Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.29
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. 9(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/03/25 09:34:53 $";
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):
9(paf 82:3): StringBody StringBody::Format(int value) {
8(paf 83:3): char local[MAX_NUMBER];
9(paf 84:3): size_t length=snprintf(local, MAX_NUMBER, "%d", value);
85:3): return StringBody(pa_strdup(local, length), length);
8(paf 86:3): }
5(paf 87:3):
3(paf 88:3): static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
89:3): uint& result=*static_cast<uint*>(client_data);
90:3): generic_hash_code(result, c);
91:3): return 0;
92:3): }
5(paf 93:3): static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
94:3): uint& result=*static_cast<uint*>(client_data);
95:3): generic_hash_code(result, s);
96:3): return 0;
97:3): };
7(paf 98:3): uint StringBody::hash_code() const {
5(paf 99:3): uint result=0;
3(paf 100:3): CORD_iter5(body, 0,
101:3): CORD_batched_iter_fn_generic_hash_code,
102:3): CORD_batched_iter_fn_generic_hash_code, &result);
5(paf 103:3): return result;
104:3): }
105:3):
(paf 106:): // String methods
1.172.2.2 paf 107:
1.172.2.21.2. (paf 108:): String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
5(paf 109:3): append(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.1 paf 110: }
1.140 paf 111:
1.172.2.21.2. (paf 112:): String::String(const String& src): body(src.body), fragments(src.fragments) {}
1.28 paf 113:
1.172.2.21.2. 6(paf 114:3): String& String::append(const char* str, size_t length, Language lang) {
3(paf 115:3): if(!str)
1.9 paf 116: return *this;
1.172.2.21.2. 6(paf 117:3): size_t use_length=length?length:strlen(str);
0(paf 118:3): if(!use_length)
1.9 paf 119: return *this;
1.122 paf 120:
1.172.2.21.2. 7(paf 121:3): body.append(str, use_length);
0(paf 122:3): fragments+=Fragment(lang, use_length);
1.1 paf 123:
124: return *this;
125: }
126:
1.172.2.21.2. 1(paf 127:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 128:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 129:): String& result=*new String;
1.33 paf 130:
1.172.2.21.2. 1(paf 131:3): size_t self_length=length();
132:3): substr_begin=max(min(substr_begin, self_length), (size_t)0);
133:3): substr_end=min(max(substr_end, substr_begin), self_length);
134:3): if(substr_begin==substr_end)
(paf 135:): return result;
1.52 paf 136:
1.172.2.21.2. (paf 137:): // first: letters themselves
7(paf 138:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 139:
1.172.2.21.2. (paf 140:): // next: their langs
3(paf 141:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 142:
1.172.2.21.2. 5(paf 143:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 144:3): //cstr(), substr_begin, substr_end, result.cstr());
1.53 paf 145: return result;
1.54 paf 146: }
147:
1.172.2.21.2. 7(paf 148:3): size_t String::pos(const StringBody substr,
5(paf 149:3): size_t this_offset, Language lang) const {
(paf 150:): // first: letters themselves
2(paf 151:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 152:3): if(substr_begin==CORD_NOT_FOUND)
(paf 153:): return STRING_NOT_FOUND;
154:):
155:): // next: check the lang when specified
156:):
5(paf 157:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 158:3): return substr_begin;
(paf 159:):
160:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 161:3): size_t substr_end=substr_begin+substr.length();
1(paf 162:3): size_t fragment_begin=0;
(paf 163:): size_t fragment_end;
1(paf 164:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
(paf 165:): const Fragment& fragment=i.next();
1(paf 166:3): fragment_end=fragment_begin+fragment.length;
(paf 167:):
1(paf 168:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 169:): continue;
2(paf 170:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
171:3): continue;
(paf 172:):
173:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 174:3): throw Exception(0, // (*) see below
175:3): this,
176:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 177:3): substr.cstr(), this_offset);
(paf 178:):
4(paf 179:3): if(fragment.lang<=lang)
2(paf 180:3): return substr_begin;
181:3): else { // bad lang...
182:3): /// WARNING: this possibly skips assert (*), but it's fast
183:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
184:3): if(substr_begin==CORD_NOT_FOUND)
185:3): return STRING_NOT_FOUND;
186:3):
7(paf 187:3): size_t substr_end=substr_begin+substr.length();
2(paf 188:3): // and continuing with next fragment
189:3): }
(paf 190:): }
191:):
192:): return STRING_NOT_FOUND;
1.58 paf 193: }
194:
1.172.2.21.2. (paf 195:): size_t String::pos(const String& substr,
5(paf 196:3): size_t this_offset, Language lang) const {
(paf 197:): return pos(substr.body, this_offset, lang);
1.60 paf 198: }
199:
1.172.2.14 paf 200: void String::split(ArrayString& result,
1.172.2.21.2. (paf 201:): size_t& pos_after,
202:): const char* delim,
2(paf 203:3): Language lang, int limit) const {
(paf 204:): size_t self_length=length();
205:): if(size_t delim_length=strlen(delim)) {
1.60 paf 206: int pos_before;
207: // while we have 'delim'...
1.172.2.21.2. (paf 208:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
209:): result+=&mid(pos_after, pos_before);
210:): pos_after=pos_before+delim_length;
1.60 paf 211: }
212: // last piece
1.172.2.21.2. (paf 213:): if(pos_after<self_length && limit) {
214:): result+=&mid(pos_after, self_length);
215:): pos_after=self_length;
1.60 paf 216: }
217: } else { // empty delim
1.172.2.21.2. (paf 218:): result+=this;
219:): pos_after+=self_length;
1.60 paf 220: }
221: }
222:
1.172.2.14 paf 223: void String::split(ArrayString& result,
1.172.2.21.2. (paf 224:): size_t& pos_after,
5(paf 225:3): const String& delim, Language lang,
(paf 226:): int limit) const {
5(paf 227:3): if(!delim.is_empty()) {
1.60 paf 228: int pos_before;
229: // while we have 'delim'...
230: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 231:): result+=&mid(pos_after, pos_before);
232:): pos_after=pos_before+delim.length();
1.60 paf 233: }
234: // last piece
1.172.2.21.2. (paf 235:): if(pos_after<length() && limit) {
236:): result+=&mid(pos_after, length());
237:): pos_after=length();
1.60 paf 238: }
239: } else { // empty delim
1.172.2.21.2. (paf 240:): result+=this;
241:): pos_after+=length();
1.60 paf 242: }
1.61 paf 243: }
244:
1.172.2.21.2. 2(paf 245:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 246: struct Regex_option {
1.172.2.11 paf 247: const char* keyL;
248: const char* keyU;
1.63 paf 249: int clear, set;
250: int *result;
1.154 paf 251: bool *flag;
1.63 paf 252: } regex_option[]={
1.153 paf 253: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
254: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
255: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
256: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
257: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 258: {"'", 0, 0, 0, 0, &need_pre_post_match},
259: {0}
1.63 paf 260: };
1.171 paf 261: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 262: result[1]=0;
263:
1.172.2.21.2. 2(paf 264:3): if(options && !options->is_empty())
1.153 paf 265: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 266:3): if(options->pos(o->keyL)>=0
267:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 268: if(o->flag)
269: *o->flag=true;
270: else { // result
271: *o->result &= ~o->clear;
272: *o->result |= o->set;
273: }
1.63 paf 274: }
275: }
276:
1.172.2.21.2. (paf 277:): Table* String::match(Charset& source_charset,
1.172.2.4 paf 278: const String& regexp,
1.172.2.21.2. 2(paf 279:3): const String* options,
1.172.2.4 paf 280: Row_action row_action, void *info,
1.172.2.20 paf 281: bool& just_matched) const {
1.172.2.21.2. 5(paf 282:3): if(regexp.is_empty())
(paf 283:): throw Exception(0,
284:): 0,
1.73 paf 285: "regexp is empty");
1.154 paf 286:
1.172.2.21.2. (paf 287:): const char* pattern=regexp.cstr();
1.172.2.11 paf 288: const char* errptr;
1.62 paf 289: int erroffset;
1.172.2.21.2. (paf 290:): bool need_pre_post_match=false;
7(paf 291:3): int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 292: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 293:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 294: &errptr, &erroffset,
1.172.2.1 paf 295: source_charset.pcre_tables);
1.62 paf 296:
1.67 paf 297: if(!code)
1.172.2.21.2. (paf 298:): throw Exception(0,
299:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 300: "regular expression syntax error - %s", errptr);
1.62 paf 301:
1.172.2.20 paf 302: int subpatterns=pcre_info(code, 0, 0);
303: if(subpatterns<0) {
1.100 parser 304: pcre_free(code);
1.149 paf 305: throw Exception(0,
1.172.2.21.2. (paf 306:): ®exp,
1.76 paf 307: "pcre_info error (%d)",
1.172.2.20 paf 308: subpatterns);
1.63 paf 309: }
310:
1.172.2.21.2. (paf 311:): const char* subject=cstr();
312:): size_t subject_length=strlen(subject);
313:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
314:): int ovector[oveclength];
1.155 paf 315:
316: // create table
1.172.2.21.2. 1(paf 317:3): Table& table=*new Table(string_match_table_template);
1.63 paf 318:
1.64 paf 319: int exec_option_bits=0;
1.154 paf 320: int prestart=0;
321: int poststart=0;
1.172.2.21.2. (paf 322:): int postfinish=length();
1.63 paf 323: while(true) {
324: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 325:): subject, subject_length, prestart,
326:): exec_option_bits, ovector, oveclength);
1.63 paf 327:
328: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 329: pcre_free(code);
1.172.2.21.2. (paf 330:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 331: if(global || subpatterns)
1.172.2.21.2. 2(paf 332:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 333: else {
1.172.2.21.2. (paf 334:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 335: }
1.63 paf 336: }
337:
338: if(exec_substrings<0) {
1.100 parser 339: pcre_free(code);
1.172.2.21.2. (paf 340:): throw Exception(0,
341:): ®exp,
1.76 paf 342: "regular expression execute error (%d)",
1.63 paf 343: exec_substrings);
344: }
345:
1.154 paf 346: int prefinish=ovector[0];
347: poststart=ovector[1];
1.172.2.21.2. (paf 348:): ArrayString* row=new ArrayString;
1.172.2.4 paf 349: if(need_pre_post_match) {
1.172.2.21.2. (paf 350:): *row+=&mid(0, prefinish); // .prematch column value
351:): *row+=&mid(prefinish, poststart); // .match
352:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 353: } else {
1.172.2.21.2. (paf 354:): *row+=0; // .prematch column value
355:): *row+=0; // .match
356:): *row+=0; // .postmatch
1.172.2.4 paf 357: }
1.63 paf 358:
359: for(int i=1; i<exec_substrings; i++) {
1.69 paf 360: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 361:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 362: }
363:
1.172.2.20 paf 364: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 365:
1.172.2.20 paf 366: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 367: pcre_free(code);
1.172.2.21.2. (paf 368:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 369:3): return &table;
1.63 paf 370: }
1.154 paf 371: prestart=poststart;
1.63 paf 372:
373: /*
374: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 375: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 376: */
377: }
1.82 parser 378: }
379:
1.172.2.21.2. 0(paf 380:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 381:): String& result=*new String();
1.172.2.4 paf 382:
1.172.2.1 paf 383: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 384:
385: const unsigned char *a;
386: const unsigned char *b;
387: switch(kind) {
388: case CC_UPPER:
389: a=tables+lcc_offset;
390: b=tables+fcc_offset;
391: break;
392: case CC_LOWER:
393: a=tables+lcc_offset;
394: b=0;
395: break;
396: default:
1.172.2.21.2. (paf 397:): throw Exception(0,
398:): this,
1.82 parser 399: "unknown change case kind #%d",
400: static_cast<int>(kind)); // never
401: a=b=0; // calm, compiler
402: break; // never
403: }
404:
1.172.2.21.2. (paf 405:): char* new_cstr=cstrm();
406:): char *dest=new_cstr;
407:): unsigned char index;
408:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
409:): unsigned char c=a[index];
410:): if(b)
411:): c=b[c];
1.82 parser 412:
1.172.2.21.2. (paf 413:): *dest++=(char)c;
414:): }
415:): result.body=new_cstr;
3(paf 416:3): result.fragments.append(fragments);
1.89 parser 417:
1.101 parser 418: return result;
419: }
420:
1.172.2.21.2. (paf 421:): const String& String::replace(const Dictionary& dict) const {
422:): String& result=*new String();
423:): const char* old_cstr=cstr();
1(paf 424:3): const char* prematch_begin=old_cstr;
(paf 425:):
426:): for(const char* current=old_cstr; *current; ) {
1(paf 427:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 428:): // prematch
1(paf 429:3): if(size_t prematch_length=current-prematch_begin) {
7(paf 430:3): result.body.append(prematch_begin, prematch_length);
4(paf 431:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 432: }
433:
1.172.2.21.2. (paf 434:): // match
435:):
436:): const String* a=row->get(0);
1(paf 437:3): // skip 'a' in 'current'; move prematch_begin
438:3): current+=a->length(); prematch_begin=current;
1.170 paf 439:
1.172.2.21.2. (paf 440:): if(row->count()>1) { // are there any b?
441:): const String* b=row->get(1);
442:): if(size_t b_length=b->length()) {
7(paf 443:3): result.body.append(b->cstr_to_string_body());
3(paf 444:3): result.fragments.append(b->fragments);
(paf 445:): }
446:): }
447:): } else // simply advance
448:): current++;
449:): }
1.156 paf 450:
1.172.2.21.2. (paf 451:): // postmatch
3(paf 452:3): if(size_t postmatch_length=current-prematch_begin) {
7(paf 453:3): result.body.append(prematch_begin, postmatch_length);
4(paf 454:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 455:): }
1.156 paf 456:
1.89 parser 457: return result;
458: }
459:
1.172.2.21.2. (paf 460:): double String::as_double() const {
1.89 parser 461: double result;
1.172.2.21.2. (paf 462:): const char *str=cstr();
463:):
464:): while(*str && isspace(*str))
465:): str++;
466:): if(!*str)
1.162 paf 467: return 0;
1.161 paf 468:
1.102 parser 469: char *error_pos;
1.89 parser 470: // 0xABC
1.172.2.21.2. (paf 471:): if(str[0]=='0')
472:): if(str[1]=='x' || str[1]=='X')
473:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 474: else
1.172.2.21.2. (paf 475:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 476: else
1.172.2.21.2. (paf 477:): result=(double)strtod(str, &error_pos);
1.89 parser 478:
1.159 paf 479: while(char c=*error_pos++)
480: if(!isspace(c))
481: throw Exception("number.format",
1.172.2.21.2. (paf 482:): this,
1.159 paf 483: "invalid number (double)");
1.89 parser 484:
485: return result;
486: }
1.172.2.21.2. (paf 487:): int String::as_int() const {
1.89 parser 488: int result;
1.172.2.21.2. (paf 489:): const char *str=cstr();
490:):
491:): while(*str && isspace(*str))
492:): str++;
493:): if(!*str)
1.162 paf 494: return 0;
1.161 paf 495:
1.102 parser 496: char *error_pos;
1.89 parser 497: // 0xABC
1.172.2.21.2. (paf 498:): if(str[0]=='0')
499:): if(str[1]=='x' || str[1]=='X')
500:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 501: else
1.172.2.21.2. (paf 502:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 503: else
1.172.2.21.2. (paf 504:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 505:
1.159 paf 506: while(char c=*error_pos++)
507: if(!isspace(c))
508: throw Exception("number.format",
1.172.2.21.2. (paf 509:): this,
1.159 paf 510: "invalid number (int)");
1.82 parser 511:
512: return result;
1.61 paf 513: }
1.113 parser 514:
1.172.2.4 paf 515: inline void uint2uchars(uint word, uchar *bytes) {
516: bytes[0]=word&0xFF;
517: bytes[1]=(word>>8)&0xFF;
518: bytes[2]=(word>>16)&0xFF;
519: bytes[3]=(word>>24)&0xFF;
520: }
521: inline uint uchars2uint(uchar *bytes) {
522: return bytes[3]<<24
523: | bytes[2]<<16
524: | bytes[1]<<8
525: | bytes[0];
526: }
527:
1.172.2.21.2. 7(paf 528:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 529:): size_t length=strlen(s);
7(paf 530:3): memcpy(*cur, s, length); *cur+=length;
(paf 531:): return 0;
532:): };
533:): String::C String::serialize(size_t prolog_length) const {
534:): size_t buf_length=
535:): prolog_length
5(paf 536:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 537:): +length();
538:): C result(new(PointerFreeGC) char[buf_length], buf_length);
539:):
540:): // 1: prolog
541:): char *cur=result.str+prolog_length;
542:):
543:):
544:): // 2: fragments.count
545:): size_t fragments_count=fragments.count();
546:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 547:
1.172.2.21.2. (paf 548:): // 3: lang info
549:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
550:): const Fragment& fragment=i.next();
1.123 paf 551: // lang
1.172.2.21.2. (paf 552:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
553:): // length
554:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
555:): }
556:):
557:): // 4: letters
7(paf 558:3): body.for_each(serialize_body_piece, &cur);
(paf 559:):
560:): return result;
1.113 parser 561: }
1.172.2.21.2. 9(paf 562:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 563:): if(buf_length<=prolog_length)
1.148 paf 564: return false;
1.172.2.21.2. (paf 565:): buf_length-=prolog_length;
1.113 parser 566:
1.172.2.21.2. (paf 567:): // 1: prolog
568:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 569:
1.172.2.21.2. (paf 570:): // 2: fragments.count
571:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
572:): return false;
573:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
574:): buf_length-=sizeof(size_t);
1.128 paf 575:
1.172.2.21.2. (paf 576:): // 3: lang info
577:): size_t total_length=0;
578:): for(size_t f=0; f<fragments_count; f++) {
5(paf 579:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
(paf 580:): if(buf_length<piece_length) // lang+length
1.148 paf 581: return false;
582:
1.172.2.21.2. 5(paf 583:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
(paf 584:): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
585:): fragments+=Fragment(lang, fragment_length);
586:): total_length+=fragment_length;
1.113 parser 587:
1.172.2.21.2. (paf 588:): buf_length-=piece_length;
1.113 parser 589: }
1.172.2.21.2. (paf 590:):
591:): // 4: letters
592:): if(buf_length!=total_length)
593:): return false;
594:):
595:): body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length);
596:):
1.148 paf 597: return true;
1.113 parser 598: }
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