Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.35
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. 5(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/03/27 14:51:29 $";
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. 5(paf 31:3): *this+=new String(StringBody::Format(1+i), String::L_CLEAN);
1.172.2.2 paf 32: }
1.172.2.21 paf 33: }
1.172.2.4 paf 34: };
35:
1.172.2.21.2. (paf 36:): Table string_match_table_template(new String_match_table_template_columns);
1.172.2.2 paf 37:
1.172.2.21.2. (paf 38:): // String::ArrayFragment methods
39:):
3(paf 40:3): void String::ArrayFragment::append_positions(const ArrayFragment& src,
41:3): size_t substr_begin, size_t substr_end) {
1(paf 42:3): if(substr_begin==substr_end)
3(paf 43:3): return;
(paf 44:):
1(paf 45:3): size_t fragment_begin=0;
(paf 46:): size_t fragment_end;
1(paf 47:3): for(Array_iterator<element_type> i(src); ; fragment_begin=fragment_end) {
3(paf 48:3): const element_type fragment=i.next();
1(paf 49:3): fragment_end=fragment_begin+fragment.length;
(paf 50:):
5(paf 51:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'substr'?
(paf 52:): continue;
53:):
6(paf 54:3): if(substr_begin>fragment_end) // not reached fragments which may include 'substr'?
55:3): continue;
56:3):
(paf 57:): // found first fragment including piece of 'substr'
58:): size_t piece_end=min(substr_end, fragment_end);
1(paf 59:3): *this+=Fragment(fragment.lang, piece_end-substr_begin);
(paf 60:):
61:): while(substr_end>fragment_end) { // are there more fragments including pieces of 'substr'?
1(paf 62:3): fragment_begin=fragment_end;
63:3):
3(paf 64:3): const element_type fragment=i.next();
1(paf 65:3): fragment_end=fragment_begin+fragment.length;
(paf 66:):
67:): if(substr_end>fragment_end) // are there still more?
3(paf 68:3): append(Fragment(fragment.lang, fragment.length)); // appending whole fragment
(paf 69:): else { // no, it was last
3(paf 70:3): append(Fragment(fragment.lang, substr_end-fragment_begin));
71:3): return;
(paf 72:): }
73:): }
74:):
75:): break;
76:): }
77:): }
78:):
8(paf 79:3): // StringBody methods
5(paf 80:3):
9(paf 81:3): StringBody StringBody::Format(int value) {
8(paf 82:3): char local[MAX_NUMBER];
9(paf 83:3): size_t length=snprintf(local, MAX_NUMBER, "%d", value);
84:3): return StringBody(pa_strdup(local, length), length);
8(paf 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. 4(paf 107:3): String::String(const char* cstr, size_t help_length, bool tainted): body(CORD_EMPTY) {
108:3): append_help_length(cstr, help_length, tainted?L_TAINTED:L_CLEAN);
1.1 paf 109: }
1.140 paf 110:
1.172.2.21.2. 0(paf 111:3): String::String(const String& src): body(src.body) {
112:3): fragments.append(src.fragments);
113:3): }
1.28 paf 114:
1.172.2.21.2. 4(paf 115:3): String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
3(paf 116:3): if(!str)
1.9 paf 117: return *this;
1.172.2.21.2. 4(paf 118:3): size_t use_length=helper_length?helper_length:strlen(str);
0(paf 119:3): if(!use_length)
1.9 paf 120: return *this;
1.122 paf 121:
1.172.2.21.2. 4(paf 122:3): body.append_help_length(str, use_length);
0(paf 123:3): fragments+=Fragment(lang, use_length);
1.1 paf 124:
125: return *this;
126: }
1.172.2.21.2. 4(paf 127:3): String& String::append_strdup(const char* str, size_t length, Language lang) {
128:3): body.append_strdup(str, length);
129:3): fragments+=Fragment(lang, length);
130:3):
131:3): return *this;
132:3): }
1.1 paf 133:
1.172.2.21.2. 1(paf 134:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 135:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 136:): String& result=*new String;
1.33 paf 137:
1.172.2.21.2. 1(paf 138:3): size_t self_length=length();
139:3): substr_begin=max(min(substr_begin, self_length), (size_t)0);
140:3): substr_end=min(max(substr_end, substr_begin), self_length);
141:3): if(substr_begin==substr_end)
(paf 142:): return result;
1.52 paf 143:
1.172.2.21.2. (paf 144:): // first: letters themselves
7(paf 145:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 146:
1.172.2.21.2. (paf 147:): // next: their langs
3(paf 148:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 149:
1.172.2.21.2. 5(paf 150:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 151:3): //cstr(), substr_begin, substr_end, result.cstr());
1.53 paf 152: return result;
1.54 paf 153: }
154:
1.172.2.21.2. 7(paf 155:3): size_t String::pos(const StringBody substr,
5(paf 156:3): size_t this_offset, Language lang) const {
(paf 157:): // first: letters themselves
2(paf 158:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 159:3): if(substr_begin==CORD_NOT_FOUND)
(paf 160:): return STRING_NOT_FOUND;
161:):
162:): // next: check the lang when specified
163:):
5(paf 164:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 165:3): return substr_begin;
(paf 166:):
167:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 168:3): size_t substr_end=substr_begin+substr.length();
1(paf 169:3): size_t fragment_begin=0;
(paf 170:): size_t fragment_end;
1(paf 171:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
3(paf 172:3): const Fragment fragment=i.next();
1(paf 173:3): fragment_end=fragment_begin+fragment.length;
(paf 174:):
1(paf 175:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 176:): continue;
2(paf 177:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
178:3): continue;
(paf 179:):
180:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 181:3): throw Exception(0, // (*) see below
182:3): this,
183:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 184:3): substr.cstr(), this_offset);
(paf 185:):
4(paf 186:3): if(fragment.lang<=lang)
2(paf 187:3): return substr_begin;
188:3): else { // bad lang...
189:3): /// WARNING: this possibly skips assert (*), but it's fast
190:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
191:3): if(substr_begin==CORD_NOT_FOUND)
192:3): return STRING_NOT_FOUND;
193:3):
7(paf 194:3): size_t substr_end=substr_begin+substr.length();
2(paf 195:3): // and continuing with next fragment
196:3): }
(paf 197:): }
198:):
199:): return STRING_NOT_FOUND;
1.58 paf 200: }
201:
1.172.2.21.2. (paf 202:): size_t String::pos(const String& substr,
5(paf 203:3): size_t this_offset, Language lang) const {
(paf 204:): return pos(substr.body, this_offset, lang);
1.60 paf 205: }
206:
1.172.2.14 paf 207: void String::split(ArrayString& result,
1.172.2.21.2. (paf 208:): size_t& pos_after,
209:): const char* delim,
2(paf 210:3): Language lang, int limit) const {
(paf 211:): size_t self_length=length();
212:): if(size_t delim_length=strlen(delim)) {
1.60 paf 213: int pos_before;
214: // while we have 'delim'...
1.172.2.21.2. (paf 215:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
216:): result+=&mid(pos_after, pos_before);
217:): pos_after=pos_before+delim_length;
1.60 paf 218: }
219: // last piece
1.172.2.21.2. (paf 220:): if(pos_after<self_length && limit) {
221:): result+=&mid(pos_after, self_length);
222:): pos_after=self_length;
1.60 paf 223: }
224: } else { // empty delim
1.172.2.21.2. (paf 225:): result+=this;
226:): pos_after+=self_length;
1.60 paf 227: }
228: }
229:
1.172.2.14 paf 230: void String::split(ArrayString& result,
1.172.2.21.2. (paf 231:): size_t& pos_after,
5(paf 232:3): const String& delim, Language lang,
(paf 233:): int limit) const {
5(paf 234:3): if(!delim.is_empty()) {
1.60 paf 235: int pos_before;
236: // while we have 'delim'...
237: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 238:): result+=&mid(pos_after, pos_before);
239:): pos_after=pos_before+delim.length();
1.60 paf 240: }
241: // last piece
1.172.2.21.2. (paf 242:): if(pos_after<length() && limit) {
243:): result+=&mid(pos_after, length());
244:): pos_after=length();
1.60 paf 245: }
246: } else { // empty delim
1.172.2.21.2. (paf 247:): result+=this;
248:): pos_after+=length();
1.60 paf 249: }
1.61 paf 250: }
251:
1.172.2.21.2. 2(paf 252:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 253: struct Regex_option {
1.172.2.11 paf 254: const char* keyL;
255: const char* keyU;
1.63 paf 256: int clear, set;
257: int *result;
1.154 paf 258: bool *flag;
1.63 paf 259: } regex_option[]={
1.153 paf 260: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
261: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
262: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
263: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
264: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 265: {"'", 0, 0, 0, 0, &need_pre_post_match},
266: {0}
1.63 paf 267: };
1.171 paf 268: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 269: result[1]=0;
270:
1.172.2.21.2. 2(paf 271:3): if(options && !options->is_empty())
1.153 paf 272: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 273:3): if(options->pos(o->keyL)>=0
274:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 275: if(o->flag)
276: *o->flag=true;
277: else { // result
278: *o->result &= ~o->clear;
279: *o->result |= o->set;
280: }
1.63 paf 281: }
282: }
283:
1.172.2.21.2. (paf 284:): Table* String::match(Charset& source_charset,
1.172.2.4 paf 285: const String& regexp,
1.172.2.21.2. 2(paf 286:3): const String* options,
1.172.2.4 paf 287: Row_action row_action, void *info,
1.172.2.20 paf 288: bool& just_matched) const {
1.172.2.21.2. 5(paf 289:3): if(regexp.is_empty())
(paf 290:): throw Exception(0,
291:): 0,
1.73 paf 292: "regexp is empty");
1.154 paf 293:
1.172.2.21.2. (paf 294:): const char* pattern=regexp.cstr();
1.172.2.11 paf 295: const char* errptr;
1.62 paf 296: int erroffset;
1.172.2.21.2. (paf 297:): bool need_pre_post_match=false;
7(paf 298:3): int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 299: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 300:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 301: &errptr, &erroffset,
1.172.2.1 paf 302: source_charset.pcre_tables);
1.62 paf 303:
1.67 paf 304: if(!code)
1.172.2.21.2. (paf 305:): throw Exception(0,
306:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 307: "regular expression syntax error - %s", errptr);
1.62 paf 308:
1.172.2.20 paf 309: int subpatterns=pcre_info(code, 0, 0);
310: if(subpatterns<0) {
1.100 parser 311: pcre_free(code);
1.149 paf 312: throw Exception(0,
1.172.2.21.2. (paf 313:): ®exp,
1.76 paf 314: "pcre_info error (%d)",
1.172.2.20 paf 315: subpatterns);
1.63 paf 316: }
317:
1.172.2.21.2. (paf 318:): const char* subject=cstr();
319:): size_t subject_length=strlen(subject);
320:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
321:): int ovector[oveclength];
1.155 paf 322:
323: // create table
1.172.2.21.2. 1(paf 324:3): Table& table=*new Table(string_match_table_template);
1.63 paf 325:
1.64 paf 326: int exec_option_bits=0;
1.154 paf 327: int prestart=0;
328: int poststart=0;
1.172.2.21.2. (paf 329:): int postfinish=length();
1.63 paf 330: while(true) {
331: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 332:): subject, subject_length, prestart,
333:): exec_option_bits, ovector, oveclength);
1.63 paf 334:
335: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 336: pcre_free(code);
1.172.2.21.2. (paf 337:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 338: if(global || subpatterns)
1.172.2.21.2. 2(paf 339:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 340: else {
1.172.2.21.2. (paf 341:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 342: }
1.63 paf 343: }
344:
345: if(exec_substrings<0) {
1.100 parser 346: pcre_free(code);
1.172.2.21.2. (paf 347:): throw Exception(0,
348:): ®exp,
1.76 paf 349: "regular expression execute error (%d)",
1.63 paf 350: exec_substrings);
351: }
352:
1.154 paf 353: int prefinish=ovector[0];
354: poststart=ovector[1];
1.172.2.21.2. (paf 355:): ArrayString* row=new ArrayString;
1.172.2.4 paf 356: if(need_pre_post_match) {
1.172.2.21.2. (paf 357:): *row+=&mid(0, prefinish); // .prematch column value
358:): *row+=&mid(prefinish, poststart); // .match
359:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 360: } else {
1.172.2.21.2. (paf 361:): *row+=0; // .prematch column value
362:): *row+=0; // .match
363:): *row+=0; // .postmatch
1.172.2.4 paf 364: }
1.63 paf 365:
366: for(int i=1; i<exec_substrings; i++) {
1.69 paf 367: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 368:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 369: }
370:
1.172.2.20 paf 371: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 372:
1.172.2.20 paf 373: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 374: pcre_free(code);
1.172.2.21.2. (paf 375:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 376:3): return &table;
1.63 paf 377: }
1.154 paf 378: prestart=poststart;
1.63 paf 379:
380: /*
381: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 382: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 383: */
384: }
1.82 parser 385: }
386:
1.172.2.21.2. 0(paf 387:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 388:): String& result=*new String();
0(paf 389:3): if(is_empty())
390:3): return result;
1.172.2.4 paf 391:
1.172.2.1 paf 392: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 393:
394: const unsigned char *a;
395: const unsigned char *b;
396: switch(kind) {
397: case CC_UPPER:
398: a=tables+lcc_offset;
399: b=tables+fcc_offset;
400: break;
401: case CC_LOWER:
402: a=tables+lcc_offset;
403: b=0;
404: break;
405: default:
1.172.2.21.2. (paf 406:): throw Exception(0,
407:): this,
1.82 parser 408: "unknown change case kind #%d",
409: static_cast<int>(kind)); // never
410: a=b=0; // calm, compiler
411: break; // never
412: }
413:
1.172.2.21.2. (paf 414:): char* new_cstr=cstrm();
415:): char *dest=new_cstr;
416:): unsigned char index;
417:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
418:): unsigned char c=a[index];
419:): if(b)
420:): c=b[c];
1.82 parser 421:
1.172.2.21.2. (paf 422:): *dest++=(char)c;
423:): }
424:): result.body=new_cstr;
3(paf 425:3): result.fragments.append(fragments);
1.89 parser 426:
1.101 parser 427: return result;
428: }
429:
1.172.2.21.2. (paf 430:): const String& String::replace(const Dictionary& dict) const {
431:): String& result=*new String();
432:): const char* old_cstr=cstr();
1(paf 433:3): const char* prematch_begin=old_cstr;
(paf 434:):
1(paf 435:3): const char* current=old_cstr;
436:3): while(*current) {
437:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 438:): // prematch
1(paf 439:3): if(size_t prematch_length=current-prematch_begin) {
4(paf 440:3): result.body.append_strdup(prematch_begin, prematch_length);
441:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 442: }
443:
1.172.2.21.2. (paf 444:): // match
445:):
446:): const String* a=row->get(0);
1(paf 447:3): // skip 'a' in 'current'; move prematch_begin
448:3): current+=a->length(); prematch_begin=current;
1.170 paf 449:
1.172.2.21.2. (paf 450:): if(row->count()>1) { // are there any b?
451:): const String* b=row->get(1);
452:): if(size_t b_length=b->length()) {
1(paf 453:3): result.body<<b->cstr_to_string_body();
3(paf 454:3): result.fragments.append(b->fragments);
(paf 455:): }
456:): }
457:): } else // simply advance
458:): current++;
459:): }
1.156 paf 460:
1.172.2.21.2. (paf 461:): // postmatch
3(paf 462:3): if(size_t postmatch_length=current-prematch_begin) {
4(paf 463:3): result.body.append_strdup(prematch_begin, postmatch_length);
464:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 465:): }
1.156 paf 466:
1.89 parser 467: return result;
468: }
469:
1.172.2.21.2. (paf 470:): double String::as_double() const {
1.89 parser 471: double result;
1.172.2.21.2. (paf 472:): const char *str=cstr();
473:):
474:): while(*str && isspace(*str))
475:): str++;
476:): if(!*str)
1.162 paf 477: return 0;
1.161 paf 478:
1.102 parser 479: char *error_pos;
1.89 parser 480: // 0xABC
1.172.2.21.2. (paf 481:): if(str[0]=='0')
482:): if(str[1]=='x' || str[1]=='X')
483:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 484: else
1.172.2.21.2. (paf 485:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 486: else
1.172.2.21.2. (paf 487:): result=(double)strtod(str, &error_pos);
1.89 parser 488:
1.159 paf 489: while(char c=*error_pos++)
490: if(!isspace(c))
491: throw Exception("number.format",
1.172.2.21.2. (paf 492:): this,
1.159 paf 493: "invalid number (double)");
1.89 parser 494:
495: return result;
496: }
1.172.2.21.2. (paf 497:): int String::as_int() const {
1.89 parser 498: int result;
1.172.2.21.2. (paf 499:): const char *str=cstr();
500:):
501:): while(*str && isspace(*str))
502:): str++;
503:): if(!*str)
1.162 paf 504: return 0;
1.161 paf 505:
1.102 parser 506: char *error_pos;
1.89 parser 507: // 0xABC
1.172.2.21.2. (paf 508:): if(str[0]=='0')
509:): if(str[1]=='x' || str[1]=='X')
510:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 511: else
1.172.2.21.2. (paf 512:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 513: else
1.172.2.21.2. (paf 514:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 515:
1.159 paf 516: while(char c=*error_pos++)
517: if(!isspace(c))
518: throw Exception("number.format",
1.172.2.21.2. (paf 519:): this,
1.159 paf 520: "invalid number (int)");
1.82 parser 521:
522: return result;
1.61 paf 523: }
1.113 parser 524:
1.172.2.4 paf 525: inline void uint2uchars(uint word, uchar *bytes) {
526: bytes[0]=word&0xFF;
527: bytes[1]=(word>>8)&0xFF;
528: bytes[2]=(word>>16)&0xFF;
529: bytes[3]=(word>>24)&0xFF;
530: }
531: inline uint uchars2uint(uchar *bytes) {
532: return bytes[3]<<24
533: | bytes[2]<<16
534: | bytes[1]<<8
535: | bytes[0];
536: }
537:
1.172.2.21.2. 7(paf 538:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 539:): size_t length=strlen(s);
7(paf 540:3): memcpy(*cur, s, length); *cur+=length;
(paf 541:): return 0;
542:): };
543:): String::C String::serialize(size_t prolog_length) const {
544:): size_t buf_length=
545:): prolog_length
5(paf 546:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 547:): +length();
548:): C result(new(PointerFreeGC) char[buf_length], buf_length);
549:):
550:): // 1: prolog
551:): char *cur=result.str+prolog_length;
552:):
553:):
554:): // 2: fragments.count
555:): size_t fragments_count=fragments.count();
556:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 557:
1.172.2.21.2. (paf 558:): // 3: lang info
559:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
3(paf 560:3): const Fragment fragment=i.next();
1.123 paf 561: // lang
1.172.2.21.2. (paf 562:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
563:): // length
564:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
565:): }
566:):
567:): // 4: letters
7(paf 568:3): body.for_each(serialize_body_piece, &cur);
(paf 569:):
570:): return result;
1.113 parser 571: }
1.172.2.21.2. 9(paf 572:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 573:): if(buf_length<=prolog_length)
1.148 paf 574: return false;
1.172.2.21.2. (paf 575:): buf_length-=prolog_length;
1.113 parser 576:
1.172.2.21.2. (paf 577:): // 1: prolog
578:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 579:
1.172.2.21.2. (paf 580:): // 2: fragments.count
581:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
582:): return false;
583:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
584:): buf_length-=sizeof(size_t);
1.128 paf 585:
1.172.2.21.2. (paf 586:): // 3: lang info
587:): size_t total_length=0;
588:): for(size_t f=0; f<fragments_count; f++) {
5(paf 589:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
(paf 590:): if(buf_length<piece_length) // lang+length
1.148 paf 591: return false;
592:
1.172.2.21.2. 5(paf 593:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
(paf 594:): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
595:): fragments+=Fragment(lang, fragment_length);
596:): total_length+=fragment_length;
1.113 parser 597:
1.172.2.21.2. (paf 598:): buf_length-=piece_length;
1.113 parser 599: }
1.172.2.21.2. (paf 600:):
601:): // 4: letters
602:): if(buf_length!=total_length)
603:): return false;
604:):
605:): body=CORD_cat_char_star(CORD_EMPTY, cur, buf_length);
606:):
1.148 paf 607: return true;
1.113 parser 608: }
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