Annotation of parser3/src/main/pa_string.C, revision 1.228
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:
1.228 ! misha 8: static const char * const IDENT_STRING_C="$Date: 2009-05-26 21:33:05 $";
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: }
1.226 misha 269: String::String(int value, char *format) : langs(L_CLEAN){
270: char buf[MAX_NUMBER];
271: body.append_strdup_know_length(buf, snprintf(buf, MAX_NUMBER, format, value));
272: }
1.174 paf 273: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
274: size_t known_length=helper_length?helper_length:strlen(str);
275: if(!known_length)
276: return *this;
1.5 paf 277:
1.176 paf 278: // first: langs
279: langs.append(body, lang, known_length);
280: // next: letters themselves
1.174 paf 281: body.append_strdup_know_length(str, known_length);
1.33 paf 282:
1.174 paf 283: ASSERT_STRING_INVARIANT(*this);
284: return *this;
1.5 paf 285: }
1.46 paf 286:
1.221 misha 287: int CORD_batched_len(const char* s, size_t* len){
288: (*len) += lengthUTF8( (const XMLByte *)s, (const XMLByte *)s+strlen(s));
289: return 0;
290: }
291:
292: // can be called only for IS_FUNCTION(CORD) which are used only in Lang
293: int CORD_batched_len(const char, size_t *len){
294: (*len)++;
1.220 misha 295: return 0;
296: }
297:
1.210 misha 298: size_t String::length(Charset& charset) const {
299: if(charset.isUTF8()){
1.220 misha 300: size_t len=0;
1.224 misha 301: body.for_each<size_t *>(CORD_batched_len, CORD_batched_len, &len);
1.220 misha 302: return len;
1.210 misha 303: } else
304: return body.length();
305: }
306:
1.174 paf 307: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
308: String& String::mid(size_t substr_begin, size_t substr_end) const {
309: String& result=*new String;
310:
311: size_t self_length=length();
312: substr_begin=min(substr_begin, self_length);
313: substr_end=min(max(substr_end, substr_begin), self_length);
1.176 paf 314: size_t substr_length=substr_end-substr_begin;
315: if(!substr_length)
1.107 parser 316: return result;
1.53 paf 317:
1.176 paf 318: // first: their langs
319: result.langs.append(result.body, langs, substr_begin, substr_length);
320: // next: letters themselves
321: result.body=body.mid(substr_begin, substr_length);
1.174 paf 322:
323: ASSERT_STRING_INVARIANT(result);
1.53 paf 324: return result;
1.54 paf 325: }
326:
1.211 misha 327: // from, to and helper_length in characters, not in bytes (it's important for utf-8)
328: String& String::mid(Charset& charset, size_t from, size_t to, size_t helper_length) const {
1.210 misha 329: String& result=*new String;
330:
1.211 misha 331: size_t self_length=(helper_length)?helper_length:length(charset);
332:
333: if(!self_length)
334: return result;
335:
336: from=min(min(to, from), self_length);
1.210 misha 337: to=min(max(to, from), self_length);
1.211 misha 338:
1.210 misha 339: size_t substr_length=to-from;
1.211 misha 340:
1.210 misha 341: if(!substr_length)
342: return result;
343:
344: if(charset.isUTF8()){
345: const XMLByte* srcPtr=(const XMLByte*)cstrm();
1.211 misha 346: const XMLByte* srcEnd=srcPtr+body.length();
1.210 misha 347:
1.212 misha 348: // convert 'from' and 'substr_length' from 'characters' to 'bytes'
1.210 misha 349: from=getUTF8BytePos(srcPtr, srcEnd, from);
350: substr_length=getUTF8BytePos(srcPtr+from, srcEnd, substr_length);
351: if(!substr_length)
352: return result;
353: }
354:
355: // first: their langs
356: result.langs.append(result.body, langs, from, substr_length);
357: // next: letters themselves
358: result.body=body.mid(from, substr_length);
359:
360: ASSERT_STRING_INVARIANT(result);
361: return result;
362: }
363:
1.176 paf 364: size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
1.183 paf 365: size_t substr_length=substr.length();
366: while(true) {
367: size_t substr_begin=body.pos(substr, this_offset);
368:
369: if(substr_begin==CORD_NOT_FOUND)
370: return STRING_NOT_FOUND;
1.174 paf 371:
1.183 paf 372: if(langs.check_lang(lang, substr_begin, substr_length))
373: return substr_begin;
374:
375: this_offset=substr_begin+substr_length;
376: }
1.58 paf 377: }
378:
1.174 paf 379: size_t String::pos(const String& substr,
380: size_t this_offset, Language lang) const {
381: return pos(substr.body, this_offset, lang);
1.60 paf 382: }
383:
1.210 misha 384: size_t String::pos(Charset& charset, const String& substr,
385: size_t this_offset, Language lang) const {
386:
1.212 misha 387: size_t result;
1.210 misha 388: if(charset.isUTF8()){
389: const XMLByte* srcPtr=(const XMLByte*)cstrm();
1.212 misha 390: const XMLByte* srcEnd=srcPtr+body.length();
391:
392: // convert 'this_offset' from 'characters' to 'bytes'
393: this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset);
394:
395: result=pos(substr.body, this_offset, lang);
396: if(result==CORD_NOT_FOUND)
397: return STRING_NOT_FOUND;
398:
399: // convert 'result' from 'bytes' to 'characters'
400: result=getUTF8CharPos(srcPtr, srcEnd, result);
401: } else {
402: result=pos(substr.body, this_offset, lang);
403: if(result==CORD_NOT_FOUND)
404: return STRING_NOT_FOUND;
1.210 misha 405: }
406:
407: return result;
408: }
409:
1.174 paf 410: void String::split(ArrayString& result,
411: size_t& pos_after,
412: const char* delim,
413: Language lang, int limit) const {
414: size_t self_length=length();
415: if(size_t delim_length=strlen(delim)) {
1.186 paf 416: size_t pos_before;
1.60 paf 417: // while we have 'delim'...
1.174 paf 418: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 419: result+=&mid(pos_after, pos_before);
1.174 paf 420: pos_after=pos_before+delim_length;
1.60 paf 421: }
422: // last piece
1.174 paf 423: if(pos_after<self_length && limit) {
424: result+=&mid(pos_after, self_length);
425: pos_after=self_length;
1.60 paf 426: }
427: } else { // empty delim
428: result+=this;
1.174 paf 429: pos_after+=self_length;
1.60 paf 430: }
431: }
432:
1.174 paf 433: void String::split(ArrayString& result,
434: size_t& pos_after,
435: const String& delim, Language lang,
436: int limit) const {
1.140 paf 437: if(!delim.is_empty()) {
1.186 paf 438: size_t pos_before;
1.60 paf 439: // while we have 'delim'...
1.174 paf 440: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 441: result+=&mid(pos_after, pos_before);
1.174 paf 442: pos_after=pos_before+delim.length();
1.60 paf 443: }
444: // last piece
1.174 paf 445: if(pos_after<length() && limit) {
446: result+=&mid(pos_after, length());
447: pos_after=length();
1.60 paf 448: }
449: } else { // empty delim
450: result+=this;
1.174 paf 451: pos_after+=length();
1.60 paf 452: }
1.61 paf 453: }
454:
1.222 misha 455: Table* String::match(VRegex* vregex,
456: Row_action row_action, void *info,
457: int& matches_count) const {
1.209 misha 458:
1.222 misha 459: // vregex->info(); // I have no idea what does it for?
1.63 paf 460:
1.222 misha 461: bool need_pre_post_match=vregex->is_pre_post_match_needed();
462: bool global=vregex->is_global_search();
1.63 paf 463:
1.174 paf 464: const char* subject=cstr();
465: size_t subject_length=strlen(subject);
1.222 misha 466: const int ovector_size=(1/*match*/+MAX_MATCH_GROUPS)*3;
467: int ovector[ovector_size];
1.155 paf 468:
1.173 paf 469: Table::Action_options table_options;
1.174 paf 470: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 471:
1.154 paf 472: int prestart=0;
473: int poststart=0;
1.174 paf 474: int postfinish=length();
1.63 paf 475: while(true) {
1.222 misha 476: int exec_result=vregex->exec(subject, subject_length, ovector, ovector_size, prestart);
1.63 paf 477:
1.222 misha 478: if(exec_result<0) // only PCRE_ERROR_NOMATCH might be here, other negative results cause an exception
479: break;
1.63 paf 480:
1.154 paf 481: int prefinish=ovector[0];
482: poststart=ovector[1];
1.174 paf 483: ArrayString* row=new ArrayString;
484: if(need_pre_post_match) {
485: *row+=&mid(0, prefinish); // .prematch column value
486: *row+=&mid(prefinish, poststart); // .match
487: *row+=&mid(poststart, postfinish); // .postmatch
488: } else {
1.185 paf 489: *row+=&Empty; // .prematch column value
490: *row+=&Empty; // .match
491: *row+=&Empty; // .postmatch
1.174 paf 492: }
1.63 paf 493:
1.222 misha 494: for(int i=1; i<exec_result; i++) {
1.69 paf 495: // -1:-1 case handled peacefully by mid() itself
1.174 paf 496: *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 497: }
498:
1.209 misha 499: matches_count++;
1.174 paf 500: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 501:
1.222 misha 502: if(!global || prestart==poststart) // last step
503: break;
504:
1.154 paf 505: prestart=poststart;
1.222 misha 506: }
1.63 paf 507:
1.222 misha 508: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
509: return vregex->is_just_count() ? 0 : &table;
1.82 parser 510: }
511:
1.174 paf 512: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
513: String& result=*new String();
514: if(is_empty())
515: return result;
516:
517: char* new_cstr=cstrm();
1.216 misha 518:
1.181 paf 519: if(source_charset.isUTF8()) {
1.216 misha 520: size_t new_cstr_len=length();
1.181 paf 521: switch(kind) {
522: case CC_UPPER:
1.192 paf 523: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
1.181 paf 524: break;
525: case CC_LOWER:
1.192 paf 526: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower);
1.181 paf 527: break;
528: default:
529: assert(!"unknown change case kind");
530: break; // never
531: }
532:
533: } else {
534: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 535:
1.181 paf 536: const unsigned char *a;
537: const unsigned char *b;
538: switch(kind) {
539: case CC_UPPER:
540: a=tables+lcc_offset;
541: b=tables+fcc_offset;
542: break;
543: case CC_LOWER:
544: a=tables+lcc_offset;
545: b=0;
546: break;
547: default:
548: assert(!"unknown change case kind");
549: a=b=0; // calm, compiler
550: break; // never
551: }
552:
1.192 paf 553: char *dest=new_cstr;
1.181 paf 554: unsigned char index;
1.190 paf 555: for(const char* current=new_cstr; (index=(unsigned char)*current); current++) {
1.181 paf 556: unsigned char c=a[index];
557: if(b)
558: c=b[c];
559:
560: *dest++=(char)c;
561: }
1.174 paf 562: }
1.176 paf 563: result.langs=langs;
1.174 paf 564: result.body=new_cstr;
1.89 parser 565:
1.101 parser 566: return result;
567: }
568:
1.213 misha 569: const String& String::escape(Charset& source_charset) const {
570: if(is_empty())
571: return *this;
572:
573: return Charset::escape(*this, source_charset);
574: }
575:
1.174 paf 576: const String& String::replace(const Dictionary& dict) const {
577: String& result=*new String();
578: const char* old_cstr=cstr();
579: const char* prematch_begin=old_cstr;
580:
581: const char* current=old_cstr;
582: while(*current) {
1.184 paf 583: if(Dictionary::Subst subst=dict.first_that_begins(current)) {
1.174 paf 584: // prematch
585: if(size_t prematch_length=current-prematch_begin) {
1.179 paf 586: result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length);
1.174 paf 587: result.body.append_strdup_know_length(prematch_begin, prematch_length);
1.101 parser 588: }
589:
1.174 paf 590: // match
591: // skip 'a' in 'current'; move prematch_begin
1.184 paf 592: current+=subst.from_length; prematch_begin=current;
1.174 paf 593:
1.184 paf 594: if(const String* b=subst.to) // are there any b?
1.174 paf 595: result<<*b;
596: } else // simply advance
597: current++;
598: }
1.156 paf 599:
1.174 paf 600: // postmatch
601: if(size_t postmatch_length=current-prematch_begin) {
1.179 paf 602: result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length);
1.174 paf 603: result.body.append_strdup_know_length(prematch_begin, postmatch_length);
604: }
1.156 paf 605:
1.174 paf 606: ASSERT_STRING_INVARIANT(result);
1.82 parser 607: return result;
1.61 paf 608: }
1.113 parser 609:
1.180 paf 610: static int serialize_body_char(char c, char** cur) {
611: *((*cur)++)=c;
612: return 0; // 0=continue
613: };
1.174 paf 614: static int serialize_body_piece(const char* s, char** cur) {
615: size_t length=strlen(s);
616: memcpy(*cur, s, length); *cur+=length;
1.178 paf 617: return 0; // 0=continue
1.174 paf 618: };
1.178 paf 619: static int serialize_lang_piece(char alang, size_t asize, char** cur) {
620: // lang
1.191 paf 621: **cur=alang; (*cur)++;
622: // length [WARNING: not cast, addresses must be %4=0 on sparc]
1.178 paf 623: memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize);
624:
625: return 0; // 0=continue
626: }
1.174 paf 627: String::Cm String::serialize(size_t prolog_length) const {
1.178 paf 628: size_t fragments_count=langs.count();
1.202 paf 629: size_t body_length=body.length();
1.174 paf 630: size_t buf_length=
1.178 paf 631: prolog_length //1
632: +sizeof(size_t) //2
1.202 paf 633: +body_length //3
634: +1 // 4 for zero terminator used in deserialize
635: +sizeof(size_t) //5
636: +fragments_count*(sizeof(char)+sizeof(size_t)); //6
637:
1.174 paf 638: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
639:
640: // 1: prolog
641: char *cur=result.str+prolog_length;
1.202 paf 642: // 2: chars.count [WARNING: not cast, addresses must be %4=0 on sparc]
643: memcpy(cur, &body_length, sizeof(body_length)); cur+=sizeof(body_length);
644: // 3: letters
645: body.for_each(serialize_body_char, serialize_body_piece, &cur);
646: // 4: zero terminator
647: *cur++=0;
648: // 5: langs.count [WARNING: not cast, addresses must be %4=0 on sparc]
1.174 paf 649: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 650: // 6: lang info
1.178 paf 651: langs.for_each(body, serialize_lang_piece, &cur);
1.113 parser 652:
1.174 paf 653: return result;
1.113 parser 654: }
1.202 paf 655: bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size) {
656: size_t in_buf=buf_size;
657: if(in_buf<=prolog_size)
1.148 paf 658: return false;
1.202 paf 659: in_buf-=prolog_size;
1.135 paf 660:
1.174 paf 661: // 1: prolog
1.202 paf 662: const char* cur=(const char* )buf+prolog_size;
1.113 parser 663:
1.207 paf 664: // 2: chars.count
1.202 paf 665: size_t body_length;
666: if(in_buf<sizeof(body_length)) // body.length don't fit?
667: return false;
668: // [WARNING: not cast, addresses must be %4=0 on sparc]
669: memcpy(&body_length, cur, sizeof(body_length)); cur+=sizeof(body_length);
670: in_buf-=sizeof(body_length);
671:
672: if(in_buf<body_length+1) // letters+terminator don't fit?
673: return false;
674: // 4: zero terminator
675: if(cur[body_length] != 0) // in place?
676: return false;
677: // 3: letters
1.227 misha 678: body=String::Body(*cur?cur:0);
1.228 ! misha 679: #ifdef STRING_LENGTH_CACHING
! 680: body.set_length(body_length);
! 681: #endif
1.202 paf 682: cur+=body_length+1;
683: in_buf-=body_length+1;
684:
685: // 5: langs.count
1.191 paf 686: size_t fragments_count;
1.202 paf 687: if(in_buf<sizeof(fragments_count)) // langs.count don't fit?
1.174 paf 688: return false;
1.191 paf 689: // [WARNING: not cast, addresses must be %4=0 on sparc]
690: memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 691: in_buf-=sizeof(fragments_count);
1.174 paf 692:
693: if(fragments_count) {
1.202 paf 694: // 6: lang info
1.174 paf 695: size_t total_length=0;
696: for(size_t f=0; f<fragments_count; f++) {
1.191 paf 697: char lang;
698: size_t fragment_length;
699: size_t piece_length=sizeof(lang)+sizeof(fragment_length);
1.202 paf 700: if(in_buf<piece_length) // lang+length
1.174 paf 701: return false;
702:
1.191 paf 703: // lang
704: lang=*cur++;
705: // length [WARNING: not cast, addresses must be %4=0 on sparc]
706: memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length);
707:
1.206 paf 708: size_t combined_length=total_length+fragment_length;
709: if(combined_length>body_length)
710: return false; // file curruption
1.191 paf 711: // uchar needed to prevent propagating 0x80 bit to upper bytes
712: langs.append(total_length, (String::Language)(uchar)lang, fragment_length);
1.206 paf 713: total_length=combined_length;
1.202 paf 714: in_buf-=piece_length;
1.174 paf 715: }
1.128 paf 716:
1.202 paf 717: if(total_length!=body_length) // length(all language fragments) vs length(letters)
1.148 paf 718: return false;
1.174 paf 719: }
1.202 paf 720: if(in_buf!=0) // some strange extra bytes
721: return false;
1.113 parser 722:
1.174 paf 723: ASSERT_STRING_INVARIANT(*this);
1.148 paf 724: return true;
1.176 paf 725: }
726:
727: const char* String::Body::v() const {
728: return CORD_to_const_char_star(body);
729: }
1.201 paf 730: void String::Body::dump() const {
731: CORD_dump(body);
732: }
733:
1.176 paf 734: const char* String::Languages::v() const {
1.177 paf 735: if(opt.is_not_just_lang)
1.176 paf 736: return CORD_to_const_char_star(langs);
737: else
738: return (const char*)&langs;
739: }
1.201 paf 740: void String::Languages::dump() const {
741: if(opt.is_not_just_lang)
742: CORD_dump(langs);
743: else
744: puts((const char*)&langs);
745: }
1.176 paf 746: const char* String::v() const {
1.198 paf 747: const uint LIMIT_VIEW=20;
1.176 paf 748: char* buf=(char*)malloc(MAX_STRING);
749: const char*body_view=body.v();
750: const char*langs_view=langs.v();
751: snprintf(buf, MAX_STRING,
1.178 paf 752: "%d:%.*s%s} "
1.176 paf 753: "{%d:%s",
1.178 paf 754: langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"",
1.176 paf 755: strlen(body_view), body_view
756: );
757:
758: return buf;
1.113 parser 759: }
1.201 paf 760: void String::dump() const {
761: body.dump();
762: langs.dump();
763: }
1.195 paf 764: const String& String::trim(String::Trim_kind kind, const char* chars) const {
1.223 misha 765: if(is_empty())
1.195 paf 766: return *this;
767:
768: size_t substr_begin, substr_length;
769: Body new_body=body.trim(kind, chars, &substr_begin, &substr_length);
770: if(new_body==body) // we received unchanged pointer, do likewise
771: return *this;
772: // new_body differs from body, adjust langs along
773:
774: String& result=*new String;
775: if(!new_body) // body.trim produced empty result
776: return result;
777: // body.trim produced nonempty result
778:
779: // first: their langs
780: result.langs.append(result.body, langs, substr_begin, substr_length);
781: // next: letters themselves
782: result.body=new_body;
783:
784: ASSERT_STRING_INVARIANT(result);
785: return result;
1.198 paf 786: }
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