Annotation of parser3/src/include/pa_string.h, revision 1.190
1.41 paf 1: /** @file
1.43 paf 2: Parser: string class decl.
3:
1.179 misha 4: Copyright (c) 2001-2009 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:
1.190 ! misha 11: static const char * const IDENT_STRING_H="$Date: 2009-07-06 08:45:09 $";
1.145 paf 12:
13: // includes
1.4 paf 14: #include "pa_types.h"
1.145 paf 15: #include "pa_array.h"
16:
17: extern "C" { // cord's author forgot to do that
18: #define CORD_NO_IO
19: #include "cord.h"
1.186 misha 20:
21: #ifdef CORD_CAT_OPTIMIZATION
22: #define CORD_cat(x, y) CORD_cat_optimized(x, y)
23: #define CORD_cat_char_star(x, y, leny) CORD_cat_char_star_optimized(x, y, leny)
24: #endif
1.145 paf 25: };
1.4 paf 26:
1.164 paf 27: // defines
28:
1.183 misha 29: // cache hash code in String::Body for faster hash access
30: #define HASH_CODE_CACHING
31:
1.187 misha 32: // cache String::Body.length() for char* strings only, CORDs have own length
33: #define STRING_LENGTH_CACHING
34:
35:
1.146 paf 36: // cord extension
37: /* Returns true if x does contain */
38: /* char not_c at positions i..i+n. Value i,i+n must be < CORD_len(x). */
39: int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c);
1.148 paf 40: size_t CORD_block_count(CORD x);
1.146 paf 41:
1.145 paf 42: // forwards
1.9 paf 43:
1.145 paf 44: class Charset;
1.135 paf 45: class Table;
1.71 paf 46: class SQL_Connection;
1.101 parser 47: class Dictionary;
1.145 paf 48: class Request_charsets;
49: class String;
50: typedef Array<const String*> ArrayString;
1.178 misha 51: class VRegex;
1.145 paf 52:
1.155 paf 53: // generally useful
54:
55: int pa_atoi(const char* str, const String* problem_source=0);
56: double pa_atod(const char* str, const String* problem_source=0);
57:
1.145 paf 58: /// this is result of pos functions which mean that substr were not found
59: #define STRING_NOT_FOUND ((size_t)-1)
60:
1.42 paf 61: /**
1.146 paf 62: String which knows the lang of all it's langs.
1.41 paf 63:
64: All pieces remember
65: - whether they are tainted or not,
1.146 paf 66: and the lang which should be used to detaint them
1.41 paf 67: */
1.176 misha 68:
69:
1.145 paf 70: class String: public PA_Object {
1.1 paf 71: public:
1.48 paf 72:
1.146 paf 73: /** piece is tainted or not. the lang to use when detaint
1.106 parser 74: remember to change String_Untaint_lang_name @ untaint.C along
1.148 paf 75:
76: WARNING WARNING WARNING WARNING WARNING WARNING
77:
78: pos function compares(<=) languages, that is used in searching
79: for table column separator being L_CLEAN or L_AS_IS.
80: they search for AS_IS, meaning AS_IS|CLEAN [doing <=L_AS_IS check].
81:
82: letters assigned for debugging, but it's important for no language-letter
83: come before L_AS_IS other then L_CLEAN
84:
85: WARNING WARNING WARNING WARNING WARNING WARNING
1.106 parser 86: */
1.145 paf 87: enum Language {
1.146 paf 88: L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty
1.145 paf 89: // these two must go before others, there are checks for >L_AS_IS
1.170 misha 90: L_CLEAN='0', ///< clean WARNING: read above warning before changing
91: L_AS_IS='A', ///< leave all characters intact WARNING: read above warning before changing
1.122 paf 92:
1.148 paf 93: L_PASS_APPENDED='P',
1.41 paf 94: /**<
1.146 paf 95: leave lang built into string being appended.
1.41 paf 96: just a flag, that value not stored
97: */
1.170 misha 98: L_TAINTED='T', ///< tainted, untaint lang as assigned later
1.146 paf 99: // untaint langs. assigned by ^untaint[lang]{...}
1.170 misha 100: L_FILE_SPEC='F', ///< file specification
101: L_HTTP_HEADER='h', ///< text in HTTP response header
102: L_MAIL_HEADER='m', ///< text in mail header
103: L_URI='U', ///< text in uri
104: L_SQL='Q', ///< ^table:sql body
105: L_JS='J', ///< JavaScript code
1.148 paf 106: L_XML='X', ///< ^dom:set xml
1.170 misha 107: L_HTML='H', ///< HTML code
108: L_REGEX='R', ///< RegEx expression
109: L_HTTP_COOKIE='C', ///< cookies encoded as %uXXXX for compartibility with js functions encode/decode
1.175 misha 110: L_FILE_POST='f', ///temporary escaping zero-char
1.148 paf 111: // READ WARNING ABOVE BEFORE ADDING ANYTHING
1.146 paf 112: L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization
1.27 paf 113: };
114:
1.157 paf 115: enum Trim_kind {
116: TRIM_BOTH,
117: TRIM_START,
118: TRIM_END
119: };
120:
1.187 misha 121: class Body;
1.176 misha 122:
1.147 paf 123: union Languages {
1.146 paf 124:
1.147 paf 125: struct {
1.160 paf 126: #ifdef PA_LITTLE_ENDIAN
1.147 paf 127: Language lang:8;
1.179 misha 128: size_t is_not_just_lang:sizeof(CORD)*8-8;
1.160 paf 129: #elif defined(PA_BIG_ENDIAN)
1.179 misha 130: size_t is_not_just_lang:sizeof(CORD)*8-8;
1.160 paf 131: Language lang:8;
132: #else
133: # error word endianness not determined for some obscure reason
134: #endif
1.147 paf 135: } opt;
136: CORD langs;
1.146 paf 137:
1.187 misha 138: CORD make_langs(const Body& current) const {
139: return opt.is_not_just_lang?langs:CORD_chars((char)opt.lang, current.length());
1.146 paf 140: }
141:
142: CORD make_langs(size_t aoffset, size_t alength) const {
1.147 paf 143: return opt.is_not_just_lang?
1.146 paf 144: CORD_substr(langs, aoffset, alength)
1.162 paf 145: :CORD_chars((char)opt.lang, alength);
1.145 paf 146: }
147:
1.157 paf 148: /// appending when 'langs' already contain something [simple cases handled elsewhere]
1.187 misha 149: void append(size_t current, const CORD to_nonempty_target_langs) {
1.146 paf 150: assert(langs);
1.187 misha 151: if(opt.is_not_just_lang)
152: langs=CORD_cat(langs, to_nonempty_target_langs);
153: else {
154: assert(current);
155: langs=CORD_cat(CORD_chars((char)opt.lang, current), to_nonempty_target_langs);
156: }
157: }
1.146 paf 158:
1.187 misha 159: void append(const Body& current, const CORD to_nonempty_target_langs) {
160: assert(langs);
1.147 paf 161: if(opt.is_not_just_lang)
1.146 paf 162: langs=CORD_cat(langs, to_nonempty_target_langs);
1.187 misha 163: else {
164: size_t current_size=current.length();
1.146 paf 165: assert(current_size);
1.187 misha 166: langs=CORD_cat(CORD_chars((char)opt.lang, current_size), to_nonempty_target_langs);
1.146 paf 167: }
1.145 paf 168: }
1.146 paf 169:
1.145 paf 170: public:
1.146 paf 171:
172: const char* v() const;
1.164 paf 173: void dump() const;
1.146 paf 174:
175: Languages(): langs(0) {}
1.147 paf 176: Languages(Language alang) {
177: opt.lang=alang;
178: opt.is_not_just_lang=0;
179: }
1.146 paf 180:
181: /// MUST be called exactly prior to modification of current [uses it's length]
1.187 misha 182: void append(size_t current, Language alang, size_t length) {
1.146 paf 183: assert(alang);
1.187 misha 184: assert(length);
1.146 paf 185:
1.147 paf 186: if(!opt.is_not_just_lang)
187: if(opt.lang) {
1.181 misha 188: if(opt.lang==alang) // same language? ignoring
1.146 paf 189: return;
190: } else {
1.147 paf 191: opt.lang=alang; // to uninitialized
1.146 paf 192: return;
1.145 paf 193: }
1.146 paf 194:
1.187 misha 195: append(current, CORD_chars((char)alang, length));
1.146 paf 196: }
197:
1.187 misha 198: void append(const Body ¤t, Language alang, size_t length) {
1.176 misha 199: assert(alang);
1.187 misha 200: assert(length);
1.176 misha 201:
202: if(!opt.is_not_just_lang)
203: if(opt.lang) {
1.183 misha 204: if(opt.lang==alang) // same language? ignoring
1.176 misha 205: return;
206: } else {
207: opt.lang=alang; // to uninitialized
208: return;
209: }
210:
1.187 misha 211: append(current, CORD_chars((char)alang, length));
1.176 misha 212: }
213:
1.187 misha 214: void appendHelper(const Body& current, Language alang, const Body &length_helper) {
215: assert(alang);
216:
217: if(!opt.is_not_just_lang)
218: if(opt.lang) {
219: if(opt.lang==alang) // same language? ignoring
220: return;
221: } else {
222: opt.lang=alang; // to uninitialized
223: return;
224: }
225:
226: append(current, CORD_chars((char)alang, length_helper.length()));
227: }
228:
229: void appendHelper(const Body& current, const Languages &src, const Body& length_helper) {
1.186 misha 230: if(!langs){
1.176 misha 231: langs=src.langs; // to uninitialized
1.186 misha 232: #ifdef CORD_CAT_OPTIMIZATION
233: if(opt.is_not_just_lang && !CORD_IS_STRING(langs))
234: CORD_concatenation_protect(langs);
235: #endif
236: }
1.176 misha 237: else if(!src.opt.is_not_just_lang)
238: appendHelper(current, src.opt.lang, length_helper); // simplifying when simple source
239: else
240: append(current, src.make_langs(length_helper));
241: }
242:
1.146 paf 243: /// MUST be called exactly prior to modification of current [uses it's length]
1.187 misha 244: void append(const Body& current, const Languages src, size_t aoffset, size_t alength) {
1.146 paf 245: assert(alength);
246:
247: if(!langs) // to uninitialized?
1.147 paf 248: if(src.opt.is_not_just_lang)
1.146 paf 249: langs=CORD_substr(src.langs, aoffset, alength); // to uninitialized complex
250: else
1.147 paf 251: opt.lang=src.opt.lang; // to uninitialized simple
1.146 paf 252: else
1.147 paf 253: if(!opt.is_not_just_lang && !src.opt.is_not_just_lang && opt.lang==src.opt.lang) // both simple & of same language?
1.146 paf 254: return; // ignoring
255: else
256: append(current, src.make_langs(aoffset, alength));
257: }
258:
259: /// checks if we have lang<=alang all from aoffset to aoffset+alength
260: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
261: if(alang==L_UNSPECIFIED) // ignore lang?
262: return true;
263:
1.147 paf 264: if(opt.is_not_just_lang)
1.146 paf 265: return CORD_range_contains_chr_greater_then(langs, aoffset, alength, (unsigned)alang)==0;
266: else
1.169 misha 267: return (unsigned)opt.lang<=(unsigned)alang;
1.146 paf 268: }
269:
1.148 paf 270: /// @returns count of blocks
271: /// @todo currently there can be adjucent blocks of same language. someday merge them
272: size_t count() const {
273: return opt.is_not_just_lang?
274: CORD_block_count(langs)
275: : opt.lang?
276: 1
277: : 0;
278: };
279:
1.187 misha 280: template<typename I>
281: void for_each(size_t current, int callback(char, size_t, I), I info) const {
282: if(opt.is_not_just_lang)
283: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
284: else
285: callback(opt.lang, current, info);
286: }
287:
288: template<typename I>
289: void for_each(const Body& current, int callback(char, size_t, I), I info) const {
1.147 paf 290: if(opt.is_not_just_lang)
1.146 paf 291: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
292: else
1.187 misha 293: callback(opt.lang, current.length(), info);
1.146 paf 294: }
295:
1.164 paf 296: bool invariant(size_t current_length) const {
1.146 paf 297: if(!langs)
298: return current_length==0;
1.147 paf 299: if(opt.is_not_just_lang)
1.146 paf 300: return CORD_len(langs)==current_length;
301: return true; // uncheckable, actually
302: }
303: };
304:
305: class Body {
306:
307: CORD body;
308:
1.183 misha 309: #ifdef HASH_CODE_CACHING
310: // cached hash code is not reseted on write operations as test shows
311: // that string body does not change after it is stored as a hash key
312: mutable uint hash_code;
313: #endif
314:
1.187 misha 315: #ifdef STRING_LENGTH_CACHING
316: // cached length is reseted on modification, used only for char*, not CORD
317: mutable size_t string_length;
318: #define INIT_LENGTH ,string_length(0)
319: #define ZERO_LENGTH string_length=0;
320: #else
321: #define INIT_LENGTH
322: #define ZERO_LENGTH
323: #endif
324:
1.146 paf 325: public:
326:
327: const char* v() const;
1.164 paf 328: void dump() const;
1.146 paf 329:
1.183 misha 330: #ifdef HASH_CODE_CACHING
1.187 misha 331: Body(): body(CORD_EMPTY), hash_code(0) INIT_LENGTH {}
332: Body(CORD abody, uint ahash_code): body(abody), hash_code(ahash_code) INIT_LENGTH {}
333: Body(CORD abody): body(abody), hash_code(0) INIT_LENGTH {
1.183 misha 334: #else
1.187 misha 335: Body(): body(CORD_EMPTY) INIT_LENGTH {}
336: Body(CORD abody): body(abody) INIT_LENGTH {
1.183 misha 337: #endif
1.186 misha 338: #ifdef CORD_CAT_OPTIMIZATION
339: assert(!body // no body
340: || *body // ordinary string
341: || body[1]==1 // CONCAT_HDR
342: || body[1]==3 // CONCAT_HDR_READ_ONLY
343: || body[1]==4 // FN_HDR
344: || body[1]==6 // SUBSTR_HDR
345: );
346: #else
1.146 paf 347: assert(!body // no body
348: || *body // ordinary string
349: || body[1]==1 // CONCAT_HDR
350: || body[1]==4 // FN_HDR
351: || body[1]==6 // SUBSTR_HDR
352: );
1.186 misha 353: #endif
1.146 paf 354: }
1.181 misha 355:
1.146 paf 356: static Body Format(int value);
357:
1.187 misha 358: void clear() { ZERO_LENGTH body=CORD_EMPTY; }
1.146 paf 359:
360: bool operator! () const { return is_empty(); }
361:
1.183 misha 362: CORD get_cord() const { return body; }
363: uint get_hash_code() const;
1.146 paf 364:
365: const char* cstr() const { return CORD_to_const_char_star(body); }
366: char* cstrm() const { return CORD_to_char_star(body); }
367:
1.187 misha 368: #ifdef STRING_LENGTH_CACHING
369: void set_length(size_t alength){ string_length = alength; }
370: size_t length() const { return body ? CORD_IS_STRING(body) ? string_length ? string_length : (string_length=strlen(body)) : CORD_len(body) : 0; }
371: #else
1.146 paf 372: size_t length() const { return CORD_len(body); }
1.187 misha 373: #endif
1.146 paf 374:
375: bool is_empty() const { return body==CORD_EMPTY; }
376:
377: void append_know_length(const char *str, size_t known_length) {
1.187 misha 378: if(known_length){
379: if(body){
380: body = CORD_cat_char_star(body, str, known_length);
381: ZERO_LENGTH
382: } else {
383: body=str;
384: #ifdef STRING_LENGTH_CACHING
385: string_length=known_length;
386: #endif
387: }
388: }
1.146 paf 389: }
390: void append_strdup_know_length(const char* str, size_t known_length) {
391: if(known_length)
392: append_know_length(pa_strdup(str, known_length), known_length);
393: }
1.187 misha 394: void append(char c) { ZERO_LENGTH body=CORD_cat_char(body, c); }
395: Body& operator << (const Body src) { ZERO_LENGTH body=CORD_cat(body, src.body); return *this; }
396:
1.146 paf 397: Body& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
398:
399: // could not figure out why this operator is needed [should do this chain: string->simple->==]
400: bool operator < (const Body src) const { return CORD_cmp(body, src.body)<0; }
401: bool operator > (const Body src) const { return CORD_cmp(body, src.body)>0; }
402: bool operator <= (const Body src) const { return CORD_cmp(body, src.body)<=0; }
403: bool operator >= (const Body src) const { return CORD_cmp(body, src.body)>=0; }
404: bool operator != (const Body src) const { return CORD_cmp(body, src.body)!=0; }
405: bool operator == (const Body src) const { return CORD_cmp(body, src.body)==0; }
406:
407: int ncmp(size_t x_begin, const Body y, size_t y_begin, size_t size) const {
408: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
1.145 paf 409: }
410:
1.146 paf 411: char fetch(size_t index) const { return CORD_fetch(body, index); }
412: Body mid(size_t index, size_t length) const { return CORD_substr(body, index, length); }
413: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr); }
414: size_t pos(const Body substr, size_t offset=0) const {
1.180 misha 415: if(substr.is_empty())
1.146 paf 416: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
1.150 paf 417:
418: // CORD_str checks for bad offset [CORD_chr does not]
1.146 paf 419: return CORD_str(body, offset, substr.body);
420: }
421: size_t pos(char c,
422: size_t offset=0) const {
1.150 paf 423: if(offset>=length()) // CORD_chr does not check that [and ABORT's in that case]
424: return STRING_NOT_FOUND;
425:
1.146 paf 426: return CORD_chr(body, offset, c);
1.145 paf 427: }
1.146 paf 428:
1.187 misha 429: template<typename I>
430: int for_each(int (*f)(char c, I), I info) const {
1.159 paf 431: return CORD_iter(body, (CORD_iter_fn)f, (void*)info);
432: }
433:
1.187 misha 434: template<typename I>
435: int for_each(int (*f1)(char c, I), int (*f2)(const char* s, I), I info) const {
1.159 paf 436: return CORD_iter5(body, 0, (CORD_iter_fn)f1, (CORD_batched_iter_fn)f2, info);
1.148 paf 437: }
1.146 paf 438:
439: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
440:
441: /*Body normalize() const {
442: return Body(CORD_balance(body));
443: }*/
1.157 paf 444:
445: /// @returns this or 0 or mid. if returns this or 0 out_* are not filled
446: Body trim(Trim_kind kind=TRIM_BOTH, const char* chars=0,
1.188 misha 447: size_t* out_start=0, size_t* out_length=0, Charset* source_charset=0) const;
1.145 paf 448: };
449:
450: struct C {
451: const char *str;
452: size_t length;
453: operator const char *() { return str; }
1.158 paf 454: C(): str(0), length(0) {}
1.145 paf 455: C(const char *astr, size_t asize): str(astr), length(asize) {}
456: };
457:
458: struct Cm {
459: char *str;
460: size_t length;
461: //operator char *() { return str; }
1.158 paf 462: Cm(): str(0), length(0) {}
1.145 paf 463: Cm(char *astr, size_t asize): str(astr), length(asize) {}
464: };
465:
466: private:
467:
1.152 paf 468: Body body; ///< all characters of string
1.146 paf 469: Languages langs; ///< string characters lang
470:
471: const char* v() const;
1.164 paf 472: void dump() const;
473: #define ASSERT_STRING_INVARIANT(string) \
474: assert((string).langs.invariant((string).body.length()))
1.145 paf 475:
1.8 paf 476: public:
1.151 paf 477:
478: static const String Empty;
1.8 paf 479:
1.181 misha 480: explicit String(){};
1.182 misha 481: explicit String(const char* cstr, Language alang=L_CLEAN){
1.181 misha 482: if(cstr && *cstr){
483: body=cstr;
1.182 misha 484: langs=alang;
1.181 misha 485: }
486: }
1.187 misha 487: explicit String(const char* cstr, Language alang, size_t alength){
488: if(cstr && *cstr){
489: body=cstr;
490: #ifdef STRING_LENGTH_CACHING
491: body.set_length(alength);
492: #endif
1.182 misha 493: langs=alang;
1.181 misha 494: }
495: }
1.187 misha 496:
1.185 misha 497: String(int value, char *format);
1.146 paf 498: String(Body abody, Language alang): body(abody), langs(alang) {
499: ASSERT_STRING_INVARIANT(*this);
500: }
501: String(const String& src): body(src.body), langs(src.langs) {
502: ASSERT_STRING_INVARIANT(*this);
1.145 paf 503: }
504:
505: /// for convinient hash lookup
1.183 misha 506: #ifdef HASH_CODE_CACHING
507: operator const Body&() const { return body; }
508: #else
1.146 paf 509: operator const Body() const { return body; }
1.183 misha 510: #endif
1.145 paf 511:
512: bool is_empty() const { return body.is_empty(); }
513: size_t length() const { return body.length(); }
1.171 misha 514: size_t length(Charset& charset) const;
1.145 paf 515:
516: /// convert to CORD. if 'lang' known, forcing 'lang' to it
1.189 misha 517: Body cstr_to_string_body(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const;
518: Body cstr_to_string_body_taint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const;
519: Body cstr_to_string_body_untaint(Language lang, SQL_Connection* connection, const Request_charsets *charsets) const;
1.145 paf 520:
1.189 misha 521: ///
522: const char* cstr() const {
523: return body.cstr();
524: }
525: ///
526: char* cstrm() const {
527: return body.cstrm();
528: }
529:
530: /// convert to constant C string forcing lang tainting
1.190 ! misha 531: const char* taint_cstr(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 532: return cstr_to_string_body_taint(lang, connection, charsets).cstr();
533: }
1.190 ! misha 534: char *taint_cstrm(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 535: return cstr_to_string_body_taint(lang, connection, charsets).cstrm();
536: }
537:
538: /// convert to constant C string with tainting dirty to lang
1.190 ! misha 539: const char* untaint_cstr(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 540: return cstr_to_string_body_untaint(lang, connection, charsets).cstr();
541: }
1.190 ! misha 542: char *untaint_cstrm(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 543: return cstr_to_string_body_untaint(lang, connection, charsets).cstrm();
544: }
1.108 parser 545: /// puts pieces to buf
1.145 paf 546: Cm serialize(size_t prolog_size) const;
1.108 parser 547: /// appends pieces from buf to self
1.145 paf 548: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
1.146 paf 549: /// @see Body::append_know_length
1.145 paf 550: String& append_know_length(const char* str, size_t known_length, Language lang);
1.146 paf 551: /// @see Body::append_help_length
1.145 paf 552: String& append_help_length(const char* str, size_t helper_length, Language lang);
553: String& append_strdup(const char* str, size_t helper_length, Language lang);
554:
1.146 paf 555: bool operator == (const char* y) const { return body==Body(y); }
556: bool operator != (const char* y) const { return body!=Body(y); }
1.145 paf 557:
558: /// this starts with y
559: bool starts_with(const char* y) const {
1.146 paf 560: return body.ncmp(0/*x_begin*/, Body(y), 0/*y_begin*/, strlen(y))==0;
1.145 paf 561: }
562: /// x starts with this
563: bool this_starts(const char* x) const {
1.146 paf 564: return Body(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 565: }
566:
1.187 misha 567: String& append_to(String& dest, Language lang, bool forced=false) const;
1.145 paf 568: String& append(const String& src, Language lang, bool forced=false) {
569: return src.append_to(*this, lang, forced);
570: }
571: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
572: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
1.187 misha 573: String& operator << (const Body& src){
574: langs.appendHelper(body, L_AS_IS, src);
575: body<<src;
576: return *this;
577: }
1.100 parser 578:
1.142 paf 579: /// extracts first char of a string, if any
580: char first_char() const {
1.145 paf 581: return is_empty()?0:body.fetch(0);
1.142 paf 582: }
1.54 paf 583:
1.145 paf 584: bool operator < (const String& src) const { return body<src.body; }
585: bool operator > (const String& src) const { return body>src.body; }
586: bool operator <= (const String& src) const { return body<=src.body; }
587: bool operator >= (const String& src) const { return body>=src.body; }
588: bool operator != (const String& src) const { return body!=src.body; }
589: bool operator == (const String& src) const { return body==src.body; }
590:
1.54 paf 591: /// extracts [start, finish) piece of string
1.145 paf 592: String& mid(size_t substr_begin, size_t substr_end) const;
1.172 misha 593: String& mid(Charset& charset, size_t from, size_t to, size_t helper_length=0) const;
1.145 paf 594:
595: /**
596: ignore lang if it's L_UNSPECIFIED
597: but when specified: look for substring that lies in ONE fragment in THAT lang
598: @return position of substr in string, -1 means "not found" [const char* version]
599: */
1.146 paf 600: size_t pos(const Body substr,
1.145 paf 601: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
602: /// String version of @see pos(const char*, int, Language)
603: size_t pos(const String& substr,
604: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
605: size_t pos(char c,
606: size_t this_offset=0) const {
607: return body.pos(c, this_offset);
608: }
1.173 misha 609: size_t pos(Charset& charset,
1.171 misha 610: const String& substr,
611: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
1.55 paf 612:
1.145 paf 613: void split(ArrayString& result,
614: size_t& pos_after,
615: const char* delim,
616: Language lang=L_UNSPECIFIED, int limit=-1) const;
617: void split(ArrayString& result,
618: size_t& pos_after,
1.62 paf 619: const String& delim,
1.145 paf 620: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 621:
1.145 paf 622: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 623: int prestart, int prefinish,
624: int poststart, int postfinish,
1.68 paf 625: void *info);
1.87 parser 626: /**
1.145 paf 627: @return table of found items, if any.
1.87 parser 628: table format is defined and fixed[can be used by others]:
629: @verbatim
630: prematch/match/postmatch/1/2/3/...
631: @endverbatim
632: */
1.178 misha 633: Table* match(VRegex* vregex,
1.99 parser 634: Row_action row_action, void *info,
1.168 misha 635: int& matches_count) const;
1.87 parser 636: enum Change_case_kind {
637: CC_UPPER,
638: CC_LOWER
639: };
1.145 paf 640: String& change_case(Charset& source_charset,
1.87 parser 641: Change_case_kind kind) const;
1.145 paf 642: const String& replace(const Dictionary& dict) const;
1.188 misha 643: const String& trim(Trim_kind kind=TRIM_BOTH, const char* chars=0, Charset* source_charset=0) const;
1.155 paf 644: double as_double() const { return pa_atod(cstr(), this); }
645: int as_int() const { return pa_atoi(cstr(), this); }
1.167 misha 646: bool as_bool() const { return as_int()!=0; }
1.174 misha 647: const String& escape(Charset& source_charset) const;
1.137 paf 648:
1.7 paf 649: private: //disabled
650:
1.12 paf 651: String& operator = (const String&) { return *this; }
1.7 paf 652:
1.1 paf 653: };
1.119 paf 654:
1.183 misha 655: #ifndef HASH_CODE_CACHING
1.145 paf 656: /// simple hash code of string. used by Hash
1.146 paf 657: inline uint hash_code(const String::Body self) {
1.184 misha 658: return self.get_hash_code();
1.119 paf 659: }
1.183 misha 660: #endif
1.147 paf 661:
662:
1.1 paf 663: #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