Annotation of parser3/src/main/pa_string.C, revision 1.241
1.45 paf 1: /** @file
1.174 paf 2: Parser: string class. @see untalength_t.C.
1.46 paf 3:
1.240 moko 4: Copyright (c) 2001-2012 Art. Lebedev Studio (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:
1.12 paf 8: #include "pa_string.h"
1.22 paf 9: #include "pa_exception.h"
1.61 paf 10: #include "pa_table.h"
1.101 parser 11: #include "pa_dictionary.h"
1.132 paf 12: #include "pa_charset.h"
1.222 misha 13: #include "pa_vregex.h"
1.215 misha 14:
1.241 ! misha 15: volatile const char * IDENT_PA_STRING_C="$Id: pa_string.C,v 1.240 2012-03-16 09:24:14 moko Exp $" IDENT_PA_STRING_H;
1.240 moko 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,
1.229 misha 169: size_t* out_start, size_t* out_length, Charset* source_charset) const {
1.195 paf 170: size_t our_length=length();
171: if(!our_length)
172: return *this;
1.229 misha 173:
174: // check if any UTF-8 in chars
175: bool fast=true;
176: if(chars && source_charset && source_charset->isUTF8()){
177: const char* pos=chars;
178: while(unsigned char c=*pos++)
179: if(c>127){
180: fast=false;
181: break;
182: }
183: }
184:
185: size_t start=0;
186: size_t end=our_length;
1.195 paf 187: if(!chars)
188: chars=" \t\n"; // white space
189:
1.229 misha 190: if(fast){
191: // from left...
192: if(kind!=TRIM_END) {
193: CORD_pos pos; set_pos(pos, 0);
194: while(true) {
195: char c=CORD_pos_fetch(pos);
196: if(strchr(chars, c)) {
197: if(++start==our_length)
198: return 0; // all chars are empty, just return empty string
199: } else
200: break;
201:
202: CORD_next(pos);
203: }
204: }
205:
206: // from right..
207: if(kind!=TRIM_START) {
208: CORD_pos pos; set_pos(pos, end-1);
209: while(true) {
210: char c=CORD_pos_fetch(pos);
211: if(strchr(chars, c)) {
212: if(--end==0) // optimization: NO need to check for 'end>=start', that's(<) impossible
213: return 0; // all chars are empty, just return empty string
214: } else
215: break;
216:
217: CORD_prev(pos);
218: }
219: }
220: } else {
1.234 misha 221: const XMLByte* src_begin=(const XMLByte*)cstr();
1.229 misha 222: const XMLByte* src_end=src_begin+our_length;
223:
224: // from left...
225: if(kind!=TRIM_END) {
226: while(src_begin<src_end){
227: uint char_length=1;
228: const XMLByte* ptr=src_begin;
1.231 misha 229: // searching first UTF-8 byte: http://tools.ietf.org/html/rfc3629#section-3
230: while(++src_begin<=src_end && (*src_begin>127 && *src_begin<0xC0))
1.229 misha 231: char_length++;
232:
233: bool found=false;
234: for(const char* chars_byte=chars; chars_byte=strchr(chars_byte, *ptr); chars_byte++)
235: if(strncmp(chars_byte, (const char*)ptr, char_length)==0){
236: found=true;
237: break;
238: }
239:
240: if(found){
241: start+=char_length;
242: if(start==our_length)
243: return 0; // all chars are empty, just return empty string
244: } else
245: break;
246: }
247: }
1.196 paf 248:
1.229 misha 249: // from right..
250: if(kind!=TRIM_START) {
251: while(src_begin<src_end){
252: uint char_length=1;
1.231 misha 253: // searching first UTF-8 byte: http://tools.ietf.org/html/rfc3629#section-3
254: while(src_begin<=--src_end && (*src_end>127 && *src_end<0xC0))
1.229 misha 255: char_length++;
256:
257: bool found=false;
258: for(const char* chars_byte=chars; chars_byte=strchr(chars_byte, *src_end); chars_byte++)
259: if(strncmp(chars_byte, (const char*)src_end, char_length)==0){
260: found=true;
261: break;
262: }
263:
264: if(found){
265: end-=char_length;
266: if(end==0)
267: return 0; // all chars are empty, just return empty string
268: } else
269: break;
270: }
1.196 paf 271: }
272: }
1.195 paf 273:
274: if(start==0 && end==our_length) // nobody moved a thing
275: return *this;
276:
277: if(out_start)
278: *out_start=start;
279: size_t new_length=end-start;
280: if(out_length)
281: *out_length=new_length;
282:
283: return mid(start, new_length);
284: }
285:
1.174 paf 286: static int CORD_batched_iter_fn_generic_hash_code(char c, void * client_data) {
287: uint& result=*static_cast<uint*>(client_data);
288: generic_hash_code(result, c);
289: return 0;
290: }
291: static int CORD_batched_iter_fn_generic_hash_code(const char* s, void * client_data) {
292: uint& result=*static_cast<uint*>(client_data);
293: generic_hash_code(result, s);
294: return 0;
295: };
1.225 misha 296: uint String::Body::get_hash_code() const {
297: #ifdef HASH_CODE_CACHING
298: if(hash_code)
299: return hash_code;
300: #else
301: uint hash_code=0;
302: #endif
1.220 misha 303: if (body && CORD_IS_STRING(body)){
1.225 misha 304: generic_hash_code(hash_code, body);
1.220 misha 305: } else {
306: CORD_iter5(body, 0,
307: CORD_batched_iter_fn_generic_hash_code,
1.225 misha 308: CORD_batched_iter_fn_generic_hash_code, &hash_code);
1.220 misha 309: }
1.225 misha 310: return hash_code;
1.94 parser 311: }
312:
1.241 ! misha 313: struct CORD_pos_info {
! 314: const char* chars;
! 315: size_t left;
! 316: size_t pos;
! 317: };
! 318:
! 319: // can be called only for IS_FUNCTION(CORD) which is used in String::Body::strrpbrk
! 320: static int CORD_iter_fn_rpos(char c, CORD_pos_info* info) {
! 321: if(info->pos < info->left){
! 322: info->pos=STRING_NOT_FOUND;
! 323: return 1;
! 324: }
! 325: if(strchr(info->chars, c))
! 326: return 1;
! 327: --(info->pos);
! 328: return 0;
! 329: }
! 330:
! 331: size_t String::Body::strrpbrk(const char* chars, size_t left, size_t right) const {
! 332: if(is_empty() || !chars || !strlen(chars))
! 333: return STRING_NOT_FOUND;
! 334: CORD_pos_info info={chars, left, right};
! 335: if(CORD_riter4(body, right, (CORD_iter_fn)CORD_iter_fn_rpos, &info))
! 336: return info.pos;
! 337: else
! 338: return STRING_NOT_FOUND;
! 339: }
! 340:
! 341:
! 342: // can be called only for IS_FUNCTION(CORD) which is used in String::Body::rskipchars
! 343: static int CORD_iter_fn_rskip(char c, CORD_pos_info* info) {
! 344: if(info->pos < info->left) {
! 345: info->pos=STRING_NOT_FOUND;
! 346: return 1;
! 347: }
! 348: if(!strchr(info->chars, c))
! 349: return 1;
! 350: --(info->pos);
! 351: return 0;
! 352: }
! 353:
! 354: size_t String::Body::rskipchars(const char* chars, size_t left, size_t right) const {
! 355: if(is_empty() || !chars || !strlen(chars))
! 356: return STRING_NOT_FOUND;
! 357: CORD_pos_info info={chars, left, right};
! 358: if(CORD_riter4(body, right, (CORD_iter_fn)CORD_iter_fn_rskip, &info))
! 359: return info.pos;
! 360: else
! 361: return STRING_NOT_FOUND;
! 362: }
! 363:
1.174 paf 364: // String methods
365:
366: String& String::append_know_length(const char* str, size_t known_length, Language lang) {
367: if(!known_length)
1.9 paf 368: return *this;
1.122 paf 369:
1.176 paf 370: // first: langs
371: langs.append(body, lang, known_length);
372: // next: letters themselves
1.174 paf 373: body.append_know_length(str, known_length);
1.1 paf 374:
1.174 paf 375: ASSERT_STRING_INVARIANT(*this);
1.1 paf 376: return *this;
377: }
1.174 paf 378: String& String::append_help_length(const char* str, size_t helper_length, Language lang) {
379: if(!str)
380: return *this;
381: size_t known_length=helper_length?helper_length:strlen(str);
382: if(!known_length)
383: return *this;
1.1 paf 384:
1.174 paf 385: return append_know_length(str, known_length, lang);
1.5 paf 386: }
1.226 misha 387: String::String(int value, char *format) : langs(L_CLEAN){
388: char buf[MAX_NUMBER];
389: body.append_strdup_know_length(buf, snprintf(buf, MAX_NUMBER, format, value));
390: }
1.174 paf 391: String& String::append_strdup(const char* str, size_t helper_length, Language lang) {
392: size_t known_length=helper_length?helper_length:strlen(str);
393: if(!known_length)
394: return *this;
1.5 paf 395:
1.176 paf 396: // first: langs
397: langs.append(body, lang, known_length);
398: // next: letters themselves
1.174 paf 399: body.append_strdup_know_length(str, known_length);
1.33 paf 400:
1.174 paf 401: ASSERT_STRING_INVARIANT(*this);
402: return *this;
1.5 paf 403: }
1.46 paf 404:
1.234 misha 405: struct CORD_length_info {
406: size_t len;
407: size_t skip;
408: };
409:
410: int CORD_batched_len(const char* s, CORD_length_info* info){
411: info->len += lengthUTF8( (const XMLByte *)s, (const XMLByte *)s+strlen(s)); return 0;
1.221 misha 412: }
413:
1.234 misha 414: // can be called only for IS_FUNCTION(CORD) which are used in large String::Body::mid
415: int CORD_batched_len(const char c, CORD_length_info* info){
416: if (info->skip==0){
417: info->len++;
418: info->skip = lengthUTF8Char(c)-1;
419: } else {
420: info->skip--;
421: }
1.220 misha 422: return 0;
423: }
424:
1.210 misha 425: size_t String::length(Charset& charset) const {
426: if(charset.isUTF8()){
1.234 misha 427: CORD_length_info info = {0, 0};
428: body.for_each<CORD_length_info *>(CORD_batched_len, CORD_batched_len, &info);
429: return info.len;
1.210 misha 430: } else
431: return body.length();
432: }
433:
1.174 paf 434: /// @todo check in doc: whether it documents NOW bad situation "abc".mid(-1, 3) =were?="ab"
435: String& String::mid(size_t substr_begin, size_t substr_end) const {
436: String& result=*new String;
437:
438: size_t self_length=length();
439: substr_begin=min(substr_begin, self_length);
440: substr_end=min(max(substr_end, substr_begin), self_length);
1.176 paf 441: size_t substr_length=substr_end-substr_begin;
442: if(!substr_length)
1.107 parser 443: return result;
1.53 paf 444:
1.176 paf 445: // first: their langs
446: result.langs.append(result.body, langs, substr_begin, substr_length);
447: // next: letters themselves
448: result.body=body.mid(substr_begin, substr_length);
1.174 paf 449:
450: ASSERT_STRING_INVARIANT(result);
1.53 paf 451: return result;
1.54 paf 452: }
453:
1.211 misha 454: // from, to and helper_length in characters, not in bytes (it's important for utf-8)
455: String& String::mid(Charset& charset, size_t from, size_t to, size_t helper_length) const {
1.210 misha 456: String& result=*new String;
457:
1.211 misha 458: size_t self_length=(helper_length)?helper_length:length(charset);
459:
460: if(!self_length)
461: return result;
462:
463: from=min(min(to, from), self_length);
1.210 misha 464: to=min(max(to, from), self_length);
1.211 misha 465:
1.210 misha 466: size_t substr_length=to-from;
1.211 misha 467:
1.210 misha 468: if(!substr_length)
469: return result;
470:
471: if(charset.isUTF8()){
1.234 misha 472: const XMLByte* src_begin=(const XMLByte*)cstr();
1.230 misha 473: const XMLByte* src_end=src_begin+body.length();
1.210 misha 474:
1.212 misha 475: // convert 'from' and 'substr_length' from 'characters' to 'bytes'
1.230 misha 476: from=getUTF8BytePos(src_begin, src_end, from);
477: substr_length=getUTF8BytePos(src_begin+from, src_end, substr_length);
1.210 misha 478: if(!substr_length)
479: return result;
480: }
481:
482: // first: their langs
483: result.langs.append(result.body, langs, from, substr_length);
484: // next: letters themselves
485: result.body=body.mid(from, substr_length);
486:
487: ASSERT_STRING_INVARIANT(result);
488: return result;
489: }
490:
1.176 paf 491: size_t String::pos(const String::Body substr, size_t this_offset, Language lang) const {
1.183 paf 492: size_t substr_length=substr.length();
493: while(true) {
494: size_t substr_begin=body.pos(substr, this_offset);
495:
496: if(substr_begin==CORD_NOT_FOUND)
497: return STRING_NOT_FOUND;
1.174 paf 498:
1.183 paf 499: if(langs.check_lang(lang, substr_begin, substr_length))
500: return substr_begin;
501:
502: this_offset=substr_begin+substr_length;
503: }
1.58 paf 504: }
505:
1.174 paf 506: size_t String::pos(const String& substr,
507: size_t this_offset, Language lang) const {
508: return pos(substr.body, this_offset, lang);
1.60 paf 509: }
510:
1.210 misha 511: size_t String::pos(Charset& charset, const String& substr,
512: size_t this_offset, Language lang) const {
513:
514: if(charset.isUTF8()){
1.234 misha 515: const XMLByte* srcPtr=(const XMLByte*)cstr();
1.212 misha 516: const XMLByte* srcEnd=srcPtr+body.length();
517:
518: // convert 'this_offset' from 'characters' to 'bytes'
519: this_offset=getUTF8BytePos(srcPtr, srcEnd, this_offset);
520:
1.229 misha 521: size_t result=pos(substr.body, this_offset, lang);
522: return (result==CORD_NOT_FOUND)
523: ? STRING_NOT_FOUND
524: : getUTF8CharPos(srcPtr, srcEnd, result); // convert 'result' from 'bytes' to 'characters'
1.212 misha 525: } else {
1.229 misha 526: size_t result=pos(substr.body, this_offset, lang);
527: return (result==CORD_NOT_FOUND)
528: ? STRING_NOT_FOUND
529: : result;
1.210 misha 530: }
531: }
532:
1.174 paf 533: void String::split(ArrayString& result,
534: size_t& pos_after,
535: const char* delim,
536: Language lang, int limit) const {
1.237 misha 537: if(is_empty())
538: return;
1.174 paf 539: size_t self_length=length();
1.237 misha 540: if(size_t delim_length=strlen(delim)) {
1.186 paf 541: size_t pos_before;
1.60 paf 542: // while we have 'delim'...
1.174 paf 543: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 544: result+=&mid(pos_after, pos_before);
1.174 paf 545: pos_after=pos_before+delim_length;
1.60 paf 546: }
547: // last piece
1.174 paf 548: if(pos_after<self_length && limit) {
549: result+=&mid(pos_after, self_length);
550: pos_after=self_length;
1.60 paf 551: }
552: } else { // empty delim
553: result+=this;
1.174 paf 554: pos_after+=self_length;
1.60 paf 555: }
556: }
557:
1.174 paf 558: void String::split(ArrayString& result,
559: size_t& pos_after,
560: const String& delim, Language lang,
561: int limit) const {
1.237 misha 562: if(is_empty())
563: return;
564: if(!delim.is_empty()) {
1.186 paf 565: size_t pos_before;
1.60 paf 566: // while we have 'delim'...
1.174 paf 567: for(; (pos_before=pos(delim, pos_after, lang))!=STRING_NOT_FOUND && limit; limit--) {
1.69 paf 568: result+=&mid(pos_after, pos_before);
1.174 paf 569: pos_after=pos_before+delim.length();
1.60 paf 570: }
571: // last piece
1.174 paf 572: if(pos_after<length() && limit) {
573: result+=&mid(pos_after, length());
574: pos_after=length();
1.60 paf 575: }
576: } else { // empty delim
577: result+=this;
1.174 paf 578: pos_after+=length();
1.60 paf 579: }
1.61 paf 580: }
581:
1.222 misha 582: Table* String::match(VRegex* vregex,
583: Row_action row_action, void *info,
584: int& matches_count) const {
1.209 misha 585:
1.222 misha 586: // vregex->info(); // I have no idea what does it for?
1.63 paf 587:
1.222 misha 588: bool need_pre_post_match=vregex->is_pre_post_match_needed();
589: bool global=vregex->is_global_search();
1.63 paf 590:
1.174 paf 591: const char* subject=cstr();
1.234 misha 592: size_t subject_length=length();
1.222 misha 593: const int ovector_size=(1/*match*/+MAX_MATCH_GROUPS)*3;
594: int ovector[ovector_size];
1.155 paf 595:
1.173 paf 596: Table::Action_options table_options;
1.174 paf 597: Table& table=*new Table(string_match_table_template, table_options);
1.63 paf 598:
1.154 paf 599: int prestart=0;
600: int poststart=0;
1.174 paf 601: int postfinish=length();
1.63 paf 602: while(true) {
1.222 misha 603: int exec_result=vregex->exec(subject, subject_length, ovector, ovector_size, prestart);
1.63 paf 604:
1.222 misha 605: if(exec_result<0) // only PCRE_ERROR_NOMATCH might be here, other negative results cause an exception
606: break;
1.63 paf 607:
1.154 paf 608: int prefinish=ovector[0];
609: poststart=ovector[1];
1.236 moko 610:
611: if (prestart==poststart && subject[poststart]=='\n'){
612: prestart++;
613: continue;
614: }
615:
1.232 misha 616: ArrayString* row=new ArrayString(3);
1.174 paf 617: if(need_pre_post_match) {
618: *row+=&mid(0, prefinish); // .prematch column value
619: *row+=&mid(prefinish, poststart); // .match
620: *row+=&mid(poststart, postfinish); // .postmatch
621: } else {
1.185 paf 622: *row+=&Empty; // .prematch column value
623: *row+=&Empty; // .match
624: *row+=&Empty; // .postmatch
1.174 paf 625: }
1.63 paf 626:
1.222 misha 627: for(int i=1; i<exec_result; i++) {
1.69 paf 628: // -1:-1 case handled peacefully by mid() itself
1.232 misha 629: *row+=(ovector[i*2+0]>=0 && ovector[i*2+1]>0)?&mid(ovector[i*2+0], ovector[i*2+1]):new String; // .i column value
1.63 paf 630: }
631:
1.209 misha 632: matches_count++;
1.174 paf 633: row_action(table, row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 634:
1.222 misha 635: if(!global || prestart==poststart) // last step
636: break;
637:
1.154 paf 638: prestart=poststart;
1.222 misha 639: }
1.63 paf 640:
1.222 misha 641: row_action(table, 0/*last time, no raw*/, 0, 0, poststart, postfinish, info);
642: return vregex->is_just_count() ? 0 : &table;
1.82 parser 643: }
644:
1.174 paf 645: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
646: String& result=*new String();
647: if(is_empty())
648: return result;
649:
650: char* new_cstr=cstrm();
1.216 misha 651:
1.181 paf 652: if(source_charset.isUTF8()) {
1.216 misha 653: size_t new_cstr_len=length();
1.181 paf 654: switch(kind) {
655: case CC_UPPER:
1.192 paf 656: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToUpper);
1.181 paf 657: break;
658: case CC_LOWER:
1.192 paf 659: change_case_UTF8((const XMLByte*)new_cstr, new_cstr_len, (XMLByte*)new_cstr, new_cstr_len, UTF8CaseToLower);
1.181 paf 660: break;
661: default:
662: assert(!"unknown change case kind");
663: break; // never
664: }
665:
666: } else {
667: const unsigned char *tables=source_charset.pcre_tables;
1.82 parser 668:
1.181 paf 669: const unsigned char *a;
670: const unsigned char *b;
671: switch(kind) {
672: case CC_UPPER:
673: a=tables+lcc_offset;
674: b=tables+fcc_offset;
675: break;
676: case CC_LOWER:
677: a=tables+lcc_offset;
678: b=0;
679: break;
680: default:
681: assert(!"unknown change case kind");
682: a=b=0; // calm, compiler
683: break; // never
684: }
685:
1.192 paf 686: char *dest=new_cstr;
1.181 paf 687: unsigned char index;
1.190 paf 688: for(const char* current=new_cstr; (index=(unsigned char)*current); current++) {
1.181 paf 689: unsigned char c=a[index];
690: if(b)
691: c=b[c];
692:
693: *dest++=(char)c;
694: }
1.174 paf 695: }
1.176 paf 696: result.langs=langs;
1.174 paf 697: result.body=new_cstr;
1.89 parser 698:
1.101 parser 699: return result;
700: }
701:
1.213 misha 702: const String& String::escape(Charset& source_charset) const {
703: if(is_empty())
704: return *this;
705:
706: return Charset::escape(*this, source_charset);
707: }
708:
1.238 misha 709: #define STRING_APPEND(result, from_cstr, langs, langs_offset, length) \
710: result.langs.append(result.body, langs, langs_offset, length); \
711: result.body.append_strdup_know_length(from_cstr, length);
712:
1.174 paf 713: const String& String::replace(const Dictionary& dict) const {
1.238 misha 714: if(!dict.count() || is_empty())
715: return *this;
716:
1.174 paf 717: String& result=*new String();
718: const char* old_cstr=cstr();
719: const char* prematch_begin=old_cstr;
720:
1.238 misha 721: if(dict.count()==1) {
722: // optimized simple case
723:
724: Dictionary::Subst subst=dict.get(0);
1.239 moko 725: while(const char* p=strstr(prematch_begin, subst.from)) {
1.174 paf 726: // prematch
1.238 misha 727: if(size_t prematch_length=p-prematch_begin) {
728: STRING_APPEND(result, prematch_begin, langs, prematch_begin-old_cstr, prematch_length)
1.101 parser 729: }
730:
1.174 paf 731: // match
1.238 misha 732: prematch_begin=p+subst.from_length;
1.174 paf 733:
1.184 paf 734: if(const String* b=subst.to) // are there any b?
1.174 paf 735: result<<*b;
1.238 misha 736: }
737:
738: } else {
739:
740: const char* current=old_cstr;
741: while(*current) {
742: if(Dictionary::Subst subst=dict.first_that_begins(current)) {
743: // prematch
744: if(size_t prematch_length=current-prematch_begin) {
745: STRING_APPEND(result, prematch_begin, langs, prematch_begin-old_cstr, prematch_length)
746: }
747:
748: // match
749: // skip 'a' in 'current'; move prematch_begin
750: current+=subst.from_length; prematch_begin=current;
751:
752: if(const String* b=subst.to) // are there any b?
753: result<<*b;
754: } else // simply advance
755: current++;
756: }
757:
1.174 paf 758: }
1.156 paf 759:
1.238 misha 760: if(prematch_begin==old_cstr) // not modified
761: return *this;
762:
1.174 paf 763: // postmatch
1.238 misha 764: if(size_t postmatch_length=old_cstr+length()-prematch_begin) {
765: STRING_APPEND(result, prematch_begin, langs, prematch_begin-old_cstr, postmatch_length)
1.174 paf 766: }
1.156 paf 767:
1.174 paf 768: ASSERT_STRING_INVARIANT(result);
1.82 parser 769: return result;
1.61 paf 770: }
1.113 parser 771:
1.180 paf 772: static int serialize_body_char(char c, char** cur) {
773: *((*cur)++)=c;
774: return 0; // 0=continue
775: };
1.174 paf 776: static int serialize_body_piece(const char* s, char** cur) {
777: size_t length=strlen(s);
778: memcpy(*cur, s, length); *cur+=length;
1.178 paf 779: return 0; // 0=continue
1.174 paf 780: };
1.178 paf 781: static int serialize_lang_piece(char alang, size_t asize, char** cur) {
782: // lang
1.191 paf 783: **cur=alang; (*cur)++;
784: // length [WARNING: not cast, addresses must be %4=0 on sparc]
1.178 paf 785: memcpy(*cur, &asize, sizeof(asize)); *cur+=sizeof(asize);
786:
787: return 0; // 0=continue
788: }
1.174 paf 789: String::Cm String::serialize(size_t prolog_length) const {
1.178 paf 790: size_t fragments_count=langs.count();
1.202 paf 791: size_t body_length=body.length();
1.174 paf 792: size_t buf_length=
1.178 paf 793: prolog_length //1
794: +sizeof(size_t) //2
1.202 paf 795: +body_length //3
796: +1 // 4 for zero terminator used in deserialize
797: +sizeof(size_t) //5
798: +fragments_count*(sizeof(char)+sizeof(size_t)); //6
799:
1.174 paf 800: String::Cm result(new(PointerFreeGC) char[buf_length], buf_length);
801:
802: // 1: prolog
803: char *cur=result.str+prolog_length;
1.202 paf 804: // 2: chars.count [WARNING: not cast, addresses must be %4=0 on sparc]
805: memcpy(cur, &body_length, sizeof(body_length)); cur+=sizeof(body_length);
806: // 3: letters
807: body.for_each(serialize_body_char, serialize_body_piece, &cur);
808: // 4: zero terminator
809: *cur++=0;
810: // 5: langs.count [WARNING: not cast, addresses must be %4=0 on sparc]
1.174 paf 811: memcpy(cur, &fragments_count, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 812: // 6: lang info
1.178 paf 813: langs.for_each(body, serialize_lang_piece, &cur);
1.113 parser 814:
1.174 paf 815: return result;
1.113 parser 816: }
1.202 paf 817: bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size) {
818: size_t in_buf=buf_size;
819: if(in_buf<=prolog_size)
1.148 paf 820: return false;
1.202 paf 821: in_buf-=prolog_size;
1.135 paf 822:
1.174 paf 823: // 1: prolog
1.202 paf 824: const char* cur=(const char* )buf+prolog_size;
1.113 parser 825:
1.207 paf 826: // 2: chars.count
1.202 paf 827: size_t body_length;
828: if(in_buf<sizeof(body_length)) // body.length don't fit?
829: return false;
830: // [WARNING: not cast, addresses must be %4=0 on sparc]
831: memcpy(&body_length, cur, sizeof(body_length)); cur+=sizeof(body_length);
832: in_buf-=sizeof(body_length);
833:
834: if(in_buf<body_length+1) // letters+terminator don't fit?
835: return false;
836: // 4: zero terminator
837: if(cur[body_length] != 0) // in place?
838: return false;
839: // 3: letters
1.227 misha 840: body=String::Body(*cur?cur:0);
1.229 misha 841: #ifdef STRING_LENGTH_CACHING
842: body.set_length(body_length);
843: #endif
1.202 paf 844: cur+=body_length+1;
845: in_buf-=body_length+1;
846:
847: // 5: langs.count
1.191 paf 848: size_t fragments_count;
1.202 paf 849: if(in_buf<sizeof(fragments_count)) // langs.count don't fit?
1.174 paf 850: return false;
1.191 paf 851: // [WARNING: not cast, addresses must be %4=0 on sparc]
852: memcpy(&fragments_count, cur, sizeof(fragments_count)); cur+=sizeof(fragments_count);
1.202 paf 853: in_buf-=sizeof(fragments_count);
1.174 paf 854:
855: if(fragments_count) {
1.202 paf 856: // 6: lang info
1.174 paf 857: size_t total_length=0;
858: for(size_t f=0; f<fragments_count; f++) {
1.191 paf 859: char lang;
860: size_t fragment_length;
861: size_t piece_length=sizeof(lang)+sizeof(fragment_length);
1.202 paf 862: if(in_buf<piece_length) // lang+length
1.174 paf 863: return false;
864:
1.191 paf 865: // lang
866: lang=*cur++;
867: // length [WARNING: not cast, addresses must be %4=0 on sparc]
868: memcpy(&fragment_length, cur, sizeof(fragment_length)); cur+=sizeof(fragment_length);
869:
1.206 paf 870: size_t combined_length=total_length+fragment_length;
871: if(combined_length>body_length)
872: return false; // file curruption
1.191 paf 873: // uchar needed to prevent propagating 0x80 bit to upper bytes
874: langs.append(total_length, (String::Language)(uchar)lang, fragment_length);
1.206 paf 875: total_length=combined_length;
1.202 paf 876: in_buf-=piece_length;
1.174 paf 877: }
1.128 paf 878:
1.202 paf 879: if(total_length!=body_length) // length(all language fragments) vs length(letters)
1.148 paf 880: return false;
1.174 paf 881: }
1.202 paf 882: if(in_buf!=0) // some strange extra bytes
883: return false;
1.113 parser 884:
1.174 paf 885: ASSERT_STRING_INVARIANT(*this);
1.148 paf 886: return true;
1.176 paf 887: }
888:
889: const char* String::Body::v() const {
1.233 misha 890: return CORD_to_const_char_star(body, length());
1.176 paf 891: }
1.201 paf 892: void String::Body::dump() const {
893: CORD_dump(body);
894: }
895:
1.176 paf 896: const char* String::Languages::v() const {
1.177 paf 897: if(opt.is_not_just_lang)
1.233 misha 898: return CORD_to_const_char_star(langs, 0);
1.176 paf 899: else
900: return (const char*)&langs;
901: }
1.201 paf 902: void String::Languages::dump() const {
903: if(opt.is_not_just_lang)
904: CORD_dump(langs);
905: else
906: puts((const char*)&langs);
907: }
1.176 paf 908: const char* String::v() const {
1.198 paf 909: const uint LIMIT_VIEW=20;
1.176 paf 910: char* buf=(char*)malloc(MAX_STRING);
911: const char*body_view=body.v();
912: const char*langs_view=langs.v();
913: snprintf(buf, MAX_STRING,
1.178 paf 914: "%d:%.*s%s} "
1.176 paf 915: "{%d:%s",
1.178 paf 916: langs.count(), LIMIT_VIEW, langs_view, strlen(langs_view)>LIMIT_VIEW?"...":"",
1.176 paf 917: strlen(body_view), body_view
918: );
919:
920: return buf;
1.113 parser 921: }
1.229 misha 922:
1.201 paf 923: void String::dump() const {
924: body.dump();
925: langs.dump();
926: }
1.229 misha 927:
928: const String& String::trim(String::Trim_kind kind, const char* chars, Charset* source_charset) const {
1.223 misha 929: if(is_empty())
1.195 paf 930: return *this;
931:
932: size_t substr_begin, substr_length;
1.229 misha 933: Body new_body=body.trim(kind, chars, &substr_begin, &substr_length, source_charset);
1.195 paf 934: if(new_body==body) // we received unchanged pointer, do likewise
935: return *this;
936: // new_body differs from body, adjust langs along
937:
938: String& result=*new String;
939: if(!new_body) // body.trim produced empty result
940: return result;
941: // body.trim produced nonempty result
942:
943: // first: their langs
944: result.langs.append(result.body, langs, substr_begin, substr_length);
945: // next: letters themselves
946: result.body=new_body;
947:
948: ASSERT_STRING_INVARIANT(result);
949: return result;
1.198 paf 950: }
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