Annotation of parser3/src/main/pa_string.C, revision 1.172.2.1
1.45 paf 1: /** @file
1.55 paf 2: Parser: string class. @see untasize_t.C.
1.46 paf 3:
1.172 paf 4: Copyright (c) 2001, 2003 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:
1.172.2.1! paf 8: static const char* IDENT_STRING_C="$Date: 2003/01/21 15:51:15 $";
1.4 paf 9:
1.70 paf 10: #include "pcre.h"
11:
1.13 paf 12: #include "pa_pool.h"
1.12 paf 13: #include "pa_string.h"
1.172.2.1! paf 14: //#include "pa_hash.h"
1.22 paf 15: #include "pa_exception.h"
1.53 paf 16: #include "pa_common.h"
1.60 paf 17: #include "pa_array.h"
18: #include "pa_globals.h"
1.61 paf 19: #include "pa_table.h"
1.101 parser 20: #include "pa_dictionary.h"
1.132 paf 21: #include "pa_charset.h"
1.60 paf 22:
1.139 paf 23: #define DEBUG_STRING_APPENDS_VS_EXPANDS
24:
25:
26: #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS
27: ulong string_piece_appends=0;
28: #endif
1.172.2.1! paf 29: /*
1.160 paf 30: String& String::OnPool(Pool& apool, const char *local_src, size_t src_size, bool tainted) {
31: if(local_src && *local_src) {
32: if(src_size==0)
33: src_size=strlen(local_src);
34:
35: char *pooled_src=(char *)apool.malloc(src_size);
36: memcpy(pooled_src, local_src, src_size);
37: return *new(apool) String(apool, pooled_src, src_size, tainted);
38: } else
39: return *new(apool) String(apool);
40: }
1.172.2.1! paf 41: */
! 42: String::String(const char *src, size_t src_size, bool tainted) {
1.151 paf 43: last_chunk=&head.chunk;
44: head.chunk.count=CR_PREALLOCATED_COUNT;
45: append_here=head.chunk.rows;
1.41 paf 46:
47: if(src)
1.75 paf 48: if(tainted)
49: APPEND_TAINTED(src, src_size, 0, 0);
1.41 paf 50: else
1.75 paf 51: APPEND_CLEAN(src, src_size, 0, 0);
1.1 paf 52: }
1.140 paf 53:
1.172.2.1! paf 54: String::String(const String& src) {
1.151 paf 55: last_chunk=&head.chunk;
56: head.chunk.count=CR_PREALLOCATED_COUNT;
57: append_here=head.chunk.rows;
1.140 paf 58:
1.169 paf 59: append(src, UL_PASS_APPENDED);
1.120 paf 60: }
1.172.2.1! paf 61: void String::~String() {
! 62: // @todonow
! 63: }
! 64:
1.120 paf 65:
66: size_t String::size() const {
67: size_t result=0;
1.123 paf 68: STRING_FOREACH_ROW(
1.120 paf 69: result+=row->item.size;
1.123 paf 70: );
1.120 paf 71: return result;
1.94 parser 72: }
73:
1.115 paf 74: /// @todo not very optimal
75: uint String::used_rows() const {
76: uint result=0;
1.123 paf 77: STRING_FOREACH_ROW(
78: result++;
79: );
1.115 paf 80: return result;
81: }
1.94 parser 82: void String::expand() {
1.143 paf 83: uint new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.139 paf 84: if(new_chunk_count>max_integral(Chunk::count_type))
85: new_chunk_count=max_integral(Chunk::count_type);
1.122 paf 86:
1.172.2.1! paf 87: Chunk *new_chunk=static_cast<Chunk *>(pa_malloc(
1.151 paf 88: sizeof(Chunk)// count+interpadding(?)+rows[CR_PREALLOCATED_COUNT]+tailpadding(??)
89: -sizeof(Chunk::rows_type) // PREALLOCATED rows
90: +sizeof(Chunk::Row)*new_chunk_count // neaded rows
91: +sizeof(Chunk *) // link size
1.172.2.1! paf 92: ));
1.141 paf 93: new_chunk->rows[new_chunk->count=new_chunk_count].link=0;
94: last_chunk->rows[last_chunk->count].link=new_chunk;
95:
96: last_chunk=new_chunk;
1.94 parser 97: append_here=last_chunk->rows;
1.5 paf 98: }
1.28 paf 99:
1.13 paf 100: String& String::real_append(STRING_APPEND_PARAMS) {
1.139 paf 101: if(!last_chunk) // growth stopped [we're appended as string to somebody]
1.149 paf 102: throw Exception(0,
1.139 paf 103: this,
1.142 paf 104: "string growth stopped (append cstr)");
1.139 paf 105:
1.9 paf 106: if(!src)
107: return *this;
1.26 paf 108: if(!size)
109: size=strlen(src);
110: if(!size)
1.9 paf 111: return *this;
1.122 paf 112:
1.139 paf 113: #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS
114: string_piece_appends++;
115: #endif
116:
117: // manually unrolled to avoid extra check
118: while(size>max_integral(Chunk::Row::item_size_type)) {
1.122 paf 119: if(chunk_is_full())
120: expand();
121:
122: append_here->item.ptr=src;
1.139 paf 123: append_here->item.size=max_integral(Chunk::Row::item_size_type);
1.122 paf 124: append_here->item.lang=lang;
125: #ifndef NO_STRING_ORIGIN
126: append_here->item.origin.file=file;
127: append_here->item.origin.line=line;
128: #endif
129: append_here++;
130:
1.139 paf 131: src+=max_integral(Chunk::Row::item_size_type);
132: size-=max_integral(Chunk::Row::item_size_type);
1.122 paf 133: }
1.9 paf 134:
1.1 paf 135: if(chunk_is_full())
136: expand();
137:
138: append_here->item.ptr=src;
1.121 paf 139: append_here->item.size=size;
1.52 paf 140: append_here->item.lang=lang;
1.13 paf 141: #ifndef NO_STRING_ORIGIN
1.14 paf 142: append_here->item.origin.file=file;
143: append_here->item.origin.line=line;
1.13 paf 144: #endif
1.115 paf 145: append_here++;
1.1 paf 146:
147: return *this;
148: }
149:
1.16 paf 150: uint String::hash_code() const {
1.7 paf 151: uint result=0;
1.123 paf 152: STRING_FOREACH_ROW(
1.6 paf 153: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.123 paf 154: );
1.5 paf 155: return result;
156: }
157:
1.60 paf 158: /// @todo move 'lang' skipping to pos
159: int String::cmp(int& partial, const String& src,
160: size_t this_offset, Untaint_lang lang) const {
1.59 paf 161: partial=-1;
1.125 paf 162: size_t a_size=size();
163: this_offset=min(this_offset, a_size-1);
1.55 paf 164:
1.151 paf 165: const Chunk *a_chunk=&head.chunk;
166: const Chunk *b_chunk=&src.head.chunk;
1.16 paf 167: const Chunk::Row *a_row=a_chunk->rows;
168: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 169: size_t a_offset=this_offset;
170: size_t b_offset=0;
1.9 paf 171: Chunk::Row *a_end=append_here;
172: Chunk::Row *b_end=src.append_here;
1.116 paf 173: uint a_countdown=a_chunk->count;
174: uint b_countdown=b_chunk->count;
175: int result;
1.60 paf 176: size_t pos=0;
1.33 paf 177:
1.125 paf 178: bool a_break=a_size==0;
1.140 paf 179: bool b_break=src.is_empty();
1.83 parser 180: if(!(a_break || b_break)) while(true) {
1.55 paf 181: if(pos+a_row->item.size > this_offset) {
1.136 paf 182: if(lang!=UL_UNSPECIFIED && a_row->item.lang>lang)
1.60 paf 183: return -1; // wrong lang -- bail out
184:
1.55 paf 185: int size_diff=
186: (a_row->item.size-a_offset)-
187: (b_row->item.size-b_offset);
188:
189: if(size_diff==0) { // a has same size as b
1.60 paf 190: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
191: a_row->item.size-a_offset);
1.55 paf 192: if(result)
193: return result;
1.60 paf 194: pos+=a_row->item.size;
1.55 paf 195: a_row++; a_countdown--; a_offset=0;
196: b_row++; b_countdown--; b_offset=0;
197: } else if (size_diff>0) { // a longer
1.60 paf 198: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
199: b_row->item.size-b_offset);
1.55 paf 200: if(result)
201: return result;
202: a_offset+=b_row->item.size-b_offset;
203: b_row++; b_countdown--; b_offset=0;
204: } else { // b longer
1.60 paf 205: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
206: a_row->item.size-a_offset);
1.55 paf 207: if(result)
208: return result;
209: b_offset+=a_row->item.size-a_offset;
1.60 paf 210: pos+=a_row->item.size;
1.55 paf 211: a_row++; a_countdown--; a_offset=0;
212: }
1.83 parser 213: if(b_break=b_row==b_end) {
214: a_break=a_row==a_end;
215: break;
216: }
1.55 paf 217: if(!b_countdown) {
218: b_chunk=b_row->link;
219: b_row=b_chunk->rows;
220: b_countdown=b_chunk->count;
221: }
222: } else {
1.60 paf 223: a_offset-=a_row->item.size;
224: pos+=a_row->item.size;
225: a_row++; a_countdown--;
1.9 paf 226: }
227:
1.83 parser 228: if(a_break=a_row==a_end) {
229: b_break=b_row==b_end;
230: break;
231: }
1.11 paf 232: if(!a_countdown) {
1.9 paf 233: a_chunk=a_row->link;
234: a_row=a_chunk->rows;
1.11 paf 235: a_countdown=a_chunk->count;
1.9 paf 236: }
1.27 paf 237: }
1.55 paf 238: if(a_break==b_break) { // ended simultaneously
239: partial=0; return 0;
240: } else if(a_break) { // first bytes equal, but a ended before b
241: partial=1; return -1;
242: } else {
243: partial=2; return +1;
244: }
1.27 paf 245: }
246:
1.60 paf 247: /// @todo move 'lang' skipping to pos
1.59 paf 248: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 249: size_t this_offset, Untaint_lang lang) const {
1.59 paf 250: partial=-1;
1.125 paf 251: size_t a_size=size();
1.50 paf 252: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.125 paf 253: this_offset=min(this_offset, a_size-1);
1.27 paf 254:
1.151 paf 255: const Chunk *a_chunk=&head.chunk;
1.27 paf 256: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 257: size_t a_offset=this_offset;
1.55 paf 258: size_t b_offset=0;
1.27 paf 259: Chunk::Row *a_end=append_here;
1.116 paf 260: uint a_countdown=a_chunk->count;
1.60 paf 261: size_t pos=0;
1.52 paf 262:
1.125 paf 263: bool a_break=a_size==0;
1.83 parser 264: bool b_break=b_size==0;
265: if(!(a_break || b_break)) while(true) {
1.59 paf 266: if(pos+a_row->item.size > this_offset) {
1.136 paf 267: if(lang!=UL_UNSPECIFIED && a_row->item.lang>lang)
1.60 paf 268: return -1; // wrong lang -- bail out
269:
1.59 paf 270: int size_diff=
271: (a_row->item.size-a_offset)-
272: (b_size-b_offset);
273:
274: if(size_diff==0) { // a has same size as b
1.116 paf 275: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 276: a_row->item.size-a_offset)!=0)
277: return result;
1.60 paf 278: pos+=a_row->item.size;
1.59 paf 279: a_row++; a_countdown--; a_offset=0;
280: b_break=true;
281: } else if (size_diff>0) { // a longer
1.116 paf 282: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 283: b_size-b_offset)!=0)
284: return result;
285: a_offset+=b_size-b_offset;
286: b_break=true;
287: } else { // b longer
1.116 paf 288: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 289: a_row->item.size-a_offset)!=0)
290: return result;
291: b_offset+=a_row->item.size-a_offset;
1.60 paf 292: pos+=a_row->item.size;
1.59 paf 293: a_row++; a_countdown--; a_offset=0;
294: }
295: } else {
1.60 paf 296: a_offset-=a_row->item.size;
297: pos+=a_row->item.size;
298: a_row++; a_countdown--;
1.27 paf 299: }
300:
1.86 parser 301: a_break=a_row==a_end;
302: if(a_break || b_break)
1.83 parser 303: break;
1.27 paf 304: if(!a_countdown) {
305: a_chunk=a_row->link;
306: a_row=a_chunk->rows;
307: a_countdown=a_chunk->count;
1.9 paf 308: }
309: }
1.55 paf 310: if(a_break==b_break) { // ended simultaneously
311: partial=0; return 0;
312: } else if(a_break) { // first bytes equal, but a ended before b
313: partial=1; return -1;
314: } else {
315: partial=2; return +1;
316: }
1.5 paf 317: }
1.46 paf 318:
319: #ifndef NO_STRING_ORIGIN
320: const Origin& String::origin() const {
1.140 paf 321: if(is_empty()) {
1.96 parser 322: static const Origin empty_origin={"empty string"};
323: return empty_origin;
324: }
1.46 paf 325:
1.147 paf 326: // determining origin by first piece or last appended piece
327: // because any of them can be constant=without origin:
1.50 paf 328: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 329: // when last peice is constant,
330: // ex: parser_root_auto_path{dynamic} / auto.p{const}
331: // using first piece
1.151 paf 332: Origin& first_origin=head.chunk.rows[0].item.origin;
1.147 paf 333: return first_origin.file ? first_origin : append_here[-1].item.origin;
1.46 paf 334: }
335: #endif
1.53 paf 336:
1.69 paf 337: String& String::mid(size_t start, size_t finish) const {
1.172.2.1! paf 338: String& result=*new String();
1.107 parser 339:
1.166 paf 340: start=min(start, size());
1.167 paf 341: finish=max(start, finish);
1.60 paf 342: if(start==finish)
1.107 parser 343: return result;
1.53 paf 344:
345: size_t pos=0;
1.123 paf 346: STRING_FOREACH_ROW(
347: size_t item_finish=pos+row->item.size;
348: if(item_finish > start) { // started now or already?
1.140 paf 349: bool started=result.is_empty(); // started now?
1.123 paf 350: bool finished=finish <= item_finish; // finished now?
351: size_t offset=started?start-pos:0;
352: size_t size=finished?finish-pos:row->item.size;
353: result.APPEND(
354: row->item.ptr+offset, size-offset,
355: row->item.lang,
356: row->item.origin.file, row->item.origin.line);
357: if(finished)
1.53 paf 358: goto break2;
359: }
1.123 paf 360: pos+=row->item.size;
361: );
1.53 paf 362: break2:
1.60 paf 363: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
364: //cstr(), start, finish, result.cstr());
1.53 paf 365: return result;
1.54 paf 366: }
367:
1.60 paf 368: int String::pos(const String& substr,
1.116 paf 369: int result, Untaint_lang lang) const {
1.125 paf 370: size_t self_size=size();
1.131 paf 371: for(; size_t(result)<self_size; result++) {
1.60 paf 372: int partial; cmp(partial, substr, result, lang);
1.58 paf 373: if(
374: partial==0 || // full match
375: partial==2) // 'substr' starts 'this'+'result'
376: return result;
377: }
378:
379: return -1;
380: }
381:
1.60 paf 382: int String::pos(const char *substr, size_t substr_size,
1.116 paf 383: int result, Untaint_lang lang) const {
1.125 paf 384: size_t self_size=size();
1.131 paf 385: for(; size_t(result)<self_size; result++) {
1.60 paf 386: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 387: if(
388: partial==0 || // full match
389: partial==2) // 'substr' starts 'this'+'result'
390: return result;
391: }
392:
393: return -1;
1.60 paf 394: }
395:
396: void String::split(Array& result,
397: size_t* pos_after_ref,
398: const char *delim, size_t delim_size,
399: Untaint_lang lang, int limit) const {
1.125 paf 400: size_t self_size=size();
1.60 paf 401: if(delim_size) {
402: size_t pos_after=pos_after_ref?*pos_after_ref:0;
403: int pos_before;
404: // while we have 'delim'...
405: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 406: result+=&mid(pos_after, pos_before);
1.60 paf 407: pos_after=pos_before+delim_size;
408: }
409: // last piece
1.124 paf 410: if(pos_after<self_size && limit) {
411: result+=&mid(pos_after, self_size);
412: pos_after=self_size;
1.60 paf 413: }
414: if(pos_after_ref)
415: *pos_after_ref=pos_after;
416: } else { // empty delim
417: result+=this;
418: if(pos_after_ref)
1.124 paf 419: *pos_after_ref+=self_size;
1.60 paf 420: }
421: }
422:
423: void String::split(Array& result,
424: size_t* pos_after_ref,
425: const String& delim, Untaint_lang lang,
426: int limit) const {
1.140 paf 427: if(!delim.is_empty()) {
1.60 paf 428: size_t pos_after=pos_after_ref?*pos_after_ref:0;
429: int pos_before;
430: // while we have 'delim'...
431: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 432: result+=&mid(pos_after, pos_before);
1.60 paf 433: pos_after=pos_before+delim.size();
434: }
435: // last piece
436: if(pos_after<size() && limit) {
1.69 paf 437: result+=&mid(pos_after, size());
1.60 paf 438: pos_after=size();
439: }
440: if(pos_after_ref)
441: *pos_after_ref=pos_after;
442: } else { // empty delim
443: result+=this;
444: if(pos_after_ref)
445: *pos_after_ref+=size();
446: }
1.61 paf 447: }
448:
1.154 paf 449: static void regex_options(const String *options, int *result, bool& need_pre_post_match){
1.63 paf 450: struct Regex_option {
1.153 paf 451: const char *keyL;
452: const char *keyU;
1.63 paf 453: int clear, set;
454: int *result;
1.154 paf 455: bool *flag;
1.63 paf 456: } regex_option[]={
1.153 paf 457: {"i", "I", 0, PCRE_CASELESS, result}, // a=A
458: {"s", "S", 0, PCRE_DOTALL, result}, // \n\n$ [default]
459: {"x", "U", 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
460: {"m", "M", PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
461: {"g", "G", 0, true, result+1}, // many rows
1.154 paf 462: {"'", 0, 0, 0, 0, &need_pre_post_match},
463: {0}
1.63 paf 464: };
1.171 paf 465: result[0]=PCRE_EXTRA | PCRE_DOTALL | PCRE_DOLLAR_ENDONLY;
1.63 paf 466: result[1]=0;
467:
468: if(options)
1.153 paf 469: for(Regex_option *o=regex_option; o->keyL; o++)
1.154 paf 470: if(options->pos(o->keyL)>=0
471: || (o->keyU && options->pos(o->keyU)>=0)) {
472: if(o->flag)
473: *o->flag=true;
474: else { // result
475: *o->result &= ~o->clear;
476: *o->result |= o->set;
477: }
1.63 paf 478: }
479: }
480:
1.155 paf 481: /// @todo make replacement Table stacked
1.172.2.1! paf 482: bool String::match(Charset& source_charset,
1.77 paf 483: const String *aorigin,
1.62 paf 484: const String& regexp,
1.63 paf 485: const String *options,
1.64 paf 486: Table **table,
1.95 parser 487: Row_action row_action, void *info,
488: bool *was_global) const {
1.64 paf 489:
1.140 paf 490: if(regexp.is_empty())
1.149 paf 491: throw Exception(0,
1.73 paf 492: aorigin,
493: "regexp is empty");
1.154 paf 494:
1.118 paf 495: const char *pattern=regexp.cstr();
1.62 paf 496: const char *errptr;
497: int erroffset;
1.154 paf 498: bool need_pre_post_match=false;
499: int option_bits[2]; regex_options(options, option_bits, need_pre_post_match);
1.95 parser 500: if(was_global)
501: *was_global=option_bits[1]!=0;
1.63 paf 502: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 503: &errptr, &erroffset,
1.172.2.1! paf 504: source_charset.pcre_tables);
1.62 paf 505:
1.67 paf 506: if(!code)
1.149 paf 507: throw Exception(0,
1.69 paf 508: ®exp.mid(erroffset, regexp.size()),
1.74 paf 509: "regular expression syntax error - %s", errptr);
1.62 paf 510:
1.63 paf 511: int info_substrings=pcre_info(code, 0, 0);
512: if(info_substrings<0) {
1.100 parser 513: pcre_free(code);
1.149 paf 514: throw Exception(0,
1.73 paf 515: aorigin,
1.76 paf 516: "pcre_info error (%d)",
1.73 paf 517: info_substrings);
1.63 paf 518: }
519:
1.158 paf 520: const char *subject=cstr();
1.62 paf 521: int length=strlen(subject);
1.155 paf 522: const int ovecsize=(1/*match*/+MAX_STRING_MATCH_TABLE_COLUMNS)*3;
523: int ovector[ovecsize];
524:
525: // create table
1.172.2.1! paf 526: *table=new Table(*string_match_table_template);
1.63 paf 527:
1.64 paf 528: int exec_option_bits=0;
1.154 paf 529: int prestart=0;
530: int poststart=0;
531: int postfinish=size();
1.63 paf 532: while(true) {
533: int exec_substrings=pcre_exec(code, 0,
1.154 paf 534: subject, length, prestart,
1.64 paf 535: exec_option_bits, ovector, ovecsize);
1.63 paf 536:
537: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 538: pcre_free(code);
1.154 paf 539: row_action(**table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
1.63 paf 540: return option_bits[1]!=0; // global=true+table, not global=false
541: }
542:
543: if(exec_substrings<0) {
1.100 parser 544: pcre_free(code);
1.149 paf 545: throw Exception(0,
1.63 paf 546: aorigin,
1.76 paf 547: "regular expression execute error (%d)",
1.63 paf 548: exec_substrings);
549: }
550:
1.154 paf 551: int prefinish=ovector[0];
552: poststart=ovector[1];
1.172.2.1! paf 553: Array& row=*new Array();
1.154 paf 554: row+=need_pre_post_match?&mid(0, prefinish):0; // .prematch column value
555: row+=need_pre_post_match?&mid(prefinish, poststart):0; // .match
556: row+=need_pre_post_match?&mid(poststart, postfinish):0; // .postmatch
1.63 paf 557:
558: for(int i=1; i<exec_substrings; i++) {
1.69 paf 559: // -1:-1 case handled peacefully by mid() itself
560: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 561: }
562:
1.154 paf 563: row_action(**table, &row, prestart, prefinish, poststart, postfinish, info);
1.63 paf 564:
1.154 paf 565: if(!option_bits[1] || prestart==poststart) { // not global | going to hang
1.100 parser 566: pcre_free(code);
1.154 paf 567: row_action(**table, 0/*last time, no row*/, 0, 0, poststart, postfinish, info);
1.63 paf 568: return true;
569: }
1.154 paf 570: prestart=poststart;
1.63 paf 571:
572: /*
573: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 574: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 575: */
576: }
1.82 parser 577: }
578:
1.172.2.1! paf 579: String& String::change_case(Charset& source_charset, Change_case_kind kind) const {
! 580: const unsigned char *tables=source_charset.pcre_tables;
! 581: String& result=*new String();
1.82 parser 582:
583: const unsigned char *a;
584: const unsigned char *b;
585: switch(kind) {
586: case CC_UPPER:
587: a=tables+lcc_offset;
588: b=tables+fcc_offset;
589: break;
590: case CC_LOWER:
591: a=tables+lcc_offset;
592: b=0;
593: break;
594: default:
1.149 paf 595: throw Exception(0,
1.82 parser 596: this,
597: "unknown change case kind #%d",
598: static_cast<int>(kind)); // never
599: a=b=0; // calm, compiler
600: break; // never
601: }
602:
1.143 paf 603: STRING_FOREACH_ROW(
1.172.2.1! paf 604: char *new_cstr=(char *)pa_malloc(row->item.size);
1.143 paf 605: char *dest=new_cstr;
606: const char *src=row->item.ptr;
607: for(int size=row->item.size; size--; src++) {
608: unsigned char c=a[(unsigned char)*src];
609: if(b)
610: c=b[c];
1.82 parser 611:
1.143 paf 612: *dest++=(char)c;
1.82 parser 613: }
1.143 paf 614:
615: result.APPEND(new_cstr, row->item.size,
616: row->item.lang,
617: row->item.origin.file, row->item.origin.line);
618: );
1.89 parser 619:
1.101 parser 620: return result;
621: }
622:
1.150 paf 623: /// @test if in some piece were found no dict words, append it, not it's duplicate
1.172.2.1! paf 624: String& String::replace(Dictionary& dict) const {
1.170 paf 625: char *lcstr=cstr();
626: const char *current=lcstr;
627:
1.172.2.1! paf 628: String& result=*new String();
1.143 paf 629: STRING_FOREACH_ROW(
1.170 paf 630: IFNDEF_NO_STRING_ORIGIN(
631: const char *joined_origin_file=row->item.origin.file;
632: const size_t joined_origin_line=row->item.origin.line;
633: );
634: uchar joined_lang=row->item.lang;
635: const char *joined_ptr=current;
636: // calc size
637: size_t joined_size=0;
638: STRING_PREPARED_FOREACH_ROW(*this,
639: if(row->item.lang==joined_lang)
640: joined_size+=row->item.size;
641: else
642: break; // before non-ours
643: );
644: current+=joined_size;
645:
646: // pointers are after joined piece
647: // & one step back, see STRING_FOREACH_ROW
648: --row; ++countdown;
649:
1.172.2.1! paf 650: char *new_cstr=(char *)pa_malloc((size_t)ceil(joined_size*dict.max_ratio()));
1.123 paf 651: char *dest=new_cstr;
1.170 paf 652: while(joined_size) {
653: // there is a row where first column starts 'joined_ptr'
654: if(Table::Item *item=dict.first_that_starts(joined_ptr, joined_size)) {
1.123 paf 655: // get a=>b values
656: const String& a=*static_cast<Array *>(item)->get_string(0);
657: const String& b=*static_cast<Array *>(item)->get_string(1);
1.170 paf 658: // skip 'a' in 'joined_ptr' && reduce work size
659: joined_ptr+=a.size(); joined_size-=a.size();
1.123 paf 660: // write 'b' to 'dest' && skip 'b' in 'dest'
661: b.store_to(dest); dest+=b.size();
662: } else {
663: // write a char to b && reduce work size
1.170 paf 664: *dest++=*joined_ptr++; joined_size--;
1.101 parser 665: }
666: }
667:
1.170 paf 668: result.APPEND(new_cstr, dest-new_cstr, joined_lang,
669: joined_origin_file, joined_origin_line);
1.156 paf 670: );
1.170 paf 671:
1.156 paf 672: return result;
673: }
674:
1.172.2.1! paf 675: String& String::join_chains(char** acstr) const {
1.156 paf 676: char *lcstr=cstr();
677: const char *current=lcstr;
678:
1.172.2.1! paf 679: String& result=*new String();
1.156 paf 680: STRING_FOREACH_ROW(
681: IFNDEF_NO_STRING_ORIGIN(
682: const char *joined_origin_file=row->item.origin.file;
683: const size_t joined_origin_line=row->item.origin.line;
684: );
685: uchar joined_lang=row->item.lang;
686: const char *joined_ptr=current;
687: // calc size
688: size_t joined_size=0;
689: STRING_PREPARED_FOREACH_ROW(*this,
690: if(row->item.lang==joined_lang)
691: joined_size+=row->item.size;
692: else
693: break; // before non-ours
694: );
695: current+=joined_size;
696:
697: // pointers are after joined piece
698: // & one step back, see STRING_FOREACH_ROW
699: --row; ++countdown;
700:
701: result.APPEND(joined_ptr, joined_size, joined_lang,
1.150 paf 702: joined_origin_file, joined_origin_line);
1.123 paf 703: );
1.156 paf 704:
705: if(acstr)
706: *acstr=lcstr;
1.89 parser 707: return result;
708: }
709:
1.90 parser 710: double String::as_double() const {
1.89 parser 711: double result;
1.114 paf 712: const char *cstr;
713: char buf[MAX_NUMBER];
1.151 paf 714: if(head.chunk.rows+1==append_here) {
715: int size=min(head.chunk.rows[0].item.size, MAX_NUMBER-1);
716: memcpy(buf, head.chunk.rows[0].item.ptr, size);
1.114 paf 717: buf[size]=0;
718: cstr=buf;
719: } else
720: cstr=this->cstr();
1.161 paf 721: while(*cstr && isspace(*cstr))
722: cstr++;
723: if(!*cstr)
1.162 paf 724: return 0;
1.161 paf 725:
1.102 parser 726: char *error_pos;
1.89 parser 727: // 0xABC
1.99 parser 728: if(cstr[0]=='0')
729: if(cstr[1]=='x' || cstr[1]=='X')
730: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
731: else
1.102 parser 732: result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos);
1.89 parser 733: else
1.99 parser 734: result=(double)strtod(cstr, &error_pos);
1.89 parser 735:
1.159 paf 736: while(char c=*error_pos++)
737: if(!isspace(c))
738: throw Exception("number.format",
739: this,
740: "invalid number (double)");
1.89 parser 741:
742: return result;
743: }
1.90 parser 744: int String::as_int() const {
1.89 parser 745: int result;
1.114 paf 746: const char *cstr;
747: char buf[MAX_NUMBER];
1.151 paf 748: if(head.chunk.rows+1==append_here) {
1.163 paf 749: size_t size=min(head.chunk.rows[0].item.size, MAX_NUMBER-1);
1.151 paf 750: memcpy(buf, head.chunk.rows[0].item.ptr, size);
1.114 paf 751: buf[size]=0;
752: cstr=buf;
753: } else
754: cstr=this->cstr();
1.161 paf 755: while(*cstr && isspace(*cstr))
756: cstr++;
757: if(!*cstr)
1.162 paf 758: return 0;
1.161 paf 759:
1.102 parser 760: char *error_pos;
1.89 parser 761: // 0xABC
1.99 parser 762: if(cstr[0]=='0')
763: if(cstr[1]=='x' || cstr[1]=='X')
764: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
765: else
1.102 parser 766: result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 767: else
768: result=(int)strtol(cstr, &error_pos, 0);
769:
1.159 paf 770: while(char c=*error_pos++)
771: if(!isspace(c))
772: throw Exception("number.format",
773: this,
774: "invalid number (int)");
1.82 parser 775:
776: return result;
1.61 paf 777: }
1.113 parser 778:
1.128 paf 779: inline void ushort2uchars(ushort word, uchar& byte1, uchar& byte2) {
780: byte1=word&0xFF;
781: byte2=word>>8;
782: }
783: inline ushort uchars2ushort(uchar byte1, uchar byte2) {
784: return (byte2<<8) | byte1;
785: }
1.113 parser 786: /* @todo maybe network order worth spending some effort?
787: don't bothering myself with network byte order,
788: am not planning to be able to move resulting file across platforms
789: for now
790: */
791: void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const {
792: buf_size=
793: prolog_size
1.126 paf 794: +used_rows()*(sizeof(uchar)+sizeof(ushort))
1.113 parser 795: +size();
1.172.2.1! paf 796: buf=pa_malloc(buf_size);
1.113 parser 797: char *cur=(char *)buf+prolog_size;
798:
1.123 paf 799: STRING_FOREACH_ROW(
800: // lang
1.126 paf 801: memcpy(cur, &row->item.lang, sizeof(uchar));
802: cur+=sizeof(uchar);
1.123 paf 803: // size
1.128 paf 804: uchar byte1; uchar byte2;
805: ushort2uchars(row->item.size, byte1, byte2);
806: memcpy(cur, &byte1, sizeof(uchar)); cur+=sizeof(uchar);
807: memcpy(cur, &byte2, sizeof(uchar)); cur+=sizeof(uchar);
1.123 paf 808: // bytes
809: memcpy(cur, row->item.ptr, row->item.size);
810: cur+=row->item.size;
811: );
1.113 parser 812: }
1.148 paf 813: bool String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) {
1.135 paf 814: if(buf_size<=prolog_size)
1.148 paf 815: return false;
1.135 paf 816:
1.126 paf 817: char *cur=(char *)buf+prolog_size;
1.113 parser 818: buf_size-=prolog_size;
819:
820: while(buf_size) {
1.148 paf 821: if(sizeof(uchar)+sizeof(ushort)>buf_size) // lang+size
822: return false;
823:
824: uchar lang=*(uchar *)(cur);
1.128 paf 825: ushort size=uchars2ushort(
826: *(uchar*)(cur+sizeof(uchar)*1),
827: *(uchar*)(cur+sizeof(uchar)*2)
828: );
829:
1.148 paf 830: size_t piece_size=sizeof(uchar)+sizeof(ushort)+size;
831: if(piece_size>buf_size) // buffer overrun, can be on incomplete cache files
832: return false;
833:
1.128 paf 834: const char *ptr=(const char*)(cur+sizeof(uchar)*3);
1.126 paf 835: APPEND(ptr, size, lang, file, 0);
1.113 parser 836:
837: cur+=piece_size;
838: buf_size-=piece_size;
839: }
1.148 paf 840: return true;
1.113 parser 841: }
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