Annotation of parser3/src/main/pa_string.C, revision 1.222
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:
1.222 ! misha 8: static const char * const IDENT_STRING_C="$Date: 2009-04-17 09:59:51 $";
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);
1.176 paf 165: return String::Body(pa_strdup(local, length), 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: };
1.176 paf 229: uint String::Body::hash_code() const {
1.174 paf 230: uint result=0;
1.220 misha 231: if (body && CORD_IS_STRING(body)){
232: generic_hash_code(result, body);
233: } else {
234: CORD_iter5(body, 0,
235: CORD_batched_iter_fn_generic_hash_code,
236: CORD_batched_iter_fn_generic_hash_code, &result);
237: }
1.120 paf 238: return result;
1.94 parser 239: }
240:
1.174 paf 241: // String methods
242:
243: String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
244: append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.115 paf 245: }
1.174 paf 246: String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) {
247: append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN);
1.5 paf 248: }
1.28 paf 249:
1.174 paf 250: String& String::append_know_length(const char* str, size_t known_length, Language lang) {
251: if(!known_length)
1.9 paf 252: return *this;
1.122 paf 253:
1.176 paf 254: // first: langs
255: langs.append(body, lang, known_length);
256: // next: letters themselves
1.174 paf 257: body.append_know_length(str, known_length);
1.1 paf 258:
1.174 paf 259: ASSERT_STRING_INVARIANT(*this);
1.1 paf 260: return *this;
261: }
1.174 paf 262: String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
263: if(!str)
264: return *this;
265: size_t known_length=helper_length?helper_length:strlen(str);
266: if(!known_length)
267: return *this;
1.1 paf 268:
1.174 paf 269: return append_know_length(str, known_length, lang);
1.5 paf 270: }
1.174 paf 271: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
272: size_t known_length=helper_length?helper_length:strlen(str);
273: if(!known_length)
274: return *this;
1.5 paf 275:
1.176 paf 276: // first: langs
277: langs.append(body, lang, known_length);
278: // next: letters themselves
1.174 paf 279: body.append_strdup_know_length(str, known_length);
1.33 paf 280:
1.174 paf 281: ASSERT_STRING_INVARIANT(*this);
282: return *this;
1.5 paf 283: }
1.46 paf 284:
1.221 misha 285: int CORD_batched_len(const char* s, size_t* len){
286: (*len) += lengthUTF8( (const XMLByte *)s, (const XMLByte *)s+strlen(s));
287: return 0;
288: }
289:
290: // can be called only for IS_FUNCTION(CORD) which are used only in Lang
291: int CORD_batched_len(const char, size_t *len){
292: (*len)++;
1.220 misha 293: return 0;
294: }
295:
1.210 misha 296: size_t String::length(Charset& charset) const {
297: if(charset.isUTF8()){
1.220 misha 298: size_t len=0;
1.221 misha 299: body.for_each(CORD_batched_len, CORD_batched_len, &len);
1.220 misha 300: return len;
1.210 misha 301: } else
302: return body.length();
303: }
304:
1.174 paf 305: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
306: String& String::mid(size_t substr_begin, size_t substr_end) const {
307: String& result=*new String;
308:
309: size_t self_length=length();
310: substr_begin=min(substr_begin, self_length);
311: substr_end=min(max(substr_end, substr_begin), self_length);
1.176 paf 312: size_t substr_length=substr_end-substr_begin;
313: if(!substr_length)
1.107 parser 314: return result;
1.53 paf 315:
1.176 paf 316: // first: their langs
317: result.langs.append(result.body, langs, substr_begin, substr_length);
318: // next: letters themselves
319: result.body=body.mid(substr_begin, substr_length);
1.174 paf 320:
321: ASSERT_STRING_INVARIANT(result);
1.53 paf 322: return result;
1.54 paf 323: }
324:
1.211 misha 325: // from, to and helper_length in characters, not in bytes (it's important for utf-8)
326: String& String::mid(Charset& charset, size_t from, size_t to, size_t helper_length) const {
1.210 misha 327: String& result=*new String;
328:
1.211 misha 329: size_t self_length=(helper_length)?helper_length:length(charset);
330:
331: if(!self_length)
332: return result;
333:
334: from=min(min(to, from), self_length);
1.210 misha 335: to=min(max(to, from), self_length);
1.211 misha 336:
1.210 misha 337: size_t substr_length=to-from;
1.211 misha 338:
1.210 misha 339: if(!substr_length)
340: return result;
341:
342: if(charset.isUTF8()){
343: const XMLByte* srcPtr=(const XMLByte*)cstrm();
1.211 misha 344: const XMLByte* srcEnd=srcPtr+body.length();
1.210 misha 345:
1.212 misha 346: // convert 'from' and 'substr_length' from 'characters' to 'bytes'
1.210 misha 347: from=getUTF8BytePos(srcPtr, srcEnd, from);
348: substr_length=getUTF8BytePos(srcPtr+from, srcEnd, substr_length);
349: if(!substr_length)
350: return result;
351: }
352:
353: // first: their langs
354: result.langs.append(result.body, langs, from, substr_length);
355: // next: letters themselves
356: result.body=body.mid(from, substr_length);
357:
358: ASSERT_STRING_INVARIANT(result);
359: return result;
360: }
361:
1.176 paf 362: size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
1.183 paf 363: size_t substr_length=substr.length();
364: while(true) {
365: size_t substr_begin=body.pos(substr, this_offset);
366:
367: if(substr_begin==CORD_NOT_FOUND)
368: return STRING_NOT_FOUND;
1.174 paf 369:
1.183 paf 370: if(langs.check_lang(lang, substr_begin, substr_length))
371: return substr_begin;
372:
373: this_offset=substr_begin+substr_length;
374: }
1.58 paf 375: }
376:
1.174 paf 377: size_t String::pos(const String& substr,
378: size_t this_offset, Language lang) const {
379: return pos(substr.body, this_offset, lang);
1.60 paf 380: }
381:
1.210 misha 382: size_t String::pos(Charset& charset, const String& substr,
383: size_t this_offset, Language lang) const {
384:
1.212 misha 385: size_t result;
1.210 misha 386: if(charset.isUTF8()){
387: const XMLByte* srcPtr=(const XMLByte*)cstrm();
1.212 misha 388: const XMLByte* srcEnd=srcPtr+body.length();
389:
390: // convert 'this_offset' from 'characters' to 'bytes'
391: this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset);
392:
393: result=pos(substr.body, this_offset, lang);
394: if(result==CORD_NOT_FOUND)
395: return STRING_NOT_FOUND;
396:
397: // convert 'result' from 'bytes' to 'characters'
398: result=getUTF8CharPos(srcPtr, srcEnd, result);
399: } else {
400: result=pos(substr.body, this_offset, lang);
401: if(result==CORD_NOT_FOUND)
402: return STRING_NOT_FOUND;
1.210 misha 403: }
404:
405: return result;
406: }
407:
1.174 paf 408: void String::split(ArrayString& result,
409: size_t& pos_after,
410: const char* delim,
411: Language lang, int limit) const {
412: size_t self_length=length();
413: if(size_t delim_length=strlen(delim)) {
1.186 paf 414: size_t pos_before;
1.60 paf 415: // while we have 'delim'...
1.174 paf 416: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 417: result+=&mid(pos_after, pos_before);
1.174 paf 418: pos_after=pos_before+delim_length;
1.60 paf 419: }
420: // last piece
1.174 paf 421: if(pos_after<self_length && limit) {
422: result+=&mid(pos_after, self_length);
423: pos_after=self_length;
1.60 paf 424: }
425: } else { // empty delim
426: result+=this;
1.174 paf 427: pos_after+=self_length;
1.60 paf 428: }
429: }
430:
1.174 paf 431: void String::split(ArrayString& result,
432: size_t& pos_after,
433: const String& delim, Language lang,
434: int limit) const {
1.140 paf 435: if(!delim.is_empty()) {
1.186 paf 436: size_t pos_before;
1.60 paf 437: // while we have 'delim'...
1.174 paf 438: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 439: result+=&mid(pos_after, pos_before);
1.174 paf 440: pos_after=pos_before+delim.length();
1.60 paf 441: }
442: // last piece
1.174 paf 443: if(pos_after<length() && limit) {
444: result+=&mid(pos_after, length());
445: pos_after=length();
1.60 paf 446: }
447: } else { // empty delim
448: result+=this;
1.174 paf 449: pos_after+=length();
1.60 paf 450: }
1.61 paf 451: }
452:
1.222 ! misha 453: Table* String::match(VRegex* vregex,
! 454: Row_action row_action, void *info,
! 455: int& matches_count) const {
1.209 misha 456:
1.222 ! misha 457: // vregex->info(); // I have no idea what does it for?
1.63 paf 458:
1.222 ! misha 459: bool need_pre_post_match=vregex->is_pre_post_match_needed();
! 460: bool global=vregex->is_global_search();
1.63 paf 461:
1.174 paf 462: const char* subject=cstr();
463: size_t subject_length=strlen(subject);
1.222 ! misha 464: const int ovector_size=(1/*match*/+MAX_MATCH_GROUPS)*3;
! 465: int ovector[ovector_size];
1.155 paf 466:
1.173 paf 467: Table::Action_options table_options;
1.174 paf 468: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 469:
1.154 paf 470: int prestart=0;
471: int poststart=0;
1.174 paf 472: int postfinish=length();
1.63 paf 473: while(true) {
1.222 ! misha 474: int exec_result=vregex->exec(subject, subject_length, ovector, ovector_size, prestart);
1.63 paf 475:
1.222 ! misha 476: if(exec_result<0) // only PCRE_ERROR_NOMATCH might be here, other negative results cause an exception
! 477: break;
1.63 paf 478:
1.154 paf 479: int prefinish=ovector[0];
480: poststart=ovector[1];
1.174 paf 481: ArrayString* row=new ArrayString;
482: if(need_pre_post_match) {
483: *row+=&mid(0, prefinish); // .prematch column value
484: *row+=&mid(prefinish, poststart); // .match
485: *row+=&mid(poststart, postfinish); // .postmatch
486: } else {
1.185 paf 487: *row+=&Empty; // .prematch column value
488: *row+=&Empty; // .match
489: *row+=&Empty; // .postmatch
1.174 paf 490: }
1.63 paf 491:
1.222 ! misha 492: for(int i=1; i<exec_result; i++) {
1.69 paf 493: // -1:-1 case handled peacefully by mid() itself
1.174 paf 494: *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 495: }
496:
1.209 misha 497: matches_count++;
1.174 paf 498: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 499:
1.222 ! misha 500: if(!global || prestart==poststart) // last step
! 501: break;
! 502:
1.154 paf 503: prestart=poststart;
1.222 ! misha 504: }
1.63 paf 505:
1.222 ! misha 506: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
! 507: return vregex->is_just_count() ? 0 : &table;
1.82 parser 508: }
509:
1.174 paf 510: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
511: String& result=*new String();
512: if(is_empty())
513: return result;
514:
515: char* new_cstr=cstrm();
1.216 misha 516:
1.181 paf 517: if(source_charset.isUTF8()) {
1.216 misha 518: size_t new_cstr_len=length();
1.181 paf 519: switch(kind) {
520: case CC_UPPER:
1.192 paf 521: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
1.181 paf 522: break;
523: case CC_LOWER:
1.192 paf 524: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower);
1.181 paf 525: break;
526: default:
527: assert(!"unknown change case kind");
528: break; // never
529: }
530:
531: } else {
532: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 533:
1.181 paf 534: const unsigned char *a;
535: const unsigned char *b;
536: switch(kind) {
537: case CC_UPPER:
538: a=tables+lcc_offset;
539: b=tables+fcc_offset;
540: break;
541: case CC_LOWER:
542: a=tables+lcc_offset;
543: b=0;
544: break;
545: default:
546: assert(!"unknown change case kind");
547: a=b=0; // calm, compiler
548: break; // never
549: }
550:
1.192 paf 551: char *dest=new_cstr;
1.181 paf 552: unsigned char index;
1.190 paf 553: for(const char* current=new_cstr; (index=(unsigned char)*current); current++) {
1.181 paf 554: unsigned char c=a[index];
555: if(b)
556: c=b[c];
557:
558: *dest++=(char)c;
559: }
1.174 paf 560: }
1.176 paf 561: result.langs=langs;
1.174 paf 562: result.body=new_cstr;
1.89 parser 563:
1.101 parser 564: return result;
565: }
566:
1.213 misha 567: const String& String::escape(Charset& source_charset) const {
568: if(is_empty())
569: return *this;
570:
571: return Charset::escape(*this, source_charset);
572: }
573:
1.174 paf 574: const String& String::replace(const Dictionary& dict) const {
575: String& result=*new String();
576: const char* old_cstr=cstr();
577: const char* prematch_begin=old_cstr;
578:
579: const char* current=old_cstr;
580: while(*current) {
1.184 paf 581: if(Dictionary::Subst subst=dict.first_that_begins(current)) {
1.174 paf 582: // prematch
583: if(size_t prematch_length=current-prematch_begin) {
1.179 paf 584: result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length);
1.174 paf 585: result.body.append_strdup_know_length(prematch_begin, prematch_length);
1.101 parser 586: }
587:
1.174 paf 588: // match
589: // skip 'a' in 'current'; move prematch_begin
1.184 paf 590: current+=subst.from_length; prematch_begin=current;
1.174 paf 591:
1.184 paf 592: if(const String* b=subst.to) // are there any b?
1.174 paf 593: result<<*b;
594: } else // simply advance
595: current++;
596: }
1.156 paf 597:
1.174 paf 598: // postmatch
599: if(size_t postmatch_length=current-prematch_begin) {
1.179 paf 600: result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length);
1.174 paf 601: result.body.append_strdup_know_length(prematch_begin, postmatch_length);
602: }
1.156 paf 603:
1.174 paf 604: ASSERT_STRING_INVARIANT(result);
1.82 parser 605: return result;
1.61 paf 606: }
1.113 parser 607:
1.180 paf 608: static int serialize_body_char(char c, char** cur) {
609: *((*cur)++)=c;
610: return 0; // 0=continue
611: };
1.174 paf 612: static int serialize_body_piece(const char* s, char** cur) {
613: size_t length=strlen(s);
614: memcpy(*cur, s, length); *cur+=length;
1.178 paf 615: return 0; // 0=continue
1.174 paf 616: };
1.178 paf 617: static int serialize_lang_piece(char alang, size_t asize, char** cur) {
618: // lang
1.191 paf 619: **cur=alang; (*cur)++;
620: // length [WARNING: not cast, addresses must be %4=0 on sparc]
1.178 paf 621: memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize);
622:
623: return 0; // 0=continue
624: }
1.174 paf 625: String::Cm String::serialize(size_t prolog_length) const {
1.178 paf 626: size_t fragments_count=langs.count();
1.202 paf 627: size_t body_length=body.length();
1.174 paf 628: size_t buf_length=
1.178 paf 629: prolog_length //1
630: +sizeof(size_t) //2
1.202 paf 631: +body_length //3
632: +1 // 4 for zero terminator used in deserialize
633: +sizeof(size_t) //5
634: +fragments_count*(sizeof(char)+sizeof(size_t)); //6
635:
1.174 paf 636: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
637:
638: // 1: prolog
639: char *cur=result.str+prolog_length;
1.202 paf 640: // 2: chars.count [WARNING: not cast, addresses must be %4=0 on sparc]
641: memcpy(cur, &body_length, sizeof(body_length)); cur+=sizeof(body_length);
642: // 3: letters
643: body.for_each(serialize_body_char, serialize_body_piece, &cur);
644: // 4: zero terminator
645: *cur++=0;
646: // 5: langs.count [WARNING: not cast, addresses must be %4=0 on sparc]
1.174 paf 647: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 648: // 6: lang info
1.178 paf 649: langs.for_each(body, serialize_lang_piece, &cur);
1.113 parser 650:
1.174 paf 651: return result;
1.113 parser 652: }
1.202 paf 653: bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size) {
654: size_t in_buf=buf_size;
655: if(in_buf<=prolog_size)
1.148 paf 656: return false;
1.202 paf 657: in_buf-=prolog_size;
1.135 paf 658:
1.174 paf 659: // 1: prolog
1.202 paf 660: const char* cur=(const char* )buf+prolog_size;
1.113 parser 661:
1.207 paf 662: // 2: chars.count
1.202 paf 663: size_t body_length;
664: if(in_buf<sizeof(body_length)) // body.length don't fit?
665: return false;
666: // [WARNING: not cast, addresses must be %4=0 on sparc]
667: memcpy(&body_length, cur, sizeof(body_length)); cur+=sizeof(body_length);
668: in_buf-=sizeof(body_length);
669:
670: if(in_buf<body_length+1) // letters+terminator don't fit?
671: return false;
672: // 4: zero terminator
673: if(cur[body_length] != 0) // in place?
674: return false;
675: // 3: letters
676: body=String::Body(cur, body_length);
677: cur+=body_length+1;
678: in_buf-=body_length+1;
679:
680: // 5: langs.count
1.191 paf 681: size_t fragments_count;
1.202 paf 682: if(in_buf<sizeof(fragments_count)) // langs.count don't fit?
1.174 paf 683: return false;
1.191 paf 684: // [WARNING: not cast, addresses must be %4=0 on sparc]
685: memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 686: in_buf-=sizeof(fragments_count);
1.174 paf 687:
688: if(fragments_count) {
1.202 paf 689: // 6: lang info
1.174 paf 690: size_t total_length=0;
691: for(size_t f=0; f<fragments_count; f++) {
1.191 paf 692: char lang;
693: size_t fragment_length;
694: size_t piece_length=sizeof(lang)+sizeof(fragment_length);
1.202 paf 695: if(in_buf<piece_length) // lang+length
1.174 paf 696: return false;
697:
1.191 paf 698: // lang
699: lang=*cur++;
700: // length [WARNING: not cast, addresses must be %4=0 on sparc]
701: memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length);
702:
1.206 paf 703: size_t combined_length=total_length+fragment_length;
704: if(combined_length>body_length)
705: return false; // file curruption
1.191 paf 706: // uchar needed to prevent propagating 0x80 bit to upper bytes
707: langs.append(total_length, (String::Language)(uchar)lang, fragment_length);
1.206 paf 708: total_length=combined_length;
1.202 paf 709: in_buf-=piece_length;
1.174 paf 710: }
1.128 paf 711:
1.202 paf 712: if(total_length!=body_length) // length(all language fragments) vs length(letters)
1.148 paf 713: return false;
1.174 paf 714: }
1.202 paf 715: if(in_buf!=0) // some strange extra bytes
716: return false;
1.113 parser 717:
1.174 paf 718: ASSERT_STRING_INVARIANT(*this);
1.148 paf 719: return true;
1.176 paf 720: }
721:
722: const char* String::Body::v() const {
723: return CORD_to_const_char_star(body);
724: }
1.201 paf 725: void String::Body::dump() const {
726: CORD_dump(body);
727: }
728:
1.176 paf 729: const char* String::Languages::v() const {
1.177 paf 730: if(opt.is_not_just_lang)
1.176 paf 731: return CORD_to_const_char_star(langs);
732: else
733: return (const char*)&langs;
734: }
1.201 paf 735: void String::Languages::dump() const {
736: if(opt.is_not_just_lang)
737: CORD_dump(langs);
738: else
739: puts((const char*)&langs);
740: }
1.176 paf 741: const char* String::v() const {
1.198 paf 742: const uint LIMIT_VIEW=20;
1.176 paf 743: char* buf=(char*)malloc(MAX_STRING);
744: const char*body_view=body.v();
745: const char*langs_view=langs.v();
746: snprintf(buf, MAX_STRING,
1.178 paf 747: "%d:%.*s%s} "
1.176 paf 748: "{%d:%s",
1.178 paf 749: langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"",
1.176 paf 750: strlen(body_view), body_view
751: );
752:
753: return buf;
1.113 parser 754: }
1.201 paf 755: void String::dump() const {
756: body.dump();
757: langs.dump();
758: }
1.195 paf 759: const String& String::trim(String::Trim_kind kind, const char* chars) const {
760: if(!length())
761: return *this;
762:
763: size_t substr_begin, substr_length;
764: Body new_body=body.trim(kind, chars, &substr_begin, &substr_length);
765: if(new_body==body) // we received unchanged pointer, do likewise
766: return *this;
767: // new_body differs from body, adjust langs along
768:
769: String& result=*new String;
770: if(!new_body) // body.trim produced empty result
771: return result;
772: // body.trim produced nonempty result
773:
774: // first: their langs
775: result.langs.append(result.body, langs, substr_begin, substr_length);
776: // next: letters themselves
777: result.body=new_body;
778:
779: ASSERT_STRING_INVARIANT(result);
780: return result;
1.198 paf 781: }
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