Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.24
1.45 paf 1: /** @file
1.172.2.21.2. (paf 2:): Parser: string class. @see untalength_t.C.
1.46 paf 3:
1.172.2.11 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:
1.172.2.21.2. 4(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/03/24 15:26:03 $";
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:
1.172.2.2 paf 18: // helpers
1.139 paf 19:
1.172.2.2 paf 20: /// String::match uses this as replace & global search table columns
1.139 paf 21:
1.172.2.4 paf 22: const int MAX_MATCH_GROUPS=100;
23:
1.172.2.14 paf 24: class String_match_table_template_columns: public ArrayString {
1.172.2.2 paf 25: public:
1.172.2.4 paf 26: String_match_table_template_columns() {
1.172.2.21.2. (paf 27:): *this+=new String("prematch");
28:): *this+=new String("match");
29:): *this+=new String("postmatch");
1.172.2.4 paf 30: for(int i=0; i<MAX_MATCH_GROUPS; i++) {
1.172.2.21.2. (paf 31:): char *cname=new(PointerFreeGC) char[3/*strlen("100")*/+1/*terminating 0*/];
32:): *this+=new String(cname, sprintf(cname, "%d", 1+i));
1.172.2.2 paf 33: }
1.172.2.21 paf 34: }
1.172.2.4 paf 35: };
36:
1.172.2.21.2. (paf 37:): Table string_match_table_template(new String_match_table_template_columns);
1.172.2.2 paf 38:
1.172.2.21.2. (paf 39:): // String::ArrayFragment methods
40:):
3(paf 41:3): void String::ArrayFragment::append_positions(const ArrayFragment& src,
42:3): size_t substr_begin, size_t substr_end) {
1(paf 43:3): if(substr_begin==substr_end)
3(paf 44:3): return;
(paf 45:):
1(paf 46:3): size_t fragment_begin=0;
(paf 47:): size_t fragment_end;
1(paf 48:3): for(Array_iterator<element_type> i(src); ; fragment_begin=fragment_end) {
(paf 49:): const element_type& fragment=i.next();
1(paf 50:3): fragment_end=fragment_begin+fragment.length;
(paf 51:):
1(paf 52:3): if(fragment_begin<substr_begin) // not reached fragments which may include 'substr'?
(paf 53:): continue;
54:):
55:): // found first fragment including piece of 'substr'
56:): size_t piece_end=min(substr_end, fragment_end);
1(paf 57:3): *this+=Fragment(fragment.lang, piece_end-substr_begin);
(paf 58:):
59:): while(substr_end>fragment_end) { // are there more fragments including pieces of 'substr'?
1(paf 60:3): fragment_begin=fragment_end;
61:3):
(paf 62:): const element_type& fragment=i.next();
1(paf 63:3): fragment_end=fragment_begin+fragment.length;
(paf 64:):
65:): if(substr_end>fragment_end) // are there still more?
3(paf 66:3): append(Fragment(fragment.lang, fragment.length)); // appending whole fragment
(paf 67:): else { // no, it was last
3(paf 68:3): append(Fragment(fragment.lang, substr_end-fragment_begin));
69:3): return;
(paf 70:): }
71:): }
72:):
73:): break;
74:): }
75:): }
76:): /*
1(paf 77:3): void String::ArrayFragment::mid(ArrayFragment& result, size_t substr_begin, size_t substr_end) {
(paf 78:): }
79:): */
80:):
5(paf 81:3): // StringSimple methods
82:3):
83:3):
3(paf 84:3): static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
85:3): uint& result=*static_cast<uint*>(client_data);
86:3): generic_hash_code(result, c);
87:3): return 0;
88:3): }
5(paf 89:3): static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
90:3): uint& result=*static_cast<uint*>(client_data);
91:3): generic_hash_code(result, s);
92:3): return 0;
93:3): };
7(paf 94:3): uint StringBody::hash_code() const {
5(paf 95:3): uint result=0;
3(paf 96:3): CORD_iter5(body, 0,
97:3): CORD_batched_iter_fn_generic_hash_code,
98:3): CORD_batched_iter_fn_generic_hash_code, &result);
5(paf 99:3): return result;
100:3): }
101:3):
(paf 102:): // String methods
1.172.2.2 paf 103:
1.172.2.21.2. (paf 104:): String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
5(paf 105:3): append(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.1 paf 106: }
1.140 paf 107:
1.172.2.21.2. (paf 108:): String::String(const String& src): body(src.body), fragments(src.fragments) {}
1.28 paf 109:
1.172.2.21.2. 6(paf 110:3): String& String::append(const char* str, size_t length, Language lang) {
3(paf 111:3): if(!str)
1.9 paf 112: return *this;
1.172.2.21.2. 6(paf 113:3): size_t use_length=length?length:strlen(str);
0(paf 114:3): if(!use_length)
1.9 paf 115: return *this;
1.122 paf 116:
1.172.2.21.2. 7(paf 117:3): body.append(str, use_length);
0(paf 118:3): fragments+=Fragment(lang, use_length);
1.1 paf 119:
120: return *this;
121: }
122:
1.172.2.21.2. 1(paf 123:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 124:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 125:): String& result=*new String;
1.33 paf 126:
1.172.2.21.2. 1(paf 127:3): size_t self_length=length();
128:3): substr_begin=max(min(substr_begin, self_length), (size_t)0);
129:3): substr_end=min(max(substr_end, substr_begin), self_length);
130:3): if(substr_begin==substr_end)
(paf 131:): return result;
1.52 paf 132:
1.172.2.21.2. (paf 133:): // first: letters themselves
7(paf 134:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 135:
1.172.2.21.2. (paf 136:): // next: their langs
3(paf 137:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 138:
1.172.2.21.2. 5(paf 139:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 140:3): //cstr(), substr_begin, substr_end, result.cstr());
1.53 paf 141: return result;
1.54 paf 142: }
143:
1.172.2.21.2. 7(paf 144:3): size_t String::pos(const StringBody substr,
5(paf 145:3): size_t this_offset, Language lang) const {
(paf 146:): // first: letters themselves
2(paf 147:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 148:3): if(substr_begin==CORD_NOT_FOUND)
(paf 149:): return STRING_NOT_FOUND;
150:):
151:): // next: check the lang when specified
152:):
5(paf 153:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 154:3): return substr_begin;
(paf 155:):
156:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 157:3): size_t substr_end=substr_begin+substr.length();
1(paf 158:3): size_t fragment_begin=0;
(paf 159:): size_t fragment_end;
1(paf 160:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
(paf 161:): const Fragment& fragment=i.next();
1(paf 162:3): fragment_end=fragment_begin+fragment.length;
(paf 163:):
1(paf 164:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 165:): continue;
2(paf 166:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
167:3): continue;
(paf 168:):
169:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 170:3): throw Exception(0, // (*) see below
171:3): this,
172:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 173:3): substr.cstr(), this_offset);
(paf 174:):
4(paf 175:3): if(fragment.lang<=lang)
2(paf 176:3): return substr_begin;
177:3): else { // bad lang...
178:3): /// WARNING: this possibly skips assert (*), but it's fast
179:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
180:3): if(substr_begin==CORD_NOT_FOUND)
181:3): return STRING_NOT_FOUND;
182:3):
7(paf 183:3): size_t substr_end=substr_begin+substr.length();
2(paf 184:3): // and continuing with next fragment
185:3): }
(paf 186:): }
187:):
188:): return STRING_NOT_FOUND;
1.58 paf 189: }
190:
1.172.2.21.2. (paf 191:): size_t String::pos(const String& substr,
5(paf 192:3): size_t this_offset, Language lang) const {
(paf 193:): return pos(substr.body, this_offset, lang);
1.60 paf 194: }
195:
1.172.2.14 paf 196: void String::split(ArrayString& result,
1.172.2.21.2. (paf 197:): size_t& pos_after,
198:): const char* delim,
2(paf 199:3): Language lang, int limit) const {
(paf 200:): size_t self_length=length();
201:): if(size_t delim_length=strlen(delim)) {
1.60 paf 202: int pos_before;
203: // while we have 'delim'...
1.172.2.21.2. (paf 204:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
205:): result+=&mid(pos_after, pos_before);
206:): pos_after=pos_before+delim_length;
1.60 paf 207: }
208: // last piece
1.172.2.21.2. (paf 209:): if(pos_after<self_length && limit) {
210:): result+=&mid(pos_after, self_length);
211:): pos_after=self_length;
1.60 paf 212: }
213: } else { // empty delim
1.172.2.21.2. (paf 214:): result+=this;
215:): pos_after+=self_length;
1.60 paf 216: }
217: }
218:
1.172.2.14 paf 219: void String::split(ArrayString& result,
1.172.2.21.2. (paf 220:): size_t& pos_after,
5(paf 221:3): const String& delim, Language lang,
(paf 222:): int limit) const {
5(paf 223:3): if(!delim.is_empty()) {
1.60 paf 224: int pos_before;
225: // while we have 'delim'...
226: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 227:): result+=&mid(pos_after, pos_before);
228:): pos_after=pos_before+delim.length();
1.60 paf 229: }
230: // last piece
1.172.2.21.2. (paf 231:): if(pos_after<length() && limit) {
232:): result+=&mid(pos_after, length());
233:): pos_after=length();
1.60 paf 234: }
235: } else { // empty delim
1.172.2.21.2. (paf 236:): result+=this;
237:): pos_after+=length();
1.60 paf 238: }
1.61 paf 239: }
240:
1.172.2.21.2. 2(paf 241:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 242: struct Regex_option {
1.172.2.11 paf 243: const char* keyL;
244: const char* keyU;
1.63 paf 245: int clear, set;
246: int *result;
1.154 paf 247: bool *flag;
1.63 paf 248: } regex_option[]={
1.153 paf 249: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
250: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
251: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
252: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
253: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 254: {"'", 0, 0, 0, 0, &need_pre_post_match},
255: {0}
1.63 paf 256: };
1.171 paf 257: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 258: result[1]=0;
259:
1.172.2.21.2. 2(paf 260:3): if(options && !options->is_empty())
1.153 paf 261: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 262:3): if(options->pos(o->keyL)>=0
263:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 264: if(o->flag)
265: *o->flag=true;
266: else { // result
267: *o->result &= ~o->clear;
268: *o->result |= o->set;
269: }
1.63 paf 270: }
271: }
272:
1.172.2.21.2. (paf 273:): Table* String::match(Charset& source_charset,
1.172.2.4 paf 274: const String& regexp,
1.172.2.21.2. 2(paf 275:3): const String* options,
1.172.2.4 paf 276: Row_action row_action, void *info,
1.172.2.20 paf 277: bool& just_matched) const {
1.172.2.21.2. 5(paf 278:3): if(regexp.is_empty())
(paf 279:): throw Exception(0,
280:): 0,
1.73 paf 281: "regexp is empty");
1.154 paf 282:
1.172.2.21.2. (paf 283:): const char* pattern=regexp.cstr();
1.172.2.11 paf 284: const char* errptr;
1.62 paf 285: int erroffset;
1.172.2.21.2. (paf 286:): bool need_pre_post_match=false;
1.154 paf 287: int option_bits[2]; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 288: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 289:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 290: &errptr, &erroffset,
1.172.2.1 paf 291: source_charset.pcre_tables);
1.62 paf 292:
1.67 paf 293: if(!code)
1.172.2.21.2. (paf 294:): throw Exception(0,
295:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 296: "regular expression syntax error - %s", errptr);
1.62 paf 297:
1.172.2.20 paf 298: int subpatterns=pcre_info(code, 0, 0);
299: if(subpatterns<0) {
1.100 parser 300: pcre_free(code);
1.149 paf 301: throw Exception(0,
1.172.2.21.2. (paf 302:): ®exp,
1.76 paf 303: "pcre_info error (%d)",
1.172.2.20 paf 304: subpatterns);
1.63 paf 305: }
306:
1.172.2.21.2. (paf 307:): const char* subject=cstr();
308:): size_t subject_length=strlen(subject);
309:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
310:): int ovector[oveclength];
1.155 paf 311:
312: // create table
1.172.2.21.2. 1(paf 313:3): Table& table=*new Table(string_match_table_template);
1.63 paf 314:
1.64 paf 315: int exec_option_bits=0;
1.154 paf 316: int prestart=0;
317: int poststart=0;
1.172.2.21.2. (paf 318:): int postfinish=length();
1.63 paf 319: while(true) {
320: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 321:): subject, subject_length, prestart,
322:): exec_option_bits, ovector, oveclength);
1.63 paf 323:
324: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 325: pcre_free(code);
1.172.2.21.2. (paf 326:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 327: if(global || subpatterns)
1.172.2.21.2. 2(paf 328:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 329: else {
1.172.2.21.2. (paf 330:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 331: }
1.63 paf 332: }
333:
334: if(exec_substrings<0) {
1.100 parser 335: pcre_free(code);
1.172.2.21.2. (paf 336:): throw Exception(0,
337:): ®exp,
1.76 paf 338: "regular expression execute error (%d)",
1.63 paf 339: exec_substrings);
340: }
341:
1.154 paf 342: int prefinish=ovector[0];
343: poststart=ovector[1];
1.172.2.21.2. (paf 344:): ArrayString* row=new ArrayString;
1.172.2.4 paf 345: if(need_pre_post_match) {
1.172.2.21.2. (paf 346:): *row+=&mid(0, prefinish); // .prematch column value
347:): *row+=&mid(prefinish, poststart); // .match
348:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 349: } else {
1.172.2.21.2. (paf 350:): *row+=0; // .prematch column value
351:): *row+=0; // .match
352:): *row+=0; // .postmatch
1.172.2.4 paf 353: }
1.63 paf 354:
355: for(int i=1; i<exec_substrings; i++) {
1.69 paf 356: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 357:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 358: }
359:
1.172.2.20 paf 360: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 361:
1.172.2.20 paf 362: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 363: pcre_free(code);
1.172.2.21.2. (paf 364:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 365:3): return &table;
1.63 paf 366: }
1.154 paf 367: prestart=poststart;
1.63 paf 368:
369: /*
370: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 371: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 372: */
373: }
1.82 parser 374: }
375:
1.172.2.21.2. 0(paf 376:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 377:): String& result=*new String();
1.172.2.4 paf 378:
1.172.2.1 paf 379: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 380:
381: const unsigned char *a;
382: const unsigned char *b;
383: switch(kind) {
384: case CC_UPPER:
385: a=tables+lcc_offset;
386: b=tables+fcc_offset;
387: break;
388: case CC_LOWER:
389: a=tables+lcc_offset;
390: b=0;
391: break;
392: default:
1.172.2.21.2. (paf 393:): throw Exception(0,
394:): this,
1.82 parser 395: "unknown change case kind #%d",
396: static_cast<int>(kind)); // never
397: a=b=0; // calm, compiler
398: break; // never
399: }
400:
1.172.2.21.2. (paf 401:): char* new_cstr=cstrm();
402:): char *dest=new_cstr;
403:): unsigned char index;
404:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
405:): unsigned char c=a[index];
406:): if(b)
407:): c=b[c];
1.82 parser 408:
1.172.2.21.2. (paf 409:): *dest++=(char)c;
410:): }
411:): result.body=new_cstr;
3(paf 412:3): result.fragments.append(fragments);
1.89 parser 413:
1.101 parser 414: return result;
415: }
416:
1.172.2.21.2. (paf 417:): const String& String::replace(const Dictionary& dict) const {
418:): String& result=*new String();
419:): const char* old_cstr=cstr();
1(paf 420:3): const char* prematch_begin=old_cstr;
(paf 421:):
422:): for(const char* current=old_cstr; *current; ) {
1(paf 423:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 424:): // prematch
1(paf 425:3): if(size_t prematch_length=current-prematch_begin) {
7(paf 426:3): result.body.append(prematch_begin, prematch_length);
4(paf 427:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 428: }
429:
1.172.2.21.2. (paf 430:): // match
431:):
432:): const String* a=row->get(0);
1(paf 433:3): // skip 'a' in 'current'; move prematch_begin
434:3): current+=a->length(); prematch_begin=current;
1.170 paf 435:
1.172.2.21.2. (paf 436:): if(row->count()>1) { // are there any b?
437:): const String* b=row->get(1);
438:): if(size_t b_length=b->length()) {
7(paf 439:3): result.body.append(b->cstr_to_string_body());
3(paf 440:3): result.fragments.append(b->fragments);
(paf 441:): }
442:): }
443:): } else // simply advance
444:): current++;
445:): }
1.156 paf 446:
1.172.2.21.2. (paf 447:): // postmatch
3(paf 448:3): if(size_t postmatch_length=current-prematch_begin) {
7(paf 449:3): result.body.append(prematch_begin, postmatch_length);
4(paf 450:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 451:): }
1.156 paf 452:
1.89 parser 453: return result;
454: }
455:
1.172.2.21.2. (paf 456:): double String::as_double() const {
1.89 parser 457: double result;
1.172.2.21.2. (paf 458:): const char *str=cstr();
459:):
460:): while(*str && isspace(*str))
461:): str++;
462:): if(!*str)
1.162 paf 463: return 0;
1.161 paf 464:
1.102 parser 465: char *error_pos;
1.89 parser 466: // 0xABC
1.172.2.21.2. (paf 467:): if(str[0]=='0')
468:): if(str[1]=='x' || str[1]=='X')
469:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 470: else
1.172.2.21.2. (paf 471:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 472: else
1.172.2.21.2. (paf 473:): result=(double)strtod(str, &error_pos);
1.89 parser 474:
1.159 paf 475: while(char c=*error_pos++)
476: if(!isspace(c))
477: throw Exception("number.format",
1.172.2.21.2. (paf 478:): this,
1.159 paf 479: "invalid number (double)");
1.89 parser 480:
481: return result;
482: }
1.172.2.21.2. (paf 483:): int String::as_int() const {
1.89 parser 484: int result;
1.172.2.21.2. (paf 485:): const char *str=cstr();
486:):
487:): while(*str && isspace(*str))
488:): str++;
489:): if(!*str)
1.162 paf 490: return 0;
1.161 paf 491:
1.102 parser 492: char *error_pos;
1.89 parser 493: // 0xABC
1.172.2.21.2. (paf 494:): if(str[0]=='0')
495:): if(str[1]=='x' || str[1]=='X')
496:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 497: else
1.172.2.21.2. (paf 498:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 499: else
1.172.2.21.2. (paf 500:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 501:
1.159 paf 502: while(char c=*error_pos++)
503: if(!isspace(c))
504: throw Exception("number.format",
1.172.2.21.2. (paf 505:): this,
1.159 paf 506: "invalid number (int)");
1.82 parser 507:
508: return result;
1.61 paf 509: }
1.113 parser 510:
1.172.2.4 paf 511: inline void uint2uchars(uint word, uchar *bytes) {
512: bytes[0]=word&0xFF;
513: bytes[1]=(word>>8)&0xFF;
514: bytes[2]=(word>>16)&0xFF;
515: bytes[3]=(word>>24)&0xFF;
516: }
517: inline uint uchars2uint(uchar *bytes) {
518: return bytes[3]<<24
519: | bytes[2]<<16
520: | bytes[1]<<8
521: | bytes[0];
522: }
523:
1.172.2.21.2. 7(paf 524:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 525:): size_t length=strlen(s);
7(paf 526:3): memcpy(*cur, s, length); *cur+=length;
(paf 527:): return 0;
528:): };
529:): String::C String::serialize(size_t prolog_length) const {
530:): size_t buf_length=
531:): prolog_length
5(paf 532:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 533:): +length();
534:): C result(new(PointerFreeGC) char[buf_length], buf_length);
535:):
536:): // 1: prolog
537:): char *cur=result.str+prolog_length;
538:):
539:):
540:): // 2: fragments.count
541:): size_t fragments_count=fragments.count();
542:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 543:
1.172.2.21.2. (paf 544:): // 3: lang info
545:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
546:): const Fragment& fragment=i.next();
1.123 paf 547: // lang
1.172.2.21.2. (paf 548:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
549:): // length
550:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
551:): }
552:):
553:): // 4: letters
7(paf 554:3): body.for_each(serialize_body_piece, &cur);
(paf 555:):
556:): return result;
1.113 parser 557: }
1.172.2.21.2. 9(paf 558:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 559:): if(buf_length<=prolog_length)
1.148 paf 560: return false;
1.172.2.21.2. (paf 561:): buf_length-=prolog_length;
1.113 parser 562:
1.172.2.21.2. (paf 563:): // 1: prolog
564:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 565:
1.172.2.21.2. (paf 566:): // 2: fragments.count
567:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
568:): return false;
569:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
570:): buf_length-=sizeof(size_t);
1.128 paf 571:
1.172.2.21.2. (paf 572:): // 3: lang info
573:): size_t total_length=0;
574:): for(size_t f=0; f<fragments_count; f++) {
5(paf 575:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
(paf 576:): if(buf_length<piece_length) // lang+length
1.148 paf 577: return false;
578:
1.172.2.21.2. 5(paf 579:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
(paf 580:): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
581:): fragments+=Fragment(lang, fragment_length);
582:): total_length+=fragment_length;
1.113 parser 583:
1.172.2.21.2. (paf 584:): buf_length-=piece_length;
1.113 parser 585: }
1.172.2.21.2. (paf 586:):
587:): // 4: letters
588:): if(buf_length!=total_length)
589:): return false;
590:):
591:): body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length);
592:):
1.148 paf 593: return true;
1.113 parser 594: }
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