Annotation of parser3/src/main/pa_string.C, revision 1.172.2.21.2.40
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. 0(paf 8:3): static const char* IDENT_STRING_C="$Date: 2003/04/10 11:35:10 $";
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:):
8(paf 45:3): // FILE *err=fopen("append.log", "wt");
6(paf 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;
8(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;
8(paf 66:3): //fprintf(err, "2end=%u\n", fragment_end);fflush(err);
6(paf 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));
8(paf 72:3): //fclose(err);
6(paf 73:3): return;
74:3): }
(paf 75:): }
76:): }
77:):
78:): break;
6(paf 79:3): }
8(paf 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: }
1.172.2.21.2. 8(paf 114:3): String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) {
115:3): append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN);
7(paf 116:3): }
1.140 paf 117:
1.172.2.21.2. 0(paf 118:3): String::String(const String& src): body(src.body) {
119:3): fragments.append(src.fragments);
8(paf 120:3): ASSERT_STRING_INVARIANT(*this);
0(paf 121:3): }
1.28 paf 122:
1.172.2.21.2. 7(paf 123:3): String& String::append_know_length(const char* str, size_t known_length, Language lang) {
9(paf 124:3): if(!known_length || !*str)
125:3): return *this;
126:3):
7(paf 127:3): body.append_know_length(str, known_length);
128:3): fragments+=Fragment(lang, known_length);
129:3):
8(paf 130:3): ASSERT_STRING_INVARIANT(*this);
7(paf 131:3): return *this;
132:3): }
4(paf 133:3): String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
3(paf 134:3): if(!str)
1.9 paf 135: return *this;
1.172.2.21.2. 6(paf 136:3): size_t known_length=helper_length?helper_length:strlen(str);
137:3): if(!known_length)
1.9 paf 138: return *this;
1.122 paf 139:
1.172.2.21.2. 7(paf 140:3): return append_know_length(str, known_length, lang);
1.1 paf 141: }
1.172.2.21.2. 6(paf 142:3): String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
143:3): size_t known_length=helper_length?helper_length:strlen(str);
144:3): if(!known_length)
145:3): return *this;
146:3):
147:3): body.append_strdup_know_length(str, known_length);
148:3): fragments+=Fragment(lang, known_length);
4(paf 149:3):
8(paf 150:3): ASSERT_STRING_INVARIANT(*this);
4(paf 151:3): return *this;
152:3): }
1.1 paf 153:
1.172.2.21.2. 1(paf 154:3): /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
8(paf 155:3): String& String::mid(size_t substr_begin, size_t substr_end) const {
(paf 156:): String& result=*new String;
1.33 paf 157:
1.172.2.21.2. 1(paf 158:3): size_t self_length=length();
6(paf 159:3): substr_begin=min(substr_begin, self_length);
1(paf 160:3): substr_end=min(max(substr_end, substr_begin), self_length);
161:3): if(substr_begin==substr_end)
(paf 162:): return result;
1.52 paf 163:
1.172.2.21.2. (paf 164:): // first: letters themselves
7(paf 165:3): result.body=body.mid(substr_begin, substr_end-substr_begin);
1.46 paf 166:
1.172.2.21.2. (paf 167:): // next: their langs
3(paf 168:3): result.fragments.append_positions(fragments, substr_begin, substr_end);
1.53 paf 169:
1.172.2.21.2. 5(paf 170:3): // SAPI::log("piece of '%s' from %d to %d is '%s'",
1(paf 171:3): //cstr(), substr_begin, substr_end, result.cstr());
8(paf 172:3): ASSERT_STRING_INVARIANT(result);
1.53 paf 173: return result;
1.54 paf 174: }
175:
1.172.2.21.2. 7(paf 176:3): size_t String::pos(const StringBody substr,
5(paf 177:3): size_t this_offset, Language lang) const {
(paf 178:): // first: letters themselves
2(paf 179:3): size_t substr_begin=body.pos(substr, this_offset);
1(paf 180:3): if(substr_begin==CORD_NOT_FOUND)
(paf 181:): return STRING_NOT_FOUND;
182:):
183:): // next: check the lang when specified
184:):
5(paf 185:3): if(lang==L_UNSPECIFIED) // ignore lang?
1(paf 186:3): return substr_begin;
(paf 187:):
188:): // substr must be in one fragment, and fragments' lang must = lang
7(paf 189:3): size_t substr_end=substr_begin+substr.length();
1(paf 190:3): size_t fragment_begin=0;
(paf 191:): size_t fragment_end;
1(paf 192:3): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); fragment_begin=fragment_end) {
3(paf 193:3): const Fragment fragment=i.next();
1(paf 194:3): fragment_end=fragment_begin+fragment.length;
(paf 195:):
1(paf 196:3): if(substr_begin<fragment_begin) // not reached fragments which may include 'result'?
(paf 197:): continue;
2(paf 198:3): if(substr_begin>=fragment_end) // begin of substr OUT of current fragment?
199:3): continue;
(paf 200:):
201:): if(substr_end>fragment_end) // end of substr OUT of current fragment?
2(paf 202:3): throw Exception(0, // (*) see below
203:3): this,
204:3): "searching for '%s' starting from %ud problem: found begin in one fragment, but end in another",
7(paf 205:3): substr.cstr(), this_offset);
(paf 206:):
4(paf 207:3): if(fragment.lang<=lang)
2(paf 208:3): return substr_begin;
209:3): else { // bad lang...
210:3): /// WARNING: this possibly skips assert (*), but it's fast
211:3): substr_begin=body.pos(substr, fragment_end/*...search AFTER for more*/);
212:3): if(substr_begin==CORD_NOT_FOUND)
213:3): return STRING_NOT_FOUND;
214:3):
7(paf 215:3): size_t substr_end=substr_begin+substr.length();
2(paf 216:3): // and continuing with next fragment
217:3): }
(paf 218:): }
219:):
220:): return STRING_NOT_FOUND;
1.58 paf 221: }
222:
1.172.2.21.2. (paf 223:): size_t String::pos(const String& substr,
5(paf 224:3): size_t this_offset, Language lang) const {
(paf 225:): return pos(substr.body, this_offset, lang);
1.60 paf 226: }
227:
1.172.2.14 paf 228: void String::split(ArrayString& result,
1.172.2.21.2. 9(paf 229:3): size_t& pos_after,
230:3): const char* delim,
231:3): Language lang, int limit) const {
(paf 232:): size_t self_length=length();
233:): if(size_t delim_length=strlen(delim)) {
1.60 paf 234: int pos_before;
235: // while we have 'delim'...
1.172.2.21.2. (paf 236:): for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
237:): result+=&mid(pos_after, pos_before);
238:): pos_after=pos_before+delim_length;
1.60 paf 239: }
240: // last piece
1.172.2.21.2. (paf 241:): if(pos_after<self_length && limit) {
242:): result+=&mid(pos_after, self_length);
243:): pos_after=self_length;
1.60 paf 244: }
245: } else { // empty delim
1.172.2.21.2. (paf 246:): result+=this;
247:): pos_after+=self_length;
1.60 paf 248: }
249: }
250:
1.172.2.14 paf 251: void String::split(ArrayString& result,
1.172.2.21.2. 9(paf 252:3): size_t& pos_after,
253:3): const String& delim, Language lang,
254:3): int limit) const {
5(paf 255:3): if(!delim.is_empty()) {
1.60 paf 256: int pos_before;
257: // while we have 'delim'...
258: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.172.2.21.2. (paf 259:): result+=&mid(pos_after, pos_before);
260:): pos_after=pos_before+delim.length();
1.60 paf 261: }
262: // last piece
1.172.2.21.2. (paf 263:): if(pos_after<length() && limit) {
264:): result+=&mid(pos_after, length());
265:): pos_after=length();
1.60 paf 266: }
267: } else { // empty delim
1.172.2.21.2. (paf 268:): result+=this;
269:): pos_after+=length();
1.60 paf 270: }
1.61 paf 271: }
272:
1.172.2.21.2. 2(paf 273:3): static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 274: struct Regex_option {
1.172.2.11 paf 275: const char* keyL;
276: const char* keyU;
1.63 paf 277: int clear, set;
278: int *result;
1.154 paf 279: bool *flag;
1.63 paf 280: } regex_option[]={
1.153 paf 281: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
282: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
283: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
284: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
285: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 286: {"'", 0, 0, 0, 0, &need_pre_post_match},
287: {0}
1.63 paf 288: };
1.171 paf 289: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 290: result[1]=0;
291:
1.172.2.21.2. 2(paf 292:3): if(options && !options->is_empty())
1.153 paf 293: for(Regex_option *o=regex_option; o->keyL; o++)
1.172.2.21.2. 2(paf 294:3): if(options->pos(o->keyL)>=0
295:3): || (o->keyU && options->pos(o->keyU)>=0)) {
1.154 paf 296: if(o->flag)
297: *o->flag=true;
298: else { // result
299: *o->result &= ~o->clear;
300: *o->result |= o->set;
301: }
1.63 paf 302: }
303: }
304:
1.172.2.21.2. (paf 305:): Table* String::match(Charset& source_charset,
9(paf 306:3): const String& regexp,
307:3): const String* options,
308:3): Row_action row_action, void *info,
309:3): bool& just_matched) const {
5(paf 310:3): if(regexp.is_empty())
(paf 311:): throw Exception(0,
312:): 0,
1.73 paf 313: "regexp is empty");
1.154 paf 314:
1.172.2.21.2. (paf 315:): const char* pattern=regexp.cstr();
1.172.2.11 paf 316: const char* errptr;
1.62 paf 317: int erroffset;
1.172.2.21.2. (paf 318:): bool need_pre_post_match=false;
7(paf 319:3): int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
1.172.2.20 paf 320: bool global=option_bits[1]!=0;
1.172.2.21.2. (paf 321:): pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 322: &errptr, &erroffset,
1.172.2.1 paf 323: source_charset.pcre_tables);
1.62 paf 324:
1.67 paf 325: if(!code)
1.172.2.21.2. (paf 326:): throw Exception(0,
327:): ®exp.mid(erroffset, regexp.length()),
1.74 paf 328: "regular expression syntax error - %s", errptr);
1.62 paf 329:
1.172.2.20 paf 330: int subpatterns=pcre_info(code, 0, 0);
331: if(subpatterns<0) {
1.100 parser 332: pcre_free(code);
1.149 paf 333: throw Exception(0,
1.172.2.21.2. (paf 334:): ®exp,
1.76 paf 335: "pcre_info error (%d)",
1.172.2.20 paf 336: subpatterns);
1.63 paf 337: }
338:
1.172.2.21.2. (paf 339:): const char* subject=cstr();
340:): size_t subject_length=strlen(subject);
341:): const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
342:): int ovector[oveclength];
1.155 paf 343:
344: // create table
1.172.2.21.2. 0(paf 345:3): Table::Action_options table_options;
346:3): Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 347:
1.64 paf 348: int exec_option_bits=0;
1.154 paf 349: int prestart=0;
350: int poststart=0;
1.172.2.21.2. (paf 351:): int postfinish=length();
1.63 paf 352: while(true) {
353: int exec_substrings=pcre_exec(code, 0,
1.172.2.21.2. (paf 354:): subject, subject_length, prestart,
355:): exec_option_bits, ovector, oveclength);
1.63 paf 356:
357: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 358: pcre_free(code);
1.172.2.21.2. (paf 359:): row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
1.172.2.20 paf 360: if(global || subpatterns)
1.172.2.21.2. 2(paf 361:3): return &table; // global or with subpatterns=true+result
1.172.2.20 paf 362: else {
1.172.2.21.2. (paf 363:): just_matched=false; return 0; // not global=no result
1.172.2.20 paf 364: }
1.63 paf 365: }
366:
367: if(exec_substrings<0) {
1.100 parser 368: pcre_free(code);
1.172.2.21.2. (paf 369:): throw Exception(0,
370:): ®exp,
1.76 paf 371: "regular expression execute error (%d)",
1.63 paf 372: exec_substrings);
373: }
374:
1.154 paf 375: int prefinish=ovector[0];
376: poststart=ovector[1];
1.172.2.21.2. (paf 377:): ArrayString* row=new ArrayString;
1.172.2.4 paf 378: if(need_pre_post_match) {
1.172.2.21.2. (paf 379:): *row+=&mid(0, prefinish); // .prematch column value
380:): *row+=&mid(prefinish, poststart); // .match
381:): *row+=&mid(poststart, postfinish); // .postmatch
1.172.2.4 paf 382: } else {
1.172.2.21.2. (paf 383:): *row+=0; // .prematch column value
384:): *row+=0; // .match
385:): *row+=0; // .postmatch
1.172.2.4 paf 386: }
1.63 paf 387:
388: for(int i=1; i<exec_substrings; i++) {
1.69 paf 389: // -1:-1 case handled peacefully by mid() itself
1.172.2.21.2. (paf 390:): *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 391: }
392:
1.172.2.20 paf 393: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 394:
1.172.2.20 paf 395: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 396: pcre_free(code);
1.172.2.21.2. (paf 397:): row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
2(paf 398:3): return &table;
1.63 paf 399: }
1.154 paf 400: prestart=poststart;
1.63 paf 401:
402: /*
403: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 404: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 405: */
406: }
1.82 parser 407: }
408:
1.172.2.21.2. 0(paf 409:3): String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
(paf 410:): String& result=*new String();
0(paf 411:3): if(is_empty())
412:3): return result;
1.172.2.4 paf 413:
1.172.2.1 paf 414: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 415:
416: const unsigned char *a;
417: const unsigned char *b;
418: switch(kind) {
419: case CC_UPPER:
420: a=tables+lcc_offset;
421: b=tables+fcc_offset;
422: break;
423: case CC_LOWER:
424: a=tables+lcc_offset;
425: b=0;
426: break;
427: default:
1.172.2.21.2. (paf 428:): throw Exception(0,
429:): this,
1.82 parser 430: "unknown change case kind #%d",
431: static_cast<int>(kind)); // never
432: a=b=0; // calm, compiler
433: break; // never
434: }
435:
1.172.2.21.2. (paf 436:): char* new_cstr=cstrm();
437:): char *dest=new_cstr;
438:): unsigned char index;
439:): for(const char* current=new_cstr; index=(unsigned char)*current; current++) {
440:): unsigned char c=a[index];
441:): if(b)
442:): c=b[c];
1.82 parser 443:
1.172.2.21.2. (paf 444:): *dest++=(char)c;
445:): }
446:): result.body=new_cstr;
3(paf 447:3): result.fragments.append(fragments);
1.89 parser 448:
1.101 parser 449: return result;
450: }
451:
1.172.2.21.2. (paf 452:): const String& String::replace(const Dictionary& dict) const {
453:): String& result=*new String();
454:): const char* old_cstr=cstr();
1(paf 455:3): const char* prematch_begin=old_cstr;
(paf 456:):
1(paf 457:3): const char* current=old_cstr;
458:3): while(*current) {
459:3): if(Table::element_type row=dict.first_that_begins(current)) {
(paf 460:): // prematch
1(paf 461:3): if(size_t prematch_length=current-prematch_begin) {
6(paf 462:3): result.body.append_strdup_know_length(prematch_begin, prematch_length);
4(paf 463:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
1.101 parser 464: }
465:
1.172.2.21.2. (paf 466:): // match
467:):
468:): const String* a=row->get(0);
1(paf 469:3): // skip 'a' in 'current'; move prematch_begin
470:3): current+=a->length(); prematch_begin=current;
1.170 paf 471:
1.172.2.21.2. (paf 472:): if(row->count()>1) { // are there any b?
473:): const String* b=row->get(1);
6(paf 474:3): result<<*b;
(paf 475:): }
476:): } else // simply advance
477:): current++;
478:): }
1.156 paf 479:
1.172.2.21.2. (paf 480:): // postmatch
3(paf 481:3): if(size_t postmatch_length=current-prematch_begin) {
6(paf 482:3): result.body.append_strdup_know_length(prematch_begin, postmatch_length);
4(paf 483:3): result.fragments.append_positions(fragments, prematch_begin-old_cstr, current-old_cstr);
(paf 484:): }
1.156 paf 485:
1.172.2.21.2. 8(paf 486:3): ASSERT_STRING_INVARIANT(result);
1.89 parser 487: return result;
488: }
489:
1.172.2.21.2. (paf 490:): double String::as_double() const {
1.89 parser 491: double result;
1.172.2.21.2. (paf 492:): const char *str=cstr();
493:):
494:): while(*str && isspace(*str))
495:): str++;
496:): if(!*str)
1.162 paf 497: return 0;
1.161 paf 498:
1.102 parser 499: char *error_pos;
1.89 parser 500: // 0xABC
1.172.2.21.2. (paf 501:): if(str[0]=='0')
502:): if(str[1]=='x' || str[1]=='X')
503:): result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 504: else
1.172.2.21.2. (paf 505:): result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
1.89 parser 506: else
1.172.2.21.2. (paf 507:): result=(double)strtod(str, &error_pos);
1.89 parser 508:
1.159 paf 509: while(char c=*error_pos++)
510: if(!isspace(c))
511: throw Exception("number.format",
1.172.2.21.2. (paf 512:): this,
1.159 paf 513: "invalid number (double)");
1.89 parser 514:
515: return result;
516: }
1.172.2.21.2. (paf 517:): int String::as_int() const {
1.89 parser 518: int result;
1.172.2.21.2. (paf 519:): const char *str=cstr();
520:):
521:): while(*str && isspace(*str))
522:): str++;
523:): if(!*str)
1.162 paf 524: return 0;
1.161 paf 525:
1.102 parser 526: char *error_pos;
1.89 parser 527: // 0xABC
1.172.2.21.2. (paf 528:): if(str[0]=='0')
529:): if(str[1]=='x' || str[1]=='X')
530:): result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.99 parser 531: else
1.172.2.21.2. (paf 532:): result=(int)strtol(str+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 533: else
1.172.2.21.2. (paf 534:): result=(int)strtol(str, &error_pos, 0);
1.89 parser 535:
1.159 paf 536: while(char c=*error_pos++)
537: if(!isspace(c))
538: throw Exception("number.format",
1.172.2.21.2. (paf 539:): this,
1.159 paf 540: "invalid number (int)");
1.82 parser 541:
542: return result;
1.61 paf 543: }
1.113 parser 544:
1.172.2.4 paf 545: inline void uint2uchars(uint word, uchar *bytes) {
546: bytes[0]=word&0xFF;
547: bytes[1]=(word>>8)&0xFF;
548: bytes[2]=(word>>16)&0xFF;
549: bytes[3]=(word>>24)&0xFF;
550: }
551: inline uint uchars2uint(uchar *bytes) {
552: return bytes[3]<<24
553: | bytes[2]<<16
554: | bytes[1]<<8
555: | bytes[0];
556: }
557:
1.172.2.21.2. 7(paf 558:3): static int serialize_body_piece(const char* s, char** cur) {
(paf 559:): size_t length=strlen(s);
7(paf 560:3): memcpy(*cur, s, length); *cur+=length;
(paf 561:): return 0;
562:): };
7(paf 563:3): String::Cm String::serialize(size_t prolog_length) const {
(paf 564:): size_t buf_length=
565:): prolog_length
5(paf 566:3): +fragments.count()*(sizeof(Language)+sizeof(size_t))
(paf 567:): +length();
7(paf 568:3): String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
(paf 569:):
570:): // 1: prolog
571:): char *cur=result.str+prolog_length;
572:):
573:):
574:): // 2: fragments.count
575:): size_t fragments_count=fragments.count();
576:): memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.113 parser 577:
1.172.2.21.2. (paf 578:): // 3: lang info
579:): for(Array_iterator<ArrayFragment::element_type> i(fragments); i.has_next(); ) {
3(paf 580:3): const Fragment fragment=i.next();
1.123 paf 581: // lang
1.172.2.21.2. (paf 582:): memcpy(cur, &fragment.lang, sizeof(fragment.lang)); cur+=sizeof(fragment.lang);
583:): // length
584:): memcpy(cur, &fragment.length, sizeof(fragment.length)); cur+=sizeof(fragment.length);
585:): }
586:):
587:): // 4: letters
7(paf 588:3): body.for_each(serialize_body_piece, &cur);
(paf 589:):
590:): return result;
1.113 parser 591: }
1.172.2.21.2. 9(paf 592:3): bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
(paf 593:): if(buf_length<=prolog_length)
1.148 paf 594: return false;
1.172.2.21.2. (paf 595:): buf_length-=prolog_length;
1.113 parser 596:
1.172.2.21.2. (paf 597:): // 1: prolog
598:): const char* cur=(const char* )buf+prolog_length;
1.148 paf 599:
1.172.2.21.2. (paf 600:): // 2: fragments.count
601:): if(buf_length<sizeof(size_t)) // fragments.count don't fit?
602:): return false;
603:): size_t fragments_count=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
604:): buf_length-=sizeof(size_t);
6(paf 605:3):
606:3): if(fragments_count) {
607:3): // 3: lang info
608:3): size_t total_length=0;
609:3): for(size_t f=0; f<fragments_count; f++) {
610:3): size_t piece_length=sizeof(Language)+sizeof(size_t);
611:3): if(buf_length<piece_length) // lang+length
612:3): return false;
613:3):
614:3): Language lang=*reinterpret_cast<const Language *>(cur); cur+=sizeof(Language);
615:3): size_t fragment_length=*reinterpret_cast<const size_t*>(cur); cur+=sizeof(size_t);
616:3): fragments+=Fragment(lang, fragment_length);
617:3): total_length+=fragment_length;
1.128 paf 618:
1.172.2.21.2. 6(paf 619:3): buf_length-=piece_length;
620:3): }
1.148 paf 621:
1.172.2.21.2. 6(paf 622:3): // 4: letters
623:3): if(buf_length!=total_length)
624:3): return false;
1.113 parser 625:
1.172.2.21.2. 6(paf 626:3): body=StringBody(cur, buf_length);
1.113 parser 627: }
1.172.2.21.2. (paf 628:):
8(paf 629:3): ASSERT_STRING_INVARIANT(*this);
1.148 paf 630: return true;
1.113 parser 631: }
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