;
321:
1.120 paf 322: switch(to_lang & ~String::L_OPTIMIZE_BIT) {
323: case String::L_CLEAN:
324: case String::L_TAINTED:
325: case String::L_AS_IS:
1.119 paf 326: // clean piece
327:
328: // tainted piece, but undefined untaint language
329: // for VString.as_double of tainted values
330: // for ^process{body} evaluation
331:
332: // tainted, untaint language: as-is
333: ec_append(info->result, optimize, whitespace, info->pos, fragment_length);
334: break;
1.120 paf 335: case String::L_FILE_SPEC:
1.119 paf 336: // tainted, untaint language: file [name]
1.140 misha 337: {
338: escape_fragment(
1.154 misha 339: encode(need_file_encode, '_', c);
1.140 misha 340: );
341: }
1.119 paf 342: break;
1.120 paf 343: case String::L_URI:
1.155 misha 344: // tainted, untaint language: uri
345: {
346: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
347: // skip source [we use recoded version]
348: pa_CORD_pos_advance(info->pos, fragment_length);
349: String::C output(fragment_str, fragment_length);
350: if(info->charsets)
351: output=Charset::transcode(output,
352: info->charsets->source(),
353: info->charsets->client());
354:
355: char c;
356: for(const char* src=output.str; (c=*src++); )
357: encode(need_uri_encode, '%', c);
358: }
359: break;
1.120 paf 360: case String::L_HTTP_HEADER:
1.119 paf 361: // tainted, untaint language: http-field-content-text
1.161 moko 362: escape_fragment(switch(c) {
363: case '\n':
364: case '\r': to_string(" "); break;
365: default: _default; break;
366: });
1.119 paf 367: break;
1.120 paf 368: case String::L_MAIL_HEADER:
1.119 paf 369: // tainted, untaint language: mail-header
370: // http://www.ietf.org/rfc/rfc2047.txt
371: if(info->charsets) {
372: size_t mail_size;
373: const char *mail_ptr=
374: info->body->mid(info->fragment_begin, mail_size=fragment_length).cstr();
375: // skip source [we use recoded version]
376: pa_CORD_pos_advance(info->pos, mail_size);
377:
378: const char* charset_name=info->charsets->mail().NAME().cstr();
379:
380: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
381: bool to_quoted_printable=false;
382:
383: bool email=false;
384: uchar c;
1.124 paf 385: for(const char* src=mail_ptr; (c=(uchar)*src++); ) {
1.137 paf 386: if(c=='\r' || c=='\n')
387: c=' ';
1.130 paf 388: if(to_quoted_printable && (c==',' || c == '"' || addr_spec_soon(src-1/*position to 'c'*/))) {
1.119 paf 389: email=c=='<';
390: to_string("?=");
391: to_quoted_printable=false;
392: }
1.130 paf 393: //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
1.119 paf 394: if(!email && (
1.166 ! moko 395: ( !to_quoted_printable && (c & 0x80) ) // starting quote-printable-encoding on first 8bit char
! 396: || ( to_quoted_printable && !mail_header_char_valid_within_Qencoded(c) )
1.119 paf 397: )) {
398: if(!to_quoted_printable) {
399: to_string("=?");
400: to_string(charset_name);
401: to_string("?Q?");
402: to_quoted_printable=true;
1.105 paf 403: }
1.119 paf 404: encode(mail_header_nonspace_char, '=', '_');
405: } else
406: to_char(c);
407: if(c=='>')
408: email=false;
409: }
410: if(to_quoted_printable) // close
411: to_string("?=");
412:
413: } else
414: ec_append(info->result, optimize, whitespace, info->pos, fragment_length);
415: break;
1.120 paf 416: case String::L_SQL:
1.119 paf 417: // tainted, untaint language: sql
418: if(info->connection) {
419: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
420: // skip source [we use recoded version]
421: pa_CORD_pos_advance(info->pos, fragment_length);
422:
423: to_string(info->connection->quote(fragment_str, fragment_length));
1.138 misha 424: } else {
425: info->exception="untaint in SQL language failed - no connection specified";
426: info->fragment_begin=fragment_end;
427: return 1; // stop processing. can't throw exception here
428: }
1.119 paf 429: break;
1.120 paf 430: case String::L_JS:
1.140 misha 431: escape_fragment(switch(c) {
432: case '\n': to_string("\\n"); break;
1.119 paf 433: case '"': to_string("\\\""); break;
434: case '\'': to_string("\\'"); break;
435: case '\\': to_string("\\\\"); break;
436: case '\xFF': to_string("\\\xFF"); break;
1.146 misha 437: case '\r': to_string("\\r"); break;
1.133 paf 438: default: _default; break;
1.119 paf 439: });
440: break;
1.120 paf 441: case String::L_XML:
1.133 paf 442: // [2] Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
1.140 misha 443: escape_fragment(switch(c) {
444: case '\x20':
445: case '\x9':
446: case '\xA':
1.133 paf 447: case '\xD': // this is usually removed on input
448: _default;
449: break;
1.119 paf 450: case '&': to_string("&"); break;
451: case '>': to_string(">"); break;
452: case '<': to_string("<"); break;
453: case '"': to_string("""); break;
454: case '\'': to_string("'"); break;
1.133 paf 455: default:
456: if(((unsigned char)c)<0x20) {
457: // fixing it, so that libxml would not result
458: // in fatal error parsing text
459: // though it really violates standard.
460: // to indicate there were an error
461: // replace bad char not to it's code,
462: // which we can do,
463: // but rather to '!' to show that input were actually
464: // invalid.
465: // life: shows that MSIE can somehow garble form values
466: // so that they contain these chars.
467: to_char('!');
468: } else {
469: _default;
470: }
471: break;
1.119 paf 472: });
473: break;
1.120 paf 474: case String::L_HTML:
1.140 misha 475: escape_fragment(switch(c) {
1.119 paf 476: case '&': to_string("&"); break;
477: case '>': to_string(">"); break;
478: case '<': to_string("<"); break;
479: case '"': to_string("""); break;
1.133 paf 480: default: _default; break;
1.119 paf 481: });
482: break;
1.136 paf 483: case String::L_REGEX:
484: // tainted, untaint language: regex
1.140 misha 485: escape_fragment(
1.136 paf 486: if(need_regex_escape(c))
487: to_char('\\')
488: _default;
489: );
490: break;
1.156 misha 491: case String::L_JSON:
492: // tainted, untaint language: json <http://json.org/>
493: // escape '"' '\' '/' '\n' '\t' '\r' '\b' '\f' chars and escape chars as \uXXXX if output charset != UTF-8
494: {
1.159 moko 495: if(info->charsets==NULL || info->charsets->client().isUTF8()){
1.156 misha 496: // escaping to \uXXXX is not needed
497: escape_fragment(switch(c) {
498: case '\n': to_string("\\n"); break;
499: case '"' : to_string("\\\""); break;
500: case '\\': to_string("\\\\"); break;
501: case '/' : to_string("\\/"); break;
502: case '\t': to_string("\\t"); break;
503: case '\r': to_string("\\r"); break;
504: case '\b': to_string("\\b"); break;
505: case '\f': to_string("\\f"); break;
1.165 moko 506: default:
507: if((unsigned char)c < 0x20){
508: to_string("\\u00");
509: to_char(hex_digits[((unsigned char)c) >> 4]);
510: to_char(hex_digits[((unsigned char)c) & 0x0F]);
511: } else {
512: _default;
513: }
514: break;
1.156 misha 515: });
516: } else {
517: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
518: // skip source [we use recoded version]
519: pa_CORD_pos_advance(info->pos, fragment_length);
520: String::C output(fragment_str, fragment_length);
521:
522: output=Charset::escape_JSON(output, info->charsets->source());
523: to_string(output);
524: }
525: }
526: break;
1.140 misha 527: case String::L_HTTP_COOKIE:
528: // tainted, untaint language: cookie (3.3.0 and higher: %uXXXX in UTF-8)
1.159 moko 529: if(info->charsets) {
1.140 misha 530: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
531: // skip source [we use recoded version]
532: pa_CORD_pos_advance(info->pos, fragment_length);
533: String::C output(fragment_str, fragment_length);
534:
1.155 misha 535: output=Charset::escape(output, info->charsets->source());
1.140 misha 536: to_string(output);
1.159 moko 537: } else
538: ec_append(info->result, optimize, whitespace, info->pos, fragment_length);
1.140 misha 539: break;
1.150 misha 540: case String::L_PARSER_CODE:
541: // for auto-untaint in process
542: escape_fragment(
543: if(need_parser_code_escape(c))
544: to_char('^');
545: _default;
546: );
547: break;
1.119 paf 548: default:
1.127 paf 549: SAPI::abort("unknown untaint language #%d",
1.119 paf 550: static_cast<int>(to_lang)); // should never
551: break; // never
1.117 paf 552: }
1.55 parser 553:
1.119 paf 554: info->fragment_begin=fragment_end;
555:
556: return 0; // 0=continue
557: }
558:
559:
1.148 misha 560: String::Body String::cstr_to_string_body_taint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
1.134 paf 561: if(is_empty())
562: return String::Body();
1.119 paf 563:
564: Cstr_to_string_body_block_info info;
565: // input
566: info.lang=lang;
567: info.connection=connection;
568: info.charsets=charsets;
569: info.body=&body;
570: // output
571: CORD_ec_init(info.result);
572: // private
573: body.set_pos(info.pos, 0);
574: info.fragment_begin=0;
1.138 misha 575: info.exception=0;
1.119 paf 576: info.whitespace=true;
577:
1.148 misha 578: cstr_to_string_body_block(lang, length(), &info);
579:
580: if(info.exception)
581: throw Exception(0,
582: 0,
583: info.exception);
584:
1.160 moko 585: return String::Body(CORD_ec_to_cord(info.result));
1.148 misha 586: }
587:
588: int cstr_to_string_body_block_untaint(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info){
589: const String::Language lang=(String::Language)(unsigned char)alang;
590: // see append_fragment_* for explanation
591: if(info->lang&String::L_OPTIMIZE_BIT)
592: return cstr_to_string_body_block(
593: lang==String::L_TAINTED?
594: info->lang
595: :lang==String::L_CLEAN?
596: (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT)
597: :lang,
598: fragment_length, info);
599: else
600: return cstr_to_string_body_block(lang==String::L_TAINTED ? info->lang:lang, fragment_length, info);
601: }
602:
603: String::Body String::cstr_to_string_body_untaint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
604: if(is_empty())
605: return String::Body();
606:
607: Cstr_to_string_body_block_info info;
608: // input
609: info.lang=lang;
610: info.connection=connection;
611: info.charsets=charsets;
612: info.body=&body;
613: // output
614: CORD_ec_init(info.result);
615: // private
616: body.set_pos(info.pos, 0);
617: info.fragment_begin=0;
618: info.exception=0;
619: info.whitespace=true;
620:
621: langs.for_each(body, cstr_to_string_body_block_untaint, &info);
1.151 misha 622:
623: if(info.exception)
624: throw Exception(0,
625: 0,
626: info.exception);
627:
1.160 moko 628: return String::Body(CORD_ec_to_cord(info.result));
1.151 misha 629: }
630:
1.155 misha 631: const char* String::untaint_and_transcode_cstr(Language lang, const Request_charsets *charsets) const {
1.151 misha 632: if(charsets && &charsets->source() != &charsets->client()){
1.155 misha 633: // Note: L_URI is allready transcoded during untaint, but transcode does not affect %XX
634: return Charset::transcode(cstr_to_string_body_untaint(lang, 0, charsets), charsets->source(), charsets->client()).cstr();
635: } else
1.151 misha 636: return cstr_to_string_body_untaint(lang, 0, charsets).cstr();
1.1 paf 637: }
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