Annotation of parser3/src/main/pa_string.C, revision 1.174
1.45 paf 1: /** @file
1.174 ! paf 2: Parser: string class. @see untalength_t.C.
1.46 paf 3:
1.174 ! paf 4: Copyright (c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com)
1.138 paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.164 paf 6: */
1.46 paf 7:
1.174 ! paf 8: static const char* IDENT_STRING_C="$Date: 2003/07/22 08:20:24 $";
1.4 paf 9:
1.70 paf 10: #include "pcre.h"
11:
1.12 paf 12: #include "pa_string.h"
1.22 paf 13: #include "pa_exception.h"
1.61 paf 14: #include "pa_table.h"
1.101 parser 15: #include "pa_dictionary.h"
1.132 paf 16: #include "pa_charset.h"
1.60 paf 17:
1.174 ! paf 18: // helpers
1.139 paf 19:
1.174 ! paf 20: /// String::match uses this as replace & global search table columns
1.139 paf 21:
1.174 ! paf 22: const int MAX_MATCH_GROUPS=100;
1.139 paf 23:
1.174 ! paf 24: class String_match_table_template_columns: public ArrayString {
! 25: public:
! 26: String_match_table_template_columns() {
! 27: *this+=new String("prematch");
! 28: *this+=new String("match");
! 29: *this+=new String("postmatch");
! 30: for(int i=0; i<MAX_MATCH_GROUPS; i++) {
! 31: *this+=new String(StringBody::Format(1+i), String::L_CLEAN);
! 32: }
! 33: }
! 34: };
! 35:
! 36: Table string_match_table_template(new String_match_table_template_columns);
! 37:
! 38: // String::ArrayFragment methods
! 39:
! 40: void String::ArrayFragment::append_positions(const ArrayFragment& src,
! 41: size_t substr_begin, size_t substr_end) {
! 42: if(substr_begin==substr_end)
! 43: return;
! 44:
! 45: // FILE *err=fopen("append.log", "wt");
! 46:
! 47: size_t fragment_begin=0;
! 48: size_t fragment_end;
! 49: for(Array_iterator<element_type> i(src); ; fragment_begin=fragment_end) {
! 50: const Fragment fragment=i.next();
! 51: fragment_end=fragment_begin+fragment.length;
! 52: //fprintf(err, "1end=%u\n", fragment_end);fflush(err);
! 53:
! 54: // not reached fragments which may include 'substr'?
! 55: if(!(substr_begin>=fragment_begin && substr_begin<fragment_end))
! 56: continue;
! 57:
! 58: // found first fragment including piece of 'substr'
! 59: if(substr_end<=fragment_end) // fits into first fragment?
! 60: *this+=Fragment(fragment.lang, substr_end-substr_begin);
! 61: else { // spans more then one fragment
! 62: *this+=Fragment(fragment.lang, fragment_end-substr_begin);
! 63: while(true) {
! 64: const Fragment fragment=i.next();
! 65: fragment_end=(fragment_begin=fragment_end)+fragment.length;
! 66: //fprintf(err, "2end=%u\n", fragment_end);fflush(err);
! 67:
! 68: if(substr_end>fragment_end) // are there still more?
! 69: append(Fragment(fragment.lang, fragment.length)); // appending whole fragment
! 70: else { // no, it was last
! 71: append(Fragment(fragment.lang, substr_end-fragment_begin));
! 72: //fclose(err);
! 73: return;
! 74: }
! 75: }
! 76: }
! 77:
! 78: break;
! 79: }
! 80: //fclose(err);
1.1 paf 81: }
1.140 paf 82:
1.174 ! paf 83: // StringBody methods
1.140 paf 84:
1.174 ! paf 85: StringBody StringBody::Format(int value) {
! 86: char local[MAX_NUMBER];
! 87: size_t length=snprintf(local, MAX_NUMBER, "%d", value);
! 88: return StringBody(pa_strdup(local, length), length);
1.120 paf 89: }
90:
1.174 ! paf 91: static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
! 92: uint& result=*static_cast<uint*>(client_data);
! 93: generic_hash_code(result, c);
! 94: return 0;
! 95: }
! 96: static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
! 97: uint& result=*static_cast<uint*>(client_data);
! 98: generic_hash_code(result, s);
! 99: return 0;
! 100: };
! 101: uint StringBody::hash_code() const {
! 102: uint result=0;
! 103: CORD_iter5(body, 0,
! 104: CORD_batched_iter_fn_generic_hash_code,
! 105: CORD_batched_iter_fn_generic_hash_code, &result);
1.120 paf 106: return result;
1.94 parser 107: }
108:
1.174 ! paf 109: // String methods
! 110:
! 111: String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
! 112: append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.115 paf 113: }
1.174 ! paf 114: String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) {
! 115: append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN);
1.5 paf 116: }
1.28 paf 117:
1.174 ! paf 118: String::String(const String& src): body(src.body) {
! 119: fragments.append(src.fragments);
! 120: ASSERT_STRING_INVARIANT(*this);
! 121: }
1.139 paf 122:
1.174 ! paf 123: String& String::append_know_length(const char* str, size_t known_length, Language lang) {
! 124: if(!known_length)
1.9 paf 125: return *this;
1.122 paf 126:
1.174 ! paf 127: body.append_know_length(str, known_length);
! 128: fragments+=Fragment(lang, known_length);
1.1 paf 129:
1.174 ! paf 130: ASSERT_STRING_INVARIANT(*this);
1.1 paf 131: return *this;
132: }
1.174 ! paf 133: String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
! 134: if(!str)
! 135: return *this;
! 136: size_t known_length=helper_length?helper_length:strlen(str);
! 137: if(!known_length)
! 138: return *this;
1.1 paf 139:
1.174 ! paf 140: return append_know_length(str, known_length, lang);
1.5 paf 141: }
1.174 ! paf 142: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
! 143: size_t known_length=helper_length?helper_length:strlen(str);
! 144: if(!known_length)
! 145: return *this;
1.5 paf 146:
1.174 ! paf 147: body.append_strdup_know_length(str, known_length);
! 148: fragments+=Fragment(lang, known_length);
1.33 paf 149:
1.174 ! paf 150: ASSERT_STRING_INVARIANT(*this);
! 151: return *this;
1.5 paf 152: }
1.46 paf 153:
1.174 ! paf 154: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
! 155: String& String::mid(size_t substr_begin, size_t substr_end) const {
! 156: String& result=*new String;
! 157:
! 158: size_t self_length=length();
! 159: substr_begin=min(substr_begin, self_length);
! 160: substr_end=min(max(substr_end, substr_begin), self_length);
! 161: if(substr_begin==substr_end)
1.107 parser 162: return result;
1.53 paf 163:
1.174 ! paf 164: // first: letters themselves
! 165: result.body=body.mid(substr_begin, substr_end-substr_begin);
! 166:
! 167: // next: their langs
! 168: result.fragments.append_positions(fragments, substr_begin, substr_end);
! 169:
! 170: // SAPI::log("piece of '%s' from %d to %d is '%s'",
! 171: //cstr(), substr_begin, substr_end, result.cstr());
! 172: ASSERT_STRING_INVARIANT(result);
1.53 paf 173: return result;
1.54 paf 174: }
175:
1.174 ! paf 176: size_t String::pos(const StringBody substr,
! 177: size_t this_offset, Language lang) const {
! 178: // first: letters themselves
! 179: size_t substr_begin=body.pos(substr, this_offset);
! 180: if(substr_begin==CORD_NOT_FOUND)
! 181: return STRING_NOT_FOUND;
! 182:
! 183: // next: check the lang when specified
! 184:
! 185: if(lang==L_UNSPECIFIED) // ignore lang?
! 186: return substr_begin;
! 187:
! 188: // substr must be in one fragment, and fragments' lang must = lang
! 189: size_t substr_end=substr_begin+substr.length();
! 190: size_t fragment_begin=0;
! 191: size_t fragment_end;
! 192: for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
! 193: const Fragment fragment=i.next();
! 194: fragment_end=fragment_begin+fragment.length;
! 195:
! 196: if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
! 197: continue;
! 198: if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
! 199: continue;
! 200:
! 201: if(substr_end>fragment_end) // end of substr OUT of current fragment?
! 202: throw Exception(0, // (*) see below
! 203: this,
! 204: "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
! 205: substr.cstr(), this_offset);
! 206:
! 207: if(fragment.lang<=lang)
! 208: return substr_begin;
! 209: else { // bad lang...
! 210: /// WARNING: this possibly skips assert (*), but it's fast
! 211: substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
! 212: if(substr_begin==CORD_NOT_FOUND)
! 213: return STRING_NOT_FOUND;
! 214:
! 215: size_t substr_end=substr_begin+substr.length();
! 216: // and continuing with next fragment
! 217: }
! 218: }
! 219:
! 220: return STRING_NOT_FOUND;
1.58 paf 221: }
222:
1.174 ! paf 223: size_t String::pos(const String& substr,
! 224: size_t this_offset, Language lang) const {
! 225: return pos(substr.body, this_offset, lang);
1.60 paf 226: }
227:
1.174 ! paf 228: void String::split(ArrayString& result,
! 229: size_t& pos_after,
! 230: const char* delim,
! 231: Language lang, int limit) const {
! 232: size_t self_length=length();
! 233: if(size_t delim_length=strlen(delim)) {
1.60 paf 234: int pos_before;
235: // while we have 'delim'...
1.174 ! paf 236: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 237: result+=&mid(pos_after, pos_before);
1.174 ! paf 238: pos_after=pos_before+delim_length;
1.60 paf 239: }
240: // last piece
1.174 ! paf 241: if(pos_after<self_length && limit) {
! 242: result+=&mid(pos_after, self_length);
! 243: pos_after=self_length;
1.60 paf 244: }
245: } else { // empty delim
246: result+=this;
1.174 ! paf 247: pos_after+=self_length;
1.60 paf 248: }
249: }
250:
1.174 ! paf 251: void String::split(ArrayString& result,
! 252: size_t& pos_after,
! 253: const String& delim, Language lang,
! 254: int limit) const {
1.140 paf 255: if(!delim.is_empty()) {
1.60 paf 256: int pos_before;
257: // while we have 'delim'...
1.174 ! paf 258: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 259: result+=&mid(pos_after, pos_before);
1.174 ! paf 260: pos_after=pos_before+delim.length();
1.60 paf 261: }
262: // last piece
1.174 ! paf 263: if(pos_after<length() && limit) {
! 264: result+=&mid(pos_after, length());
! 265: pos_after=length();
1.60 paf 266: }
267: } else { // empty delim
268: result+=this;
1.174 ! paf 269: pos_after+=length();
1.60 paf 270: }
1.61 paf 271: }
272:
1.174 ! paf 273: static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 274: struct Regex_option {
1.174 ! paf 275: const char* keyL;
! 276: const char* keyU;
1.63 paf 277: int clear, set;
278: int *result;
1.154 paf 279: bool *flag;
1.63 paf 280: } regex_option[]={
1.153 paf 281: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
282: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
283: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
284: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
285: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 286: {"'", 0, 0, 0, 0, &need_pre_post_match},
287: {0}
1.63 paf 288: };
1.171 paf 289: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 290: result[1]=0;
291:
1.174 ! paf 292: if(options && !options->is_empty())
1.153 paf 293: for(Regex_option *o=regex_option; o->keyL; o++)
1.174 ! paf 294: if(options->pos(o->keyL)!=STRING_NOT_FOUND
! 295: || (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)) {
1.154 paf 296: if(o->flag)
297: *o->flag=true;
298: else { // result
299: *o->result &= ~o->clear;
300: *o->result |= o->set;
301: }
1.63 paf 302: }
303: }
304:
1.174 ! paf 305: Table* String::match(Charset& source_charset,
! 306: const String& regexp,
! 307: const String* options,
! 308: Row_action row_action, void *info,
! 309: bool& just_matched) const {
1.140 paf 310: if(regexp.is_empty())
1.149 paf 311: throw Exception(0,
1.174 ! paf 312: 0,
1.73 paf 313: "regexp is empty");
1.154 paf 314:
1.174 ! paf 315: const char* pattern=regexp.cstr();
! 316: const char* errptr;
1.62 paf 317: int erroffset;
1.173 paf 318: bool need_pre_post_match=false;
1.174 ! paf 319: int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
! 320: bool global=option_bits[1]!=0;
1.63 paf 321: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 322: &errptr, &erroffset,
1.174 ! paf 323: source_charset.pcre_tables);
1.62 paf 324:
1.67 paf 325: if(!code)
1.149 paf 326: throw Exception(0,
1.174 ! paf 327: ®exp.mid(erroffset, regexp.length()),
1.74 paf 328: "regular expression syntax error - %s", errptr);
1.62 paf 329:
1.174 ! paf 330: int subpatterns=pcre_info(code, 0, 0);
! 331: if(subpatterns<0) {
1.100 parser 332: pcre_free(code);
1.149 paf 333: throw Exception(0,
1.174 ! paf 334: ®exp,
1.76 paf 335: "pcre_info error (%d)",
1.174 ! paf 336: subpatterns);
1.63 paf 337: }
338:
1.174 ! paf 339: const char* subject=cstr();
! 340: size_t subject_length=strlen(subject);
! 341: const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
! 342: int ovector[oveclength];
1.155 paf 343:
344: // create table
1.173 paf 345: Table::Action_options table_options;
1.174 ! paf 346: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 347:
1.64 paf 348: int exec_option_bits=0;
1.154 paf 349: int prestart=0;
350: int poststart=0;
1.174 ! paf 351: int postfinish=length();
1.63 paf 352: while(true) {
353: int exec_substrings=pcre_exec(code, 0,
1.174 ! paf 354: subject, subject_length, prestart,
! 355: exec_option_bits, ovector, oveclength);
1.63 paf 356:
357: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 358: pcre_free(code);
1.174 ! paf 359: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
! 360: if(global || subpatterns)
! 361: return &table; // global or with subpatterns=true+result
! 362: else {
! 363: just_matched=false; return 0; // not global=no result
! 364: }
1.63 paf 365: }
366:
367: if(exec_substrings<0) {
1.100 parser 368: pcre_free(code);
1.149 paf 369: throw Exception(0,
1.174 ! paf 370: ®exp,
1.76 paf 371: "regular expression execute error (%d)",
1.63 paf 372: exec_substrings);
373: }
374:
1.154 paf 375: int prefinish=ovector[0];
376: poststart=ovector[1];
1.174 ! paf 377: ArrayString* row=new ArrayString;
! 378: if(need_pre_post_match) {
! 379: *row+=&mid(0, prefinish); // .prematch column value
! 380: *row+=&mid(prefinish, poststart); // .match
! 381: *row+=&mid(poststart, postfinish); // .postmatch
! 382: } else {
! 383: *row+=0; // .prematch column value
! 384: *row+=0; // .match
! 385: *row+=0; // .postmatch
! 386: }
1.63 paf 387:
388: for(int i=1; i<exec_substrings; i++) {
1.69 paf 389: // -1:-1 case handled peacefully by mid() itself
1.174 ! paf 390: *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 391: }
392:
1.174 ! paf 393: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 394:
1.174 ! paf 395: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 396: pcre_free(code);
1.174 ! paf 397: row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
! 398: return &table;
1.63 paf 399: }
1.154 paf 400: prestart=poststart;
1.63 paf 401:
402: /*
403: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 404: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 405: */
406: }
1.82 parser 407: }
408:
1.174 ! paf 409: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
! 410: String& result=*new String();
! 411: if(is_empty())
! 412: return result;
! 413:
! 414: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 415:
416: const unsigned char *a;
417: const unsigned char *b;
418: switch(kind) {
419: case CC_UPPER:
420: a=tables+lcc_offset;
421: b=tables+fcc_offset;
422: break;
423: case CC_LOWER:
424: a=tables+lcc_offset;
425: b=0;
426: break;
427: default:
1.149 paf 428: throw Exception(0,
1.82 parser 429: this,
430: "unknown change case kind #%d",
431: static_cast<int>(kind)); // never
432: a=b=0; // calm, compiler
433: break; // never
434: }
435:
1.174 ! paf 436: char* new_cstr=cstrm();
! 437: char *dest=new_cstr;
! 438: unsigned char index;
! 439: for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
! 440: unsigned char c=a[index];
! 441: if(b)
! 442: c=b[c];
1.82 parser 443:
1.174 ! paf 444: *dest++=(char)c;
! 445: }
! 446: result.body=new_cstr;
! 447: result.fragments.append(fragments);
1.89 parser 448:
1.101 parser 449: return result;
450: }
451:
1.174 ! paf 452: const String& String::replace(const Dictionary& dict) const {
! 453: String& result=*new String();
! 454: const char* old_cstr=cstr();
! 455: const char* prematch_begin=old_cstr;
! 456:
! 457: const char* current=old_cstr;
! 458: while(*current) {
! 459: if(Table::element_type row=dict.first_that_begins(current)) {
! 460: // prematch
! 461: if(size_t prematch_length=current-prematch_begin) {
! 462: result.body.append_strdup_know_length(prematch_begin, prematch_length);
! 463: result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 464: }
465:
1.174 ! paf 466: // match
1.170 paf 467:
1.174 ! paf 468: const String* a=row->get(0);
! 469: // skip 'a' in 'current'; move prematch_begin
! 470: current+=a->length(); prematch_begin=current;
! 471:
! 472: if(row->count()>1) { // are there any b?
! 473: const String* b=row->get(1);
! 474: result<<*b;
! 475: }
! 476: } else // simply advance
! 477: current++;
! 478: }
1.156 paf 479:
1.174 ! paf 480: // postmatch
! 481: if(size_t postmatch_length=current-prematch_begin) {
! 482: result.body.append_strdup_know_length(prematch_begin, postmatch_length);
! 483: result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
! 484: }
1.156 paf 485:
1.174 ! paf 486: ASSERT_STRING_INVARIANT(result);
1.89 parser 487: return result;
488: }
489:
1.90 parser 490: double String::as_double() const {
1.89 parser 491: double result;
1.174 ! paf 492: const char *str=cstr();
! 493:
! 494: while(*str && isspace(*str))
! 495: str++;
! 496: if(!*str)
1.162 paf 497: return 0;
1.161 paf 498:
1.102 parser 499: char *error_pos;
1.89 parser 500: // 0xABC
1.174 ! paf 501: if(str[0]=='0')
! 502: if(str[1]=='x' || str[1]=='X')
! 503: result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 504: else
1.174 ! paf 505: result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 506: else
1.174 ! paf 507: result=(double)strtod(str, &error_pos);
1.89 parser 508:
1.159 paf 509: while(char c=*error_pos++)
510: if(!isspace(c))
511: throw Exception("number.format",
512: this,
513: "invalid number (double)");
1.89 parser 514:
515: return result;
516: }
1.90 parser 517: int String::as_int() const {
1.89 parser 518: int result;
1.174 ! paf 519: const char *str=cstr();
! 520:
! 521: while(*str && isspace(*str))
! 522: str++;
! 523: if(!*str)
1.162 paf 524: return 0;
1.161 paf 525:
1.102 parser 526: char *error_pos;
1.89 parser 527: // 0xABC
1.174 ! paf 528: if(str[0]=='0')
! 529: if(str[1]=='x' || str[1]=='X')
! 530: result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 531: else
1.174 ! paf 532: result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 533: else
1.174 ! paf 534: result=(int)strtol(str, &error_pos, 0);
1.89 parser 535:
1.159 paf 536: while(char c=*error_pos++)
537: if(!isspace(c))
538: throw Exception("number.format",
539: this,
540: "invalid number (int)");
1.82 parser 541:
542: return result;
1.61 paf 543: }
1.113 parser 544:
1.174 ! paf 545: inline void uint2uchars(uint word, uchar *bytes) {
! 546: bytes[0]=word&0xFF;
! 547: bytes[1]=(word>>8)&0xFF;
! 548: bytes[2]=(word>>16)&0xFF;
! 549: bytes[3]=(word>>24)&0xFF;
! 550: }
! 551: inline uint uchars2uint(uchar *bytes) {
! 552: return bytes[3]<<24
! 553: | bytes[2]<<16
! 554: | bytes[1]<<8
! 555: | bytes[0];
! 556: }
! 557:
! 558: static int serialize_body_piece(const char* s, char** cur) {
! 559: size_t length=strlen(s);
! 560: memcpy(*cur, s, length); *cur+=length;
! 561: return 0;
! 562: };
! 563: String::Cm String::serialize(size_t prolog_length) const {
! 564: size_t buf_length=
! 565: prolog_length
! 566: +fragments.count()*(sizeof(Language)+sizeof(size_t))
! 567: +length();
! 568: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
! 569:
! 570: // 1: prolog
! 571: char *cur=result.str+prolog_length;
! 572:
! 573:
! 574: // 2: fragments.count
! 575: size_t fragments_count=fragments.count();
! 576: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
! 577:
! 578: // 3: lang info
! 579: for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
! 580: const Fragment fragment=i.next();
! 581: // lang
! 582: memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
! 583: // length
! 584: memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
! 585: }
! 586:
! 587: // 4: letters
! 588: body.for_each(serialize_body_piece, &cur);
1.113 parser 589:
1.174 ! paf 590: return result;
1.113 parser 591: }
1.174 ! paf 592: bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
! 593: if(buf_length<=prolog_length)
1.148 paf 594: return false;
1.174 ! paf 595: buf_length-=prolog_length;
1.135 paf 596:
1.174 ! paf 597: // 1: prolog
! 598: const char* cur=(const char* )buf+prolog_length;
1.113 parser 599:
1.174 ! paf 600: // 2: fragments.count
! 601: if(buf_length<sizeof(size_t)) // fragments.count don't fit?
! 602: return false;
! 603: size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
! 604: buf_length-=sizeof(size_t);
! 605:
! 606: if(fragments_count) {
! 607: // 3: lang info
! 608: size_t total_length=0;
! 609: for(size_t f=0; f<fragments_count; f++) {
! 610: size_t piece_length=sizeof(Language)+sizeof(size_t);
! 611: if(buf_length<piece_length) // lang+length
! 612: return false;
! 613:
! 614: Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
! 615: size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
! 616: fragments+=Fragment(lang, fragment_length);
! 617: total_length+=fragment_length;
1.148 paf 618:
1.174 ! paf 619: buf_length-=piece_length;
! 620: }
1.128 paf 621:
1.174 ! paf 622: // 4: letters
! 623: if(buf_length!=total_length)
1.148 paf 624: return false;
625:
1.174 ! paf 626: body=StringBody(cur, buf_length);
! 627: }
1.113 parser 628:
1.174 ! paf 629: ASSERT_STRING_INVARIANT(*this);
1.148 paf 630: return true;
1.113 parser 631: }
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_string.C CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / main