Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.28
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. 8(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/03/24 16:43:27 $";
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:):
5(paf 52:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'substr'?
(paf 53:): continue;
54:):
6(paf 55:3): if(substr_begin>fragment_end) // not reached fragments which may include 'substr'?
56:3): continue;
57:3):
(paf 58:): // found first fragment including piece of 'substr'
59:): size_t piece_end=min(substr_end, fragment_end);
1(paf 60:3): *this+=Fragment(fragment.lang, piece_end-substr_begin);
(paf 61:):
62:): while(substr_end>fragment_end) { // are there more fragments including pieces of 'substr'?
1(paf 63:3): fragment_begin=fragment_end;
64:3):
(paf 65:): const element_type& fragment=i.next();
1(paf 66:3): fragment_end=fragment_begin+fragment.length;
(paf 67:):
68:): if(substr_end>fragment_end) // are there still more?
3(paf 69:3): append(Fragment(fragment.lang, fragment.length)); // appending whole fragment
(paf 70:): else { // no, it was last
3(paf 71:3): append(Fragment(fragment.lang, substr_end-fragment_begin));
72:3): return;
(paf 73:): }
74:): }
75:):
76:): break;
77:): }
78:): }
79:):
8(paf 80:3): // StringBody methods
5(paf 81:3):
8(paf 82:3): StringBody::StringBody(int value) {
83:3): char local[MAX_NUMBER];
84:3): body=pa_strdup(local, snprintf(local, MAX_NUMBER, "%d", value));
85:3): }
5(paf 86:3):
3(paf 87:3): static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
88:3): uint& result=*static_cast<uint*>(client_data);
89:3): generic_hash_code(result, c);
90:3): return 0;
91:3): }
5(paf 92:3): static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
93:3): uint& result=*static_cast<uint*>(client_data);
94:3): generic_hash_code(result, s);
95:3): return 0;
96:3): };
7(paf 97:3): uint StringBody::hash_code() const {
5(paf 98:3): uint result=0;
3(paf 99:3): CORD_iter5(body, 0,
100:3): CORD_batched_iter_fn_generic_hash_code,
101:3): CORD_batched_iter_fn_generic_hash_code, &result);
5(paf 102:3): return result;
103:3): }
104:3):
(paf 105:): // String methods
1.172.2.2 paf 106:
1.172.2.21.2. (paf 107:): String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
5(paf 108:3): append(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.1 paf 109: }
1.140 paf 110:
1.172.2.21.2. (paf 111:): String::String(const String& src): body(src.body), fragments(src.fragments) {}
1.28 paf 112:
1.172.2.21.2. 6(paf 113:3): String& String::append(const char* str, size_t length, Language lang) {
3(paf 114:3): if(!str)
1.9 paf 115: return *this;
1.172.2.21.2. 6(paf 116:3): size_t use_length=length?length:strlen(str);
0(paf 117:3): if(!use_length)
1.9 paf 118: return *this;
1.122 paf 119:
1.172.2.21.2. 7(paf 120:3): body.append(str, use_length);
0(paf 121:3): fragments+=Fragment(lang, use_length);
1.1 paf 122:
123: return *this;
124: }
125:
1.172.2.21.2. 1(paf 126:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 127:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 128:): String& result=*new String;
1.33 paf 129:
1.172.2.21.2. 1(paf 130:3): size_t self_length=length();
131:3): substr_begin=max(min(substr_begin, self_length), (size_t)0);
132:3): substr_end=min(max(substr_end, substr_begin), self_length);
133:3): if(substr_begin==substr_end)
(paf 134:): return result;
1.52 paf 135:
1.172.2.21.2. (paf 136:): // first: letters themselves
7(paf 137:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 138:
1.172.2.21.2. (paf 139:): // next: their langs
3(paf 140:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 141:
1.172.2.21.2. 5(paf 142:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 143:3): //cstr(), substr_begin, substr_end, result.cstr());
1.53 paf 144: return result;
1.54 paf 145: }
146:
1.172.2.21.2. 7(paf 147:3): size_t String::pos(const StringBody substr,
5(paf 148:3): size_t this_offset, Language lang) const {
(paf 149:): // first: letters themselves
2(paf 150:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 151:3): if(substr_begin==CORD_NOT_FOUND)
(paf 152:): return STRING_NOT_FOUND;
153:):
154:): // next: check the lang when specified
155:):
5(paf 156:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 157:3): return substr_begin;
(paf 158:):
159:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 160:3): size_t substr_end=substr_begin+substr.length();
1(paf 161:3): size_t fragment_begin=0;
(paf 162:): size_t fragment_end;
1(paf 163:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
(paf 164:): const Fragment& fragment=i.next();
1(paf 165:3): fragment_end=fragment_begin+fragment.length;
(paf 166:):
1(paf 167:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 168:): continue;
2(paf 169:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
170:3): continue;
(paf 171:):
172:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 173:3): throw Exception(0, // (*) see below
174:3): this,
175:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 176:3): substr.cstr(), this_offset);
(paf 177:):
4(paf 178:3): if(fragment.lang<=lang)
2(paf 179:3): return substr_begin;
180:3): else { // bad lang...
181:3): /// WARNING: this possibly skips assert (*), but it's fast
182:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
183:3): if(substr_begin==CORD_NOT_FOUND)
184:3): return STRING_NOT_FOUND;
185:3):
7(paf 186:3): size_t substr_end=substr_begin+substr.length();
2(paf 187:3): // and continuing with next fragment
188:3): }
(paf 189:): }
190:):
191:): return STRING_NOT_FOUND;
1.58 paf 192: }
193:
1.172.2.21.2. (paf 194:): size_t String::pos(const String& substr,
5(paf 195:3): size_t this_offset, Language lang) const {
(paf 196:): return pos(substr.body, this_offset, lang);
1.60 paf 197: }
198:
1.172.2.14 paf 199: void String::split(ArrayString& result,
1.172.2.21.2. (paf 200:): size_t& pos_after,
201:): const char* delim,
2(paf 202:3): Language lang, int limit) const {
(paf 203:): size_t self_length=length();
204:): if(size_t delim_length=strlen(delim)) {
1.60 paf 205: int pos_before;
206: // while we have 'delim'...
1.172.2.21.2. (paf 207:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
208:): result+=&mid(pos_after, pos_before);
209:): pos_after=pos_before+delim_length;
1.60 paf 210: }
211: // last piece
1.172.2.21.2. (paf 212:): if(pos_after<self_length && limit) {
213:): result+=&mid(pos_after, self_length);
214:): pos_after=self_length;
1.60 paf 215: }
216: } else { // empty delim
1.172.2.21.2. (paf 217:): result+=this;
218:): pos_after+=self_length;
1.60 paf 219: }
220: }
221:
1.172.2.14 paf 222: void String::split(ArrayString& result,
1.172.2.21.2. (paf 223:): size_t& pos_after,
5(paf 224:3): const String& delim, Language lang,
(paf 225:): int limit) const {
5(paf 226:3): if(!delim.is_empty()) {
1.60 paf 227: int pos_before;
228: // while we have 'delim'...
229: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 230:): result+=&mid(pos_after, pos_before);
231:): pos_after=pos_before+delim.length();
1.60 paf 232: }
233: // last piece
1.172.2.21.2. (paf 234:): if(pos_after<length() && limit) {
235:): result+=&mid(pos_after, length());
236:): pos_after=length();
1.60 paf 237: }
238: } else { // empty delim
1.172.2.21.2. (paf 239:): result+=this;
240:): pos_after+=length();
1.60 paf 241: }
1.61 paf 242: }
243:
1.172.2.21.2. 2(paf 244:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 245: struct Regex_option {
1.172.2.11 paf 246: const char* keyL;
247: const char* keyU;
1.63 paf 248: int clear, set;
249: int *result;
1.154 paf 250: bool *flag;
1.63 paf 251: } regex_option[]={
1.153 paf 252: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
253: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
254: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
255: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
256: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 257: {"'", 0, 0, 0, 0, &need_pre_post_match},
258: {0}
1.63 paf 259: };
1.171 paf 260: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 261: result[1]=0;
262:
1.172.2.21.2. 2(paf 263:3): if(options && !options->is_empty())
1.153 paf 264: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 265:3): if(options->pos(o->keyL)>=0
266:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 267: if(o->flag)
268: *o->flag=true;
269: else { // result
270: *o->result &= ~o->clear;
271: *o->result |= o->set;
272: }
1.63 paf 273: }
274: }
275:
1.172.2.21.2. (paf 276:): Table* String::match(Charset& source_charset,
1.172.2.4 paf 277: const String& regexp,
1.172.2.21.2. 2(paf 278:3): const String* options,
1.172.2.4 paf 279: Row_action row_action, void *info,
1.172.2.20 paf 280: bool& just_matched) const {
1.172.2.21.2. 5(paf 281:3): if(regexp.is_empty())
(paf 282:): throw Exception(0,
283:): 0,
1.73 paf 284: "regexp is empty");
1.154 paf 285:
1.172.2.21.2. (paf 286:): const char* pattern=regexp.cstr();
1.172.2.11 paf 287: const char* errptr;
1.62 paf 288: int erroffset;
1.172.2.21.2. (paf 289:): bool need_pre_post_match=false;
7(paf 290:3): int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 291: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 292:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 293: &errptr, &erroffset,
1.172.2.1 paf 294: source_charset.pcre_tables);
1.62 paf 295:
1.67 paf 296: if(!code)
1.172.2.21.2. (paf 297:): throw Exception(0,
298:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 299: "regular expression syntax error - %s", errptr);
1.62 paf 300:
1.172.2.20 paf 301: int subpatterns=pcre_info(code, 0, 0);
302: if(subpatterns<0) {
1.100 parser 303: pcre_free(code);
1.149 paf 304: throw Exception(0,
1.172.2.21.2. (paf 305:): ®exp,
1.76 paf 306: "pcre_info error (%d)",
1.172.2.20 paf 307: subpatterns);
1.63 paf 308: }
309:
1.172.2.21.2. (paf 310:): const char* subject=cstr();
311:): size_t subject_length=strlen(subject);
312:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
313:): int ovector[oveclength];
1.155 paf 314:
315: // create table
1.172.2.21.2. 1(paf 316:3): Table& table=*new Table(string_match_table_template);
1.63 paf 317:
1.64 paf 318: int exec_option_bits=0;
1.154 paf 319: int prestart=0;
320: int poststart=0;
1.172.2.21.2. (paf 321:): int postfinish=length();
1.63 paf 322: while(true) {
323: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 324:): subject, subject_length, prestart,
325:): exec_option_bits, ovector, oveclength);
1.63 paf 326:
327: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 328: pcre_free(code);
1.172.2.21.2. (paf 329:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 330: if(global || subpatterns)
1.172.2.21.2. 2(paf 331:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 332: else {
1.172.2.21.2. (paf 333:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 334: }
1.63 paf 335: }
336:
337: if(exec_substrings<0) {
1.100 parser 338: pcre_free(code);
1.172.2.21.2. (paf 339:): throw Exception(0,
340:): ®exp,
1.76 paf 341: "regular expression execute error (%d)",
1.63 paf 342: exec_substrings);
343: }
344:
1.154 paf 345: int prefinish=ovector[0];
346: poststart=ovector[1];
1.172.2.21.2. (paf 347:): ArrayString* row=new ArrayString;
1.172.2.4 paf 348: if(need_pre_post_match) {
1.172.2.21.2. (paf 349:): *row+=&mid(0, prefinish); // .prematch column value
350:): *row+=&mid(prefinish, poststart); // .match
351:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 352: } else {
1.172.2.21.2. (paf 353:): *row+=0; // .prematch column value
354:): *row+=0; // .match
355:): *row+=0; // .postmatch
1.172.2.4 paf 356: }
1.63 paf 357:
358: for(int i=1; i<exec_substrings; i++) {
1.69 paf 359: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 360:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 361: }
362:
1.172.2.20 paf 363: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 364:
1.172.2.20 paf 365: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 366: pcre_free(code);
1.172.2.21.2. (paf 367:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 368:3): return &table;
1.63 paf 369: }
1.154 paf 370: prestart=poststart;
1.63 paf 371:
372: /*
373: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 374: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 375: */
376: }
1.82 parser 377: }
378:
1.172.2.21.2. 0(paf 379:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 380:): String& result=*new String();
1.172.2.4 paf 381:
1.172.2.1 paf 382: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 383:
384: const unsigned char *a;
385: const unsigned char *b;
386: switch(kind) {
387: case CC_UPPER:
388: a=tables+lcc_offset;
389: b=tables+fcc_offset;
390: break;
391: case CC_LOWER:
392: a=tables+lcc_offset;
393: b=0;
394: break;
395: default:
1.172.2.21.2. (paf 396:): throw Exception(0,
397:): this,
1.82 parser 398: "unknown change case kind #%d",
399: static_cast<int>(kind)); // never
400: a=b=0; // calm, compiler
401: break; // never
402: }
403:
1.172.2.21.2. (paf 404:): char* new_cstr=cstrm();
405:): char *dest=new_cstr;
406:): unsigned char index;
407:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
408:): unsigned char c=a[index];
409:): if(b)
410:): c=b[c];
1.82 parser 411:
1.172.2.21.2. (paf 412:): *dest++=(char)c;
413:): }
414:): result.body=new_cstr;
3(paf 415:3): result.fragments.append(fragments);
1.89 parser 416:
1.101 parser 417: return result;
418: }
419:
1.172.2.21.2. (paf 420:): const String& String::replace(const Dictionary& dict) const {
421:): String& result=*new String();
422:): const char* old_cstr=cstr();
1(paf 423:3): const char* prematch_begin=old_cstr;
(paf 424:):
425:): for(const char* current=old_cstr; *current; ) {
1(paf 426:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 427:): // prematch
1(paf 428:3): if(size_t prematch_length=current-prematch_begin) {
7(paf 429:3): result.body.append(prematch_begin, prematch_length);
4(paf 430:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 431: }
432:
1.172.2.21.2. (paf 433:): // match
434:):
435:): const String* a=row->get(0);
1(paf 436:3): // skip 'a' in 'current'; move prematch_begin
437:3): current+=a->length(); prematch_begin=current;
1.170 paf 438:
1.172.2.21.2. (paf 439:): if(row->count()>1) { // are there any b?
440:): const String* b=row->get(1);
441:): if(size_t b_length=b->length()) {
7(paf 442:3): result.body.append(b->cstr_to_string_body());
3(paf 443:3): result.fragments.append(b->fragments);
(paf 444:): }
445:): }
446:): } else // simply advance
447:): current++;
448:): }
1.156 paf 449:
1.172.2.21.2. (paf 450:): // postmatch
3(paf 451:3): if(size_t postmatch_length=current-prematch_begin) {
7(paf 452:3): result.body.append(prematch_begin, postmatch_length);
4(paf 453:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 454:): }
1.156 paf 455:
1.89 parser 456: return result;
457: }
458:
1.172.2.21.2. (paf 459:): double String::as_double() const {
1.89 parser 460: double result;
1.172.2.21.2. (paf 461:): const char *str=cstr();
462:):
463:): while(*str && isspace(*str))
464:): str++;
465:): if(!*str)
1.162 paf 466: return 0;
1.161 paf 467:
1.102 parser 468: char *error_pos;
1.89 parser 469: // 0xABC
1.172.2.21.2. (paf 470:): if(str[0]=='0')
471:): if(str[1]=='x' || str[1]=='X')
472:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 473: else
1.172.2.21.2. (paf 474:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 475: else
1.172.2.21.2. (paf 476:): result=(double)strtod(str, &error_pos);
1.89 parser 477:
1.159 paf 478: while(char c=*error_pos++)
479: if(!isspace(c))
480: throw Exception("number.format",
1.172.2.21.2. (paf 481:): this,
1.159 paf 482: "invalid number (double)");
1.89 parser 483:
484: return result;
485: }
1.172.2.21.2. (paf 486:): int String::as_int() const {
1.89 parser 487: int result;
1.172.2.21.2. (paf 488:): const char *str=cstr();
489:):
490:): while(*str && isspace(*str))
491:): str++;
492:): if(!*str)
1.162 paf 493: return 0;
1.161 paf 494:
1.102 parser 495: char *error_pos;
1.89 parser 496: // 0xABC
1.172.2.21.2. (paf 497:): if(str[0]=='0')
498:): if(str[1]=='x' || str[1]=='X')
499:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 500: else
1.172.2.21.2. (paf 501:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 502: else
1.172.2.21.2. (paf 503:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 504:
1.159 paf 505: while(char c=*error_pos++)
506: if(!isspace(c))
507: throw Exception("number.format",
1.172.2.21.2. (paf 508:): this,
1.159 paf 509: "invalid number (int)");
1.82 parser 510:
511: return result;
1.61 paf 512: }
1.113 parser 513:
1.172.2.4 paf 514: inline void uint2uchars(uint word, uchar *bytes) {
515: bytes[0]=word&0xFF;
516: bytes[1]=(word>>8)&0xFF;
517: bytes[2]=(word>>16)&0xFF;
518: bytes[3]=(word>>24)&0xFF;
519: }
520: inline uint uchars2uint(uchar *bytes) {
521: return bytes[3]<<24
522: | bytes[2]<<16
523: | bytes[1]<<8
524: | bytes[0];
525: }
526:
1.172.2.21.2. 7(paf 527:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 528:): size_t length=strlen(s);
7(paf 529:3): memcpy(*cur, s, length); *cur+=length;
(paf 530:): return 0;
531:): };
532:): String::C String::serialize(size_t prolog_length) const {
533:): size_t buf_length=
534:): prolog_length
5(paf 535:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 536:): +length();
537:): C result(new(PointerFreeGC) char[buf_length], buf_length);
538:):
539:): // 1: prolog
540:): char *cur=result.str+prolog_length;
541:):
542:):
543:): // 2: fragments.count
544:): size_t fragments_count=fragments.count();
545:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 546:
1.172.2.21.2. (paf 547:): // 3: lang info
548:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
549:): const Fragment& fragment=i.next();
1.123 paf 550: // lang
1.172.2.21.2. (paf 551:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
552:): // length
553:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
554:): }
555:):
556:): // 4: letters
7(paf 557:3): body.for_each(serialize_body_piece, &cur);
(paf 558:):
559:): return result;
1.113 parser 560: }
1.172.2.21.2. 9(paf 561:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 562:): if(buf_length<=prolog_length)
1.148 paf 563: return false;
1.172.2.21.2. (paf 564:): buf_length-=prolog_length;
1.113 parser 565:
1.172.2.21.2. (paf 566:): // 1: prolog
567:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 568:
1.172.2.21.2. (paf 569:): // 2: fragments.count
570:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
571:): return false;
572:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
573:): buf_length-=sizeof(size_t);
1.128 paf 574:
1.172.2.21.2. (paf 575:): // 3: lang info
576:): size_t total_length=0;
577:): for(size_t f=0; f<fragments_count; f++) {
5(paf 578:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
(paf 579:): if(buf_length<piece_length) // lang+length
1.148 paf 580: return false;
581:
1.172.2.21.2. 5(paf 582:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
(paf 583:): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
584:): fragments+=Fragment(lang, fragment_length);
585:): total_length+=fragment_length;
1.113 parser 586:
1.172.2.21.2. (paf 587:): buf_length-=piece_length;
1.113 parser 588: }
1.172.2.21.2. (paf 589:):
590:): // 4: letters
591:): if(buf_length!=total_length)
592:): return false;
593:):
594:): body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length);
595:):
1.148 paf 596: return true;
1.113 parser 597: }
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