Annotation of parser3/src/main/pa_string.C, revision 1.197
1.45 paf 1: /** @file
1.174 paf 2: Parser: string class. @see untalength_t.C.
1.46 paf 3:
1.194 paf 4: Copyright (c) 2001-2004 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:
1.197 ! paf 8: static const char * const IDENT_STRING_C="$Date: 2004/02/27 15:24:03 $";
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:
1.185 paf 18: const String String::Empty;
19:
1.193 paf 20: int pa_atoi(const char* str, const String* problem_source) {
21: if(!str)
22: return 0;
23:
24: while(*str && isspace(*str))
25: str++;
26: if(!*str)
27: return 0;
28:
29: int result;
30: char *error_pos;
31: // 0xABC
32: if(str[0]=='0')
33: if(str[1]=='x' || str[1]=='X')
34: result=(int)(unsigned long)strtol(str, &error_pos, 0);
1.197 ! paf 35: else {
! 36: // skip leading 0000, to disable octal interpretation
! 37: do str++; while(*str=='0');
! 38: result=(int)strtol(str, &error_pos, 0);
! 39: }
1.193 paf 40: else
41: result=(int)strtol(str, &error_pos, 0);
42:
43: while(char c=*error_pos++)
44: if(!isspace(c))
45: throw Exception("number.format",
46: problem_source,
47: problem_source?"invalid number (int)": "'%s' is invalid number (int)", str);
48:
49: return result;
50: }
51:
52: double pa_atod(const char* str, const String* problem_source) {
53: if(!str)
54: return 0;
55:
56: while(*str && isspace(*str))
57: str++;
58: if(!*str)
59: return 0;
60:
61: double result;
62: char *error_pos;
63: // 0xABC
64: if(str[0]=='0')
65: if(str[1]=='x' || str[1]=='X')
66: result=(double)(unsigned long)strtol(str, &error_pos, 0);
67: else
68: result=(double)strtod(str+1/*skip leading 0*/, &error_pos);
69: else
70: result=(double)strtod(str, &error_pos);
71:
72: while(char c=*error_pos++)
73: if(!isspace(c))
74: throw Exception("number.format",
75: problem_source,
76: problem_source?"invalid number (double)": "'%s' is invalid number (double)", str);
77:
78: return result;
79: }
80:
1.176 paf 81: // cord lib extension
82:
83: #ifndef DOXYGEN
84: typedef struct {
85: ssize_t countdown;
86: char target; /* Character we're looking for */
87: } chr_data;
88: #endif
89: static int CORD_range_contains_chr_greater_then_proc(char c, size_t size, void* client_data)
90: {
91: register chr_data * d = (chr_data *)client_data;
92:
93: if (d -> countdown<=0) return(2);
94: d -> countdown -= size;
95: if (c > d -> target) return(1);
96: return(0);
97: }
98: int CORD_range_contains_chr_greater_then(CORD x, size_t i, size_t n, int c)
99: {
100: chr_data d;
101:
102: d.countdown = n;
103: d.target = c;
104: return(CORD_block_iter(x, i, CORD_range_contains_chr_greater_then_proc, &d) == 1/*alternatives: 0 normally ended, 2=struck 'n'*/);
105: }
106:
1.187 paf 107: static int CORD_block_count_proc(char /*c*/, size_t /*size*/, void* client_data)
1.178 paf 108: {
109: int* result=(int*)client_data;
110: (*result)++;
111: return(0); // 0=continue
112: }
113: size_t CORD_block_count(CORD x)
114: {
115: size_t result=0;
116: CORD_block_iter(x, 0, CORD_block_count_proc, &result);
117: return result;
118: }
119:
1.174 paf 120: // helpers
1.139 paf 121:
1.174 paf 122: /// String::match uses this as replace & global search table columns
1.139 paf 123:
1.174 paf 124: const int MAX_MATCH_GROUPS=100;
1.139 paf 125:
1.174 paf 126: class String_match_table_template_columns: public ArrayString {
127: public:
128: String_match_table_template_columns() {
129: *this+=new String("prematch");
130: *this+=new String("match");
131: *this+=new String("postmatch");
132: for(int i=0; i<MAX_MATCH_GROUPS; i++) {
1.176 paf 133: *this+=new String(String::Body::Format(1+i), String::L_CLEAN);
1.174 paf 134: }
135: }
136: };
137:
138: Table string_match_table_template(new String_match_table_template_columns);
139:
1.176 paf 140: // String::Body methods
1.140 paf 141:
1.176 paf 142: String::Body String::Body::Format(int value) {
1.174 paf 143: char local[MAX_NUMBER];
144: size_t length=snprintf(local, MAX_NUMBER, "%d", value);
1.176 paf 145: return String::Body(pa_strdup(local, length), length);
1.120 paf 146: }
147:
1.195 paf 148: String::Body String::Body::trim(String::Trim_kind kind, const char* chars,
149: size_t* out_start, size_t* out_length) const {
150: size_t our_length=length();
151: if(!our_length)
152: return *this;
153: if(!chars)
154: chars=" \t\n"; // white space
155: Body result=*this;
156:
157: size_t start=0;
158: size_t end=our_length;
1.196 paf 159: // from left...
160: if(kind!=TRIM_END) {
1.195 paf 161: CORD_pos pos; set_pos(pos, 0);
162: while(true) {
163: char c=CORD_pos_fetch(pos);
164: if(strchr(chars, c)) {
165: if(++start==our_length)
166: return 0; // all chars are empty, just return empty string
167: } else
168: break;
169:
170: CORD_next(pos);
171: }
172: }
1.196 paf 173: // from right..
174: if(kind!=TRIM_START) {
175: CORD_pos pos; set_pos(pos, end-1);
176: while(true) {
177: char c=CORD_pos_fetch(pos);
178: if(strchr(chars, c)) {
179: if(--end==0) // optimization: NO need to check for 'end>=start', that's(<) impossible
180: return 0; // all chars are empty, just return empty string
181: } else
182: break;
183:
184: CORD_prev(pos);
185: }
186: }
1.195 paf 187:
188: if(start==0 && end==our_length) // nobody moved a thing
189: return *this;
190:
191: if(out_start)
192: *out_start=start;
193: size_t new_length=end-start;
194: if(out_length)
195: *out_length=new_length;
196:
197: return mid(start, new_length);
198: }
199:
1.174 paf 200: static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
201: uint& result=*static_cast<uint*>(client_data);
202: generic_hash_code(result, c);
203: return 0;
204: }
205: static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
206: uint& result=*static_cast<uint*>(client_data);
207: generic_hash_code(result, s);
208: return 0;
209: };
1.176 paf 210: uint String::Body::hash_code() const {
1.174 paf 211: uint result=0;
212: CORD_iter5(body, 0,
213: CORD_batched_iter_fn_generic_hash_code,
214: CORD_batched_iter_fn_generic_hash_code, &result);
1.120 paf 215: return result;
1.94 parser 216: }
217:
1.174 paf 218: // String methods
219:
220: String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
221: append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.115 paf 222: }
1.174 paf 223: String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) {
224: append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN);
1.5 paf 225: }
1.28 paf 226:
1.174 paf 227: String& String::append_know_length(const char* str, size_t known_length, Language lang) {
228: if(!known_length)
1.9 paf 229: return *this;
1.122 paf 230:
1.176 paf 231: // first: langs
232: langs.append(body, lang, known_length);
233: // next: letters themselves
1.174 paf 234: body.append_know_length(str, known_length);
1.1 paf 235:
1.174 paf 236: ASSERT_STRING_INVARIANT(*this);
1.1 paf 237: return *this;
238: }
1.174 paf 239: String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
240: if(!str)
241: return *this;
242: size_t known_length=helper_length?helper_length:strlen(str);
243: if(!known_length)
244: return *this;
1.1 paf 245:
1.174 paf 246: return append_know_length(str, known_length, lang);
1.5 paf 247: }
1.174 paf 248: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
249: size_t known_length=helper_length?helper_length:strlen(str);
250: if(!known_length)
251: return *this;
1.5 paf 252:
1.176 paf 253: // first: langs
254: langs.append(body, lang, known_length);
255: // next: letters themselves
1.174 paf 256: body.append_strdup_know_length(str, known_length);
1.33 paf 257:
1.174 paf 258: ASSERT_STRING_INVARIANT(*this);
259: return *this;
1.5 paf 260: }
1.46 paf 261:
1.174 paf 262: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
263: String& String::mid(size_t substr_begin, size_t substr_end) const {
264: String& result=*new String;
265:
266: size_t self_length=length();
267: substr_begin=min(substr_begin, self_length);
268: substr_end=min(max(substr_end, substr_begin), self_length);
1.176 paf 269: size_t substr_length=substr_end-substr_begin;
270: if(!substr_length)
1.107 parser 271: return result;
1.53 paf 272:
1.176 paf 273: // first: their langs
274: result.langs.append(result.body, langs, substr_begin, substr_length);
275: // next: letters themselves
276: result.body=body.mid(substr_begin, substr_length);
1.174 paf 277:
278: // SAPI::log("piece of '%s' from %d to %d is '%s'",
279: //cstr(), substr_begin, substr_end, result.cstr());
280: ASSERT_STRING_INVARIANT(result);
1.53 paf 281: return result;
1.54 paf 282: }
283:
1.176 paf 284: size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
1.183 paf 285: size_t substr_length=substr.length();
286: while(true) {
287: size_t substr_begin=body.pos(substr, this_offset);
288:
289: if(substr_begin==CORD_NOT_FOUND)
290: return STRING_NOT_FOUND;
1.174 paf 291:
1.183 paf 292: if(langs.check_lang(lang, substr_begin, substr_length))
293: return substr_begin;
294:
295: this_offset=substr_begin+substr_length;
296: }
1.58 paf 297: }
298:
1.174 paf 299: size_t String::pos(const String& substr,
300: size_t this_offset, Language lang) const {
301: return pos(substr.body, this_offset, lang);
1.60 paf 302: }
303:
1.174 paf 304: void String::split(ArrayString& result,
305: size_t& pos_after,
306: const char* delim,
307: Language lang, int limit) const {
308: size_t self_length=length();
309: if(size_t delim_length=strlen(delim)) {
1.186 paf 310: size_t pos_before;
1.60 paf 311: // while we have 'delim'...
1.174 paf 312: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 313: result+=&mid(pos_after, pos_before);
1.174 paf 314: pos_after=pos_before+delim_length;
1.60 paf 315: }
316: // last piece
1.174 paf 317: if(pos_after<self_length && limit) {
318: result+=&mid(pos_after, self_length);
319: pos_after=self_length;
1.60 paf 320: }
321: } else { // empty delim
322: result+=this;
1.174 paf 323: pos_after+=self_length;
1.60 paf 324: }
325: }
326:
1.174 paf 327: void String::split(ArrayString& result,
328: size_t& pos_after,
329: const String& delim, Language lang,
330: int limit) const {
1.140 paf 331: if(!delim.is_empty()) {
1.186 paf 332: size_t pos_before;
1.60 paf 333: // while we have 'delim'...
1.174 paf 334: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 335: result+=&mid(pos_after, pos_before);
1.174 paf 336: pos_after=pos_before+delim.length();
1.60 paf 337: }
338: // last piece
1.174 paf 339: if(pos_after<length() && limit) {
340: result+=&mid(pos_after, length());
341: pos_after=length();
1.60 paf 342: }
343: } else { // empty delim
344: result+=this;
1.174 paf 345: pos_after+=length();
1.60 paf 346: }
1.61 paf 347: }
348:
1.174 paf 349: static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 350: struct Regex_option {
1.174 paf 351: const char* keyL;
352: const char* keyU;
1.63 paf 353: int clear, set;
354: int *result;
1.154 paf 355: bool *flag;
1.63 paf 356: } regex_option[]={
1.189 paf 357: {"i", "I", 0, PCRE_CASELESS, result, 0}, // a=A
358: {"s", "S", 0, PCRE_DOTALL, result, 0}, // \n\n$ [default]
359: {"x", "U", 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored
360: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result, 0}, // ^aaa\n$^bbb\n$
361: {"g", "G", 0, 1, result+1, 0}, // many rows
1.154 paf 362: {"'", 0, 0, 0, 0, &need_pre_post_match},
1.189 paf 363: {0, 0, 0, 0, 0, 0}
1.63 paf 364: };
1.171 paf 365: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 366: result[1]=0;
367:
1.174 paf 368: if(options && !options->is_empty())
1.153 paf 369: for(Regex_option *o=regex_option; o->keyL; o++)
1.174 paf 370: if(options->pos(o->keyL)!=STRING_NOT_FOUND
371: || (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)) {
1.154 paf 372: if(o->flag)
373: *o->flag=true;
374: else { // result
375: *o->result &= ~o->clear;
376: *o->result |= o->set;
377: }
1.63 paf 378: }
379: }
380:
1.174 paf 381: Table* String::match(Charset& source_charset,
382: const String& regexp,
383: const String* options,
384: Row_action row_action, void *info,
385: bool& just_matched) const {
1.140 paf 386: if(regexp.is_empty())
1.149 paf 387: throw Exception(0,
1.174 paf 388: 0,
1.73 paf 389: "regexp is empty");
1.154 paf 390:
1.174 paf 391: const char* pattern=regexp.cstr();
392: const char* errptr;
1.62 paf 393: int erroffset;
1.173 paf 394: bool need_pre_post_match=false;
1.174 paf 395: int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
396: bool global=option_bits[1]!=0;
1.63 paf 397: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 398: &errptr, &erroffset,
1.174 paf 399: source_charset.pcre_tables);
1.62 paf 400:
1.67 paf 401: if(!code)
1.149 paf 402: throw Exception(0,
1.174 paf 403: ®exp.mid(erroffset, regexp.length()),
1.74 paf 404: "regular expression syntax error - %s", errptr);
1.62 paf 405:
1.174 paf 406: int subpatterns=pcre_info(code, 0, 0);
407: if(subpatterns<0) {
1.100 parser 408: pcre_free(code);
1.149 paf 409: throw Exception(0,
1.174 paf 410: ®exp,
1.76 paf 411: "pcre_info error (%d)",
1.174 paf 412: subpatterns);
1.63 paf 413: }
414:
1.174 paf 415: const char* subject=cstr();
416: size_t subject_length=strlen(subject);
417: const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
418: int ovector[oveclength];
1.155 paf 419:
420: // create table
1.173 paf 421: Table::Action_options table_options;
1.174 paf 422: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 423:
1.64 paf 424: int exec_option_bits=0;
1.154 paf 425: int prestart=0;
426: int poststart=0;
1.174 paf 427: int postfinish=length();
1.63 paf 428: while(true) {
429: int exec_substrings=pcre_exec(code, 0,
1.174 paf 430: subject, subject_length, prestart,
431: exec_option_bits, ovector, oveclength);
1.63 paf 432:
433: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 434: pcre_free(code);
1.174 paf 435: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
436: if(global || subpatterns)
437: return &table; // global or with subpatterns=true+result
438: else {
439: just_matched=false; return 0; // not global=no result
440: }
1.63 paf 441: }
442:
443: if(exec_substrings<0) {
1.100 parser 444: pcre_free(code);
1.149 paf 445: throw Exception(0,
1.174 paf 446: ®exp,
1.76 paf 447: "regular expression execute error (%d)",
1.63 paf 448: exec_substrings);
449: }
450:
1.154 paf 451: int prefinish=ovector[0];
452: poststart=ovector[1];
1.174 paf 453: ArrayString* row=new ArrayString;
454: if(need_pre_post_match) {
455: *row+=&mid(0, prefinish); // .prematch column value
456: *row+=&mid(prefinish, poststart); // .match
457: *row+=&mid(poststart, postfinish); // .postmatch
458: } else {
1.185 paf 459: *row+=&Empty; // .prematch column value
460: *row+=&Empty; // .match
461: *row+=&Empty; // .postmatch
1.174 paf 462: }
1.63 paf 463:
464: for(int i=1; i<exec_substrings; i++) {
1.69 paf 465: // -1:-1 case handled peacefully by mid() itself
1.174 paf 466: *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 467: }
468:
1.174 paf 469: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 470:
1.174 paf 471: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 472: pcre_free(code);
1.174 paf 473: row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
474: return &table;
1.63 paf 475: }
1.154 paf 476: prestart=poststart;
1.63 paf 477:
478: /*
479: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 480: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 481: */
482: }
1.82 parser 483: }
484:
1.174 paf 485: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
486: String& result=*new String();
487: if(is_empty())
488: return result;
489:
490: char* new_cstr=cstrm();
1.192 paf 491: size_t new_cstr_len=length();
1.181 paf 492: if(source_charset.isUTF8()) {
493: switch(kind) {
494: case CC_UPPER:
1.192 paf 495: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
1.181 paf 496: break;
497: case CC_LOWER:
1.192 paf 498: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower);
1.181 paf 499: break;
500: default:
501: assert(!"unknown change case kind");
502: break; // never
503: }
504:
505: } else {
506: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 507:
1.181 paf 508: const unsigned char *a;
509: const unsigned char *b;
510: switch(kind) {
511: case CC_UPPER:
512: a=tables+lcc_offset;
513: b=tables+fcc_offset;
514: break;
515: case CC_LOWER:
516: a=tables+lcc_offset;
517: b=0;
518: break;
519: default:
520: assert(!"unknown change case kind");
521: a=b=0; // calm, compiler
522: break; // never
523: }
524:
1.192 paf 525: char *dest=new_cstr;
1.181 paf 526: unsigned char index;
1.190 paf 527: for(const char* current=new_cstr; (index=(unsigned char)*current); current++) {
1.181 paf 528: unsigned char c=a[index];
529: if(b)
530: c=b[c];
531:
532: *dest++=(char)c;
533: }
1.174 paf 534: }
1.176 paf 535: result.langs=langs;
1.174 paf 536: result.body=new_cstr;
1.89 parser 537:
1.101 parser 538: return result;
539: }
540:
1.174 paf 541: const String& String::replace(const Dictionary& dict) const {
542: String& result=*new String();
543: const char* old_cstr=cstr();
544: const char* prematch_begin=old_cstr;
545:
546: const char* current=old_cstr;
547: while(*current) {
1.184 paf 548: if(Dictionary::Subst subst=dict.first_that_begins(current)) {
1.174 paf 549: // prematch
550: if(size_t prematch_length=current-prematch_begin) {
1.179 paf 551: result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length);
1.174 paf 552: result.body.append_strdup_know_length(prematch_begin, prematch_length);
1.101 parser 553: }
554:
1.174 paf 555: // match
556: // skip 'a' in 'current'; move prematch_begin
1.184 paf 557: current+=subst.from_length; prematch_begin=current;
1.174 paf 558:
1.184 paf 559: if(const String* b=subst.to) // are there any b?
1.174 paf 560: result<<*b;
561: } else // simply advance
562: current++;
563: }
1.156 paf 564:
1.174 paf 565: // postmatch
566: if(size_t postmatch_length=current-prematch_begin) {
1.179 paf 567: result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length);
1.174 paf 568: result.body.append_strdup_know_length(prematch_begin, postmatch_length);
569: }
1.156 paf 570:
1.174 paf 571: ASSERT_STRING_INVARIANT(result);
1.82 parser 572: return result;
1.61 paf 573: }
1.113 parser 574:
1.180 paf 575: static int serialize_body_char(char c, char** cur) {
576: *((*cur)++)=c;
577: return 0; // 0=continue
578: };
1.174 paf 579: static int serialize_body_piece(const char* s, char** cur) {
580: size_t length=strlen(s);
581: memcpy(*cur, s, length); *cur+=length;
1.178 paf 582: return 0; // 0=continue
1.174 paf 583: };
1.178 paf 584: static int serialize_lang_piece(char alang, size_t asize, char** cur) {
585: // lang
1.191 paf 586: **cur=alang; (*cur)++;
587: // length [WARNING: not cast, addresses must be %4=0 on sparc]
1.178 paf 588: memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize);
589:
590: return 0; // 0=continue
591: }
1.174 paf 592: String::Cm String::serialize(size_t prolog_length) const {
1.178 paf 593: size_t fragments_count=langs.count();
1.174 paf 594: size_t buf_length=
1.178 paf 595: prolog_length //1
596: +sizeof(size_t) //2
597: +fragments_count*(sizeof(char)+sizeof(size_t)) //3
598: +body.length() //4
599: +1; // for zero terminator used in deserialize
1.174 paf 600: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
601:
602: // 1: prolog
603: char *cur=result.str+prolog_length;
1.191 paf 604: // 2: langs.count [WARNING: not cast, addresses must be %4=0 on sparc]
1.174 paf 605: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
606: // 3: lang info
1.178 paf 607: langs.for_each(body, serialize_lang_piece, &cur);
1.174 paf 608: // 4: letters
1.180 paf 609: body.for_each(serialize_body_char, serialize_body_piece, &cur);
1.182 paf 610: // 5: zero terminator
611: *cur=0;
1.113 parser 612:
1.174 paf 613: return result;
1.113 parser 614: }
1.174 paf 615: bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
616: if(buf_length<=prolog_length)
1.148 paf 617: return false;
1.174 paf 618: buf_length-=prolog_length;
1.178 paf 619: buf_length-=1; // 5: zero terminator
1.135 paf 620:
1.174 paf 621: // 1: prolog
622: const char* cur=(const char* )buf+prolog_length;
1.113 parser 623:
1.176 paf 624: // 2: langs.count
1.191 paf 625: size_t fragments_count;
626: if(buf_length<sizeof(fragments_count)) // langs.count don't fit?
1.174 paf 627: return false;
1.191 paf 628: // [WARNING: not cast, addresses must be %4=0 on sparc]
629: memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count);
630: buf_length-=sizeof(fragments_count);
1.174 paf 631:
632: if(fragments_count) {
633: // 3: lang info
634: size_t total_length=0;
635: for(size_t f=0; f<fragments_count; f++) {
1.191 paf 636: char lang;
637: size_t fragment_length;
638: size_t piece_length=sizeof(lang)+sizeof(fragment_length);
1.174 paf 639: if(buf_length<piece_length) // lang+length
640: return false;
641:
1.191 paf 642: // lang
643: lang=*cur++;
644: // length [WARNING: not cast, addresses must be %4=0 on sparc]
645: memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length);
646:
647: // uchar needed to prevent propagating 0x80 bit to upper bytes
648: langs.append(total_length, (String::Language)(uchar)lang, fragment_length);
1.174 paf 649: total_length+=fragment_length;
1.148 paf 650:
1.174 paf 651: buf_length-=piece_length;
652: }
1.128 paf 653:
1.174 paf 654: // 4: letters
655: if(buf_length!=total_length)
1.148 paf 656: return false;
657:
1.178 paf 658: // serialize wrote extra zero byte there, we can rely on that
1.176 paf 659: body=String::Body(cur, buf_length);
1.174 paf 660: }
1.113 parser 661:
1.174 paf 662: ASSERT_STRING_INVARIANT(*this);
1.148 paf 663: return true;
1.176 paf 664: }
665:
666: const char* String::Body::v() const {
667: return CORD_to_const_char_star(body);
668: }
669: const char* String::Languages::v() const {
1.177 paf 670: if(opt.is_not_just_lang)
1.176 paf 671: return CORD_to_const_char_star(langs);
672: else
673: return (const char*)&langs;
674: }
675: const char* String::v() const {
1.195 paf 676: const int LIMIT_VIEW=20;
1.176 paf 677: char* buf=(char*)malloc(MAX_STRING);
678: const char*body_view=body.v();
679: const char*langs_view=langs.v();
680: snprintf(buf, MAX_STRING,
1.178 paf 681: "%d:%.*s%s} "
1.176 paf 682: "{%d:%s",
1.178 paf 683: langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"",
1.176 paf 684: strlen(body_view), body_view
685: );
686:
687: return buf;
1.113 parser 688: }
1.195 paf 689:
690: const String& String::trim(String::Trim_kind kind, const char* chars) const {
691: if(!length())
692: return *this;
693:
694: size_t substr_begin, substr_length;
695: Body new_body=body.trim(kind, chars, &substr_begin, &substr_length);
696: if(new_body==body) // we received unchanged pointer, do likewise
697: return *this;
698: // new_body differs from body, adjust langs along
699:
700: String& result=*new String;
701: if(!new_body) // body.trim produced empty result
702: return result;
703: // body.trim produced nonempty result
704:
705: // first: their langs
706: result.langs.append(result.body, langs, substr_begin, substr_length);
707: // next: letters themselves
708: result.body=new_body;
709:
710: ASSERT_STRING_INVARIANT(result);
711: return result;
712: }
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