Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.36
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. 6(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/04/02 10:10:01 $";
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:):
6(paf 45:3): FILE *err=fopen("append.log", "wt");
46:3):
1(paf 47:3): size_t fragment_begin=0;
(paf 48:): size_t fragment_end;
1(paf 49:3): for(Array_iterator<element_type> i(src); ; fragment_begin=fragment_end) {
6(paf 50:3): const Fragment fragment=i.next();
1(paf 51:3): fragment_end=fragment_begin+fragment.length;
6(paf 52:3): fprintf(err, "1end=%u\n", fragment_end);fflush(err);
(paf 53:):
6(paf 54:3): // not reached fragments which may include 'substr'?
55:3): if(!(substr_begin>=fragment_begin && substr_begin<=fragment_end))
56:3): continue;
57:3):
(paf 58:): // found first fragment including piece of 'substr'
6(paf 59:3): if(substr_end<=fragment_end) // fits into first fragment?
60:3): *this+=Fragment(fragment.lang, substr_end-substr_begin);
61:3): else { // spans more then one fragment
62:3): *this+=Fragment(fragment.lang, fragment_end-substr_begin);
63:3): while(true) {
64:3): const Fragment fragment=i.next();
65:3): fragment_end=(fragment_begin=fragment_end)+fragment.length;
66:3): fprintf(err, "2end=%u\n", fragment_end);fflush(err);
67:3):
68:3): if(substr_end>fragment_end) // are there still more?
69:3): append(Fragment(fragment.lang, fragment.length)); // appending whole fragment
70:3): else { // no, it was last
71:3): append(Fragment(fragment.lang, substr_end-fragment_begin));
72:3): fclose(err);
73:3): return;
74:3): }
(paf 75:): }
76:): }
77:):
78:): break;
6(paf 79:3): }
80:3): fclose(err);
(paf 81:): }
82:):
8(paf 83:3): // StringBody methods
5(paf 84:3):
9(paf 85:3): StringBody StringBody::Format(int value) {
8(paf 86:3): char local[MAX_NUMBER];
9(paf 87:3): size_t length=snprintf(local, MAX_NUMBER, "%d", value);
88:3): return StringBody(pa_strdup(local, length), length);
8(paf 89:3): }
5(paf 90:3):
3(paf 91:3): static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
92:3): uint& result=*static_cast<uint*>(client_data);
93:3): generic_hash_code(result, c);
94:3): return 0;
95:3): }
5(paf 96:3): static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
97:3): uint& result=*static_cast<uint*>(client_data);
98:3): generic_hash_code(result, s);
99:3): return 0;
100:3): };
7(paf 101:3): uint StringBody::hash_code() const {
5(paf 102:3): uint result=0;
3(paf 103:3): CORD_iter5(body, 0,
104:3): CORD_batched_iter_fn_generic_hash_code,
105:3): CORD_batched_iter_fn_generic_hash_code, &result);
5(paf 106:3): return result;
107:3): }
108:3):
(paf 109:): // String methods
1.172.2.2 paf 110:
1.172.2.21.2. 6(paf 111:3): String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
112:3): append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.1 paf 113: }
1.140 paf 114:
1.172.2.21.2. 0(paf 115:3): String::String(const String& src): body(src.body) {
116:3): fragments.append(src.fragments);
117:3): }
1.28 paf 118:
1.172.2.21.2. 4(paf 119:3): String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
3(paf 120:3): if(!str)
1.9 paf 121: return *this;
1.172.2.21.2. 6(paf 122:3): size_t known_length=helper_length?helper_length:strlen(str);
123:3): if(!known_length)
1.9 paf 124: return *this;
1.122 paf 125:
1.172.2.21.2. 6(paf 126:3): body.append_know_length(str, known_length);
127:3): fragments+=Fragment(lang, known_length);
1.1 paf 128:
129: return *this;
130: }
1.172.2.21.2. 6(paf 131:3): String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
132:3): size_t known_length=helper_length?helper_length:strlen(str);
133:3): if(!known_length)
134:3): return *this;
135:3):
136:3): body.append_strdup_know_length(str, known_length);
137:3): fragments+=Fragment(lang, known_length);
4(paf 138:3):
139:3): return *this;
140:3): }
1.1 paf 141:
1.172.2.21.2. 1(paf 142:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 143:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 144:): String& result=*new String;
1.33 paf 145:
1.172.2.21.2. 1(paf 146:3): size_t self_length=length();
6(paf 147:3): substr_begin=min(substr_begin, self_length);
1(paf 148:3): substr_end=min(max(substr_end, substr_begin), self_length);
149:3): if(substr_begin==substr_end)
(paf 150:): return result;
1.52 paf 151:
1.172.2.21.2. (paf 152:): // first: letters themselves
7(paf 153:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 154:
1.172.2.21.2. (paf 155:): // next: their langs
3(paf 156:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 157:
1.172.2.21.2. 5(paf 158:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 159:3): //cstr(), substr_begin, substr_end, result.cstr());
1.53 paf 160: return result;
1.54 paf 161: }
162:
1.172.2.21.2. 7(paf 163:3): size_t String::pos(const StringBody substr,
5(paf 164:3): size_t this_offset, Language lang) const {
(paf 165:): // first: letters themselves
2(paf 166:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 167:3): if(substr_begin==CORD_NOT_FOUND)
(paf 168:): return STRING_NOT_FOUND;
169:):
170:): // next: check the lang when specified
171:):
5(paf 172:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 173:3): return substr_begin;
(paf 174:):
175:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 176:3): size_t substr_end=substr_begin+substr.length();
1(paf 177:3): size_t fragment_begin=0;
(paf 178:): size_t fragment_end;
1(paf 179:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
3(paf 180:3): const Fragment fragment=i.next();
1(paf 181:3): fragment_end=fragment_begin+fragment.length;
(paf 182:):
1(paf 183:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 184:): continue;
2(paf 185:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
186:3): continue;
(paf 187:):
188:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 189:3): throw Exception(0, // (*) see below
190:3): this,
191:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 192:3): substr.cstr(), this_offset);
(paf 193:):
4(paf 194:3): if(fragment.lang<=lang)
2(paf 195:3): return substr_begin;
196:3): else { // bad lang...
197:3): /// WARNING: this possibly skips assert (*), but it's fast
198:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
199:3): if(substr_begin==CORD_NOT_FOUND)
200:3): return STRING_NOT_FOUND;
201:3):
7(paf 202:3): size_t substr_end=substr_begin+substr.length();
2(paf 203:3): // and continuing with next fragment
204:3): }
(paf 205:): }
206:):
207:): return STRING_NOT_FOUND;
1.58 paf 208: }
209:
1.172.2.21.2. (paf 210:): size_t String::pos(const String& substr,
5(paf 211:3): size_t this_offset, Language lang) const {
(paf 212:): return pos(substr.body, this_offset, lang);
1.60 paf 213: }
214:
1.172.2.14 paf 215: void String::split(ArrayString& result,
1.172.2.21.2. (paf 216:): size_t& pos_after,
217:): const char* delim,
2(paf 218:3): Language lang, int limit) const {
(paf 219:): size_t self_length=length();
220:): if(size_t delim_length=strlen(delim)) {
1.60 paf 221: int pos_before;
222: // while we have 'delim'...
1.172.2.21.2. (paf 223:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
224:): result+=&mid(pos_after, pos_before);
225:): pos_after=pos_before+delim_length;
1.60 paf 226: }
227: // last piece
1.172.2.21.2. (paf 228:): if(pos_after<self_length && limit) {
229:): result+=&mid(pos_after, self_length);
230:): pos_after=self_length;
1.60 paf 231: }
232: } else { // empty delim
1.172.2.21.2. (paf 233:): result+=this;
234:): pos_after+=self_length;
1.60 paf 235: }
236: }
237:
1.172.2.14 paf 238: void String::split(ArrayString& result,
1.172.2.21.2. (paf 239:): size_t& pos_after,
5(paf 240:3): const String& delim, Language lang,
(paf 241:): int limit) const {
5(paf 242:3): if(!delim.is_empty()) {
1.60 paf 243: int pos_before;
244: // while we have 'delim'...
245: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 246:): result+=&mid(pos_after, pos_before);
247:): pos_after=pos_before+delim.length();
1.60 paf 248: }
249: // last piece
1.172.2.21.2. (paf 250:): if(pos_after<length() && limit) {
251:): result+=&mid(pos_after, length());
252:): pos_after=length();
1.60 paf 253: }
254: } else { // empty delim
1.172.2.21.2. (paf 255:): result+=this;
256:): pos_after+=length();
1.60 paf 257: }
1.61 paf 258: }
259:
1.172.2.21.2. 2(paf 260:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 261: struct Regex_option {
1.172.2.11 paf 262: const char* keyL;
263: const char* keyU;
1.63 paf 264: int clear, set;
265: int *result;
1.154 paf 266: bool *flag;
1.63 paf 267: } regex_option[]={
1.153 paf 268: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
269: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
270: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
271: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
272: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 273: {"'", 0, 0, 0, 0, &need_pre_post_match},
274: {0}
1.63 paf 275: };
1.171 paf 276: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 277: result[1]=0;
278:
1.172.2.21.2. 2(paf 279:3): if(options && !options->is_empty())
1.153 paf 280: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 281:3): if(options->pos(o->keyL)>=0
282:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 283: if(o->flag)
284: *o->flag=true;
285: else { // result
286: *o->result &= ~o->clear;
287: *o->result |= o->set;
288: }
1.63 paf 289: }
290: }
291:
1.172.2.21.2. (paf 292:): Table* String::match(Charset& source_charset,
1.172.2.4 paf 293: const String& regexp,
1.172.2.21.2. 2(paf 294:3): const String* options,
1.172.2.4 paf 295: Row_action row_action, void *info,
1.172.2.20 paf 296: bool& just_matched) const {
1.172.2.21.2. 5(paf 297:3): if(regexp.is_empty())
(paf 298:): throw Exception(0,
299:): 0,
1.73 paf 300: "regexp is empty");
1.154 paf 301:
1.172.2.21.2. (paf 302:): const char* pattern=regexp.cstr();
1.172.2.11 paf 303: const char* errptr;
1.62 paf 304: int erroffset;
1.172.2.21.2. (paf 305:): bool need_pre_post_match=false;
7(paf 306:3): int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 307: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 308:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 309: &errptr, &erroffset,
1.172.2.1 paf 310: source_charset.pcre_tables);
1.62 paf 311:
1.67 paf 312: if(!code)
1.172.2.21.2. (paf 313:): throw Exception(0,
314:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 315: "regular expression syntax error - %s", errptr);
1.62 paf 316:
1.172.2.20 paf 317: int subpatterns=pcre_info(code, 0, 0);
318: if(subpatterns<0) {
1.100 parser 319: pcre_free(code);
1.149 paf 320: throw Exception(0,
1.172.2.21.2. (paf 321:): ®exp,
1.76 paf 322: "pcre_info error (%d)",
1.172.2.20 paf 323: subpatterns);
1.63 paf 324: }
325:
1.172.2.21.2. (paf 326:): const char* subject=cstr();
327:): size_t subject_length=strlen(subject);
328:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
329:): int ovector[oveclength];
1.155 paf 330:
331: // create table
1.172.2.21.2. 1(paf 332:3): Table& table=*new Table(string_match_table_template);
1.63 paf 333:
1.64 paf 334: int exec_option_bits=0;
1.154 paf 335: int prestart=0;
336: int poststart=0;
1.172.2.21.2. (paf 337:): int postfinish=length();
1.63 paf 338: while(true) {
339: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 340:): subject, subject_length, prestart,
341:): exec_option_bits, ovector, oveclength);
1.63 paf 342:
343: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 344: pcre_free(code);
1.172.2.21.2. (paf 345:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 346: if(global || subpatterns)
1.172.2.21.2. 2(paf 347:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 348: else {
1.172.2.21.2. (paf 349:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 350: }
1.63 paf 351: }
352:
353: if(exec_substrings<0) {
1.100 parser 354: pcre_free(code);
1.172.2.21.2. (paf 355:): throw Exception(0,
356:): ®exp,
1.76 paf 357: "regular expression execute error (%d)",
1.63 paf 358: exec_substrings);
359: }
360:
1.154 paf 361: int prefinish=ovector[0];
362: poststart=ovector[1];
1.172.2.21.2. (paf 363:): ArrayString* row=new ArrayString;
1.172.2.4 paf 364: if(need_pre_post_match) {
1.172.2.21.2. (paf 365:): *row+=&mid(0, prefinish); // .prematch column value
366:): *row+=&mid(prefinish, poststart); // .match
367:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 368: } else {
1.172.2.21.2. (paf 369:): *row+=0; // .prematch column value
370:): *row+=0; // .match
371:): *row+=0; // .postmatch
1.172.2.4 paf 372: }
1.63 paf 373:
374: for(int i=1; i<exec_substrings; i++) {
1.69 paf 375: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 376:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 377: }
378:
1.172.2.20 paf 379: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 380:
1.172.2.20 paf 381: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 382: pcre_free(code);
1.172.2.21.2. (paf 383:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 384:3): return &table;
1.63 paf 385: }
1.154 paf 386: prestart=poststart;
1.63 paf 387:
388: /*
389: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 390: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 391: */
392: }
1.82 parser 393: }
394:
1.172.2.21.2. 0(paf 395:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 396:): String& result=*new String();
0(paf 397:3): if(is_empty())
398:3): return result;
1.172.2.4 paf 399:
1.172.2.1 paf 400: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 401:
402: const unsigned char *a;
403: const unsigned char *b;
404: switch(kind) {
405: case CC_UPPER:
406: a=tables+lcc_offset;
407: b=tables+fcc_offset;
408: break;
409: case CC_LOWER:
410: a=tables+lcc_offset;
411: b=0;
412: break;
413: default:
1.172.2.21.2. (paf 414:): throw Exception(0,
415:): this,
1.82 parser 416: "unknown change case kind #%d",
417: static_cast<int>(kind)); // never
418: a=b=0; // calm, compiler
419: break; // never
420: }
421:
1.172.2.21.2. (paf 422:): char* new_cstr=cstrm();
423:): char *dest=new_cstr;
424:): unsigned char index;
425:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
426:): unsigned char c=a[index];
427:): if(b)
428:): c=b[c];
1.82 parser 429:
1.172.2.21.2. (paf 430:): *dest++=(char)c;
431:): }
432:): result.body=new_cstr;
3(paf 433:3): result.fragments.append(fragments);
1.89 parser 434:
1.101 parser 435: return result;
436: }
437:
1.172.2.21.2. (paf 438:): const String& String::replace(const Dictionary& dict) const {
439:): String& result=*new String();
440:): const char* old_cstr=cstr();
1(paf 441:3): const char* prematch_begin=old_cstr;
(paf 442:):
1(paf 443:3): const char* current=old_cstr;
444:3): while(*current) {
445:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 446:): // prematch
1(paf 447:3): if(size_t prematch_length=current-prematch_begin) {
6(paf 448:3): result.body.append_strdup_know_length(prematch_begin, prematch_length);
4(paf 449:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 450: }
451:
1.172.2.21.2. (paf 452:): // match
453:):
454:): const String* a=row->get(0);
1(paf 455:3): // skip 'a' in 'current'; move prematch_begin
456:3): current+=a->length(); prematch_begin=current;
1.170 paf 457:
1.172.2.21.2. (paf 458:): if(row->count()>1) { // are there any b?
459:): const String* b=row->get(1);
6(paf 460:3): result<<*b;
(paf 461:): }
462:): } else // simply advance
463:): current++;
464:): }
1.156 paf 465:
1.172.2.21.2. (paf 466:): // postmatch
3(paf 467:3): if(size_t postmatch_length=current-prematch_begin) {
6(paf 468:3): result.body.append_strdup_know_length(prematch_begin, postmatch_length);
4(paf 469:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 470:): }
1.156 paf 471:
1.89 parser 472: return result;
473: }
474:
1.172.2.21.2. (paf 475:): double String::as_double() const {
1.89 parser 476: double result;
1.172.2.21.2. (paf 477:): const char *str=cstr();
478:):
479:): while(*str && isspace(*str))
480:): str++;
481:): if(!*str)
1.162 paf 482: return 0;
1.161 paf 483:
1.102 parser 484: char *error_pos;
1.89 parser 485: // 0xABC
1.172.2.21.2. (paf 486:): if(str[0]=='0')
487:): if(str[1]=='x' || str[1]=='X')
488:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 489: else
1.172.2.21.2. (paf 490:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 491: else
1.172.2.21.2. (paf 492:): result=(double)strtod(str, &error_pos);
1.89 parser 493:
1.159 paf 494: while(char c=*error_pos++)
495: if(!isspace(c))
496: throw Exception("number.format",
1.172.2.21.2. (paf 497:): this,
1.159 paf 498: "invalid number (double)");
1.89 parser 499:
500: return result;
501: }
1.172.2.21.2. (paf 502:): int String::as_int() const {
1.89 parser 503: int result;
1.172.2.21.2. (paf 504:): const char *str=cstr();
505:):
506:): while(*str && isspace(*str))
507:): str++;
508:): if(!*str)
1.162 paf 509: return 0;
1.161 paf 510:
1.102 parser 511: char *error_pos;
1.89 parser 512: // 0xABC
1.172.2.21.2. (paf 513:): if(str[0]=='0')
514:): if(str[1]=='x' || str[1]=='X')
515:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 516: else
1.172.2.21.2. (paf 517:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 518: else
1.172.2.21.2. (paf 519:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 520:
1.159 paf 521: while(char c=*error_pos++)
522: if(!isspace(c))
523: throw Exception("number.format",
1.172.2.21.2. (paf 524:): this,
1.159 paf 525: "invalid number (int)");
1.82 parser 526:
527: return result;
1.61 paf 528: }
1.113 parser 529:
1.172.2.4 paf 530: inline void uint2uchars(uint word, uchar *bytes) {
531: bytes[0]=word&0xFF;
532: bytes[1]=(word>>8)&0xFF;
533: bytes[2]=(word>>16)&0xFF;
534: bytes[3]=(word>>24)&0xFF;
535: }
536: inline uint uchars2uint(uchar *bytes) {
537: return bytes[3]<<24
538: | bytes[2]<<16
539: | bytes[1]<<8
540: | bytes[0];
541: }
542:
1.172.2.21.2. 7(paf 543:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 544:): size_t length=strlen(s);
7(paf 545:3): memcpy(*cur, s, length); *cur+=length;
(paf 546:): return 0;
547:): };
548:): String::C String::serialize(size_t prolog_length) const {
549:): size_t buf_length=
550:): prolog_length
5(paf 551:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 552:): +length();
553:): C result(new(PointerFreeGC) char[buf_length], buf_length);
554:):
555:): // 1: prolog
556:): char *cur=result.str+prolog_length;
557:):
558:):
559:): // 2: fragments.count
560:): size_t fragments_count=fragments.count();
561:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 562:
1.172.2.21.2. (paf 563:): // 3: lang info
564:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
3(paf 565:3): const Fragment fragment=i.next();
1.123 paf 566: // lang
1.172.2.21.2. (paf 567:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
568:): // length
569:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
570:): }
571:):
572:): // 4: letters
7(paf 573:3): body.for_each(serialize_body_piece, &cur);
(paf 574:):
575:): return result;
1.113 parser 576: }
1.172.2.21.2. 9(paf 577:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 578:): if(buf_length<=prolog_length)
1.148 paf 579: return false;
1.172.2.21.2. (paf 580:): buf_length-=prolog_length;
1.113 parser 581:
1.172.2.21.2. (paf 582:): // 1: prolog
583:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 584:
1.172.2.21.2. (paf 585:): // 2: fragments.count
586:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
587:): return false;
588:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
589:): buf_length-=sizeof(size_t);
6(paf 590:3):
591:3): if(fragments_count) {
592:3): // 3: lang info
593:3): size_t total_length=0;
594:3): for(size_t f=0; f<fragments_count; f++) {
595:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
596:3): if(buf_length<piece_length) // lang+length
597:3): return false;
598:3):
599:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
600:3): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
601:3): fragments+=Fragment(lang, fragment_length);
602:3): total_length+=fragment_length;
1.128 paf 603:
1.172.2.21.2. 6(paf 604:3): buf_length-=piece_length;
605:3): }
1.148 paf 606:
1.172.2.21.2. 6(paf 607:3): // 4: letters
608:3): if(buf_length!=total_length)
609:3): return false;
1.113 parser 610:
1.172.2.21.2. 6(paf 611:3): body=StringBody(cur, buf_length);
1.113 parser 612: }
1.172.2.21.2. (paf 613:):
1.148 paf 614: return true;
1.113 parser 615: }
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