Annotation of parser3/src/main/pa_string.C, revision 1.225
1.45 paf 1: /** @file
1.174 paf 2: Parser: string class. @see untalength_t.C.
1.46 paf 3:
1.219 misha 4: Copyright (c) 2001-2009 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:
1.225 ! misha 8: static const char * const IDENT_STRING_C="$Date: 2009-05-13 07:35:42 $";
1.70 paf 9:
1.12 paf 10: #include "pa_string.h"
1.22 paf 11: #include "pa_exception.h"
1.61 paf 12: #include "pa_table.h"
1.101 parser 13: #include "pa_dictionary.h"
1.132 paf 14: #include "pa_charset.h"
1.222 misha 15: #include "pa_vregex.h"
1.215 misha 16:
1.185 paf 17: const String String::Empty;
18:
1.193 paf 19: int pa_atoi(const char* str, const String* problem_source) {
20: if(!str)
21: return 0;
22:
1.199 paf 23: while(*str && isspace((unsigned char)*str))
1.193 paf 24: str++;
25: if(!*str)
26: return 0;
27:
28: int result;
29: char *error_pos;
1.200 paf 30: bool negative=false;
31: if(str[0]=='-') {
32: negative=true;
33: str++;
34: } else if(str[0]=='+') {
35: str++;
36: }
1.193 paf 37: // 0xABC
38: if(str[0]=='0')
39: if(str[1]=='x' || str[1]=='X')
40: result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.197 paf 41: else {
42: // skip leading 0000, to disable octal interpretation
43: do str++; while(*str=='0');
44: result=(int)strtol(str, &error_pos, 0);
45: }
1.193 paf 46: else
47: result=(int)strtol(str, &error_pos, 0);
1.200 paf 48: if(negative)
49: result=-result;
1.193 paf 50:
51: while(char c=*error_pos++)
1.199 paf 52: if(!isspace((unsigned char)c))
1.193 paf 53: throw Exception("number.format",
54: problem_source,
55: problem_source?"invalid number (int)": "'%s' is invalid number (int)", str);
56:
57: return result;
58: }
59:
60: double pa_atod(const char* str, const String* problem_source) {
61: if(!str)
62: return 0;
63:
1.199 paf 64: while(*str && isspace((unsigned char)*str))
1.193 paf 65: str++;
66: if(!*str)
67: return 0;
68:
69: double result;
70: char *error_pos;
1.200 paf 71: bool negative=false;
72: if(str[0]=='-') {
73: negative=true;
74: str++;
75: } else if(str[0]=='+') {
76: str++;
77: }
1.193 paf 78: // 0xABC
79: if(str[0]=='0')
80: if(str[1]=='x' || str[1]=='X')
81: result=(double)(unsigned long)strtol(str, &error_pos, 0);
1.200 paf 82: else {
83: // skip leading 0000, to disable octal interpretation
84: do str++; while(*str=='0');
85: result=(double)strtod(str, &error_pos);
86: }
1.193 paf 87: else
88: result=(double)strtod(str, &error_pos);
1.200 paf 89: if(negative)
90: result=-result;
1.193 paf 91:
92: while(char c=*error_pos++)
1.199 paf 93: if(!isspace((unsigned char)c))
1.193 paf 94: throw Exception("number.format",
95: problem_source,
96: problem_source?"invalid number (double)": "'%s' is invalid number (double)", str);
97:
98: return result;
99: }
100:
1.176 paf 101: // cord lib extension
102:
103: #ifndef DOXYGEN
104: typedef struct {
105: ssize_t countdown;
1.204 paf 106: int target; /* Character we're looking for */
1.176 paf 107: } chr_data;
108: #endif
109: static int CORD_range_contains_chr_greater_then_proc(char c, size_t size, void* client_data)
110: {
111: register chr_data * d = (chr_data *)client_data;
112:
113: if (d -> countdown<=0) return(2);
114: d -> countdown -= size;
115: if (c > d -> target) return(1);
116: return(0);
117: }
118: int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c)
119: {
120: chr_data d;
121:
122: d.countdown = n;
123: d.target = c;
124: return(CORD_block_iter(x, i, CORD_range_contains_chr_greater_then_proc, &d) == 1/*alternatives: 0 normally ended, 2=struck 'n'*/);
125: }
126:
1.187 paf 127: static int CORD_block_count_proc(char /*c*/, size_t /*size*/, void* client_data)
1.178 paf 128: {
129: int* result=(int*)client_data;
130: (*result)++;
131: return(0); // 0=continue
132: }
133: size_t CORD_block_count(CORD x)
134: {
135: size_t result=0;
136: CORD_block_iter(x, 0, CORD_block_count_proc, &result);
137: return result;
138: }
139:
1.174 paf 140: // helpers
1.139 paf 141:
1.174 paf 142: /// String::match uses this as replace & global search table columns
1.139 paf 143:
1.174 paf 144: const int MAX_MATCH_GROUPS=100;
1.139 paf 145:
1.174 paf 146: class String_match_table_template_columns: public ArrayString {
147: public:
148: String_match_table_template_columns() {
149: *this+=new String("prematch");
150: *this+=new String("match");
151: *this+=new String("postmatch");
152: for(int i=0; i<MAX_MATCH_GROUPS; i++) {
1.176 paf 153: *this+=new String(String::Body::Format(1+i), String::L_CLEAN);
1.174 paf 154: }
155: }
156: };
157:
158: Table string_match_table_template(new String_match_table_template_columns);
159:
1.176 paf 160: // String::Body methods
1.140 paf 161:
1.176 paf 162: String::Body String::Body::Format(int value) {
1.174 paf 163: char local[MAX_NUMBER];
164: size_t length=snprintf(local, MAX_NUMBER, "%d", value);
1.224 misha 165: return String::Body(pa_strdup(local, length));
1.120 paf 166: }
167:
1.195 paf 168: String::Body String::Body::trim(String::Trim_kind kind, const char* chars,
169: size_t* out_start, size_t* out_length) const {
170: size_t our_length=length();
171: if(!our_length)
172: return *this;
173: if(!chars)
174: chars=" \t\n"; // white space
175:
176: size_t start=0;
177: size_t end=our_length;
1.196 paf 178: // from left...
179: if(kind!=TRIM_END) {
1.195 paf 180: CORD_pos pos; set_pos(pos, 0);
181: while(true) {
182: char c=CORD_pos_fetch(pos);
183: if(strchr(chars, c)) {
184: if(++start==our_length)
185: return 0; // all chars are empty, just return empty string
186: } else
187: break;
188:
189: CORD_next(pos);
190: }
191: }
1.196 paf 192: // from right..
193: if(kind!=TRIM_START) {
194: CORD_pos pos; set_pos(pos, end-1);
195: while(true) {
196: char c=CORD_pos_fetch(pos);
197: if(strchr(chars, c)) {
198: if(--end==0) // optimization: NO need to check for 'end>=start', that's(<) impossible
199: return 0; // all chars are empty, just return empty string
200: } else
201: break;
202:
203: CORD_prev(pos);
204: }
205: }
1.195 paf 206:
207: if(start==0 && end==our_length) // nobody moved a thing
208: return *this;
209:
210: if(out_start)
211: *out_start=start;
212: size_t new_length=end-start;
213: if(out_length)
214: *out_length=new_length;
215:
216: return mid(start, new_length);
217: }
218:
1.174 paf 219: static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
220: uint& result=*static_cast<uint*>(client_data);
221: generic_hash_code(result, c);
222: return 0;
223: }
224: static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
225: uint& result=*static_cast<uint*>(client_data);
226: generic_hash_code(result, s);
227: return 0;
228: };
1.225 ! misha 229: uint String::Body::get_hash_code() const {
! 230: #ifdef HASH_CODE_CACHING
! 231: if(hash_code)
! 232: return hash_code;
! 233: #else
! 234: uint hash_code=0;
! 235: #endif
1.220 misha 236: if (body && CORD_IS_STRING(body)){
1.225 ! misha 237: generic_hash_code(hash_code, body);
1.220 misha 238: } else {
239: CORD_iter5(body, 0,
240: CORD_batched_iter_fn_generic_hash_code,
1.225 ! misha 241: CORD_batched_iter_fn_generic_hash_code, &hash_code);
1.220 misha 242: }
1.225 ! misha 243: return hash_code;
1.94 parser 244: }
245:
1.174 paf 246: // String methods
247:
248: String& String::append_know_length(const char* str, size_t known_length, Language lang) {
249: if(!known_length)
1.9 paf 250: return *this;
1.122 paf 251:
1.176 paf 252: // first: langs
253: langs.append(body, lang, known_length);
254: // next: letters themselves
1.174 paf 255: body.append_know_length(str, known_length);
1.1 paf 256:
1.174 paf 257: ASSERT_STRING_INVARIANT(*this);
1.1 paf 258: return *this;
259: }
1.174 paf 260: String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
261: if(!str)
262: return *this;
263: size_t known_length=helper_length?helper_length:strlen(str);
264: if(!known_length)
265: return *this;
1.1 paf 266:
1.174 paf 267: return append_know_length(str, known_length, lang);
1.5 paf 268: }
1.174 paf 269: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
270: size_t known_length=helper_length?helper_length:strlen(str);
271: if(!known_length)
272: return *this;
1.5 paf 273:
1.176 paf 274: // first: langs
275: langs.append(body, lang, known_length);
276: // next: letters themselves
1.174 paf 277: body.append_strdup_know_length(str, known_length);
1.33 paf 278:
1.174 paf 279: ASSERT_STRING_INVARIANT(*this);
280: return *this;
1.5 paf 281: }
1.46 paf 282:
1.221 misha 283: int CORD_batched_len(const char* s, size_t* len){
284: (*len) += lengthUTF8( (const XMLByte *)s, (const XMLByte *)s+strlen(s));
285: return 0;
286: }
287:
288: // can be called only for IS_FUNCTION(CORD) which are used only in Lang
289: int CORD_batched_len(const char, size_t *len){
290: (*len)++;
1.220 misha 291: return 0;
292: }
293:
1.210 misha 294: size_t String::length(Charset& charset) const {
295: if(charset.isUTF8()){
1.220 misha 296: size_t len=0;
1.224 misha 297: body.for_each<size_t *>(CORD_batched_len, CORD_batched_len, &len);
1.220 misha 298: return len;
1.210 misha 299: } else
300: return body.length();
301: }
302:
1.174 paf 303: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
304: String& String::mid(size_t substr_begin, size_t substr_end) const {
305: String& result=*new String;
306:
307: size_t self_length=length();
308: substr_begin=min(substr_begin, self_length);
309: substr_end=min(max(substr_end, substr_begin), self_length);
1.176 paf 310: size_t substr_length=substr_end-substr_begin;
311: if(!substr_length)
1.107 parser 312: return result;
1.53 paf 313:
1.176 paf 314: // first: their langs
315: result.langs.append(result.body, langs, substr_begin, substr_length);
316: // next: letters themselves
317: result.body=body.mid(substr_begin, substr_length);
1.174 paf 318:
319: ASSERT_STRING_INVARIANT(result);
1.53 paf 320: return result;
1.54 paf 321: }
322:
1.211 misha 323: // from, to and helper_length in characters, not in bytes (it's important for utf-8)
324: String& String::mid(Charset& charset, size_t from, size_t to, size_t helper_length) const {
1.210 misha 325: String& result=*new String;
326:
1.211 misha 327: size_t self_length=(helper_length)?helper_length:length(charset);
328:
329: if(!self_length)
330: return result;
331:
332: from=min(min(to, from), self_length);
1.210 misha 333: to=min(max(to, from), self_length);
1.211 misha 334:
1.210 misha 335: size_t substr_length=to-from;
1.211 misha 336:
1.210 misha 337: if(!substr_length)
338: return result;
339:
340: if(charset.isUTF8()){
341: const XMLByte* srcPtr=(const XMLByte*)cstrm();
1.211 misha 342: const XMLByte* srcEnd=srcPtr+body.length();
1.210 misha 343:
1.212 misha 344: // convert 'from' and 'substr_length' from 'characters' to 'bytes'
1.210 misha 345: from=getUTF8BytePos(srcPtr, srcEnd, from);
346: substr_length=getUTF8BytePos(srcPtr+from, srcEnd, substr_length);
347: if(!substr_length)
348: return result;
349: }
350:
351: // first: their langs
352: result.langs.append(result.body, langs, from, substr_length);
353: // next: letters themselves
354: result.body=body.mid(from, substr_length);
355:
356: ASSERT_STRING_INVARIANT(result);
357: return result;
358: }
359:
1.176 paf 360: size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
1.183 paf 361: size_t substr_length=substr.length();
362: while(true) {
363: size_t substr_begin=body.pos(substr, this_offset);
364:
365: if(substr_begin==CORD_NOT_FOUND)
366: return STRING_NOT_FOUND;
1.174 paf 367:
1.183 paf 368: if(langs.check_lang(lang, substr_begin, substr_length))
369: return substr_begin;
370:
371: this_offset=substr_begin+substr_length;
372: }
1.58 paf 373: }
374:
1.174 paf 375: size_t String::pos(const String& substr,
376: size_t this_offset, Language lang) const {
377: return pos(substr.body, this_offset, lang);
1.60 paf 378: }
379:
1.210 misha 380: size_t String::pos(Charset& charset, const String& substr,
381: size_t this_offset, Language lang) const {
382:
1.212 misha 383: size_t result;
1.210 misha 384: if(charset.isUTF8()){
385: const XMLByte* srcPtr=(const XMLByte*)cstrm();
1.212 misha 386: const XMLByte* srcEnd=srcPtr+body.length();
387:
388: // convert 'this_offset' from 'characters' to 'bytes'
389: this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset);
390:
391: result=pos(substr.body, this_offset, lang);
392: if(result==CORD_NOT_FOUND)
393: return STRING_NOT_FOUND;
394:
395: // convert 'result' from 'bytes' to 'characters'
396: result=getUTF8CharPos(srcPtr, srcEnd, result);
397: } else {
398: result=pos(substr.body, this_offset, lang);
399: if(result==CORD_NOT_FOUND)
400: return STRING_NOT_FOUND;
1.210 misha 401: }
402:
403: return result;
404: }
405:
1.174 paf 406: void String::split(ArrayString& result,
407: size_t& pos_after,
408: const char* delim,
409: Language lang, int limit) const {
410: size_t self_length=length();
411: if(size_t delim_length=strlen(delim)) {
1.186 paf 412: size_t pos_before;
1.60 paf 413: // while we have 'delim'...
1.174 paf 414: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 415: result+=&mid(pos_after, pos_before);
1.174 paf 416: pos_after=pos_before+delim_length;
1.60 paf 417: }
418: // last piece
1.174 paf 419: if(pos_after<self_length && limit) {
420: result+=&mid(pos_after, self_length);
421: pos_after=self_length;
1.60 paf 422: }
423: } else { // empty delim
424: result+=this;
1.174 paf 425: pos_after+=self_length;
1.60 paf 426: }
427: }
428:
1.174 paf 429: void String::split(ArrayString& result,
430: size_t& pos_after,
431: const String& delim, Language lang,
432: int limit) const {
1.140 paf 433: if(!delim.is_empty()) {
1.186 paf 434: size_t pos_before;
1.60 paf 435: // while we have 'delim'...
1.174 paf 436: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 437: result+=&mid(pos_after, pos_before);
1.174 paf 438: pos_after=pos_before+delim.length();
1.60 paf 439: }
440: // last piece
1.174 paf 441: if(pos_after<length() && limit) {
442: result+=&mid(pos_after, length());
443: pos_after=length();
1.60 paf 444: }
445: } else { // empty delim
446: result+=this;
1.174 paf 447: pos_after+=length();
1.60 paf 448: }
1.61 paf 449: }
450:
1.222 misha 451: Table* String::match(VRegex* vregex,
452: Row_action row_action, void *info,
453: int& matches_count) const {
1.209 misha 454:
1.222 misha 455: // vregex->info(); // I have no idea what does it for?
1.63 paf 456:
1.222 misha 457: bool need_pre_post_match=vregex->is_pre_post_match_needed();
458: bool global=vregex->is_global_search();
1.63 paf 459:
1.174 paf 460: const char* subject=cstr();
461: size_t subject_length=strlen(subject);
1.222 misha 462: const int ovector_size=(1/*match*/+MAX_MATCH_GROUPS)*3;
463: int ovector[ovector_size];
1.155 paf 464:
1.173 paf 465: Table::Action_options table_options;
1.174 paf 466: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 467:
1.154 paf 468: int prestart=0;
469: int poststart=0;
1.174 paf 470: int postfinish=length();
1.63 paf 471: while(true) {
1.222 misha 472: int exec_result=vregex->exec(subject, subject_length, ovector, ovector_size, prestart);
1.63 paf 473:
1.222 misha 474: if(exec_result<0) // only PCRE_ERROR_NOMATCH might be here, other negative results cause an exception
475: break;
1.63 paf 476:
1.154 paf 477: int prefinish=ovector[0];
478: poststart=ovector[1];
1.174 paf 479: ArrayString* row=new ArrayString;
480: if(need_pre_post_match) {
481: *row+=&mid(0, prefinish); // .prematch column value
482: *row+=&mid(prefinish, poststart); // .match
483: *row+=&mid(poststart, postfinish); // .postmatch
484: } else {
1.185 paf 485: *row+=&Empty; // .prematch column value
486: *row+=&Empty; // .match
487: *row+=&Empty; // .postmatch
1.174 paf 488: }
1.63 paf 489:
1.222 misha 490: for(int i=1; i<exec_result; i++) {
1.69 paf 491: // -1:-1 case handled peacefully by mid() itself
1.174 paf 492: *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 493: }
494:
1.209 misha 495: matches_count++;
1.174 paf 496: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 497:
1.222 misha 498: if(!global || prestart==poststart) // last step
499: break;
500:
1.154 paf 501: prestart=poststart;
1.222 misha 502: }
1.63 paf 503:
1.222 misha 504: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
505: return vregex->is_just_count() ? 0 : &table;
1.82 parser 506: }
507:
1.174 paf 508: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
509: String& result=*new String();
510: if(is_empty())
511: return result;
512:
513: char* new_cstr=cstrm();
1.216 misha 514:
1.181 paf 515: if(source_charset.isUTF8()) {
1.216 misha 516: size_t new_cstr_len=length();
1.181 paf 517: switch(kind) {
518: case CC_UPPER:
1.192 paf 519: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
1.181 paf 520: break;
521: case CC_LOWER:
1.192 paf 522: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower);
1.181 paf 523: break;
524: default:
525: assert(!"unknown change case kind");
526: break; // never
527: }
528:
529: } else {
530: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 531:
1.181 paf 532: const unsigned char *a;
533: const unsigned char *b;
534: switch(kind) {
535: case CC_UPPER:
536: a=tables+lcc_offset;
537: b=tables+fcc_offset;
538: break;
539: case CC_LOWER:
540: a=tables+lcc_offset;
541: b=0;
542: break;
543: default:
544: assert(!"unknown change case kind");
545: a=b=0; // calm, compiler
546: break; // never
547: }
548:
1.192 paf 549: char *dest=new_cstr;
1.181 paf 550: unsigned char index;
1.190 paf 551: for(const char* current=new_cstr; (index=(unsigned char)*current); current++) {
1.181 paf 552: unsigned char c=a[index];
553: if(b)
554: c=b[c];
555:
556: *dest++=(char)c;
557: }
1.174 paf 558: }
1.176 paf 559: result.langs=langs;
1.174 paf 560: result.body=new_cstr;
1.89 parser 561:
1.101 parser 562: return result;
563: }
564:
1.213 misha 565: const String& String::escape(Charset& source_charset) const {
566: if(is_empty())
567: return *this;
568:
569: return Charset::escape(*this, source_charset);
570: }
571:
1.174 paf 572: const String& String::replace(const Dictionary& dict) const {
573: String& result=*new String();
574: const char* old_cstr=cstr();
575: const char* prematch_begin=old_cstr;
576:
577: const char* current=old_cstr;
578: while(*current) {
1.184 paf 579: if(Dictionary::Subst subst=dict.first_that_begins(current)) {
1.174 paf 580: // prematch
581: if(size_t prematch_length=current-prematch_begin) {
1.179 paf 582: result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length);
1.174 paf 583: result.body.append_strdup_know_length(prematch_begin, prematch_length);
1.101 parser 584: }
585:
1.174 paf 586: // match
587: // skip 'a' in 'current'; move prematch_begin
1.184 paf 588: current+=subst.from_length; prematch_begin=current;
1.174 paf 589:
1.184 paf 590: if(const String* b=subst.to) // are there any b?
1.174 paf 591: result<<*b;
592: } else // simply advance
593: current++;
594: }
1.156 paf 595:
1.174 paf 596: // postmatch
597: if(size_t postmatch_length=current-prematch_begin) {
1.179 paf 598: result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length);
1.174 paf 599: result.body.append_strdup_know_length(prematch_begin, postmatch_length);
600: }
1.156 paf 601:
1.174 paf 602: ASSERT_STRING_INVARIANT(result);
1.82 parser 603: return result;
1.61 paf 604: }
1.113 parser 605:
1.180 paf 606: static int serialize_body_char(char c, char** cur) {
607: *((*cur)++)=c;
608: return 0; // 0=continue
609: };
1.174 paf 610: static int serialize_body_piece(const char* s, char** cur) {
611: size_t length=strlen(s);
612: memcpy(*cur, s, length); *cur+=length;
1.178 paf 613: return 0; // 0=continue
1.174 paf 614: };
1.178 paf 615: static int serialize_lang_piece(char alang, size_t asize, char** cur) {
616: // lang
1.191 paf 617: **cur=alang; (*cur)++;
618: // length [WARNING: not cast, addresses must be %4=0 on sparc]
1.178 paf 619: memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize);
620:
621: return 0; // 0=continue
622: }
1.174 paf 623: String::Cm String::serialize(size_t prolog_length) const {
1.178 paf 624: size_t fragments_count=langs.count();
1.202 paf 625: size_t body_length=body.length();
1.174 paf 626: size_t buf_length=
1.178 paf 627: prolog_length //1
628: +sizeof(size_t) //2
1.202 paf 629: +body_length //3
630: +1 // 4 for zero terminator used in deserialize
631: +sizeof(size_t) //5
632: +fragments_count*(sizeof(char)+sizeof(size_t)); //6
633:
1.174 paf 634: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
635:
636: // 1: prolog
637: char *cur=result.str+prolog_length;
1.202 paf 638: // 2: chars.count [WARNING: not cast, addresses must be %4=0 on sparc]
639: memcpy(cur, &body_length, sizeof(body_length)); cur+=sizeof(body_length);
640: // 3: letters
641: body.for_each(serialize_body_char, serialize_body_piece, &cur);
642: // 4: zero terminator
643: *cur++=0;
644: // 5: langs.count [WARNING: not cast, addresses must be %4=0 on sparc]
1.174 paf 645: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 646: // 6: lang info
1.178 paf 647: langs.for_each(body, serialize_lang_piece, &cur);
1.113 parser 648:
1.174 paf 649: return result;
1.113 parser 650: }
1.202 paf 651: bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size) {
652: size_t in_buf=buf_size;
653: if(in_buf<=prolog_size)
1.148 paf 654: return false;
1.202 paf 655: in_buf-=prolog_size;
1.135 paf 656:
1.174 paf 657: // 1: prolog
1.202 paf 658: const char* cur=(const char* )buf+prolog_size;
1.113 parser 659:
1.207 paf 660: // 2: chars.count
1.202 paf 661: size_t body_length;
662: if(in_buf<sizeof(body_length)) // body.length don't fit?
663: return false;
664: // [WARNING: not cast, addresses must be %4=0 on sparc]
665: memcpy(&body_length, cur, sizeof(body_length)); cur+=sizeof(body_length);
666: in_buf-=sizeof(body_length);
667:
668: if(in_buf<body_length+1) // letters+terminator don't fit?
669: return false;
670: // 4: zero terminator
671: if(cur[body_length] != 0) // in place?
672: return false;
673: // 3: letters
1.224 misha 674: body=String::Body(cur);
1.202 paf 675: cur+=body_length+1;
676: in_buf-=body_length+1;
677:
678: // 5: langs.count
1.191 paf 679: size_t fragments_count;
1.202 paf 680: if(in_buf<sizeof(fragments_count)) // langs.count don't fit?
1.174 paf 681: return false;
1.191 paf 682: // [WARNING: not cast, addresses must be %4=0 on sparc]
683: memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 684: in_buf-=sizeof(fragments_count);
1.174 paf 685:
686: if(fragments_count) {
1.202 paf 687: // 6: lang info
1.174 paf 688: size_t total_length=0;
689: for(size_t f=0; f<fragments_count; f++) {
1.191 paf 690: char lang;
691: size_t fragment_length;
692: size_t piece_length=sizeof(lang)+sizeof(fragment_length);
1.202 paf 693: if(in_buf<piece_length) // lang+length
1.174 paf 694: return false;
695:
1.191 paf 696: // lang
697: lang=*cur++;
698: // length [WARNING: not cast, addresses must be %4=0 on sparc]
699: memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length);
700:
1.206 paf 701: size_t combined_length=total_length+fragment_length;
702: if(combined_length>body_length)
703: return false; // file curruption
1.191 paf 704: // uchar needed to prevent propagating 0x80 bit to upper bytes
705: langs.append(total_length, (String::Language)(uchar)lang, fragment_length);
1.206 paf 706: total_length=combined_length;
1.202 paf 707: in_buf-=piece_length;
1.174 paf 708: }
1.128 paf 709:
1.202 paf 710: if(total_length!=body_length) // length(all language fragments) vs length(letters)
1.148 paf 711: return false;
1.174 paf 712: }
1.202 paf 713: if(in_buf!=0) // some strange extra bytes
714: return false;
1.113 parser 715:
1.174 paf 716: ASSERT_STRING_INVARIANT(*this);
1.148 paf 717: return true;
1.176 paf 718: }
719:
720: const char* String::Body::v() const {
721: return CORD_to_const_char_star(body);
722: }
1.201 paf 723: void String::Body::dump() const {
724: CORD_dump(body);
725: }
726:
1.176 paf 727: const char* String::Languages::v() const {
1.177 paf 728: if(opt.is_not_just_lang)
1.176 paf 729: return CORD_to_const_char_star(langs);
730: else
731: return (const char*)&langs;
732: }
1.201 paf 733: void String::Languages::dump() const {
734: if(opt.is_not_just_lang)
735: CORD_dump(langs);
736: else
737: puts((const char*)&langs);
738: }
1.176 paf 739: const char* String::v() const {
1.198 paf 740: const uint LIMIT_VIEW=20;
1.176 paf 741: char* buf=(char*)malloc(MAX_STRING);
742: const char*body_view=body.v();
743: const char*langs_view=langs.v();
744: snprintf(buf, MAX_STRING,
1.178 paf 745: "%d:%.*s%s} "
1.176 paf 746: "{%d:%s",
1.178 paf 747: langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"",
1.176 paf 748: strlen(body_view), body_view
749: );
750:
751: return buf;
1.113 parser 752: }
1.201 paf 753: void String::dump() const {
754: body.dump();
755: langs.dump();
756: }
1.195 paf 757: const String& String::trim(String::Trim_kind kind, const char* chars) const {
1.223 misha 758: if(is_empty())
1.195 paf 759: return *this;
760:
761: size_t substr_begin, substr_length;
762: Body new_body=body.trim(kind, chars, &substr_begin, &substr_length);
763: if(new_body==body) // we received unchanged pointer, do likewise
764: return *this;
765: // new_body differs from body, adjust langs along
766:
767: String& result=*new String;
768: if(!new_body) // body.trim produced empty result
769: return result;
770: // body.trim produced nonempty result
771:
772: // first: their langs
773: result.langs.append(result.body, langs, substr_begin, substr_length);
774: // next: letters themselves
775: result.body=new_body;
776:
777: ASSERT_STRING_INVARIANT(result);
778: return result;
1.198 paf 779: }
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