Annotation of parser3/src/main/untaint.C, revision 1.156
1.7 paf 1: /** @file
1.8 paf 2: Parser: String class part: untaint mechanizm.
3:
1.135 paf 4: Copyright(c) 2001-2005 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:
1.156 ! misha 8: static const char * const IDENT_UNTAINT_C="$Date: 2009-10-15 01:07:54 $";
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)) { \
64: static const char* hex="0123456789ABCDEF"; \
1.119 paf 65: CORD_ec_append(info->result, prefix); \
66: CORD_ec_append(info->result, hex[((unsigned char)c)/0x10]); \
67: CORD_ec_append(info->result, hex[((unsigned char)c)%0x10]); \
1.117 paf 68: } else \
1.119 paf 69: CORD_ec_append(info->result, otherwise);
1.126 paf 70: #define to_char(c) { CORD_ec_append(info->result, c); whitespace=false; }
71: #define to_string(s) { CORD_ec_append_cord(info->result, s); whitespace=false; }
1.4 paf 72:
1.9 paf 73: inline bool need_file_encode(unsigned char c){
1.108 paf 74: // russian letters and space ENABLED
75: // encoding only these...
76: return strchr(
1.143 misha 77: "*?'\"<>|"
1.108 paf 78: #ifndef WIN32
1.143 misha 79: ":\\"
1.31 paf 80: #endif
1.143 misha 81: , c)!=0;
1.9 paf 82: }
1.143 misha 83:
1.5 paf 84: inline bool need_uri_encode(unsigned char c){
1.152 misha 85: if((c>='0') && (c<='9') || (c>='A') && (c<='Z') || (c>='a') && (c<='z'))
1.4 paf 86: return false;
87:
1.143 misha 88: return !strchr("_-./", c);
1.5 paf 89: }
1.143 misha 90:
1.36 paf 91: inline bool need_http_header_encode(unsigned char c){
1.143 misha 92: if(strchr(" , :", c))
1.5 paf 93: return false;
94:
95: return need_uri_encode(c);
1.4 paf 96: }
1.143 misha 97:
1.136 paf 98: inline bool need_regex_escape(unsigned char c){
1.142 misha 99: return strchr("\\^$.[]|()?*+{}-", c)!=0;
1.136 paf 100: }
1.1 paf 101:
1.150 misha 102: inline bool need_parser_code_escape(unsigned char c){
103: return strchr("^$;@()[]{}:#\"", c)!=0;
104: }
105:
1.1 paf 106: // String
107:
1.41 paf 108: /*
109: HTTP-header = field-name ":" [ field-value ] CRLF
110:
1.152 misha 111: field-name = token
112: field-value = *( field-content | LWS )
1.41 paf 113:
1.152 misha 114: field-content = <the OCTETs making up the field-value
1.41 paf 115: and consisting of either *TEXT or combinations
116: of token, tspecials, and quoted-string>
117:
118:
1.152 misha 119: token = 1*<any CHAR except CTLs or tspecials>
1.41 paf 120: word = token | quoted-string
1.152 misha 121: quoted-string = ( <"> *(qdtext | quoted-pair ) <"> )
122: qdtext = <any TEXT except <">>
123: quoted-pair = "\" CHAR
1.41 paf 124:
1.152 misha 125: OCTET = <any 8-bit sequence of data>
126: CHAR = <any US-ASCII character (octets 0 - 127)>
1.41 paf 127:
128: tspecials = "(" | ")" | "<" | ">" | "@"
1.152 misha 129: | "," | ";" | ":" | "\" | <">
130: | "/" | "[" | "]" | "?" | "="
131: | "{" | "}" | SP | HT
1.41 paf 132:
133: SP = <US-ASCII SP, space (32)>
134: HT = <US-ASCII HT, horizontal-tab (9)>
135:
136: LWS = [CRLF] 1*( SP | HT )
1.152 misha 137: TEXT = <any OCTET except CTLs, but including LWS>
138: CTL = <any US-ASCII control character (octets 0 - 31) and DEL (127)>
1.41 paf 139:
140:
141: if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
142: */
1.117 paf 143: inline bool need_quote_http_header(const char* ptr, size_t size) {
1.41 paf 144: for(; size--; ptr++)
1.42 paf 145: if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
1.41 paf 146: return true;
147: return false;
148: }
149:
1.119 paf 150: #ifndef DOXYGEN
151: struct Append_fragment_info {
152: String::Language lang;
153: String::Languages* dest_languages;
154: size_t dest_body_plan_length;
155: };
156: #endif
157: int append_fragment_optimizing(char alang, size_t asize, Append_fragment_info* info) {
158: const String::Language lang=(String::Language)(unsigned char)alang;
159: // main idea here:
160: // tainted piece would get OPTIMIZED bit from 'lang'
161: // clean piece would be marked OPTIMIZED manually
162: // pieces with determined languages [not tainted|clean] would retain theirs langs
163: info->dest_languages->append(info->dest_body_plan_length,
1.120 paf 164: lang==String::L_TAINTED?
1.119 paf 165: info->lang
1.120 paf 166: :lang==String::L_CLEAN?
167: (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT) // ORing with OPTIMIZED flag
1.119 paf 168: :lang,
169: asize);
170: info->dest_body_plan_length+=asize;
171:
172: return 0; // 0=continue
173: }
174: int append_fragment_nonoptimizing(char alang, size_t asize, Append_fragment_info* info) {
175: const String::Language lang=(String::Language)(unsigned char)alang;
176: // The core idea: tainted pieces got marked with context's lang
177: info->dest_languages->append(info->dest_body_plan_length,
1.120 paf 178: lang==String::L_TAINTED?
1.119 paf 179: info->lang
180: :lang,
181: asize);
182: info->dest_body_plan_length+=asize;
183:
184: return 0; // 0=continue
185: }
1.117 paf 186:
1.91 paf 187: /**
1.117 paf 188: appends to other String,
1.91 paf 189: marking all tainted pieces of it with @a lang.
1.92 paf 190: or marking ALL pieces of it with a @a lang when @a forced to,
191: and propagating OPTIMIZE language bit.
1.117 paf 192: */
1.147 misha 193: String& String::append_to(String& dest, Language ilang, bool forced) const {
1.117 paf 194: if(is_empty())
195: return dest;
1.91 paf 196:
1.119 paf 197: // first: fragment infos
1.117 paf 198:
1.147 misha 199: if(ilang==L_PASS_APPENDED) // without language-change?
200: dest.langs.appendHelper(dest.body, langs, body);
1.117 paf 201: else if(forced) //forcing passed lang?
1.147 misha 202: dest.langs.appendHelper(dest.body, ilang, body);
203: else {
204: if(langs.opt.is_not_just_lang){
205: Append_fragment_info info={ilang, &dest.langs, dest.body.length()};
206: langs.for_each(body, ilang&L_OPTIMIZE_BIT?
207: append_fragment_optimizing
208: :append_fragment_nonoptimizing, &info);
209: } else {
210: Language lang=langs.opt.lang;
211: // see append_fragment_* for explanation
212: if(ilang&L_OPTIMIZE_BIT){
213: dest.langs.appendHelper(dest.body,
214: lang==String::L_TAINTED?
215: ilang
216: :lang==String::L_CLEAN?
217: (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT)
218: :lang,
219: body);
220: } else {
221: dest.langs.appendHelper(dest.body, lang==String::L_TAINTED ? ilang:lang, body);
222: }
223: }
1.91 paf 224: }
225:
1.119 paf 226: // next: letters
227: dest.body<<body;
228:
1.117 paf 229: ASSERT_STRING_INVARIANT(dest);
1.75 paf 230: return dest;
1.51 parser 231: }
232:
1.109 paf 233: /** http://www.ietf.org/rfc/rfc2047.txt
234: RFC
235: (3) As a replacement for a 'word' entity within a 'phrase', for example,
236: one that precedes an address in a From, To, or Cc header. The ABNF
237: definition for 'phrase' from RFC 822 thus becomes:
238:
239: phrase = 1*( encoded-word / word )
240:
241: In this case the set of characters that may be used in a "Q"-encoded
242: 'encoded-word' is restricted to: <upper and lower case ASCII
243: letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_"
244: (underscore, ASCII 95.)>. An 'encoded-word' that appears within a
245: 'phrase' MUST be separated from any adjacent 'word', 'text' or
246: 'special' by 'linear-white-space'.
247: ...
248: (2) The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE) may be
249: represented as "_" (underscore, ASCII 95.). (This character may
250: not pass through some internetwork mail gateways, but its use
251: will greatly enhance readability of "Q" encoded data with mail
252: readers that do not support this encoding.) Note that the "_"
253: always represents hexadecimal 20, even if the SPACE character
254: occupies a different code position in the character set in use.
255:
256: paf: obviously,
257: without "=", or one could not differ "=E0" and "russian letter a"
258: and without "_", or in would mean 0x20
259: */
1.117 paf 260: inline bool mail_header_char_valid_within_Qencoded(char c) {
1.109 paf 261: return c>='A' && c<='Z'
262: || c>='a' && c<='Z'
263: || c>='0' && c<='9'
264: || strchr("!*+-/", c);
265: }
1.118 paf 266: inline bool addr_spec_soon(const char *src) {
1.124 paf 267: for(char c; (c=*src); src++)
1.118 paf 268: if(c=='<')
269: return true;
270: else if(!(c==' ' || c=='\t'))
271: return false;
272: return false;
273: }
1.117 paf 274: /**
275: RFC
276: Upper case should be used for hexadecimal digits "A" through "F"
277: The 8-bit hexadecimal value 20 (e.g., ISO-8859-1 SPACE)
278: may be represented as "_"
279: */
280: inline bool mail_header_nonspace_char(char c) {
281: return c != 0x20;
282: }
1.125 paf 283:
1.117 paf 284: inline void ec_append(CORD_ec& result, bool& optimize, bool& whitespace, CORD_pos pos, size_t size) {
285: while(size--) {
286: CORD_ec_append(result, CORD_pos_fetch(pos));
287: CORD_next(pos);
288: }
289: }
290: inline void pa_CORD_pos_advance(CORD_pos pos, size_t n) {
291: while(true) {
1.125 paf 292: long avail=CORD_pos_chars_left(pos);
293: if(avail<=0) {
1.117 paf 294: CORD_next(pos);
295: if(!--n)
296: break;
1.125 paf 297: } else if((size_t)avail<n) {
1.117 paf 298: CORD_pos_advance(pos, avail);
299: n-=avail;
300: } else { // avail>=n
301: CORD_pos_advance(pos, n);
302: break;
303: }
304: }
305: }
306:
1.119 paf 307: #ifndef DOXYGEN
308: struct Cstr_to_string_body_block_info {
309: // input
310: String::Language lang;
311: SQL_Connection* connection;
312: const Request_charsets* charsets;
313: const String::Body* body;
314:
315: // output
316: CORD_ec result;
317:
318: // private
319: CORD_pos pos;
320: size_t fragment_begin;
321: bool whitespace;
1.138 misha 322: const char* exception;
1.119 paf 323: };
324: #endif
1.151 misha 325:
326: // @todo: replace info->body->mid with something that uses info->pos
1.148 misha 327: int cstr_to_string_body_block(String::Language to_lang, size_t fragment_length, Cstr_to_string_body_block_info* info) {
1.119 paf 328: bool& whitespace=info->whitespace;
329: size_t fragment_end=info->fragment_begin+fragment_length;
1.148 misha 330: //fprintf(stderr, "%d, %d =%s=\n", to_lang, fragment_length, info->body->cstr());
1.119 paf 331:
1.120 paf 332: bool optimize=(to_lang & String::L_OPTIMIZE_BIT)!=0;
1.119 paf 333: if(!optimize)
334: whitespace=false;
335:
1.120 paf 336: switch(to_lang & ~String::L_OPTIMIZE_BIT) {
337: case String::L_CLEAN:
338: case String::L_TAINTED:
339: case String::L_AS_IS:
1.119 paf 340: // clean piece
341:
342: // tainted piece, but undefined untaint language
343: // for VString.as_double of tainted values
344: // for ^process{body} evaluation
345:
346: // tainted, untaint language: as-is
347: ec_append(info->result, optimize, whitespace, info->pos, fragment_length);
348: break;
1.120 paf 349: case String::L_FILE_SPEC:
1.119 paf 350: // tainted, untaint language: file [name]
1.140 misha 351: {
352: escape_fragment(
1.154 misha 353: encode(need_file_encode, '_', c);
1.140 misha 354: );
355: }
1.119 paf 356: break;
1.144 misha 357: case String::L_FILE_POST:
358: {
359: escape_fragment(switch(c) {
360: case '\0': to_string("\\0"); break;
361: case '\\': to_string("\\\\"); break;
362: default: _default; break;
363: });
364: }
365: break;
1.120 paf 366: case String::L_URI:
1.155 misha 367: // tainted, untaint language: uri
368: {
369: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
370: // skip source [we use recoded version]
371: pa_CORD_pos_advance(info->pos, fragment_length);
372: String::C output(fragment_str, fragment_length);
373: if(info->charsets)
374: output=Charset::transcode(output,
375: info->charsets->source(),
376: info->charsets->client());
377:
378: char c;
379: for(const char* src=output.str; (c=*src++); )
380: encode(need_uri_encode, '%', c);
381: }
382: break;
1.120 paf 383: case String::L_HTTP_HEADER:
1.119 paf 384: // tainted, untaint language: http-field-content-text
1.140 misha 385: escape_fragment(
1.119 paf 386: encode(need_uri_encode, '%', c);
387: );
388: break;
1.120 paf 389: case String::L_MAIL_HEADER:
1.119 paf 390: // tainted, untaint language: mail-header
391: // http://www.ietf.org/rfc/rfc2047.txt
392: if(info->charsets) {
393: size_t mail_size;
394: const char *mail_ptr=
395: info->body->mid(info->fragment_begin, mail_size=fragment_length).cstr();
396: // skip source [we use recoded version]
397: pa_CORD_pos_advance(info->pos, mail_size);
398:
399: const char* charset_name=info->charsets->mail().NAME().cstr();
400:
401: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
402: bool to_quoted_printable=false;
403:
404: bool email=false;
405: uchar c;
1.124 paf 406: for(const char* src=mail_ptr; (c=(uchar)*src++); ) {
1.137 paf 407: if(c=='\r' || c=='\n')
408: c=' ';
1.130 paf 409: if(to_quoted_printable && (c==',' || c == '"' || addr_spec_soon(src-1/*position to 'c'*/))) {
1.119 paf 410: email=c=='<';
411: to_string("?=");
412: to_quoted_printable=false;
413: }
1.130 paf 414: //RFC + An 'encoded-word' MUST NOT appear in any portion of an 'addr-spec'.
1.119 paf 415: if(!email && (
416: !to_quoted_printable && (c & 0x80) // starting quote-printable-encoding on first 8bit char
417: || to_quoted_printable && !mail_header_char_valid_within_Qencoded(c)
418: )) {
419: if(!to_quoted_printable) {
420: to_string("=?");
421: to_string(charset_name);
422: to_string("?Q?");
423: to_quoted_printable=true;
1.105 paf 424: }
1.119 paf 425: encode(mail_header_nonspace_char, '=', '_');
426: } else
427: to_char(c);
428: if(c=='>')
429: email=false;
430: }
431: if(to_quoted_printable) // close
432: to_string("?=");
433:
434: } else
435: ec_append(info->result, optimize, whitespace, info->pos, fragment_length);
436: break;
1.120 paf 437: case String::L_SQL:
1.119 paf 438: // tainted, untaint language: sql
439: if(info->connection) {
440: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
441: // skip source [we use recoded version]
442: pa_CORD_pos_advance(info->pos, fragment_length);
443:
444: to_string(info->connection->quote(fragment_str, fragment_length));
1.138 misha 445: } else {
446: info->exception="untaint in SQL language failed - no connection specified";
447: info->fragment_begin=fragment_end;
448: return 1; // stop processing. can't throw exception here
449: }
1.119 paf 450: break;
1.120 paf 451: case String::L_JS:
1.140 misha 452: escape_fragment(switch(c) {
453: case '\n': to_string("\\n"); break;
1.119 paf 454: case '"': to_string("\\\""); break;
455: case '\'': to_string("\\'"); break;
456: case '\\': to_string("\\\\"); break;
457: case '\xFF': to_string("\\\xFF"); break;
1.146 misha 458: case '\r': to_string("\\r"); break;
1.133 paf 459: default: _default; break;
1.119 paf 460: });
461: break;
1.120 paf 462: case String::L_XML:
1.133 paf 463: // [2] Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
1.140 misha 464: escape_fragment(switch(c) {
465: case '\x20':
466: case '\x9':
467: case '\xA':
1.133 paf 468: case '\xD': // this is usually removed on input
469: _default;
470: break;
1.119 paf 471: case '&': to_string("&"); break;
472: case '>': to_string(">"); break;
473: case '<': to_string("<"); break;
474: case '"': to_string("""); break;
475: case '\'': to_string("'"); break;
1.133 paf 476: default:
477: if(((unsigned char)c)<0x20) {
478: // fixing it, so that libxml would not result
479: // in fatal error parsing text
480: // though it really violates standard.
481: // to indicate there were an error
482: // replace bad char not to it's code,
483: // which we can do,
484: // but rather to '!' to show that input were actually
485: // invalid.
486: // life: shows that MSIE can somehow garble form values
487: // so that they contain these chars.
488: to_char('!');
489: } else {
490: _default;
491: }
492: break;
1.119 paf 493: });
494: break;
1.120 paf 495: case String::L_HTML:
1.140 misha 496: escape_fragment(switch(c) {
1.119 paf 497: case '&': to_string("&"); break;
498: case '>': to_string(">"); break;
499: case '<': to_string("<"); break;
500: case '"': to_string("""); break;
1.133 paf 501: default: _default; break;
1.119 paf 502: });
503: break;
1.136 paf 504: case String::L_REGEX:
505: // tainted, untaint language: regex
1.140 misha 506: escape_fragment(
1.136 paf 507: if(need_regex_escape(c))
508: to_char('\\')
509: _default;
510: );
511: break;
1.156 ! misha 512: case String::L_JSON:
! 513: // tainted, untaint language: json <http://json.org/>
! 514: // escape '"' '\' '/' '\n' '\t' '\r' '\b' '\f' chars and escape chars as \uXXXX if output charset != UTF-8
! 515: {
! 516: if(info->charsets->client().isUTF8()){
! 517: // escaping to \uXXXX is not needed
! 518: escape_fragment(switch(c) {
! 519: case '\n': to_string("\\n"); break;
! 520: case '"' : to_string("\\\""); break;
! 521: case '\\': to_string("\\\\"); break;
! 522: case '/' : to_string("\\/"); break;
! 523: case '\t': to_string("\\t"); break;
! 524: case '\r': to_string("\\r"); break;
! 525: case '\b': to_string("\\b"); break;
! 526: case '\f': to_string("\\f"); break;
! 527: default : _default; break;
! 528: });
! 529: } else {
! 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:
! 535: output=Charset::escape_JSON(output, info->charsets->source());
! 536: to_string(output);
! 537: }
! 538: }
! 539: break;
1.140 misha 540: case String::L_HTTP_COOKIE:
541: // tainted, untaint language: cookie (3.3.0 and higher: %uXXXX in UTF-8)
542: {
543: const char *fragment_str=info->body->mid(info->fragment_begin, fragment_length).cstr();
544: // skip source [we use recoded version]
545: pa_CORD_pos_advance(info->pos, fragment_length);
546: String::C output(fragment_str, fragment_length);
547:
1.155 misha 548: output=Charset::escape(output, info->charsets->source());
1.140 misha 549: //throw Exception(0, 0, output);
550: to_string(output);
551:
552: }
553: break;
1.150 misha 554: case String::L_PARSER_CODE:
555: // for auto-untaint in process
556: escape_fragment(
557: if(need_parser_code_escape(c))
558: to_char('^');
559: _default;
560: );
561: break;
1.119 paf 562: default:
1.127 paf 563: SAPI::abort("unknown untaint language #%d",
1.119 paf 564: static_cast<int>(to_lang)); // should never
565: break; // never
1.117 paf 566: }
1.55 parser 567:
1.119 paf 568: info->fragment_begin=fragment_end;
569:
570: return 0; // 0=continue
571: }
572:
573:
1.148 misha 574: String::Body String::cstr_to_string_body_taint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
1.134 paf 575: if(is_empty())
576: return String::Body();
1.119 paf 577:
578: Cstr_to_string_body_block_info info;
579: // input
580: info.lang=lang;
581: info.connection=connection;
582: info.charsets=charsets;
583: info.body=&body;
584: // output
585: CORD_ec_init(info.result);
586: // private
587: body.set_pos(info.pos, 0);
588: info.fragment_begin=0;
1.138 misha 589: info.exception=0;
1.119 paf 590: info.whitespace=true;
591:
1.148 misha 592: cstr_to_string_body_block(lang, length(), &info);
593:
594: if(info.exception)
595: throw Exception(0,
596: 0,
597: info.exception);
598:
1.155 misha 599: return String::Body(CORD_ec_to_cord(info.result), info.fragment_begin);
1.148 misha 600: }
601:
602: int cstr_to_string_body_block_untaint(char alang, size_t fragment_length, Cstr_to_string_body_block_info* info){
603: const String::Language lang=(String::Language)(unsigned char)alang;
604: // see append_fragment_* for explanation
605: if(info->lang&String::L_OPTIMIZE_BIT)
606: return cstr_to_string_body_block(
607: lang==String::L_TAINTED?
608: info->lang
609: :lang==String::L_CLEAN?
610: (String::Language)(String::L_CLEAN|String::L_OPTIMIZE_BIT)
611: :lang,
612: fragment_length, info);
613: else
614: return cstr_to_string_body_block(lang==String::L_TAINTED ? info->lang:lang, fragment_length, info);
615: }
616:
617: String::Body String::cstr_to_string_body_untaint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const {
618: if(is_empty())
619: return String::Body();
620:
621: Cstr_to_string_body_block_info info;
622: // input
623: info.lang=lang;
624: info.connection=connection;
625: info.charsets=charsets;
626: info.body=&body;
627: // output
628: CORD_ec_init(info.result);
629: // private
630: body.set_pos(info.pos, 0);
631: info.fragment_begin=0;
632: info.exception=0;
633: info.whitespace=true;
634:
635: langs.for_each(body, cstr_to_string_body_block_untaint, &info);
1.151 misha 636:
637: if(info.exception)
638: throw Exception(0,
639: 0,
640: info.exception);
641:
1.155 misha 642: return String::Body(CORD_ec_to_cord(info.result), info.fragment_begin);
1.151 misha 643: }
644:
1.155 misha 645: const char* String::untaint_and_transcode_cstr(Language lang, const Request_charsets *charsets) const {
1.151 misha 646: if(charsets && &charsets->source() != &charsets->client()){
1.155 misha 647: // Note: L_URI is allready transcoded during untaint, but transcode does not affect %XX
648: return Charset::transcode(cstr_to_string_body_untaint(lang, 0, charsets), charsets->source(), charsets->client()).cstr();
649: } else
1.151 misha 650: return cstr_to_string_body_untaint(lang, 0, charsets).cstr();
1.1 paf 651: }
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