Annotation of parser3/src/include/pa_string.h, revision 1.145.4.4
1.41 paf 1: /** @file
1.43 paf 2: Parser: string class decl.
3:
1.145 paf 4: Copyright (c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com)
1.124 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.1 paf 6: */
7:
8: #ifndef PA_STRING_H
9: #define PA_STRING_H
1.140 paf 10:
1.145.4.4! paf 11: static const char* IDENT_STRING_H="$Date: 2003/09/23 12:14:00 $";
1.145 paf 12:
13: // includes
1.1 paf 14:
1.4 paf 15: #include "pa_types.h"
1.145 paf 16: #include "pa_array.h"
17:
18: extern "C" { // cord's author forgot to do that
19: #define CORD_NO_IO
20: #include "cord.h"
21: };
1.4 paf 22:
1.145.4.3 paf 23: /// must use simple version, without optimization for short pieces
24: /// it's much more convinient for iterating. we rely on this here: CORD_block_iter
25: #define CORD_chars(c, i) CORD_chars_block((c), (i))
26: extern "C" CORD CORD_chars_block(char c, size_t i);
1.145.4.1 paf 27:
28: // cord extension
29: /* Returns true if x does contain */
30: /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */
31: int CORD_nnchr(CORD x, size_t i, size_t n, int not_c);
32:
1.145 paf 33: // forwards
1.9 paf 34:
1.145 paf 35: class Charset;
1.135 paf 36: class Table;
1.71 paf 37: class SQL_Connection;
1.101 parser 38: class Dictionary;
1.145 paf 39: class Request_charsets;
40: class String;
41: typedef Array<const String*> ArrayString;
42:
43: /// this is result of pos functions which mean that substr were not found
44: #define STRING_NOT_FOUND ((size_t)-1)
45:
46: class StringBody {
47:
48: CORD body;
49:
50: public:
51:
52: StringBody(): body(CORD_EMPTY) {}
53: StringBody(CORD abody): body(abody) {
54: assert(!body // no body
55: || *body // ordinary string
56: || body[1]==1 // CONCAT_HDR
57: || body[1]==4 // FN_HDR
58: || body[1]==6 // SUBSTR_HDR
59: );
60: }
61: /// WARNING: length is only HELPER length, str in ANY case should be zero-terminated
62: StringBody(const char* str, size_t helper_length): body(CORD_EMPTY) {
63: append_know_length(str, helper_length?helper_length:strlen(str));
64: }
65: static StringBody Format(int value);
66:
67: void clear() { body=CORD_EMPTY; }
68:
69: bool operator! () const { return is_empty(); }
70:
71: uint hash_code() const;
72:
73: const char* cstr() const { return CORD_to_const_char_star(body); }
74: char* cstrm() const { return CORD_to_char_star(body); }
75:
76: size_t length() const { return CORD_len(body); }
77:
78: bool is_empty() const { return body==CORD_EMPTY; }
79:
80: void append_know_length(const char *str, size_t known_length) {
81: if(known_length)
82: body=CORD_cat_char_star(body, str, known_length);
83: }
84: void append_strdup_know_length(const char* str, size_t known_length) {
85: if(known_length)
86: append_know_length(pa_strdup(str, known_length), known_length);
87: }
88: void append(char c) { body=CORD_cat_char(body, c); }
89: StringBody& operator << (const StringBody src) { body=CORD_cat(body, src.body); return *this; }
90: StringBody& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
91:
92: // could not figure out why this operator is needed [should do this chain: string->simple->==]
93: bool operator < (const StringBody src) const { return CORD_cmp(body, src.body)<0; }
94: bool operator > (const StringBody src) const { return CORD_cmp(body, src.body)>0; }
95: bool operator <= (const StringBody src) const { return CORD_cmp(body, src.body)<=0; }
96: bool operator >= (const StringBody src) const { return CORD_cmp(body, src.body)>=0; }
97: bool operator != (const StringBody src) const { return CORD_cmp(body, src.body)!=0; }
98: bool operator == (const StringBody src) const { return CORD_cmp(body, src.body)==0; }
99:
100: int ncmp(size_t x_begin, const StringBody y, size_t y_begin, size_t size) const {
101: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
102: }
103:
104: char fetch(size_t index) const { return CORD_fetch(body, index); }
105: StringBody mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
106: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
107: size_t pos(const StringBody substr, size_t offset=0) const {
108: if(!substr.length())
109: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
110: return CORD_str(body, offset, substr.body);
111: }
112: size_t pos(char c,
113: size_t offset=0) const {
114: return CORD_chr(body, offset, c);
115: }
116:
1.145.4.1 paf 117: /* template<typename I> void for_each(int (*callback)(const char* s, I), I info) const {
1.145 paf 118: CORD_iter5(body, 0, 0, (CORD_batched_iter_fn)callback, info);
1.145.4.1 paf 119: }*/
1.145 paf 120:
121: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
122:
1.145.4.2 paf 123: /*StringBody normalize() const {
1.145 paf 124: return StringBody(CORD_balance(body));
1.145.4.2 paf 125: }*/
1.145 paf 126:
127: void dump() const {
128: CORD_dump(body);
129: }
130: };
1.62 paf 131:
1.145.4.4! paf 132: template<typename T>
! 133: inline size_t get_length(T current) {
! 134: return current;
! 135: }
1.42 paf 136: /**
1.145.4.1 paf 137: String which knows the lang of all it's langs.
1.41 paf 138:
139: All pieces remember
140: - whether they are tainted or not,
1.145.4.1 paf 141: and the lang which should be used to detaint them
1.41 paf 142: */
1.145 paf 143: class String: public PA_Object {
1.1 paf 144: public:
1.48 paf 145:
1.145.4.1 paf 146: /** piece is tainted or not. the lang to use when detaint
1.106 parser 147: remember to change String_Untaint_lang_name @ untaint.C along
148: */
1.145 paf 149: enum Language {
1.145.4.1 paf 150: L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty
1.145 paf 151: // these two must go before others, there are checks for >L_AS_IS
152: L_CLEAN, ///< clean
153: L_AS_IS, ///< leave all characters intact
1.122 paf 154:
1.145 paf 155: L_PASS_APPENDED,
1.41 paf 156: /**<
1.145.4.1 paf 157: leave lang built into string being appended.
1.41 paf 158: just a flag, that value not stored
159: */
1.145.4.1 paf 160: L_TAINTED, ///< tainted, untaint lang as assigned later
161: // untaint langs. assigned by ^untaint[lang]{...}
1.145 paf 162: L_FILE_SPEC, ///< file specification
163: L_HTTP_HEADER, ///< text in HTTP response header
164: L_MAIL_HEADER, ///< text in mail header
165: L_URI, ///< text in uri
166: L_TABLE, ///< ^table:set body
167: L_SQL, ///< ^table:sql body
168: L_JS, ///< JavaScript code
169: L_XML, ///< ^dom:set xml
170: L_HTML, ///< HTML code (for editing)
1.145.4.1 paf 171: L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization
1.27 paf 172: };
173:
1.145.4.1 paf 174: class Languages {
175:
176: union {
177: struct {
178: Language lang:8;
179: int is_not_just_lang:16-8;
180: };
181: CORD langs;
182: };
183:
1.145.4.4! paf 184: template<typename C>
! 185: CORD make_langs(C current) const {
1.145.4.1 paf 186: return is_not_just_lang?
187: langs
1.145.4.4! paf 188: :CORD_chars((char)lang, get_length(current));
1.145 paf 189: }
190:
1.145.4.1 paf 191: CORD make_langs(size_t aoffset, size_t alength) const {
192: return is_not_just_lang?
193: CORD_substr(langs, aoffset, alength)
1.145.4.3 paf 194: :CORD_chars((char)lang, alength);
1.145 paf 195: }
1.145.4.1 paf 196:
1.145.4.3 paf 197: /// appending when 'langs' already contain something [simple cases hanled elsewhere]
1.145.4.4! paf 198: template<typename C>
! 199: void append(C current,
1.145.4.1 paf 200: const CORD to_nonempty_target_langs) {
1.145.4.3 paf 201: assert(langs);
1.145.4.1 paf 202:
203: if(is_not_just_lang)
204: langs=CORD_cat(langs, to_nonempty_target_langs);
205: else { // we were "just lang"
1.145.4.4! paf 206: size_t current_size=get_length(current);
1.145.4.1 paf 207: assert(current_size);
208: langs=CORD_cat(
1.145.4.3 paf 209: CORD_chars((char)lang, current_size), // first piece [making from just 'lang']
1.145.4.1 paf 210: to_nonempty_target_langs); // new piece
211: }
1.145 paf 212: }
1.145.4.1 paf 213:
1.145 paf 214: public:
1.145.4.1 paf 215:
216: Languages(): langs(0) {}
217: Languages(Language alang): lang(alang), is_not_just_lang(0) {}
218:
1.145.4.4! paf 219: /// MUST be called exactly prior to modification of current [uses it's length]
! 220: template<typename C>
! 221: void append(C current, Language alang, size_t asize) {
1.145.4.1 paf 222: assert(alang);
223: assert(asize);
224:
1.145.4.3 paf 225: if(!is_not_just_lang)
226: if(lang) {
227: if(lang==alang) // same length? ignoring
228: return;
229: } else {
230: lang=alang; // to uninitialized
231: return;
232: }
1.145.4.1 paf 233:
1.145.4.4! paf 234: append(current, CORD_chars((char)alang, asize));
1.145 paf 235: }
236:
1.145.4.4! paf 237: /// MUST be called exactly prior to modification of current [uses it's length]
! 238: template<typename C>
! 239: void append(C current, size_t appending_length,
1.145.4.1 paf 240: const Languages src) {
1.145.4.4! paf 241: assert(appending_length);
1.145.4.1 paf 242:
1.145.4.3 paf 243: if(!langs)
244: langs=src.langs; // to uninitialized
245: else if(!src.is_not_just_lang)
1.145.4.4! paf 246: append(current, src.lang, appending_length); // simplifying when simple source
1.145.4.3 paf 247: else
1.145.4.4! paf 248: append(current, src.make_langs(appending_length));
1.145.4.1 paf 249: }
250:
1.145.4.4! paf 251: /// MUST be called exactly prior to modification of current [uses it's length]
! 252: template<typename C>
! 253: void append(C current,
1.145.4.1 paf 254: const Languages src, size_t aoffset, size_t alength) {
255: assert(alength);
256:
1.145.4.3 paf 257: if(!langs) // to uninitialized?
1.145.4.1 paf 258: if(src.is_not_just_lang)
1.145.4.3 paf 259: langs=CORD_substr(src.langs, aoffset, alength); // to uninitialized complex
260: else
261: lang=src.lang; // to uninitialized simple
262: else
263: if(!is_not_just_lang && !src.is_not_just_lang && lang==src.lang) // both simple & of same language?
264: return; // ignoring
1.145.4.1 paf 265: else
1.145.4.4! paf 266: append(current, src.make_langs(aoffset, alength));
1.145.4.1 paf 267: }
268:
269: /// checks if we have alang all from aoffset to aoffset+alength
270: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
271: if(alang==L_UNSPECIFIED) // ignore lang?
272: return true;
273:
274: if(is_not_just_lang)
275: return CORD_nnchr(langs, aoffset, alength, (unsigned)alang)!=0;
276: else
277: return lang==alang;
278: }
279:
1.145.4.4! paf 280: template<typename C, typename I>
! 281: void for_each(C current,
1.145.4.2 paf 282: int callback(char, size_t, I), I info) const {
1.145.4.1 paf 283:
284: if(is_not_just_lang)
1.145.4.2 paf 285: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
1.145.4.1 paf 286: else
1.145.4.4! paf 287: callback(lang, get_length(current), info);
! 288: }
! 289:
! 290: bool invariant(size_t current_length) {
! 291: if(!langs)
! 292: return current_length==0;
! 293: if(is_not_just_lang)
! 294: return CORD_len(langs)==current_length;
! 295: return true; // uncheckable, actually
1.145 paf 296: }
297: };
298:
299: struct C {
300: const char *str;
301: size_t length;
302: operator const char *() { return str; }
303: C(const char *astr, size_t asize): str(astr), length(asize) {}
304: };
305:
306: struct Cm {
307: char *str;
308: size_t length;
309: //operator char *() { return str; }
310: Cm(char *astr, size_t asize): str(astr), length(asize) {}
311: };
312:
313: private:
314:
1.145.4.1 paf 315: Languages langs; ///< string characters lang
1.145 paf 316: StringBody body; ///< all characters of string
317:
1.145.4.4! paf 318: #define ASSERT_STRING_INVARIANT(string) \
! 319: assert((string).langs.invariant((string).body.length()))
! 320:
1.8 paf 321: public:
322:
1.145 paf 323: explicit String(const char* cstr=0, size_t helper_length=0, bool tainted=false);
324: explicit String(const C cstr, bool tainted=false);
1.145.4.4! paf 325: String(StringBody abody, Language alang): body(abody), langs(alang) {
! 326: assert(!body.is_empty());
! 327: ASSERT_STRING_INVARIANT(*this);
! 328: }
! 329: String(const String& src): body(src.body), langs(src.langs) {
! 330: ASSERT_STRING_INVARIANT(*this);
! 331: }
1.145 paf 332:
333: /// for convinient hash lookup
334: operator const StringBody() const { return body; }
335:
336: bool is_empty() const { return body.is_empty(); }
337: size_t length() const { return body.length(); }
338:
339: /// convert to CORD. if 'lang' known, forcing 'lang' to it
340: StringBody cstr_to_string_body(Language lang=L_AS_IS,
341: SQL_Connection* connection=0,
342: const Request_charsets *charsets=0) const;
343:
344: /// convert to constant C string. if 'lang' known, forcing 'lang' to it
345: const char* cstr(Language lang=L_AS_IS,
346: SQL_Connection* connection=0,
347: const Request_charsets *charsets=0) const {
348: return cstr_to_string_body(lang, connection, charsets).cstr();
349: }
350: /// convert to Modifiable C string. if 'lang' known, forcing 'lang' to it
351: char *cstrm(Language lang=L_AS_IS,
352: SQL_Connection* connection=0,
353: const Request_charsets *charsets=0) const {
354: return cstr_to_string_body(lang, connection, charsets).cstrm();
1.50 paf 355: }
1.108 parser 356: /// puts pieces to buf
1.145 paf 357: Cm serialize(size_t prolog_size) const;
1.108 parser 358: /// appends pieces from buf to self
1.145 paf 359: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
360: /// @see StringBody::append_know_length
361: String& append_know_length(const char* str, size_t known_length, Language lang);
362: /// @see StringBody::append_help_length
363: String& append_help_length(const char* str, size_t helper_length, Language lang);
364: String& append_strdup(const char* str, size_t helper_length, Language lang);
365:
366: bool operator == (const char* y) const { return body==StringBody(y); }
367: bool operator != (const char* y) const { return body!=StringBody(y); }
368:
369: /// this starts with y
370: bool starts_with(const char* y) const {
371: return body.ncmp(0/*x_begin*/, StringBody(y), 0/*y_begin*/, strlen(y))==0;
372: }
373: /// x starts with this
374: bool this_starts(const char* x) const {
375: return StringBody(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 376: }
377:
1.145 paf 378: String& append_to(String& dest, Language lang, bool forced) const;
379: String& append(const String& src, Language lang, bool forced=false) {
380: return src.append_to(*this, lang, forced);
381: }
382: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
383: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
384: String& operator << (const StringBody src) {
1.145.4.1 paf 385: langs.append(body, L_AS_IS, src.length());
1.145 paf 386: body<<src;
387: return *this;
388: }
1.100 parser 389:
1.142 paf 390: /// extracts first char of a string, if any
391: char first_char() const {
1.145 paf 392: return is_empty()?0:body.fetch(0);
1.142 paf 393: }
1.54 paf 394:
1.145 paf 395: bool operator < (const String& src) const { return body<src.body; }
396: bool operator > (const String& src) const { return body>src.body; }
397: bool operator <= (const String& src) const { return body<=src.body; }
398: bool operator >= (const String& src) const { return body>=src.body; }
399: bool operator != (const String& src) const { return body!=src.body; }
400: bool operator == (const String& src) const { return body==src.body; }
401:
1.54 paf 402: /// extracts [start, finish) piece of string
1.145 paf 403: String& mid(size_t substr_begin, size_t substr_end) const;
404:
405: /**
406: ignore lang if it's L_UNSPECIFIED
407: but when specified: look for substring that lies in ONE fragment in THAT lang
408: @return position of substr in string, -1 means "not found" [const char* version]
409: */
410: size_t pos(const StringBody substr,
411: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
412: /// String version of @see pos(const char*, int, Language)
413: size_t pos(const String& substr,
414: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
415: size_t pos(char c,
416: size_t this_offset=0) const {
417: return body.pos(c, this_offset);
418: }
1.55 paf 419:
1.145 paf 420: void split(ArrayString& result,
421: size_t& pos_after,
422: const char* delim,
423: Language lang=L_UNSPECIFIED, int limit=-1) const;
424: void split(ArrayString& result,
425: size_t& pos_after,
1.62 paf 426: const String& delim,
1.145 paf 427: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 428:
1.145 paf 429: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 430: int prestart, int prefinish,
431: int poststart, int postfinish,
1.68 paf 432: void *info);
1.87 parser 433: /**
1.145 paf 434: @return table of found items, if any.
1.87 parser 435: table format is defined and fixed[can be used by others]:
436: @verbatim
437: prematch/match/postmatch/1/2/3/...
438: @endverbatim
439: */
1.145 paf 440: Table* match(Charset& source_charset,
1.64 paf 441: const String& regexp,
1.145 paf 442: const String* options,
1.99 parser 443: Row_action row_action, void *info,
1.145 paf 444: bool& just_matched) const;
1.87 parser 445: enum Change_case_kind {
446: CC_UPPER,
447: CC_LOWER
448: };
1.145 paf 449: String& change_case(Charset& source_charset,
1.87 parser 450: Change_case_kind kind) const;
1.145 paf 451: const String& replace(const Dictionary& dict) const;
1.96 parser 452: double as_double() const;
453: int as_int() const;
1.137 paf 454:
1.7 paf 455: private: //disabled
456:
1.12 paf 457: String& operator = (const String&) { return *this; }
1.7 paf 458:
1.1 paf 459: };
1.145.4.4! paf 460:
! 461: template<>
! 462: inline size_t get_length<StringBody>(StringBody body) {
! 463: return body.length();
! 464: }
1.119 paf 465:
1.145 paf 466: /// simple hash code of string. used by Hash
467: inline uint hash_code(const StringBody self) {
468: return self.hash_code();
1.119 paf 469: }
1.1 paf 470:
471: #endif
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_string.h CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / include