Annotation of parser3/src/include/pa_string.h, revision 1.205
1.41 paf 1: /** @file
1.43 paf 2: Parser: string class decl.
3:
1.201 moko 4: Copyright (c) 2001-2012 Art. Lebedev Studio (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:
1.205 ! moko 11: #define IDENT_PA_STRING_H "$Id: pa_string.h,v 1.204 2012-05-28 10:33:18 moko Exp $"
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);
1.204 moko 57: unsigned int pa_atoui(const char *str, int base, const String* problem_source=0);
1.155 paf 58:
1.145 paf 59: /// this is result of pos functions which mean that substr were not found
60: #define STRING_NOT_FOUND ((size_t)-1)
61:
1.42 paf 62: /**
1.146 paf 63: String which knows the lang of all it's langs.
1.41 paf 64:
65: All pieces remember
66: - whether they are tainted or not,
1.146 paf 67: and the lang which should be used to detaint them
1.41 paf 68: */
1.176 misha 69:
70:
1.145 paf 71: class String: public PA_Object {
1.1 paf 72: public:
1.48 paf 73:
1.146 paf 74: /** piece is tainted or not. the lang to use when detaint
1.106 parser 75: remember to change String_Untaint_lang_name @ untaint.C along
1.148 paf 76:
77: WARNING WARNING WARNING WARNING WARNING WARNING
78:
79: pos function compares(<=) languages, that is used in searching
80: for table column separator being L_CLEAN or L_AS_IS.
81: they search for AS_IS, meaning AS_IS|CLEAN [doing <=L_AS_IS check].
82:
83: letters assigned for debugging, but it's important for no language-letter
84: come before L_AS_IS other then L_CLEAN
85:
86: WARNING WARNING WARNING WARNING WARNING WARNING
1.106 parser 87: */
1.145 paf 88: enum Language {
1.146 paf 89: L_UNSPECIFIED=0, ///< no real string has parts of this lange: it's just convinient to check when string's empty
1.145 paf 90: // these two must go before others, there are checks for >L_AS_IS
1.170 misha 91: L_CLEAN='0', ///< clean WARNING: read above warning before changing
92: L_AS_IS='A', ///< leave all characters intact WARNING: read above warning before changing
1.122 paf 93:
1.148 paf 94: L_PASS_APPENDED='P',
1.41 paf 95: /**<
1.146 paf 96: leave lang built into string being appended.
1.41 paf 97: just a flag, that value not stored
98: */
1.170 misha 99: L_TAINTED='T', ///< tainted, untaint lang as assigned later
1.146 paf 100: // untaint langs. assigned by ^untaint[lang]{...}
1.170 misha 101: L_FILE_SPEC='F', ///< file specification
102: L_HTTP_HEADER='h', ///< text in HTTP response header
103: L_MAIL_HEADER='m', ///< text in mail header
1.192 misha 104: L_URI='U', ///< text in uri
105: L_SQL='Q', ///< ^table:sql body
106: L_JS='J', ///< JavaScript code
1.198 misha 107: L_XML='X', ///< ^xdoc:create xml
1.192 misha 108: L_HTML='H', ///< HTML code
1.198 misha 109: L_REGEX='R', ///< RegExp
110: L_JSON='S', ///< JSON code
1.170 misha 111: L_HTTP_COOKIE='C', ///< cookies encoded as %uXXXX for compartibility with js functions encode/decode
1.192 misha 112: L_PARSER_CODE='p', ///< ^process body
1.148 paf 113: // READ WARNING ABOVE BEFORE ADDING ANYTHING
1.146 paf 114: L_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization
1.27 paf 115: };
116:
1.157 paf 117: enum Trim_kind {
118: TRIM_BOTH,
119: TRIM_START,
120: TRIM_END
121: };
122:
1.187 misha 123: class Body;
1.176 misha 124:
1.147 paf 125: union Languages {
1.146 paf 126:
1.147 paf 127: struct {
1.160 paf 128: #ifdef PA_LITTLE_ENDIAN
1.147 paf 129: Language lang:8;
1.179 misha 130: size_t is_not_just_lang:sizeof(CORD)*8-8;
1.160 paf 131: #elif defined(PA_BIG_ENDIAN)
1.179 misha 132: size_t is_not_just_lang:sizeof(CORD)*8-8;
1.160 paf 133: Language lang:8;
134: #else
135: # error word endianness not determined for some obscure reason
136: #endif
1.147 paf 137: } opt;
138: CORD langs;
1.146 paf 139:
1.187 misha 140: CORD make_langs(const Body& current) const {
141: return opt.is_not_just_lang?langs:CORD_chars((char)opt.lang, current.length());
1.146 paf 142: }
143:
144: CORD make_langs(size_t aoffset, size_t alength) const {
1.147 paf 145: return opt.is_not_just_lang?
1.194 misha 146: CORD_substr(langs, aoffset, alength, 0)
1.162 paf 147: :CORD_chars((char)opt.lang, alength);
1.145 paf 148: }
149:
1.157 paf 150: /// appending when 'langs' already contain something [simple cases handled elsewhere]
1.187 misha 151: void append(size_t current, const CORD to_nonempty_target_langs) {
1.146 paf 152: assert(langs);
1.187 misha 153: if(opt.is_not_just_lang)
154: langs=CORD_cat(langs, to_nonempty_target_langs);
155: else {
156: assert(current);
157: langs=CORD_cat(CORD_chars((char)opt.lang, current), to_nonempty_target_langs);
158: }
159: }
1.146 paf 160:
1.187 misha 161: void append(const Body& current, const CORD to_nonempty_target_langs) {
162: assert(langs);
1.147 paf 163: if(opt.is_not_just_lang)
1.146 paf 164: langs=CORD_cat(langs, to_nonempty_target_langs);
1.187 misha 165: else {
166: size_t current_size=current.length();
1.146 paf 167: assert(current_size);
1.187 misha 168: langs=CORD_cat(CORD_chars((char)opt.lang, current_size), to_nonempty_target_langs);
1.146 paf 169: }
1.145 paf 170: }
1.146 paf 171:
1.145 paf 172: public:
1.146 paf 173:
174: const char* v() const;
1.164 paf 175: void dump() const;
1.146 paf 176:
177: Languages(): langs(0) {}
1.147 paf 178: Languages(Language alang) {
179: opt.lang=alang;
180: opt.is_not_just_lang=0;
181: }
1.146 paf 182:
183: /// MUST be called exactly prior to modification of current [uses it's length]
1.187 misha 184: void append(size_t current, Language alang, size_t length) {
1.146 paf 185: assert(alang);
1.187 misha 186: assert(length);
1.146 paf 187:
1.147 paf 188: if(!opt.is_not_just_lang)
189: if(opt.lang) {
1.181 misha 190: if(opt.lang==alang) // same language? ignoring
1.146 paf 191: return;
192: } else {
1.147 paf 193: opt.lang=alang; // to uninitialized
1.146 paf 194: return;
1.145 paf 195: }
1.146 paf 196:
1.187 misha 197: append(current, CORD_chars((char)alang, length));
1.146 paf 198: }
199:
1.187 misha 200: void append(const Body ¤t, Language alang, size_t length) {
1.176 misha 201: assert(alang);
1.187 misha 202: assert(length);
1.176 misha 203:
204: if(!opt.is_not_just_lang)
205: if(opt.lang) {
1.183 misha 206: if(opt.lang==alang) // same language? ignoring
1.176 misha 207: return;
208: } else {
209: opt.lang=alang; // to uninitialized
210: return;
211: }
212:
1.187 misha 213: append(current, CORD_chars((char)alang, length));
1.176 misha 214: }
215:
1.187 misha 216: void appendHelper(const Body& current, Language alang, const Body &length_helper) {
217: assert(alang);
218:
219: if(!opt.is_not_just_lang)
220: if(opt.lang) {
221: if(opt.lang==alang) // same language? ignoring
222: return;
223: } else {
224: opt.lang=alang; // to uninitialized
225: return;
226: }
227:
228: append(current, CORD_chars((char)alang, length_helper.length()));
229: }
230:
231: void appendHelper(const Body& current, const Languages &src, const Body& length_helper) {
1.186 misha 232: if(!langs){
1.176 misha 233: langs=src.langs; // to uninitialized
1.186 misha 234: #ifdef CORD_CAT_OPTIMIZATION
235: if(opt.is_not_just_lang && !CORD_IS_STRING(langs))
236: CORD_concatenation_protect(langs);
237: #endif
238: }
1.176 misha 239: else if(!src.opt.is_not_just_lang)
240: appendHelper(current, src.opt.lang, length_helper); // simplifying when simple source
241: else
242: append(current, src.make_langs(length_helper));
243: }
244:
1.146 paf 245: /// MUST be called exactly prior to modification of current [uses it's length]
1.187 misha 246: void append(const Body& current, const Languages src, size_t aoffset, size_t alength) {
1.146 paf 247: assert(alength);
248:
249: if(!langs) // to uninitialized?
1.147 paf 250: if(src.opt.is_not_just_lang)
1.194 misha 251: langs=CORD_substr(src.langs, aoffset, alength, 0); // to uninitialized complex
1.146 paf 252: else
1.147 paf 253: opt.lang=src.opt.lang; // to uninitialized simple
1.146 paf 254: else
1.147 paf 255: 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 256: return; // ignoring
257: else
258: append(current, src.make_langs(aoffset, alength));
259: }
260:
261: /// checks if we have lang<=alang all from aoffset to aoffset+alength
262: bool check_lang(Language alang, size_t aoffset, size_t alength) const {
263: if(alang==L_UNSPECIFIED) // ignore lang?
264: return true;
265:
1.147 paf 266: if(opt.is_not_just_lang)
1.146 paf 267: return CORD_range_contains_chr_greater_then(langs, aoffset, alength, (unsigned)alang)==0;
268: else
1.169 misha 269: return (unsigned)opt.lang<=(unsigned)alang;
1.146 paf 270: }
271:
1.148 paf 272: /// @returns count of blocks
273: /// @todo currently there can be adjucent blocks of same language. someday merge them
274: size_t count() const {
275: return opt.is_not_just_lang?
276: CORD_block_count(langs)
277: : opt.lang?
278: 1
279: : 0;
280: };
281:
1.187 misha 282: template<typename I>
283: void for_each(size_t current, int callback(char, size_t, I), I info) const {
284: if(opt.is_not_just_lang)
285: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
286: else
287: callback(opt.lang, current, info);
288: }
289:
290: template<typename I>
291: void for_each(const Body& current, int callback(char, size_t, I), I info) const {
1.147 paf 292: if(opt.is_not_just_lang)
1.146 paf 293: CORD_block_iter(langs, 0, (CORD_block_iter_fn)callback, info);
294: else
1.187 misha 295: callback(opt.lang, current.length(), info);
1.146 paf 296: }
297:
1.164 paf 298: bool invariant(size_t current_length) const {
1.146 paf 299: if(!langs)
300: return current_length==0;
1.147 paf 301: if(opt.is_not_just_lang)
1.146 paf 302: return CORD_len(langs)==current_length;
303: return true; // uncheckable, actually
304: }
305: };
306:
307: class Body {
308:
309: CORD body;
310:
1.183 misha 311: #ifdef HASH_CODE_CACHING
312: // cached hash code is not reseted on write operations as test shows
313: // that string body does not change after it is stored as a hash key
314: mutable uint hash_code;
315: #endif
316:
1.187 misha 317: #ifdef STRING_LENGTH_CACHING
318: // cached length is reseted on modification, used only for char*, not CORD
319: mutable size_t string_length;
320: #define INIT_LENGTH ,string_length(0)
321: #define ZERO_LENGTH string_length=0;
322: #else
323: #define INIT_LENGTH
324: #define ZERO_LENGTH
325: #endif
326:
1.146 paf 327: public:
328:
329: const char* v() const;
1.164 paf 330: void dump() const;
1.146 paf 331:
1.183 misha 332: #ifdef HASH_CODE_CACHING
1.187 misha 333: Body(): body(CORD_EMPTY), hash_code(0) INIT_LENGTH {}
334: Body(CORD abody, uint ahash_code): body(abody), hash_code(ahash_code) INIT_LENGTH {}
335: Body(CORD abody): body(abody), hash_code(0) INIT_LENGTH {
1.183 misha 336: #else
1.187 misha 337: Body(): body(CORD_EMPTY) INIT_LENGTH {}
338: Body(CORD abody): body(abody) INIT_LENGTH {
1.183 misha 339: #endif
1.186 misha 340: #ifdef CORD_CAT_OPTIMIZATION
341: assert(!body // no body
342: || *body // ordinary string
343: || body[1]==1 // CONCAT_HDR
344: || body[1]==3 // CONCAT_HDR_READ_ONLY
345: || body[1]==4 // FN_HDR
346: || body[1]==6 // SUBSTR_HDR
347: );
348: #else
1.146 paf 349: assert(!body // no body
350: || *body // ordinary string
351: || body[1]==1 // CONCAT_HDR
352: || body[1]==4 // FN_HDR
353: || body[1]==6 // SUBSTR_HDR
354: );
1.186 misha 355: #endif
1.146 paf 356: }
1.181 misha 357:
1.146 paf 358: static Body Format(int value);
359:
1.187 misha 360: void clear() { ZERO_LENGTH body=CORD_EMPTY; }
1.146 paf 361:
362: bool operator! () const { return is_empty(); }
363:
1.183 misha 364: CORD get_cord() const { return body; }
365: uint get_hash_code() const;
1.146 paf 366:
1.197 misha 367: const char* cstr() const {
368: #ifdef STRING_LENGTH_CACHING
369: string_length = length();
370: if(string_length)
371: return const_cast<Body*>(this)->body=CORD_to_const_char_star(body, string_length);
372: #endif
373: return CORD_to_const_char_star(body, length());
374: }
375:
1.194 misha 376: char* cstrm() const { return CORD_to_char_star(body, length()); }
1.146 paf 377:
1.187 misha 378: #ifdef STRING_LENGTH_CACHING
379: void set_length(size_t alength){ string_length = alength; }
380: size_t length() const { return body ? CORD_IS_STRING(body) ? string_length ? string_length : (string_length=strlen(body)) : CORD_len(body) : 0; }
381: #else
1.146 paf 382: size_t length() const { return CORD_len(body); }
1.187 misha 383: #endif
1.146 paf 384:
385: bool is_empty() const { return body==CORD_EMPTY; }
386:
387: void append_know_length(const char *str, size_t known_length) {
1.187 misha 388: if(known_length){
389: if(body){
390: body = CORD_cat_char_star(body, str, known_length);
391: ZERO_LENGTH
392: } else {
393: body=str;
394: #ifdef STRING_LENGTH_CACHING
395: string_length=known_length;
396: #endif
397: }
398: }
1.146 paf 399: }
400: void append_strdup_know_length(const char* str, size_t known_length) {
401: if(known_length)
402: append_know_length(pa_strdup(str, known_length), known_length);
403: }
1.187 misha 404: void append(char c) { ZERO_LENGTH body=CORD_cat_char(body, c); }
405: Body& operator << (const Body src) { ZERO_LENGTH body=CORD_cat(body, src.body); return *this; }
406:
1.146 paf 407: Body& operator << (const char* str) { append_know_length(str, strlen(str)); return *this; }
408:
409: // could not figure out why this operator is needed [should do this chain: string->simple->==]
410: bool operator < (const Body src) const { return CORD_cmp(body, src.body)<0; }
411: bool operator > (const Body src) const { return CORD_cmp(body, src.body)>0; }
412: bool operator <= (const Body src) const { return CORD_cmp(body, src.body)<=0; }
413: bool operator >= (const Body src) const { return CORD_cmp(body, src.body)>=0; }
414: bool operator != (const Body src) const { return CORD_cmp(body, src.body)!=0; }
415: bool operator == (const Body src) const { return CORD_cmp(body, src.body)==0; }
416:
417: int ncmp(size_t x_begin, const Body y, size_t y_begin, size_t size) const {
418: return CORD_ncmp(body, x_begin, y.body, y_begin, size);
1.145 paf 419: }
420:
1.146 paf 421: char fetch(size_t index) const { return CORD_fetch(body, index); }
1.194 misha 422: Body mid(size_t aindex, size_t alength) const { return CORD_substr(body, aindex, alength, length()); }
423: size_t pos(const char* substr, size_t offset=0) const { return CORD_str(body, offset, substr, length()); }
1.146 paf 424: size_t pos(const Body substr, size_t offset=0) const {
1.180 misha 425: if(substr.is_empty())
1.146 paf 426: return STRING_NOT_FOUND; // in this case CORD_str returns 0 [parser users got used to -1]
1.150 paf 427:
428: // CORD_str checks for bad offset [CORD_chr does not]
1.194 misha 429: return CORD_str(body, offset, substr.body, length());
1.146 paf 430: }
431: size_t pos(char c,
432: size_t offset=0) const {
1.150 paf 433: if(offset>=length()) // CORD_chr does not check that [and ABORT's in that case]
434: return STRING_NOT_FOUND;
435:
1.146 paf 436: return CORD_chr(body, offset, c);
1.145 paf 437: }
1.146 paf 438:
1.203 misha 439: size_t strrpbrk(const char* chars, size_t left, size_t right) const;
440:
441: size_t rskipchars(const char* chars, size_t left, size_t right) const;
442:
1.187 misha 443: template<typename I>
444: int for_each(int (*f)(char c, I), I info) const {
1.159 paf 445: return CORD_iter(body, (CORD_iter_fn)f, (void*)info);
446: }
447:
1.187 misha 448: template<typename I>
449: int for_each(int (*f1)(char c, I), int (*f2)(const char* s, I), I info) const {
1.159 paf 450: return CORD_iter5(body, 0, (CORD_iter_fn)f1, (CORD_batched_iter_fn)f2, info);
1.148 paf 451: }
1.146 paf 452:
453: void set_pos(CORD_pos& pos, size_t index) const { CORD_set_pos(pos, body, index); }
454:
455: /*Body normalize() const {
456: return Body(CORD_balance(body));
457: }*/
1.157 paf 458:
459: /// @returns this or 0 or mid. if returns this or 0 out_* are not filled
460: Body trim(Trim_kind kind=TRIM_BOTH, const char* chars=0,
1.188 misha 461: size_t* out_start=0, size_t* out_length=0, Charset* source_charset=0) const;
1.145 paf 462: };
463:
464: struct C {
465: const char *str;
466: size_t length;
467: operator const char *() { return str; }
1.158 paf 468: C(): str(0), length(0) {}
1.145 paf 469: C(const char *astr, size_t asize): str(astr), length(asize) {}
470: };
471:
472: struct Cm {
473: char *str;
474: size_t length;
475: //operator char *() { return str; }
1.158 paf 476: Cm(): str(0), length(0) {}
1.145 paf 477: Cm(char *astr, size_t asize): str(astr), length(asize) {}
478: };
479:
480: private:
481:
1.152 paf 482: Body body; ///< all characters of string
1.146 paf 483: Languages langs; ///< string characters lang
484:
485: const char* v() const;
1.164 paf 486: void dump() const;
487: #define ASSERT_STRING_INVARIANT(string) \
488: assert((string).langs.invariant((string).body.length()))
1.145 paf 489:
1.8 paf 490: public:
1.151 paf 491:
492: static const String Empty;
1.8 paf 493:
1.181 misha 494: explicit String(){};
1.182 misha 495: explicit String(const char* cstr, Language alang=L_CLEAN){
1.181 misha 496: if(cstr && *cstr){
497: body=cstr;
1.182 misha 498: langs=alang;
1.181 misha 499: }
500: }
1.187 misha 501: explicit String(const char* cstr, Language alang, size_t alength){
502: if(cstr && *cstr){
503: body=cstr;
504: #ifdef STRING_LENGTH_CACHING
505: body.set_length(alength);
506: #endif
1.182 misha 507: langs=alang;
1.181 misha 508: }
509: }
1.187 misha 510:
1.185 misha 511: String(int value, char *format);
1.146 paf 512: String(Body abody, Language alang): body(abody), langs(alang) {
513: ASSERT_STRING_INVARIANT(*this);
514: }
515: String(const String& src): body(src.body), langs(src.langs) {
516: ASSERT_STRING_INVARIANT(*this);
1.145 paf 517: }
518:
519: /// for convinient hash lookup
1.183 misha 520: #ifdef HASH_CODE_CACHING
521: operator const Body&() const { return body; }
522: #else
1.146 paf 523: operator const Body() const { return body; }
1.183 misha 524: #endif
1.145 paf 525:
526: bool is_empty() const { return body.is_empty(); }
527: size_t length() const { return body.length(); }
1.171 misha 528: size_t length(Charset& charset) const;
1.145 paf 529:
1.191 misha 530: /// convert to CORD forcing lang tainting
531: Body cstr_to_string_body_taint(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const;
532: /// convert to CORD with tainting dirty to lang
533: Body cstr_to_string_body_untaint(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const;
1.145 paf 534:
1.189 misha 535: ///
536: const char* cstr() const {
537: return body.cstr();
538: }
539: ///
540: char* cstrm() const {
541: return body.cstrm();
542: }
543:
544: /// convert to constant C string forcing lang tainting
1.190 misha 545: const char* taint_cstr(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 546: return cstr_to_string_body_taint(lang, connection, charsets).cstr();
547: }
1.190 misha 548: char *taint_cstrm(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 549: return cstr_to_string_body_taint(lang, connection, charsets).cstrm();
550: }
551:
552: /// convert to constant C string with tainting dirty to lang
1.190 misha 553: const char* untaint_cstr(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 554: return cstr_to_string_body_untaint(lang, connection, charsets).cstr();
555: }
1.190 misha 556: char *untaint_cstrm(Language lang, SQL_Connection* connection=0, const Request_charsets *charsets=0) const {
1.189 misha 557: return cstr_to_string_body_untaint(lang, connection, charsets).cstrm();
558: }
1.195 misha 559:
1.196 misha 560: const char* untaint_and_transcode_cstr(Language lang, const Request_charsets *charsets) const;
1.195 misha 561:
1.202 moko 562: bool is_not_just_lang() const {
563: return langs.opt.is_not_just_lang !=0;
564: }
565:
566: Language just_lang() const {
567: return langs.opt.lang;
568: }
569:
1.108 parser 570: /// puts pieces to buf
1.145 paf 571: Cm serialize(size_t prolog_size) const;
1.108 parser 572: /// appends pieces from buf to self
1.145 paf 573: bool deserialize(size_t prolog_size, void *buf, size_t buf_size);
1.146 paf 574: /// @see Body::append_know_length
1.145 paf 575: String& append_know_length(const char* str, size_t known_length, Language lang);
1.146 paf 576: /// @see Body::append_help_length
1.145 paf 577: String& append_help_length(const char* str, size_t helper_length, Language lang);
578: String& append_strdup(const char* str, size_t helper_length, Language lang);
579:
1.146 paf 580: bool operator == (const char* y) const { return body==Body(y); }
581: bool operator != (const char* y) const { return body!=Body(y); }
1.145 paf 582:
583: /// this starts with y
584: bool starts_with(const char* y) const {
1.146 paf 585: return body.ncmp(0/*x_begin*/, Body(y), 0/*y_begin*/, strlen(y))==0;
1.145 paf 586: }
587: /// x starts with this
588: bool this_starts(const char* x) const {
1.146 paf 589: return Body(x).ncmp(0/*x_begin*/, body, 0/*y_begin*/, length())==0;
1.26 paf 590: }
591:
1.187 misha 592: String& append_to(String& dest, Language lang, bool forced=false) const;
1.145 paf 593: String& append(const String& src, Language lang, bool forced=false) {
594: return src.append_to(*this, lang, forced);
595: }
1.199 misha 596: String& append_quoted(const String* src, Language lang=L_JSON){
597: *this << "\"";
598: if(src)
599: this->append(*src, lang, true/*forced lang*/);
600: *this << "\"";
601: return *this;
602: }
603:
1.145 paf 604: String& operator << (const String& src) { return append(src, L_PASS_APPENDED); }
605: String& operator << (const char* src) { return append_help_length(src, 0, L_AS_IS); }
1.187 misha 606: String& operator << (const Body& src){
607: langs.appendHelper(body, L_AS_IS, src);
608: body<<src;
609: return *this;
610: }
1.100 parser 611:
1.142 paf 612: char first_char() const {
1.145 paf 613: return is_empty()?0:body.fetch(0);
1.142 paf 614: }
1.54 paf 615:
1.205 ! moko 616: char last_char() const {
! 617: return is_empty()?0:body.fetch(body.length()-1);
! 618: }
! 619:
1.145 paf 620: bool operator < (const String& src) const { return body<src.body; }
621: bool operator > (const String& src) const { return body>src.body; }
622: bool operator <= (const String& src) const { return body<=src.body; }
623: bool operator >= (const String& src) const { return body>=src.body; }
624: bool operator != (const String& src) const { return body!=src.body; }
625: bool operator == (const String& src) const { return body==src.body; }
626:
1.54 paf 627: /// extracts [start, finish) piece of string
1.145 paf 628: String& mid(size_t substr_begin, size_t substr_end) const;
1.172 misha 629: String& mid(Charset& charset, size_t from, size_t to, size_t helper_length=0) const;
1.145 paf 630:
631: /**
632: ignore lang if it's L_UNSPECIFIED
633: but when specified: look for substring that lies in ONE fragment in THAT lang
634: @return position of substr in string, -1 means "not found" [const char* version]
635: */
1.146 paf 636: size_t pos(const Body substr,
1.145 paf 637: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
638: /// String version of @see pos(const char*, int, Language)
639: size_t pos(const String& substr,
640: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
641: size_t pos(char c,
642: size_t this_offset=0) const {
643: return body.pos(c, this_offset);
644: }
1.173 misha 645: size_t pos(Charset& charset,
1.171 misha 646: const String& substr,
647: size_t this_offset=0, Language lang=L_UNSPECIFIED) const;
1.55 paf 648:
1.203 misha 649: size_t strrpbrk(const char* chars, size_t left=0) const {
650: return (length()) ? body.strrpbrk(chars, left, length()-1) : STRING_NOT_FOUND;
651: }
652: size_t strrpbrk(const char* chars, size_t left, size_t right) const {
653: return body.strrpbrk(chars, left, right);
654: }
655:
656: size_t rskipchars(const char* chars, size_t left=0) const {
657: return (length()) ? body.rskipchars(chars, left, length()-1) : STRING_NOT_FOUND;
658: }
659: size_t rskipchars(const char* chars, size_t left, size_t right) const {
660: return body.rskipchars(chars, left, right);
661: }
662:
1.145 paf 663: void split(ArrayString& result,
664: size_t& pos_after,
665: const char* delim,
666: Language lang=L_UNSPECIFIED, int limit=-1) const;
667: void split(ArrayString& result,
668: size_t& pos_after,
1.62 paf 669: const String& delim,
1.145 paf 670: Language lang=L_UNSPECIFIED, int limit=-1) const;
1.62 paf 671:
1.145 paf 672: typedef void (*Row_action)(Table& table, ArrayString* row,
1.136 paf 673: int prestart, int prefinish,
674: int poststart, int postfinish,
1.68 paf 675: void *info);
1.87 parser 676: /**
1.145 paf 677: @return table of found items, if any.
1.87 parser 678: table format is defined and fixed[can be used by others]:
679: @verbatim
680: prematch/match/postmatch/1/2/3/...
681: @endverbatim
682: */
1.178 misha 683: Table* match(VRegex* vregex,
1.99 parser 684: Row_action row_action, void *info,
1.168 misha 685: int& matches_count) const;
1.87 parser 686: enum Change_case_kind {
687: CC_UPPER,
688: CC_LOWER
689: };
1.145 paf 690: String& change_case(Charset& source_charset,
1.87 parser 691: Change_case_kind kind) const;
1.145 paf 692: const String& replace(const Dictionary& dict) const;
1.188 misha 693: const String& trim(Trim_kind kind=TRIM_BOTH, const char* chars=0, Charset* source_charset=0) const;
1.155 paf 694: double as_double() const { return pa_atod(cstr(), this); }
695: int as_int() const { return pa_atoi(cstr(), this); }
1.167 misha 696: bool as_bool() const { return as_int()!=0; }
1.174 misha 697: const String& escape(Charset& source_charset) const;
1.137 paf 698:
1.7 paf 699: private: //disabled
700:
1.12 paf 701: String& operator = (const String&) { return *this; }
1.7 paf 702:
1.1 paf 703: };
1.119 paf 704:
1.183 misha 705: #ifndef HASH_CODE_CACHING
1.145 paf 706: /// simple hash code of string. used by Hash
1.146 paf 707: inline uint hash_code(const String::Body self) {
1.184 misha 708: return self.get_hash_code();
1.119 paf 709: }
1.183 misha 710: #endif
1.147 paf 711:
712:
1.1 paf 713: #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