Annotation of parser3/src/main/pa_string.C, revision 1.198
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:
1.198 ! paf 8: static const char * const IDENT_STRING_C="$Date: 2004/04/02 13:57:07 $";
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:
156: size_t start=0;
157: size_t end=our_length;
1.196 paf 158: // from left...
159: if(kind!=TRIM_END) {
1.195 paf 160: CORD_pos pos; set_pos(pos, 0);
161: while(true) {
162: char c=CORD_pos_fetch(pos);
163: if(strchr(chars, c)) {
164: if(++start==our_length)
165: return 0; // all chars are empty, just return empty string
166: } else
167: break;
168:
169: CORD_next(pos);
170: }
171: }
1.196 paf 172: // from right..
173: if(kind!=TRIM_START) {
174: CORD_pos pos; set_pos(pos, end-1);
175: while(true) {
176: char c=CORD_pos_fetch(pos);
177: if(strchr(chars, c)) {
178: if(--end==0) // optimization: NO need to check for 'end>=start', that's(<) impossible
179: return 0; // all chars are empty, just return empty string
180: } else
181: break;
182:
183: CORD_prev(pos);
184: }
185: }
1.195 paf 186:
187: if(start==0 && end==our_length) // nobody moved a thing
188: return *this;
189:
190: if(out_start)
191: *out_start=start;
192: size_t new_length=end-start;
193: if(out_length)
194: *out_length=new_length;
195:
196: return mid(start, new_length);
197: }
198:
1.174 paf 199: static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
200: uint& result=*static_cast<uint*>(client_data);
201: generic_hash_code(result, c);
202: return 0;
203: }
204: static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
205: uint& result=*static_cast<uint*>(client_data);
206: generic_hash_code(result, s);
207: return 0;
208: };
1.176 paf 209: uint String::Body::hash_code() const {
1.174 paf 210: uint result=0;
211: CORD_iter5(body, 0,
212: CORD_batched_iter_fn_generic_hash_code,
213: CORD_batched_iter_fn_generic_hash_code, &result);
1.120 paf 214: return result;
1.94 parser 215: }
216:
1.174 paf 217: // String methods
218:
219: String::String(const char* cstr, size_t helper_length, bool tainted): body(CORD_EMPTY) {
220: append_help_length(cstr, helper_length, tainted?L_TAINTED:L_CLEAN);
1.115 paf 221: }
1.174 paf 222: String::String(const String::C cstr, bool tainted): body(CORD_EMPTY) {
223: append_know_length(cstr.str, cstr.length, tainted?L_TAINTED:L_CLEAN);
1.5 paf 224: }
1.28 paf 225:
1.174 paf 226: String& String::append_know_length(const char* str, size_t known_length, Language lang) {
227: if(!known_length)
1.9 paf 228: return *this;
1.122 paf 229:
1.176 paf 230: // first: langs
231: langs.append(body, lang, known_length);
232: // next: letters themselves
1.174 paf 233: body.append_know_length(str, known_length);
1.1 paf 234:
1.174 paf 235: ASSERT_STRING_INVARIANT(*this);
1.1 paf 236: return *this;
237: }
1.174 paf 238: String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
239: if(!str)
240: return *this;
241: size_t known_length=helper_length?helper_length:strlen(str);
242: if(!known_length)
243: return *this;
1.1 paf 244:
1.174 paf 245: return append_know_length(str, known_length, lang);
1.5 paf 246: }
1.174 paf 247: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
248: size_t known_length=helper_length?helper_length:strlen(str);
249: if(!known_length)
250: return *this;
1.5 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_strdup_know_length(str, known_length);
1.33 paf 256:
1.174 paf 257: ASSERT_STRING_INVARIANT(*this);
258: return *this;
1.5 paf 259: }
1.46 paf 260:
1.174 paf 261: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
262: String& String::mid(size_t substr_begin, size_t substr_end) const {
263: String& result=*new String;
264:
265: size_t self_length=length();
266: substr_begin=min(substr_begin, self_length);
267: substr_end=min(max(substr_end, substr_begin), self_length);
1.176 paf 268: size_t substr_length=substr_end-substr_begin;
269: if(!substr_length)
1.107 parser 270: return result;
1.53 paf 271:
1.176 paf 272: // first: their langs
273: result.langs.append(result.body, langs, substr_begin, substr_length);
274: // next: letters themselves
275: result.body=body.mid(substr_begin, substr_length);
1.174 paf 276:
277: // SAPI::log("piece of '%s' from %d to %d is '%s'",
278: //cstr(), substr_begin, substr_end, result.cstr());
279: ASSERT_STRING_INVARIANT(result);
1.53 paf 280: return result;
1.54 paf 281: }
282:
1.176 paf 283: size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
1.183 paf 284: size_t substr_length=substr.length();
285: while(true) {
286: size_t substr_begin=body.pos(substr, this_offset);
287:
288: if(substr_begin==CORD_NOT_FOUND)
289: return STRING_NOT_FOUND;
1.174 paf 290:
1.183 paf 291: if(langs.check_lang(lang, substr_begin, substr_length))
292: return substr_begin;
293:
294: this_offset=substr_begin+substr_length;
295: }
1.58 paf 296: }
297:
1.174 paf 298: size_t String::pos(const String& substr,
299: size_t this_offset, Language lang) const {
300: return pos(substr.body, this_offset, lang);
1.60 paf 301: }
302:
1.174 paf 303: void String::split(ArrayString& result,
304: size_t& pos_after,
305: const char* delim,
306: Language lang, int limit) const {
307: size_t self_length=length();
308: if(size_t delim_length=strlen(delim)) {
1.186 paf 309: size_t pos_before;
1.60 paf 310: // while we have 'delim'...
1.174 paf 311: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 312: result+=&mid(pos_after, pos_before);
1.174 paf 313: pos_after=pos_before+delim_length;
1.60 paf 314: }
315: // last piece
1.174 paf 316: if(pos_after<self_length && limit) {
317: result+=&mid(pos_after, self_length);
318: pos_after=self_length;
1.60 paf 319: }
320: } else { // empty delim
321: result+=this;
1.174 paf 322: pos_after+=self_length;
1.60 paf 323: }
324: }
325:
1.174 paf 326: void String::split(ArrayString& result,
327: size_t& pos_after,
328: const String& delim, Language lang,
329: int limit) const {
1.140 paf 330: if(!delim.is_empty()) {
1.186 paf 331: size_t pos_before;
1.60 paf 332: // while we have 'delim'...
1.174 paf 333: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 334: result+=&mid(pos_after, pos_before);
1.174 paf 335: pos_after=pos_before+delim.length();
1.60 paf 336: }
337: // last piece
1.174 paf 338: if(pos_after<length() && limit) {
339: result+=&mid(pos_after, length());
340: pos_after=length();
1.60 paf 341: }
342: } else { // empty delim
343: result+=this;
1.174 paf 344: pos_after+=length();
1.60 paf 345: }
1.61 paf 346: }
347:
1.174 paf 348: static void regex_options(const String* options, int *result, bool& need_pre_post_match){
1.63 paf 349: struct Regex_option {
1.174 paf 350: const char* keyL;
351: const char* keyU;
1.63 paf 352: int clear, set;
353: int *result;
1.154 paf 354: bool *flag;
1.63 paf 355: } regex_option[]={
1.189 paf 356: {"i", "I", 0, PCRE_CASELESS, result, 0}, // a=A
357: {"s", "S", 0, PCRE_DOTALL, result, 0}, // \n\n$ [default]
358: {"x", "U", 0, PCRE_EXTENDED, result, 0}, // whitespace in regex ignored
359: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result, 0}, // ^aaa\n$^bbb\n$
360: {"g", "G", 0, 1, result+1, 0}, // many rows
1.154 paf 361: {"'", 0, 0, 0, 0, &need_pre_post_match},
1.189 paf 362: {0, 0, 0, 0, 0, 0}
1.63 paf 363: };
1.171 paf 364: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 365: result[1]=0;
366:
1.174 paf 367: if(options && !options->is_empty())
1.153 paf 368: for(Regex_option *o=regex_option; o->keyL; o++)
1.174 paf 369: if(options->pos(o->keyL)!=STRING_NOT_FOUND
370: || (o->keyU && options->pos(o->keyU)!=STRING_NOT_FOUND)) {
1.154 paf 371: if(o->flag)
372: *o->flag=true;
373: else { // result
374: *o->result &= ~o->clear;
375: *o->result |= o->set;
376: }
1.63 paf 377: }
378: }
379:
1.174 paf 380: Table* String::match(Charset& source_charset,
381: const String& regexp,
382: const String* options,
383: Row_action row_action, void *info,
384: bool& just_matched) const {
1.140 paf 385: if(regexp.is_empty())
1.149 paf 386: throw Exception(0,
1.174 paf 387: 0,
1.73 paf 388: "regexp is empty");
1.154 paf 389:
1.174 paf 390: const char* pattern=regexp.cstr();
391: const char* errptr;
1.62 paf 392: int erroffset;
1.173 paf 393: bool need_pre_post_match=false;
1.174 paf 394: int option_bits[2]={0}; regex_options(options, option_bits, need_pre_post_match);
395: bool global=option_bits[1]!=0;
1.63 paf 396: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 397: &errptr, &erroffset,
1.174 paf 398: source_charset.pcre_tables);
1.62 paf 399:
1.67 paf 400: if(!code)
1.149 paf 401: throw Exception(0,
1.174 paf 402: ®exp.mid(erroffset, regexp.length()),
1.74 paf 403: "regular expression syntax error - %s", errptr);
1.62 paf 404:
1.174 paf 405: int subpatterns=pcre_info(code, 0, 0);
406: if(subpatterns<0) {
1.100 parser 407: pcre_free(code);
1.149 paf 408: throw Exception(0,
1.174 paf 409: ®exp,
1.76 paf 410: "pcre_info error (%d)",
1.174 paf 411: subpatterns);
1.63 paf 412: }
413:
1.174 paf 414: const char* subject=cstr();
415: size_t subject_length=strlen(subject);
416: const int oveclength=(1/*match*/+MAX_MATCH_GROUPS)*3;
417: int ovector[oveclength];
1.155 paf 418:
419: // create table
1.173 paf 420: Table::Action_options table_options;
1.174 paf 421: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 422:
1.64 paf 423: int exec_option_bits=0;
1.154 paf 424: int prestart=0;
425: int poststart=0;
1.174 paf 426: int postfinish=length();
1.63 paf 427: while(true) {
428: int exec_substrings=pcre_exec(code, 0,
1.174 paf 429: subject, subject_length, prestart,
430: exec_option_bits, ovector, oveclength);
1.63 paf 431:
432: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 433: pcre_free(code);
1.174 paf 434: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
435: if(global || subpatterns)
436: return &table; // global or with subpatterns=true+result
437: else {
438: just_matched=false; return 0; // not global=no result
439: }
1.63 paf 440: }
441:
442: if(exec_substrings<0) {
1.100 parser 443: pcre_free(code);
1.149 paf 444: throw Exception(0,
1.174 paf 445: ®exp,
1.76 paf 446: "regular expression execute error (%d)",
1.63 paf 447: exec_substrings);
448: }
449:
1.154 paf 450: int prefinish=ovector[0];
451: poststart=ovector[1];
1.174 paf 452: ArrayString* row=new ArrayString;
453: if(need_pre_post_match) {
454: *row+=&mid(0, prefinish); // .prematch column value
455: *row+=&mid(prefinish, poststart); // .match
456: *row+=&mid(poststart, postfinish); // .postmatch
457: } else {
1.185 paf 458: *row+=&Empty; // .prematch column value
459: *row+=&Empty; // .match
460: *row+=&Empty; // .postmatch
1.174 paf 461: }
1.63 paf 462:
463: for(int i=1; i<exec_substrings; i++) {
1.69 paf 464: // -1:-1 case handled peacefully by mid() itself
1.174 paf 465: *row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 466: }
467:
1.174 paf 468: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 469:
1.174 paf 470: if(!global || prestart==poststart) { // not global | going to hang
1.100 parser 471: pcre_free(code);
1.174 paf 472: row_action(table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
473: return &table;
1.63 paf 474: }
1.154 paf 475: prestart=poststart;
1.63 paf 476:
477: /*
478: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 479: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 480: */
481: }
1.82 parser 482: }
483:
1.174 paf 484: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
485: String& result=*new String();
486: if(is_empty())
487: return result;
488:
489: char* new_cstr=cstrm();
1.192 paf 490: size_t new_cstr_len=length();
1.181 paf 491: if(source_charset.isUTF8()) {
492: switch(kind) {
493: case CC_UPPER:
1.192 paf 494: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
1.181 paf 495: break;
496: case CC_LOWER:
1.192 paf 497: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower);
1.181 paf 498: break;
499: default:
500: assert(!"unknown change case kind");
501: break; // never
502: }
503:
504: } else {
505: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 506:
1.181 paf 507: const unsigned char *a;
508: const unsigned char *b;
509: switch(kind) {
510: case CC_UPPER:
511: a=tables+lcc_offset;
512: b=tables+fcc_offset;
513: break;
514: case CC_LOWER:
515: a=tables+lcc_offset;
516: b=0;
517: break;
518: default:
519: assert(!"unknown change case kind");
520: a=b=0; // calm, compiler
521: break; // never
522: }
523:
1.192 paf 524: char *dest=new_cstr;
1.181 paf 525: unsigned char index;
1.190 paf 526: for(const char* current=new_cstr; (index=(unsigned char)*current); current++) {
1.181 paf 527: unsigned char c=a[index];
528: if(b)
529: c=b[c];
530:
531: *dest++=(char)c;
532: }
1.174 paf 533: }
1.176 paf 534: result.langs=langs;
1.174 paf 535: result.body=new_cstr;
1.89 parser 536:
1.101 parser 537: return result;
538: }
539:
1.174 paf 540: const String& String::replace(const Dictionary& dict) const {
541: String& result=*new String();
542: const char* old_cstr=cstr();
543: const char* prematch_begin=old_cstr;
544:
545: const char* current=old_cstr;
546: while(*current) {
1.184 paf 547: if(Dictionary::Subst subst=dict.first_that_begins(current)) {
1.174 paf 548: // prematch
549: if(size_t prematch_length=current-prematch_begin) {
1.179 paf 550: result.langs.append(result.body, langs, prematch_begin-old_cstr, prematch_length);
1.174 paf 551: result.body.append_strdup_know_length(prematch_begin, prematch_length);
1.101 parser 552: }
553:
1.174 paf 554: // match
555: // skip 'a' in 'current'; move prematch_begin
1.184 paf 556: current+=subst.from_length; prematch_begin=current;
1.174 paf 557:
1.184 paf 558: if(const String* b=subst.to) // are there any b?
1.174 paf 559: result<<*b;
560: } else // simply advance
561: current++;
562: }
1.156 paf 563:
1.174 paf 564: // postmatch
565: if(size_t postmatch_length=current-prematch_begin) {
1.179 paf 566: result.langs.append(result.body, langs, prematch_begin-old_cstr, postmatch_length);
1.174 paf 567: result.body.append_strdup_know_length(prematch_begin, postmatch_length);
568: }
1.156 paf 569:
1.174 paf 570: ASSERT_STRING_INVARIANT(result);
1.82 parser 571: return result;
1.61 paf 572: }
1.113 parser 573:
1.180 paf 574: static int serialize_body_char(char c, char** cur) {
575: *((*cur)++)=c;
576: return 0; // 0=continue
577: };
1.174 paf 578: static int serialize_body_piece(const char* s, char** cur) {
579: size_t length=strlen(s);
580: memcpy(*cur, s, length); *cur+=length;
1.178 paf 581: return 0; // 0=continue
1.174 paf 582: };
1.178 paf 583: static int serialize_lang_piece(char alang, size_t asize, char** cur) {
584: // lang
1.191 paf 585: **cur=alang; (*cur)++;
586: // length [WARNING: not cast, addresses must be %4=0 on sparc]
1.178 paf 587: memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize);
588:
589: return 0; // 0=continue
590: }
1.174 paf 591: String::Cm String::serialize(size_t prolog_length) const {
1.178 paf 592: size_t fragments_count=langs.count();
1.174 paf 593: size_t buf_length=
1.178 paf 594: prolog_length //1
595: +sizeof(size_t) //2
596: +fragments_count*(sizeof(char)+sizeof(size_t)) //3
597: +body.length() //4
598: +1; // for zero terminator used in deserialize
1.174 paf 599: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
600:
601: // 1: prolog
602: char *cur=result.str+prolog_length;
1.191 paf 603: // 2: langs.count [WARNING: not cast, addresses must be %4=0 on sparc]
1.174 paf 604: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
605: // 3: lang info
1.178 paf 606: langs.for_each(body, serialize_lang_piece, &cur);
1.174 paf 607: // 4: letters
1.180 paf 608: body.for_each(serialize_body_char, serialize_body_piece, &cur);
1.182 paf 609: // 5: zero terminator
610: *cur=0;
1.113 parser 611:
1.174 paf 612: return result;
1.113 parser 613: }
1.174 paf 614: bool String::deserialize(size_t prolog_length, void *buf, size_t buf_length) {
615: if(buf_length<=prolog_length)
1.148 paf 616: return false;
1.174 paf 617: buf_length-=prolog_length;
1.178 paf 618: buf_length-=1; // 5: zero terminator
1.135 paf 619:
1.174 paf 620: // 1: prolog
621: const char* cur=(const char* )buf+prolog_length;
1.113 parser 622:
1.176 paf 623: // 2: langs.count
1.191 paf 624: size_t fragments_count;
625: if(buf_length<sizeof(fragments_count)) // langs.count don't fit?
1.174 paf 626: return false;
1.191 paf 627: // [WARNING: not cast, addresses must be %4=0 on sparc]
628: memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count);
629: buf_length-=sizeof(fragments_count);
1.174 paf 630:
631: if(fragments_count) {
632: // 3: lang info
633: size_t total_length=0;
634: for(size_t f=0; f<fragments_count; f++) {
1.191 paf 635: char lang;
636: size_t fragment_length;
637: size_t piece_length=sizeof(lang)+sizeof(fragment_length);
1.174 paf 638: if(buf_length<piece_length) // lang+length
639: return false;
640:
1.191 paf 641: // lang
642: lang=*cur++;
643: // length [WARNING: not cast, addresses must be %4=0 on sparc]
644: memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length);
645:
646: // uchar needed to prevent propagating 0x80 bit to upper bytes
647: langs.append(total_length, (String::Language)(uchar)lang, fragment_length);
1.174 paf 648: total_length+=fragment_length;
1.148 paf 649:
1.174 paf 650: buf_length-=piece_length;
651: }
1.128 paf 652:
1.174 paf 653: // 4: letters
654: if(buf_length!=total_length)
1.148 paf 655: return false;
656:
1.178 paf 657: // serialize wrote extra zero byte there, we can rely on that
1.176 paf 658: body=String::Body(cur, buf_length);
1.174 paf 659: }
1.113 parser 660:
1.174 paf 661: ASSERT_STRING_INVARIANT(*this);
1.148 paf 662: return true;
1.176 paf 663: }
664:
665: const char* String::Body::v() const {
666: return CORD_to_const_char_star(body);
667: }
668: const char* String::Languages::v() const {
1.177 paf 669: if(opt.is_not_just_lang)
1.176 paf 670: return CORD_to_const_char_star(langs);
671: else
672: return (const char*)&langs;
673: }
674: const char* String::v() const {
1.198 ! paf 675: const uint LIMIT_VIEW=20;
1.176 paf 676: char* buf=(char*)malloc(MAX_STRING);
677: const char*body_view=body.v();
678: const char*langs_view=langs.v();
679: snprintf(buf, MAX_STRING,
1.178 paf 680: "%d:%.*s%s} "
1.176 paf 681: "{%d:%s",
1.178 paf 682: langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"",
1.176 paf 683: strlen(body_view), body_view
684: );
685:
686: return buf;
1.113 parser 687: }
1.195 paf 688:
689: const String& String::trim(String::Trim_kind kind, const char* chars) const {
690: if(!length())
691: return *this;
692:
693: size_t substr_begin, substr_length;
694: Body new_body=body.trim(kind, chars, &substr_begin, &substr_length);
695: if(new_body==body) // we received unchanged pointer, do likewise
696: return *this;
697: // new_body differs from body, adjust langs along
698:
699: String& result=*new String;
700: if(!new_body) // body.trim produced empty result
701: return result;
702: // body.trim produced nonempty result
703:
704: // first: their langs
705: result.langs.append(result.body, langs, substr_begin, substr_length);
706: // next: letters themselves
707: result.body=new_body;
708:
709: ASSERT_STRING_INVARIANT(result);
710: return result;
1.198 ! paf 711: }
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