Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.31
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. 1(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/03/26 07:45:24 $";
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:
1.172.2.2 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. 0(paf 112:3): String::String(const String& src): body(src.body) {
113:3): fragments.append(src.fragments);
114:3): }
1.28 paf 115:
1.172.2.21.2. 6(paf 116:3): String& String::append(const char* str, size_t length, Language lang) {
3(paf 117:3): if(!str)
1.9 paf 118: return *this;
1.172.2.21.2. 6(paf 119:3): size_t use_length=length?length:strlen(str);
0(paf 120:3): if(!use_length)
1.9 paf 121: return *this;
1.122 paf 122:
1.172.2.21.2. 7(paf 123:3): body.append(str, use_length);
0(paf 124:3): fragments+=Fragment(lang, use_length);
1.1 paf 125:
126: return *this;
127: }
128:
1.172.2.21.2. 1(paf 129:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 130:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 131:): String& result=*new String;
1.33 paf 132:
1.172.2.21.2. 1(paf 133:3): size_t self_length=length();
134:3): substr_begin=max(min(substr_begin, self_length), (size_t)0);
135:3): substr_end=min(max(substr_end, substr_begin), self_length);
136:3): if(substr_begin==substr_end)
(paf 137:): return result;
1.52 paf 138:
1.172.2.21.2. (paf 139:): // first: letters themselves
7(paf 140:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 141:
1.172.2.21.2. (paf 142:): // next: their langs
3(paf 143:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 144:
1.172.2.21.2. 5(paf 145:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 146:3): //cstr(), substr_begin, substr_end, result.cstr());
1.53 paf 147: return result;
1.54 paf 148: }
149:
1.172.2.21.2. 7(paf 150:3): size_t String::pos(const StringBody substr,
5(paf 151:3): size_t this_offset, Language lang) const {
(paf 152:): // first: letters themselves
2(paf 153:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 154:3): if(substr_begin==CORD_NOT_FOUND)
(paf 155:): return STRING_NOT_FOUND;
156:):
157:): // next: check the lang when specified
158:):
5(paf 159:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 160:3): return substr_begin;
(paf 161:):
162:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 163:3): size_t substr_end=substr_begin+substr.length();
1(paf 164:3): size_t fragment_begin=0;
(paf 165:): size_t fragment_end;
1(paf 166:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
(paf 167:): const Fragment& fragment=i.next();
1(paf 168:3): fragment_end=fragment_begin+fragment.length;
(paf 169:):
1(paf 170:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 171:): continue;
2(paf 172:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
173:3): continue;
(paf 174:):
175:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 176:3): throw Exception(0, // (*) see below
177:3): this,
178:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 179:3): substr.cstr(), this_offset);
(paf 180:):
4(paf 181:3): if(fragment.lang<=lang)
2(paf 182:3): return substr_begin;
183:3): else { // bad lang...
184:3): /// WARNING: this possibly skips assert (*), but it's fast
185:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
186:3): if(substr_begin==CORD_NOT_FOUND)
187:3): return STRING_NOT_FOUND;
188:3):
7(paf 189:3): size_t substr_end=substr_begin+substr.length();
2(paf 190:3): // and continuing with next fragment
191:3): }
(paf 192:): }
193:):
194:): return STRING_NOT_FOUND;
1.58 paf 195: }
196:
1.172.2.21.2. (paf 197:): size_t String::pos(const String& substr,
5(paf 198:3): size_t this_offset, Language lang) const {
(paf 199:): return pos(substr.body, this_offset, lang);
1.60 paf 200: }
201:
1.172.2.14 paf 202: void String::split(ArrayString& result,
1.172.2.21.2. (paf 203:): size_t& pos_after,
204:): const char* delim,
2(paf 205:3): Language lang, int limit) const {
(paf 206:): size_t self_length=length();
207:): if(size_t delim_length=strlen(delim)) {
1.60 paf 208: int pos_before;
209: // while we have 'delim'...
1.172.2.21.2. (paf 210:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
211:): result+=&mid(pos_after, pos_before);
212:): pos_after=pos_before+delim_length;
1.60 paf 213: }
214: // last piece
1.172.2.21.2. (paf 215:): if(pos_after<self_length && limit) {
216:): result+=&mid(pos_after, self_length);
217:): pos_after=self_length;
1.60 paf 218: }
219: } else { // empty delim
1.172.2.21.2. (paf 220:): result+=this;
221:): pos_after+=self_length;
1.60 paf 222: }
223: }
224:
1.172.2.14 paf 225: void String::split(ArrayString& result,
1.172.2.21.2. (paf 226:): size_t& pos_after,
5(paf 227:3): const String& delim, Language lang,
(paf 228:): int limit) const {
5(paf 229:3): if(!delim.is_empty()) {
1.60 paf 230: int pos_before;
231: // while we have 'delim'...
232: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 233:): result+=&mid(pos_after, pos_before);
234:): pos_after=pos_before+delim.length();
1.60 paf 235: }
236: // last piece
1.172.2.21.2. (paf 237:): if(pos_after<length() && limit) {
238:): result+=&mid(pos_after, length());
239:): pos_after=length();
1.60 paf 240: }
241: } else { // empty delim
1.172.2.21.2. (paf 242:): result+=this;
243:): pos_after+=length();
1.60 paf 244: }
1.61 paf 245: }
246:
1.172.2.21.2. 2(paf 247:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 248: struct Regex_option {
1.172.2.11 paf 249: const char* keyL;
250: const char* keyU;
1.63 paf 251: int clear, set;
252: int *result;
1.154 paf 253: bool *flag;
1.63 paf 254: } regex_option[]={
1.153 paf 255: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
256: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
257: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
258: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
259: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 260: {"'", 0, 0, 0, 0, &need_pre_post_match},
261: {0}
1.63 paf 262: };
1.171 paf 263: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 264: result[1]=0;
265:
1.172.2.21.2. 2(paf 266:3): if(options && !options->is_empty())
1.153 paf 267: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 268:3): if(options->pos(o->keyL)>=0
269:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 270: if(o->flag)
271: *o->flag=true;
272: else { // result
273: *o->result &= ~o->clear;
274: *o->result |= o->set;
275: }
1.63 paf 276: }
277: }
278:
1.172.2.21.2. (paf 279:): Table* String::match(Charset& source_charset,
1.172.2.4 paf 280: const String& regexp,
1.172.2.21.2. 2(paf 281:3): const String* options,
1.172.2.4 paf 282: Row_action row_action, void *info,
1.172.2.20 paf 283: bool& just_matched) const {
1.172.2.21.2. 5(paf 284:3): if(regexp.is_empty())
(paf 285:): throw Exception(0,
286:): 0,
1.73 paf 287: "regexp is empty");
1.154 paf 288:
1.172.2.21.2. (paf 289:): const char* pattern=regexp.cstr();
1.172.2.11 paf 290: const char* errptr;
1.62 paf 291: int erroffset;
1.172.2.21.2. (paf 292:): bool need_pre_post_match=false;
7(paf 293:3): int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 294: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 295:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 296: &errptr, &erroffset,
1.172.2.1 paf 297: source_charset.pcre_tables);
1.62 paf 298:
1.67 paf 299: if(!code)
1.172.2.21.2. (paf 300:): throw Exception(0,
301:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 302: "regular expression syntax error - %s", errptr);
1.62 paf 303:
1.172.2.20 paf 304: int subpatterns=pcre_info(code, 0, 0);
305: if(subpatterns<0) {
1.100 parser 306: pcre_free(code);
1.149 paf 307: throw Exception(0,
1.172.2.21.2. (paf 308:): ®exp,
1.76 paf 309: "pcre_info error (%d)",
1.172.2.20 paf 310: subpatterns);
1.63 paf 311: }
312:
1.172.2.21.2. (paf 313:): const char* subject=cstr();
314:): size_t subject_length=strlen(subject);
315:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
316:): int ovector[oveclength];
1.155 paf 317:
318: // create table
1.172.2.21.2. 1(paf 319:3): Table& table=*new Table(string_match_table_template);
1.63 paf 320:
1.64 paf 321: int exec_option_bits=0;
1.154 paf 322: int prestart=0;
323: int poststart=0;
1.172.2.21.2. (paf 324:): int postfinish=length();
1.63 paf 325: while(true) {
326: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 327:): subject, subject_length, prestart,
328:): exec_option_bits, ovector, oveclength);
1.63 paf 329:
330: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 331: pcre_free(code);
1.172.2.21.2. (paf 332:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 333: if(global || subpatterns)
1.172.2.21.2. 2(paf 334:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 335: else {
1.172.2.21.2. (paf 336:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 337: }
1.63 paf 338: }
339:
340: if(exec_substrings<0) {
1.100 parser 341: pcre_free(code);
1.172.2.21.2. (paf 342:): throw Exception(0,
343:): ®exp,
1.76 paf 344: "regular expression execute error (%d)",
1.63 paf 345: exec_substrings);
346: }
347:
1.154 paf 348: int prefinish=ovector[0];
349: poststart=ovector[1];
1.172.2.21.2. (paf 350:): ArrayString* row=new ArrayString;
1.172.2.4 paf 351: if(need_pre_post_match) {
1.172.2.21.2. (paf 352:): *row+=&mid(0, prefinish); // .prematch column value
353:): *row+=&mid(prefinish, poststart); // .match
354:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 355: } else {
1.172.2.21.2. (paf 356:): *row+=0; // .prematch column value
357:): *row+=0; // .match
358:): *row+=0; // .postmatch
1.172.2.4 paf 359: }
1.63 paf 360:
361: for(int i=1; i<exec_substrings; i++) {
1.69 paf 362: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 363:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 364: }
365:
1.172.2.20 paf 366: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 367:
1.172.2.20 paf 368: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 369: pcre_free(code);
1.172.2.21.2. (paf 370:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 371:3): return &table;
1.63 paf 372: }
1.154 paf 373: prestart=poststart;
1.63 paf 374:
375: /*
376: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 377: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 378: */
379: }
1.82 parser 380: }
381:
1.172.2.21.2. 0(paf 382:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 383:): String& result=*new String();
0(paf 384:3): if(is_empty())
385:3): return result;
1.172.2.4 paf 386:
1.172.2.1 paf 387: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 388:
389: const unsigned char *a;
390: const unsigned char *b;
391: switch(kind) {
392: case CC_UPPER:
393: a=tables+lcc_offset;
394: b=tables+fcc_offset;
395: break;
396: case CC_LOWER:
397: a=tables+lcc_offset;
398: b=0;
399: break;
400: default:
1.172.2.21.2. (paf 401:): throw Exception(0,
402:): this,
1.82 parser 403: "unknown change case kind #%d",
404: static_cast<int>(kind)); // never
405: a=b=0; // calm, compiler
406: break; // never
407: }
408:
1.172.2.21.2. (paf 409:): char* new_cstr=cstrm();
410:): char *dest=new_cstr;
411:): unsigned char index;
412:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
413:): unsigned char c=a[index];
414:): if(b)
415:): c=b[c];
1.82 parser 416:
1.172.2.21.2. (paf 417:): *dest++=(char)c;
418:): }
419:): result.body=new_cstr;
3(paf 420:3): result.fragments.append(fragments);
1.89 parser 421:
1.101 parser 422: return result;
423: }
424:
1.172.2.21.2. (paf 425:): const String& String::replace(const Dictionary& dict) const {
426:): String& result=*new String();
427:): const char* old_cstr=cstr();
1(paf 428:3): const char* prematch_begin=old_cstr;
(paf 429:):
1(paf 430:3): const char* current=old_cstr;
431:3): while(*current) {
432:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 433:): // prematch
1(paf 434:3): if(size_t prematch_length=current-prematch_begin) {
7(paf 435:3): result.body.append(prematch_begin, prematch_length);
4(paf 436:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 437: }
438:
1.172.2.21.2. (paf 439:): // match
440:):
441:): const String* a=row->get(0);
1(paf 442:3): // skip 'a' in 'current'; move prematch_begin
443:3): current+=a->length(); prematch_begin=current;
1.170 paf 444:
1.172.2.21.2. (paf 445:): if(row->count()>1) { // are there any b?
446:): const String* b=row->get(1);
447:): if(size_t b_length=b->length()) {
1(paf 448:3): result.body<<b->cstr_to_string_body();
3(paf 449:3): result.fragments.append(b->fragments);
(paf 450:): }
451:): }
452:): } else // simply advance
453:): current++;
454:): }
1.156 paf 455:
1.172.2.21.2. (paf 456:): // postmatch
3(paf 457:3): if(size_t postmatch_length=current-prematch_begin) {
7(paf 458:3): result.body.append(prematch_begin, postmatch_length);
4(paf 459:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 460:): }
1.156 paf 461:
1.89 parser 462: return result;
463: }
464:
1.172.2.21.2. (paf 465:): double String::as_double() const {
1.89 parser 466: double result;
1.172.2.21.2. (paf 467:): const char *str=cstr();
468:):
469:): while(*str && isspace(*str))
470:): str++;
471:): if(!*str)
1.162 paf 472: return 0;
1.161 paf 473:
1.102 parser 474: char *error_pos;
1.89 parser 475: // 0xABC
1.172.2.21.2. (paf 476:): if(str[0]=='0')
477:): if(str[1]=='x' || str[1]=='X')
478:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 479: else
1.172.2.21.2. (paf 480:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 481: else
1.172.2.21.2. (paf 482:): result=(double)strtod(str, &error_pos);
1.89 parser 483:
1.159 paf 484: while(char c=*error_pos++)
485: if(!isspace(c))
486: throw Exception("number.format",
1.172.2.21.2. (paf 487:): this,
1.159 paf 488: "invalid number (double)");
1.89 parser 489:
490: return result;
491: }
1.172.2.21.2. (paf 492:): int String::as_int() const {
1.89 parser 493: int result;
1.172.2.21.2. (paf 494:): const char *str=cstr();
495:):
496:): while(*str && isspace(*str))
497:): str++;
498:): if(!*str)
1.162 paf 499: return 0;
1.161 paf 500:
1.102 parser 501: char *error_pos;
1.89 parser 502: // 0xABC
1.172.2.21.2. (paf 503:): if(str[0]=='0')
504:): if(str[1]=='x' || str[1]=='X')
505:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 506: else
1.172.2.21.2. (paf 507:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 508: else
1.172.2.21.2. (paf 509:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 510:
1.159 paf 511: while(char c=*error_pos++)
512: if(!isspace(c))
513: throw Exception("number.format",
1.172.2.21.2. (paf 514:): this,
1.159 paf 515: "invalid number (int)");
1.82 parser 516:
517: return result;
1.61 paf 518: }
1.113 parser 519:
1.172.2.4 paf 520: inline void uint2uchars(uint word, uchar *bytes) {
521: bytes[0]=word&0xFF;
522: bytes[1]=(word>>8)&0xFF;
523: bytes[2]=(word>>16)&0xFF;
524: bytes[3]=(word>>24)&0xFF;
525: }
526: inline uint uchars2uint(uchar *bytes) {
527: return bytes[3]<<24
528: | bytes[2]<<16
529: | bytes[1]<<8
530: | bytes[0];
531: }
532:
1.172.2.21.2. 7(paf 533:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 534:): size_t length=strlen(s);
7(paf 535:3): memcpy(*cur, s, length); *cur+=length;
(paf 536:): return 0;
537:): };
538:): String::C String::serialize(size_t prolog_length) const {
539:): size_t buf_length=
540:): prolog_length
5(paf 541:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 542:): +length();
543:): C result(new(PointerFreeGC) char[buf_length], buf_length);
544:):
545:): // 1: prolog
546:): char *cur=result.str+prolog_length;
547:):
548:):
549:): // 2: fragments.count
550:): size_t fragments_count=fragments.count();
551:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 552:
1.172.2.21.2. (paf 553:): // 3: lang info
554:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
555:): const Fragment& fragment=i.next();
1.123 paf 556: // lang
1.172.2.21.2. (paf 557:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
558:): // length
559:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
560:): }
561:):
562:): // 4: letters
7(paf 563:3): body.for_each(serialize_body_piece, &cur);
(paf 564:):
565:): return result;
1.113 parser 566: }
1.172.2.21.2. 9(paf 567:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 568:): if(buf_length<=prolog_length)
1.148 paf 569: return false;
1.172.2.21.2. (paf 570:): buf_length-=prolog_length;
1.113 parser 571:
1.172.2.21.2. (paf 572:): // 1: prolog
573:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 574:
1.172.2.21.2. (paf 575:): // 2: fragments.count
576:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
577:): return false;
578:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
579:): buf_length-=sizeof(size_t);
1.128 paf 580:
1.172.2.21.2. (paf 581:): // 3: lang info
582:): size_t total_length=0;
583:): for(size_t f=0; f<fragments_count; f++) {
5(paf 584:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
(paf 585:): if(buf_length<piece_length) // lang+length
1.148 paf 586: return false;
587:
1.172.2.21.2. 5(paf 588:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
(paf 589:): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
590:): fragments+=Fragment(lang, fragment_length);
591:): total_length+=fragment_length;
1.113 parser 592:
1.172.2.21.2. (paf 593:): buf_length-=piece_length;
1.113 parser 594: }
1.172.2.21.2. (paf 595:):
596:): // 4: letters
597:): if(buf_length!=total_length)
598:): return false;
599:):
600:): body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length);
601:):
1.148 paf 602: return true;
1.113 parser 603: }
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