![[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|
Return to untaint.C CVS log | Up to [parser3project] / parser3 / src / main |
1.7 paf 1: /** @file
1.8 paf 2: Parser: String class part: untaint mechanizm.
3:
1.13 paf 4: Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com)
1.74 paf 5: Author: Alexander Petrosyan <paf@design.ru>(http://paf.design.ru)
1.8 paf 6:
1.82 ! paf 7: $Id: untaint.C,v 1.81 2001/11/22 10:28:46 paf Exp $
1.1 paf 8: */
9:
10: #include "pa_pool.h"
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"
1.34 paf 16: #include "pa_sql_connection.h"
1.58 parser 17: #include "pa_dictionary.h"
1.66 parser 18: #include "pa_common.h"
1.1 paf 19:
1.18 paf 20: #define escape(action) \
1.1 paf 21: { \
1.13 paf 22: const char *src=row->item.ptr; \
23: for(int size=row->item.size; size--; src++) \
1.18 paf 24: action \
1.1 paf 25: }
1.13 paf 26: #define _default default: *dest++=*src; break
27: #define encode(need_encode_func, prefix) \
1.5 paf 28: default: \
1.13 paf 29: if(need_encode_func(*src)) { \
1.5 paf 30: static const char *hex="0123456789ABCDEF"; \
1.9 paf 31: char chunk[3]={prefix}; \
1.13 paf 32: chunk[1]=hex[((unsigned char)*src)/0x10]; \
33: chunk[2]=hex[((unsigned char)*src)%0x10]; \
1.60 parser 34: memcpy(dest, chunk, 3); dest+=3; \
1.5 paf 35: } else \
1.13 paf 36: *dest++=*src; \
1.5 paf 37: break
1.18 paf 38: #define to_char(c) *dest++=c
39: #define to_string(b, bsize) \
1.60 parser 40: memcpy(dest, b, bsize); \
1.18 paf 41: dest+=bsize; \
1.4 paf 42:
1.9 paf 43: inline bool need_file_encode(unsigned char c){
1.13 paf 44: if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z'))
1.9 paf 45: return false;
46:
1.31 paf 47: return !strchr(
48: #ifdef WIN32
1.37 paf 49: ":\\~"
1.31 paf 50: #endif
1.39 paf 51: "./()_-", c);
1.9 paf 52: }
1.5 paf 53: inline bool need_uri_encode(unsigned char c){
1.13 paf 54: if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z'))
1.4 paf 55: return false;
56:
1.5 paf 57: return !strchr("_-./", c);
58: }
1.36 paf 59: inline bool need_http_header_encode(unsigned char c){
1.18 paf 60: if(strchr(" , :", c))
1.5 paf 61: return false;
62:
63: return need_uri_encode(c);
1.4 paf 64: }
1.1 paf 65:
1.56 parser 66: //
67:
68: static const char * String_Untaint_lang_name[]={
69: "U", ///< zero value handy for hash lookup @see untaint_lang_name2enum
70: "C", ///< clean
71: "T", ///< tainted, untaint language as assigned later
72: // untaint languages. assigned by ^untaint[lang]{...}
73: "P",
74: /**<
75: leave language built into string being appended.
76: just a flag, that value not stored
77: */
78: "A", ///< leave all characters intact
1.68 parser 79: "F", ///< file specification
80: "H", ///< ext in HTTP response header
1.56 parser 81: "M", ///< text in mail header
82: "URI", ///< text in uri
83: "T", ///< ^table:set body
84: "SQL", ///< ^table:sql body
85: "JS", ///< JavaScript code
1.68 parser 86: "XML", ///< ^dom:set xml
1.82 ! paf 87: "HTML" ///< HTML code (for editing)
1.56 parser 88: };
89:
90:
1.1 paf 91: // String
92:
1.41 paf 93: /*
94:
95: HTTP-header = field-name ":" [ field-value ] CRLF
96:
97: field-name = token
98: field-value = *( field-content | LWS )
99:
100: field-content = <the OCTETs making up the field-value
101: and consisting of either *TEXT or combinations
102: of token, tspecials, and quoted-string>
103:
104:
105:
106: word = token | quoted-string
107:
108: token = 1*<any CHAR except CTLs or tspecials>
109:
110:
111:
112: tspecials = "(" | ")" | "<" | ">" | "@"
113: | "," | ";" | ":" | "\" | <">
114: | "/" | "[" | "]" | "?" | "="
115: | "{" | "}" | SP | HT
116:
117: SP = <US-ASCII SP, space (32)>
118: HT = <US-ASCII HT, horizontal-tab (9)>
119:
120: LWS = [CRLF] 1*( SP | HT )
121: TEXT = <any OCTET except CTLs,
122: but including LWS>
123:
124: quoted-pair = "\" CHAR
125:
126: if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr))
127: */
128: inline bool need_quote_http_header(const char *ptr, size_t size) {
129: for(; size--; ptr++)
1.42 paf 130: if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr))
1.41 paf 131: return true;
132: return false;
133: }
134:
1.77 paf 135: String& String::append(const String& src, uchar lang, bool forced) {
136: // manually unrolled code to avoid do{if(const)} constructs
137: if(forced)
138: STRING_SRC_FOREACH_ROW(
139: APPEND(row->item.ptr, row->item.size,
140: lang, //forcing passed lang
141: row->item.origin.file, row->item.origin.line);
142: )
143: else if(lang==UL_PASS_APPENDED)
144: STRING_SRC_FOREACH_ROW(
145: APPEND(row->item.ptr, row->item.size,
146: row->item.lang, // passing item's lang
147: row->item.origin.file, row->item.origin.line);
148: )
149: else if(lang&UL_OPTIMIZE_BIT) // main idea here
150: // tainted piece would get OPTIMIZED bit from 'lang'
151: // clean piece would be marked OPTIMIZED manually
152: // pieces with determined languages [not tainted|clean] would retain theirs langs
153: STRING_SRC_FOREACH_ROW(
154: APPEND(row->item.ptr, row->item.size,
155: row->item.lang==UL_TAINTED?lang:(
156: row->item.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag
157: row->item.lang
158: ),
159: row->item.origin.file, row->item.origin.line);
160: )
161: else
162: STRING_SRC_FOREACH_ROW(
163: APPEND(row->item.ptr, row->item.size,
164: row->item.lang==UL_TAINTED?lang:row->item.lang,
165: row->item.origin.file, row->item.origin.line);
166: );
167: break2:
168: return *this;
169: }
170:
1.75 paf 171: size_t String::cstr_bufsize(Untaint_lang lang,
172: SQL_Connection *connection,
173: const char *charset) const {
1.77 paf 174: size_t dest=1; // for terminating 0
175: STRING_FOREACH_ROW(
176: uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang;
177:
178: switch(to_lang & ~UL_OPTIMIZE_BIT) {
179: case UL_CLEAN:
180: case UL_TAINTED:
181: case UL_AS_IS:
182: // clean piece
183:
184: // tainted piece, but undefined untaint language
185: // for VString.as_double of tainted values
186: // for ^process{body} evaluation
187:
188: // tainted, untaint language: as-is
189: dest+=row->item.size;
190: break;
191: case UL_FILE_SPEC:
192: // tainted, untaint language: file [name]
193: dest+=row->item.size*3/* worst: Z->%XX */;
194: break;
195: case UL_URI:
196: // tainted, untaint language: uri
197: dest+=row->item.size*3/* worst: Z->%XX */;
198: break;
199: case UL_HTTP_HEADER:
200: // tainted, untaint language: http-field-content-text
201: dest+=row->item.size*3/* worst: Z->%XX */;
202: break;
203: case UL_MAIL_HEADER:
204: // tainted, untaint language: mail-header
205: if(charset) {
206: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
207: dest+=row->item.size*3+MAX_STRING/* worst: =?charset?Q?=%XX?= */;
208: } else {
1.75 paf 209: dest+=row->item.size;
210: }
1.77 paf 211: break;
212: case UL_TABLE:
213: // tainted, untaint language: table
214: dest+=row->item.size;
215: break;
216: case UL_SQL:
217: // tainted, untaint language: sql
218: if(connection)
219: dest+=connection->quote(0, row->item.ptr, row->item.size);
220: break;
221: case UL_JS:
222: escape(switch(*src) {
223: case '"': case '\'': case '\n': case '\\': case '\xFF':
224: dest+=2; break;
225: default:
226: dest++; break;
227: });
228: break;
229: case UL_XML:
230: escape(switch(*src) {
231: case '&': case '>': case '<': case '"': case '\'':
232: dest+= 6; break;
233: default:
234: dest++; break;
235: });
236: break;
237: case UL_HTML:
238: escape(switch(*src) {
239: case '&':
240: case '>':
241: case '<':
242: case '"':
243: dest+=6; break;
244: default:
245: dest++; break;
246: });
247: break;
1.75 paf 248: }
1.77 paf 249: );
1.75 paf 250: break2:
251: return dest;
1.51 parser 252: }
253:
1.43 paf 254: char *String::store_to(char *dest, Untaint_lang lang,
255: SQL_Connection *connection,
256: const char *charset) const {
1.75 paf 257: // WARNING:
258: // before any changes check cstr_bufsize first!!!
1.44 paf 259: bool whitespace=true;
1.79 paf 260: // expanded STRING_FOREACH_ROW here for debugging purposes
1.78 paf 261: const Chunk *chunk=&head; \
262: do { \
263: const Chunk::Row *row=chunk->rows; \
264: for(uint i=0; i<chunk->count; i++, row++) { \
265: if(row==append_here) \
266: goto break2; \
267: \
1.77 paf 268: uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang;
269:
270: char *start=dest;
271:
272: switch(to_lang & ~UL_OPTIMIZE_BIT) {
273: case UL_CLEAN:
274: case UL_TAINTED:
275: case UL_AS_IS:
276: // clean piece
277:
278: // tainted piece, but undefined untaint language
279: // for VString.as_double of tainted values
280: // for ^process{body} evaluation
281:
282: // tainted, untaint language: as-is
283: memcpy(dest, row->item.ptr, row->item.size);
284: dest+=row->item.size;
285: break;
286: case UL_FILE_SPEC:
287: // tainted, untaint language: file [name]
288: escape(switch(*src) {
289: case ' ': to_char('_'); break;
290: encode(need_file_encode, '+');
291: });
292: break;
293: case UL_URI:
294: // tainted, untaint language: uri
295: escape(switch(*src) {
296: case ' ': to_char('+'); break;
297: encode(need_uri_encode, '%');
298: });
299: break;
300: case UL_HTTP_HEADER:
301: // tainted, untaint language: http-field-content-text
302: escape(switch(*src) {
303: case ' ': to_char('+'); break;
304: encode(need_uri_encode, '%');
305: });
306: break;
307: case UL_MAIL_HEADER:
308: // tainted, untaint language: mail-header
309: if(charset) {
310: // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?=
311: const char *src=row->item.ptr;
312: bool to_quoted_printable=false;
313: for(int size=row->item.size; size--; src++) {
314: if(*src & 0x80) {
315: if(!to_quoted_printable) {
316: dest+=sprintf(dest, "=?%.15s?Q?", charset);
317: to_quoted_printable=true;
318: }
319: dest+=sprintf(dest, "=%02X", *src & 0xFF);
320: } else {
321: *dest++=*src;
322: }
1.44 paf 323: }
1.77 paf 324: if(to_quoted_printable) // close
325: dest+=sprintf(dest, "?=");
326: } else {
1.13 paf 327: memcpy(dest, row->item.ptr, row->item.size);
328: dest+=row->item.size;
1.1 paf 329: }
1.77 paf 330: break;
331: case UL_TABLE:
332: // tainted, untaint language: table
333: escape(switch(*src) {
334: case '\t': to_char(' '); break;
335: case '\n': to_char(' '); break;
336: _default;
337: });
338: break;
339: case UL_SQL:
340: // tainted, untaint language: sql
341: if(connection)
342: dest+=connection->quote(dest, row->item.ptr, row->item.size);
343: else
344: throw Exception(0, 0,
345: this,
346: "untaint in SQL language failed - no connection specified");
347: break;
348: case UL_JS:
349: escape(switch(*src) {
350: case '"': to_string("\\\"", 2); break;
351: case '\'': to_string("\\'", 2); break;
352: case '\n': to_string("\\n", 2); break;
353: case '\\': to_string("\\\\", 2); break;
354: case '\xFF': to_string("\\\xFF", 2); break;
355: _default;
356: });
357: break;
358: case UL_XML:
359: escape(switch(*src) {
360: case '&': to_string("&", 5); break;
361: case '>': to_string(">", 4); break;
362: case '<': to_string("<", 4); break;
363: case '"': to_string(""", 6); break;
364: case '\'': to_string("'", 6); break;
365: _default;
366: });
367: break;
368: case UL_HTML:
369: escape(switch(*src) {
370: case '&': to_string("&", 5); break;
371: case '>': to_string(">", 4); break;
372: case '<': to_string("<", 4); break;
373: case '"': to_string(""", 6); break;
374: _default;
375: });
376: break;
377: default:
378: throw Exception(0, 0,
379: this,
1.81 paf 380: "unknown untaint language #%d",
381: static_cast<int>(row->item.lang)); // sould never
1.77 paf 382: break; // never
1.76 paf 383: }
1.55 parser 384:
1.77 paf 385: if(to_lang & UL_OPTIMIZE_BIT) {
386: // optimizing whitespace
387: char *stop=dest; dest=start;
388: for(char *src=start; src<stop; src++)
389: switch(*src) {
390: // of all consequent white space chars leaving only first one
1.80 paf 391: case ' ': case '\r': case '\n': case '\t':
1.77 paf 392: if(!whitespace) {
393: *dest++=*src;
394: whitespace=true;
395: }
396: break;
397: default:
398: whitespace=false;
399: *dest++=*src;
400: break;
401: };
402: } else // piece without optimization
403: whitespace=false;
1.78 paf 404:
405: } \
406: chunk=row->link; \
407: } while(chunk); \
408:
1.76 paf 409: break2:
410: return dest;
411: }
412:
413: char *String::cstr_debug_origins() const {
1.81 paf 414: //_asm int 3;
1.76 paf 415: char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2);
416: char *dest=result;
417:
418: const Chunk *chunk=&head;
419: do {
420: const Chunk::Row *row=chunk->rows;
421: for(uint i=0; i<chunk->count; i++, row++) {
422: if(row==append_here)
423: goto break2;
1.55 parser 424:
1.76 paf 425: #ifndef NO_STRING_ORIGIN
426: if(row->item.origin.file)
427: dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT,
428: row->item.origin.file,
429: 1+row->item.origin.line);
430: else
431: dest+=sprintf(dest, "<unknown>");
432: #endif
1.81 paf 433: uchar show_lang=row->item.lang & ~UL_OPTIMIZE_BIT;
434: if(show_lang>=sizeof(String_Untaint_lang_name)/sizeof(String_Untaint_lang_name[0]))
435: throw Exception(0, 0,
436: this,
437: "unknown untaint language #%d",
438: static_cast<int>(show_lang)); // sould never
439:
440: dest+=sprintf(dest, "#%s%s: ",
441: String_Untaint_lang_name[show_lang],
442: row->item.lang & UL_OPTIMIZE_BIT?".O":"");
1.76 paf 443: char *dest_after_origins=dest;
444:
445: memcpy(dest, row->item.ptr, row->item.size);
446: dest+=row->item.size;
447:
448: remove_crlf(dest_after_origins, dest);
449: to_char('\n');
1.1 paf 450: }
451: chunk=row->link;
452: } while(chunk);
1.64 parser 453:
1.1 paf 454: break2:
1.76 paf 455: *dest=0;
456: return result;
1.1 paf 457: }