Annotation of parser3/src/main/pa_string.C, revision 1.116
1.45 paf 1: /** @file
1.55 paf 2: Parser: string class. @see untasize_t.C.
1.46 paf 3:
1.36 paf 4: Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com)
1.104 parser 5: Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf)
1.46 paf 6:
1.116 ! paf 7: $Id: pa_string.C,v 1.115 2001/10/29 15:15:11 paf Exp $
1.4 paf 8: */
9:
1.48 paf 10: #include "pa_config_includes.h"
1.1 paf 11:
1.70 paf 12: #include "pcre.h"
1.82 parser 13: #include "internal.h"
1.70 paf 14:
1.13 paf 15: #include "pa_pool.h"
1.12 paf 16: #include "pa_string.h"
1.5 paf 17: #include "pa_hash.h"
1.22 paf 18: #include "pa_exception.h"
1.53 paf 19: #include "pa_common.h"
1.60 paf 20: #include "pa_array.h"
21: #include "pa_globals.h"
1.61 paf 22: #include "pa_table.h"
1.101 parser 23: #include "pa_dictionary.h"
1.60 paf 24:
1.75 paf 25: String::String(Pool& apool, const char *src, size_t src_size, bool tainted) :
1.94 parser 26: Pooled(apool),
27: forigins_mode(false) {
1.28 paf 28: last_chunk=&head;
29: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 30: append_here=head.rows;
1.2 paf 31: head.preallocated_link=0;
1.28 paf 32: link_row=&head.rows[head.count];
1.115 paf 33: fsize=0;
1.41 paf 34:
35: if(src)
1.75 paf 36: if(tainted)
37: APPEND_TAINTED(src, src_size, 0, 0);
1.41 paf 38: else
1.75 paf 39: APPEND_CLEAN(src, src_size, 0, 0);
1.1 paf 40: }
41:
1.94 parser 42: String::String(const String& src) :
43: Pooled(src.pool()),
44: forigins_mode(false) {
1.8 paf 45: head.count=CR_PREALLOCATED_COUNT;
46:
1.116 ! paf 47: uint src_used_rows=src.used_rows();
1.8 paf 48: if(src_used_rows<=head.count) {
1.55 paf 49: // all new rows fit size_to preallocated area
1.98 parser 50: last_chunk=&head;
1.116 ! paf 51: uint curr_chunk_rows=head.count;
1.8 paf 52: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
53: append_here=&head.rows[src_used_rows];
54: link_row=&head.rows[curr_chunk_rows];
55: } else {
56: // warning:
1.10 paf 57: // heavily relies on the fact
58: // "preallocated area is the same for all strings"
1.8 paf 59: //
60: // info:
61: // allocating only enough mem to fit src string rows
62: // next append would allocate a new chunk
63: //
1.55 paf 64: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 65: // preallocated chunk src to constructing head
66: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 paf 67: // remaining rows size_to new_chunk
1.116 ! paf 68: uint curr_chunk_rows=src_used_rows-head.count;
1.98 parser 69: last_chunk=static_cast<Chunk *>(
1.116 ! paf 70: malloc(sizeof(uint)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *), 9));
1.98 parser 71: last_chunk->count=curr_chunk_rows;
72: head.preallocated_link=last_chunk;
73: append_here=link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 74:
75: Chunk *old_chunk=src.head.preallocated_link;
1.98 parser 76: Chunk::Row *new_rows=last_chunk->rows;
1.116 ! paf 77: uint rows_left_to_copy=last_chunk->count;
1.8 paf 78: while(true) {
1.116 ! paf 79: uint old_count=old_chunk->count;
1.8 paf 80: Chunk *next_chunk=old_chunk->rows[old_count].link;
81: if(next_chunk) {
82: // not last source chunk
83: // taking it all
84: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
85: new_rows+=old_count;
86: rows_left_to_copy-=old_count;
87:
88: old_chunk=next_chunk;
89: } else {
90: // the last source chunk
91: // taking only those rows of chunk that _left_to_copy
92: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
93: break;
94: }
95: }
1.5 paf 96: }
1.8 paf 97: link_row->link=0;
1.115 paf 98: src_used_rows;
1.8 paf 99: fsize=src.fsize;
1.94 parser 100: }
101:
1.115 paf 102: /// @todo not very optimal
103: uint String::used_rows() const {
104: uint result=0;
105: const Chunk *chunk=&head;
106: do {
107: const Chunk::Row *row=chunk->rows;
1.116 ! paf 108: for(uint i=0; i<chunk->count; i++, row++) {
1.115 paf 109: if(row==append_here)
110: goto break2;
111:
112: result++;
113: }
114: chunk=row->link;
115: } while(chunk);
116:
117: break2:
118: return result;
119: }
1.94 parser 120: void String::expand() {
1.116 ! paf 121: uint new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.94 parser 122: last_chunk=static_cast<Chunk *>(
1.116 ! paf 123: malloc(sizeof(uint)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *), 10));
1.94 parser 124: last_chunk->count=new_chunk_count;
125: link_row->link=last_chunk;
126: append_here=last_chunk->rows;
127: link_row=&last_chunk->rows[last_chunk->count];
128: link_row->link=0;
1.5 paf 129: }
1.28 paf 130:
1.42 paf 131: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.60 paf 132: const Chunk *chunk=&src.head;
1.40 paf 133: do {
1.60 paf 134: const Chunk::Row *row=chunk->rows;
1.116 ! paf 135: for(uint i=0; i<chunk->count; i++, row++) {
1.60 paf 136: if(row==src.append_here)
1.40 paf 137: goto break2;
1.60 paf 138:
139: APPEND(row->item.ptr, row->item.size,
140: (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang,
141: row->item.origin.file, row->item.origin.line);
1.40 paf 142: }
143: chunk=row->link;
144: } while(chunk);
145: break2:
1.60 paf 146: return *this;
1.34 paf 147: }
1.60 paf 148:
1.13 paf 149: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 150: if(!src)
151: return *this;
1.26 paf 152: if(!size)
153: size=strlen(src);
154: if(!size)
1.9 paf 155: return *this;
156:
1.1 paf 157: if(chunk_is_full())
158: expand();
159:
160: append_here->item.ptr=src;
1.26 paf 161: fsize+=append_here->item.size=size;
1.52 paf 162: append_here->item.lang=lang;
1.13 paf 163: #ifndef NO_STRING_ORIGIN
1.14 paf 164: append_here->item.origin.file=file;
165: append_here->item.origin.line=line;
1.13 paf 166: #endif
1.115 paf 167: append_here++;
1.1 paf 168:
169: return *this;
1.97 parser 170: }
171:
172: char String::first_char() const {
1.115 paf 173: if(!fsize)
1.112 parser 174: throw Exception(0, 0,
1.97 parser 175: this,
176: "getting first char of empty string");
177:
178: return *head.rows[0].item.ptr;
1.1 paf 179: }
180:
1.16 paf 181: uint String::hash_code() const {
1.7 paf 182: uint result=0;
1.5 paf 183:
1.16 paf 184: const Chunk *chunk=&head;
1.5 paf 185: do {
1.16 paf 186: const Chunk::Row *row=chunk->rows;
1.116 ! paf 187: for(uint i=0; i<chunk->count; i++) {
1.5 paf 188: if(row==append_here)
189: goto break2;
190:
1.6 paf 191: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 192: row++;
193: }
194: chunk=row->link;
195: } while(chunk);
196: break2:
197: return result;
198: }
199:
1.60 paf 200: /// @todo move 'lang' skipping to pos
201: int String::cmp(int& partial, const String& src,
202: size_t this_offset, Untaint_lang lang) const {
1.59 paf 203: partial=-1;
1.55 paf 204: this_offset=min(this_offset, size()-1);
205:
1.16 paf 206: const Chunk *a_chunk=&head;
207: const Chunk *b_chunk=&src.head;
208: const Chunk::Row *a_row=a_chunk->rows;
209: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 210: size_t a_offset=this_offset;
211: size_t b_offset=0;
1.9 paf 212: Chunk::Row *a_end=append_here;
213: Chunk::Row *b_end=src.append_here;
1.116 ! paf 214: uint a_countdown=a_chunk->count;
! 215: uint b_countdown=b_chunk->count;
! 216: int result;
1.60 paf 217: size_t pos=0;
1.33 paf 218:
1.83 parser 219: bool a_break=size()==0;
1.91 parser 220: bool b_break=src.size()==0;
1.83 parser 221: if(!(a_break || b_break)) while(true) {
1.55 paf 222: if(pos+a_row->item.size > this_offset) {
1.71 paf 223: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 224: return -1; // wrong lang -- bail out
225:
1.55 paf 226: int size_diff=
227: (a_row->item.size-a_offset)-
228: (b_row->item.size-b_offset);
229:
230: if(size_diff==0) { // a has same size as b
1.60 paf 231: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
232: a_row->item.size-a_offset);
1.55 paf 233: if(result)
234: return result;
1.60 paf 235: pos+=a_row->item.size;
1.55 paf 236: a_row++; a_countdown--; a_offset=0;
237: b_row++; b_countdown--; b_offset=0;
238: } else if (size_diff>0) { // a longer
1.60 paf 239: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
240: b_row->item.size-b_offset);
1.55 paf 241: if(result)
242: return result;
243: a_offset+=b_row->item.size-b_offset;
244: b_row++; b_countdown--; b_offset=0;
245: } else { // b longer
1.60 paf 246: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
247: a_row->item.size-a_offset);
1.55 paf 248: if(result)
249: return result;
250: b_offset+=a_row->item.size-a_offset;
1.60 paf 251: pos+=a_row->item.size;
1.55 paf 252: a_row++; a_countdown--; a_offset=0;
253: }
1.83 parser 254: if(b_break=b_row==b_end) {
255: a_break=a_row==a_end;
256: break;
257: }
1.55 paf 258: if(!b_countdown) {
259: b_chunk=b_row->link;
260: b_row=b_chunk->rows;
261: b_countdown=b_chunk->count;
262: }
263: } else {
1.60 paf 264: a_offset-=a_row->item.size;
265: pos+=a_row->item.size;
266: a_row++; a_countdown--;
1.9 paf 267: }
268:
1.83 parser 269: if(a_break=a_row==a_end) {
270: b_break=b_row==b_end;
271: break;
272: }
1.11 paf 273: if(!a_countdown) {
1.9 paf 274: a_chunk=a_row->link;
275: a_row=a_chunk->rows;
1.11 paf 276: a_countdown=a_chunk->count;
1.9 paf 277: }
1.27 paf 278: }
1.55 paf 279: if(a_break==b_break) { // ended simultaneously
280: partial=0; return 0;
281: } else if(a_break) { // first bytes equal, but a ended before b
282: partial=1; return -1;
283: } else {
284: partial=2; return +1;
285: }
1.27 paf 286: }
287:
1.60 paf 288: /// @todo move 'lang' skipping to pos
1.59 paf 289: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 290: size_t this_offset, Untaint_lang lang) const {
1.59 paf 291: partial=-1;
1.50 paf 292: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 293: this_offset=min(this_offset, size()-1);
1.27 paf 294:
295: const Chunk *a_chunk=&head;
296: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 297: size_t a_offset=this_offset;
1.55 paf 298: size_t b_offset=0;
1.27 paf 299: Chunk::Row *a_end=append_here;
1.116 ! paf 300: uint a_countdown=a_chunk->count;
1.60 paf 301: size_t pos=0;
1.52 paf 302:
1.83 parser 303: bool a_break=size()==0;
304: bool b_break=b_size==0;
305: if(!(a_break || b_break)) while(true) {
1.59 paf 306: if(pos+a_row->item.size > this_offset) {
1.71 paf 307: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 308: return -1; // wrong lang -- bail out
309:
1.59 paf 310: int size_diff=
311: (a_row->item.size-a_offset)-
312: (b_size-b_offset);
313:
314: if(size_diff==0) { // a has same size as b
1.116 ! paf 315: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 316: a_row->item.size-a_offset)!=0)
317: return result;
1.60 paf 318: pos+=a_row->item.size;
1.59 paf 319: a_row++; a_countdown--; a_offset=0;
320: b_break=true;
321: } else if (size_diff>0) { // a longer
1.116 ! paf 322: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 323: b_size-b_offset)!=0)
324: return result;
325: a_offset+=b_size-b_offset;
326: b_break=true;
327: } else { // b longer
1.116 ! paf 328: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 329: a_row->item.size-a_offset)!=0)
330: return result;
331: b_offset+=a_row->item.size-a_offset;
1.60 paf 332: pos+=a_row->item.size;
1.59 paf 333: a_row++; a_countdown--; a_offset=0;
334: }
335: } else {
1.60 paf 336: a_offset-=a_row->item.size;
337: pos+=a_row->item.size;
338: a_row++; a_countdown--;
1.27 paf 339: }
340:
1.86 parser 341: a_break=a_row==a_end;
342: if(a_break || b_break)
1.83 parser 343: break;
1.27 paf 344: if(!a_countdown) {
345: a_chunk=a_row->link;
346: a_row=a_chunk->rows;
347: a_countdown=a_chunk->count;
1.9 paf 348: }
349: }
1.55 paf 350: if(a_break==b_break) { // ended simultaneously
351: partial=0; return 0;
352: } else if(a_break) { // first bytes equal, but a ended before b
353: partial=1; return -1;
354: } else {
355: partial=2; return +1;
356: }
1.5 paf 357: }
1.46 paf 358:
359: #ifndef NO_STRING_ORIGIN
360: const Origin& String::origin() const {
1.115 paf 361: if(!fsize) {
1.96 parser 362: static const Origin empty_origin={"empty string"};
363: return empty_origin;
364: }
1.46 paf 365:
1.49 paf 366: // determining origin by last appended piece
1.50 paf 367: // because first one frequently constant.
368: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 369: // when last peice is constant,
370: // ex: parser_root_auto_path{dynamic} / auto.p{const}
371: // using first piece
372: Origin& last_origin=append_here[-1].item.origin;
373: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 374: }
375: #endif
1.53 paf 376:
1.69 paf 377: String& String::mid(size_t start, size_t finish) const {
1.107 parser 378: String& result=*NEW String(pool());
379:
1.53 paf 380: start=max(0, start);
1.111 parser 381: finish=min(size(), finish);
1.60 paf 382: if(start==finish)
1.107 parser 383: return result;
1.53 paf 384:
385: size_t pos=0;
386: const Chunk *chunk=&head;
387: do {
388: const Chunk::Row *row=chunk->rows;
1.116 ! paf 389: for(uint i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 390: if(row==append_here)
391: goto break2;
392:
1.60 paf 393: size_t item_finish=pos+row->item.size;
394: if(item_finish > start) { // started now or already?
395: bool started=result.size()==0; // started now?
396: bool finished=finish <= item_finish; // finished now?
1.53 paf 397: size_t offset=started?start-pos:0;
398: size_t size=finished?finish-pos:row->item.size;
399: result.APPEND(
400: row->item.ptr+offset, size-offset,
401: row->item.lang,
402: row->item.origin.file, row->item.origin.line);
403: if(finished)
404: goto break2;
405: }
406: }
407: chunk=row->link;
408: } while(chunk);
409: break2:
1.60 paf 410: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
411: //cstr(), start, finish, result.cstr());
1.53 paf 412: return result;
1.54 paf 413: }
414:
1.60 paf 415: int String::pos(const String& substr,
1.116 ! paf 416: int result, Untaint_lang lang) const {
1.58 paf 417: for(; result<size(); result++) {
1.60 paf 418: int partial; cmp(partial, substr, result, lang);
1.58 paf 419: if(
420: partial==0 || // full match
421: partial==2) // 'substr' starts 'this'+'result'
422: return result;
423: }
424:
425: return -1;
426: }
427:
1.60 paf 428: int String::pos(const char *substr, size_t substr_size,
1.116 ! paf 429: int result, Untaint_lang lang) const {
1.57 paf 430: for(; result<size(); result++) {
1.60 paf 431: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 432: if(
433: partial==0 || // full match
434: partial==2) // 'substr' starts 'this'+'result'
435: return result;
436: }
437:
438: return -1;
1.60 paf 439: }
440:
441: void String::split(Array& result,
442: size_t* pos_after_ref,
443: const char *delim, size_t delim_size,
444: Untaint_lang lang, int limit) const {
445: if(delim_size) {
446: size_t pos_after=pos_after_ref?*pos_after_ref:0;
447: int pos_before;
448: // while we have 'delim'...
449: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 450: result+=&mid(pos_after, pos_before);
1.60 paf 451: pos_after=pos_before+delim_size;
452: }
453: // last piece
454: if(pos_after<size() && limit) {
1.69 paf 455: result+=&mid(pos_after, size());
1.60 paf 456: pos_after=size();
457: }
458: if(pos_after_ref)
459: *pos_after_ref=pos_after;
460: } else { // empty delim
461: result+=this;
462: if(pos_after_ref)
463: *pos_after_ref+=size();
464: }
465: }
466:
467: void String::split(Array& result,
468: size_t* pos_after_ref,
469: const String& delim, Untaint_lang lang,
470: int limit) const {
471: if(delim.size()) {
472: size_t pos_after=pos_after_ref?*pos_after_ref:0;
473: int pos_before;
474: // while we have 'delim'...
475: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 476: result+=&mid(pos_after, pos_before);
1.60 paf 477: pos_after=pos_before+delim.size();
478: }
479: // last piece
480: if(pos_after<size() && limit) {
1.69 paf 481: result+=&mid(pos_after, size());
1.60 paf 482: pos_after=size();
483: }
484: if(pos_after_ref)
485: *pos_after_ref=pos_after;
486: } else { // empty delim
487: result+=this;
488: if(pos_after_ref)
489: *pos_after_ref+=size();
490: }
1.61 paf 491: }
492:
1.63 paf 493: static void regex_options(char *options, int *result){
494: struct Regex_option {
495: char key;
496: int clear, set;
497: int *result;
498: } regex_option[]={
499: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 500: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 501: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
502: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
503: {'g', 0, true, result+1}, // many rows
504: {0},
505: };
506: result[0]=PCRE_EXTRA | PCRE_DOTALL;
507: result[1]=0;
508:
509: if(options)
510: for(Regex_option *o=regex_option; o->key; o++)
511: if(
512: strchr(options, o->key) ||
513: strchr(options, toupper(o->key))) {
514: *(o->result)&=~o->clear;
515: *(o->result)|=o->set;
516: }
517: }
518:
1.88 parser 519: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.77 paf 520: bool String::match(const unsigned char *pcre_tables,
521: const String *aorigin,
1.62 paf 522: const String& regexp,
1.63 paf 523: const String *options,
1.64 paf 524: Table **table,
1.95 parser 525: Row_action row_action, void *info,
526: bool *was_global) const {
1.64 paf 527:
1.73 paf 528: if(!regexp.size())
1.112 parser 529: throw Exception(0, 0,
1.73 paf 530: aorigin,
531: "regexp is empty");
1.68 paf 532: const char *pattern=regexp.cstr(UL_AS_IS);
1.62 paf 533: const char *errptr;
534: int erroffset;
1.63 paf 535: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
1.95 parser 536: if(was_global)
537: *was_global=option_bits[1]!=0;
1.63 paf 538: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 539: &errptr, &erroffset,
1.74 paf 540: pcre_tables);
1.62 paf 541:
1.67 paf 542: if(!code)
1.112 parser 543: throw Exception(0, 0,
1.69 paf 544: ®exp.mid(erroffset, regexp.size()),
1.74 paf 545: "regular expression syntax error - %s", errptr);
1.62 paf 546:
1.63 paf 547: int info_substrings=pcre_info(code, 0, 0);
548: if(info_substrings<0) {
1.100 parser 549: pcre_free(code);
1.112 parser 550: throw Exception(0, 0,
1.73 paf 551: aorigin,
1.76 paf 552: "pcre_info error (%d)",
1.73 paf 553: info_substrings);
1.63 paf 554: }
555:
556: int startoffset=0;
1.68 paf 557: const char *subject=cstr(UL_AS_IS);
1.62 paf 558: int length=strlen(subject);
1.63 paf 559: int ovecsize;
560: int *ovector=(int *)malloc(sizeof(int)*
1.114 paf 561: (ovecsize=(1/*match*/+info_substrings)*3), 11);
1.62 paf 562:
1.64 paf 563: { // create table
564: Array& columns=*NEW Array(pool());
565: columns+=string_pre_match_name;
566: columns+=string_match_name;
567: columns+=string_post_match_name;
568: for(int i=1; i<=info_substrings; i++) {
569: char *column=(char *)malloc(MAX_NUMBER);
570: snprintf(column, MAX_NUMBER, "%d", i);
571: columns+=NEW String(pool(), column); // .i column name
572: }
573: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 574: }
1.63 paf 575:
1.64 paf 576: int exec_option_bits=0;
1.63 paf 577: while(true) {
578: int exec_substrings=pcre_exec(code, 0,
579: subject, length, startoffset,
1.64 paf 580: exec_option_bits, ovector, ovecsize);
1.63 paf 581:
582: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 583: pcre_free(code);
584: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 585: return option_bits[1]!=0; // global=true+table, not global=false
586: }
587:
588: if(exec_substrings<0) {
1.100 parser 589: pcre_free(code);
1.112 parser 590: throw Exception(0, 0,
1.63 paf 591: aorigin,
1.76 paf 592: "regular expression execute error (%d)",
1.63 paf 593: exec_substrings);
594: }
595:
596: Array& row=*NEW Array(pool());
1.81 paf 597: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 598: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 599: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 600:
601: for(int i=1; i<exec_substrings; i++) {
1.69 paf 602: // -1:-1 case handled peacefully by mid() itself
603: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 604: }
605:
1.100 parser 606: row_action(**table, &row, startoffset, ovector[0], info);
1.63 paf 607:
1.100 parser 608: if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang
609: pcre_free(code);
610: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 611: return true;
612: }
1.100 parser 613: startoffset=ovector[1];
1.63 paf 614:
615: /*
616: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 617: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 618: */
619: }
1.82 parser 620: }
621:
622: String& String::change_case(Pool& pool, const unsigned char *tables,
623: Change_case_kind kind) const {
624: String& result=*new(pool) String(pool);
625:
626: const unsigned char *a;
627: const unsigned char *b;
628: switch(kind) {
629: case CC_UPPER:
630: a=tables+lcc_offset;
631: b=tables+fcc_offset;
632: break;
633: case CC_LOWER:
634: a=tables+lcc_offset;
635: b=0;
636: break;
637: default:
1.112 parser 638: throw Exception(0, 0,
1.82 parser 639: this,
640: "unknown change case kind #%d",
641: static_cast<int>(kind)); // never
642: a=b=0; // calm, compiler
643: break; // never
644: }
645:
646: const Chunk *chunk=&head;
647: do {
648: const Chunk::Row *row=chunk->rows;
1.116 ! paf 649: for(uint i=0; i<chunk->count; i++, row++) {
1.82 parser 650: if(row==append_here)
651: goto break2;
652:
1.114 paf 653: char *new_cstr=(char *)pool.malloc(row->item.size, 12);
1.82 parser 654: char *dest=new_cstr;
655: const char *src=row->item.ptr;
656: for(int size=row->item.size; size--; src++) {
657: unsigned char c=a[(unsigned char)*src];
658: if(b)
659: c=b[c];
660:
661: *dest++=(char)c;
662: }
663:
664: result.APPEND(new_cstr, row->item.size,
665: row->item.lang,
666: row->item.origin.file, row->item.origin.line);
667: }
668: chunk=row->link;
669: } while(chunk);
670: break2:
1.89 parser 671:
1.101 parser 672: return result;
673: }
674:
1.108 parser 675: void String::join_chain(Pool& pool,
1.116 ! paf 676: uint& ai, const Chunk*& achunk, const Chunk::Row*& arow,
1.108 parser 677: Untaint_lang& joined_lang, const char *& joined_ptr, size_t& joined_size) const {
678: joined_lang=arow->item.lang;
679:
680: // calc size
681: joined_size=0;
682: {
1.116 ! paf 683: uint start_i=ai;
1.108 parser 684: const Chunk::Row *start_row=arow;
685: const Chunk *chunk=achunk;
686: do {
687: const Chunk::Row *row=start_row;
1.116 ! paf 688: for(uint i=start_i; i<chunk->count; i++, row++) {
1.108 parser 689: if(row==append_here)
690: goto break21;
691:
692: if(row->item.lang==joined_lang)
693: joined_size+=row->item.size;
694: else
695: break;
696: }
697: if(chunk=row->link) {
698: start_i=0;
699: start_row=chunk->rows;
700: } else
701: break;
702: } while(true);
703: break21:;
704: }
705:
706: // if one row, return simply itself
707: if(joined_size==arow->item.size) {
708: joined_ptr=arow->item.ptr;
709: ai++; arow++;
710: if(ai==achunk->count)
711: achunk=arow->link;
712: } else {
713: // join adjacent rows
1.114 paf 714: char *ptr=(char *)pool.malloc(joined_size,13);
1.108 parser 715: joined_ptr=ptr;
1.116 ! paf 716: uint start_i=ai;
1.108 parser 717: const Chunk::Row *start_row=arow;
718: const Chunk *chunk=achunk;
1.116 ! paf 719: uint i;
1.108 parser 720: const Chunk::Row *row;
721: do {
722: row=start_row;
723: for(i=start_i; i<chunk->count; i++, row++) {
724: if(row==append_here)
725: goto break22;
726:
727: if(row->item.lang==joined_lang) {
728: memcpy(ptr, row->item.ptr, row->item.size);
729: ptr+=row->item.size;
730: } else
731: break;
732: }
733: if(chunk=row->link) {
734: start_i=0;
735: start_row=chunk->rows;
736: } else
737: break;
738: } while(true);
739: break22:;
740:
741: // return joined rows
742: ai=i;
743: arow=row;
744: achunk=chunk;
745: }
746: }
747:
748: String& String::reconstruct(Pool& pool) const {
749: //_asm int 3;
750: String& result=*new(pool) String(pool);
751: const Chunk *chunk=&head;
752: do {
753: const Chunk::Row *row=chunk->rows;
1.116 ! paf 754: for(uint i=0; i<chunk->count; ) {
1.108 parser 755: if(row==append_here)
756: goto break2;
757:
758: Untaint_lang joined_lang;
1.109 parser 759: const char *joined_ptr;
1.108 parser 760: size_t joined_size;
761: join_chain(pool, i, chunk, row,
762: joined_lang, joined_ptr, joined_size);
763:
764: result.APPEND(joined_ptr, joined_size,
765: joined_lang,
766: row->item.origin.file, row->item.origin.line);
767: if(!chunk)
768: goto break2;
769: }
770: } while(true);
771: break2:
772:
773: return result;
774: };
775:
776: String& String::replace_in_reconstructed(Pool& pool, Dictionary& dict) const {
1.106 parser 777: //_asm int 3;
1.101 parser 778: String& result=*new(pool) String(pool);
779: const Chunk *chunk=&head;
780: do {
781: const Chunk::Row *row=chunk->rows;
1.116 ! paf 782: for(uint i=0; i<chunk->count; i++, row++) {
1.101 parser 783: if(row==append_here)
784: goto break2;
785:
786: const char *src=row->item.ptr;
787: size_t src_size=row->item.size;
1.114 paf 788: char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14);
1.101 parser 789: char *dest=new_cstr;
790: while(src_size) {
791: // there is a row where first column starts 'src'
1.106 parser 792: if(Table::Item *item=dict.first_that_starts(src, src_size)) {
1.101 parser 793: // get a=>b values
794: const String& a=*static_cast<Array *>(item)->get_string(0);
795: const String& b=*static_cast<Array *>(item)->get_string(1);
1.105 parser 796: // skip 'a' in 'src' && reduce work size
797: src+=a.size(); src_size-=a.size();
798: // write 'b' to 'dest' && skip 'b' in 'dest'
799: b.store_to(dest); dest+=b.size();
1.101 parser 800: } else {
1.105 parser 801: // write a char to b && reduce work size
802: *dest++=*src++; src_size--;
1.101 parser 803: }
804: }
805:
806: result.APPEND(new_cstr, dest-new_cstr,
807: row->item.lang,
808: row->item.origin.file, row->item.origin.line);
809: }
810: chunk=row->link;
811: } while(chunk);
812:
813: break2:
1.89 parser 814: return result;
1.108 parser 815: }
816:
817: String& String::replace(Pool& pool, Dictionary& dict) const {
818: return reconstruct(pool).replace_in_reconstructed(pool, dict);
1.89 parser 819: }
820:
1.90 parser 821: double String::as_double() const {
1.89 parser 822: double result;
1.114 paf 823: const char *cstr;
824: char buf[MAX_NUMBER];
1.115 paf 825: if(head.rows+1==append_here) {
1.114 paf 826: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
827: memcpy(buf, head.rows[0].item.ptr, size);
828: buf[size]=0;
829: cstr=buf;
830: } else
831: cstr=this->cstr();
1.102 parser 832: char *error_pos;
1.89 parser 833: // 0xABC
1.99 parser 834: if(cstr[0]=='0')
835: if(cstr[1]=='x' || cstr[1]=='X')
836: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
837: else
1.102 parser 838: result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos);
1.89 parser 839: else
1.99 parser 840: result=(double)strtod(cstr, &error_pos);
1.89 parser 841:
1.103 parser 842: if(*error_pos/*not EOS*/)
1.112 parser 843: throw Exception(0, 0,
1.89 parser 844: this,
845: "invalid number (double)");
846:
847: return result;
848: }
1.90 parser 849: int String::as_int() const {
1.89 parser 850: int result;
1.114 paf 851: const char *cstr;
852: char buf[MAX_NUMBER];
1.115 paf 853: if(head.rows+1==append_here) {
1.114 paf 854: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
855: memcpy(buf, head.rows[0].item.ptr, size);
856: buf[size]=0;
857: cstr=buf;
858: } else
859: cstr=this->cstr();
1.102 parser 860: char *error_pos;
1.89 parser 861: // 0xABC
1.99 parser 862: if(cstr[0]=='0')
863: if(cstr[1]=='x' || cstr[1]=='X')
864: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
865: else
1.102 parser 866: result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 867: else
868: result=(int)strtol(cstr, &error_pos, 0);
869:
1.103 parser 870: if(*error_pos/*not EOS*/)
1.112 parser 871: throw Exception(0, 0,
1.89 parser 872: this,
873: "invalid number (int)");
1.82 parser 874:
875: return result;
1.61 paf 876: }
1.113 parser 877:
878: /* @todo maybe network order worth spending some effort?
879: don't bothering myself with network byte order,
880: am not planning to be able to move resulting file across platforms
881: for now
882: */
883: void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const {
884: buf_size=
885: prolog_size
1.115 paf 886: +used_rows()*(sizeof(Untaint_lang)+sizeof(size_t))
1.113 parser 887: +size();
1.114 paf 888: buf=malloc(buf_size,15);
1.113 parser 889: char *cur=(char *)buf+prolog_size;
890:
891: const Chunk *chunk=&head;
892: do {
893: const Chunk::Row *row=chunk->rows;
1.116 ! paf 894: for(uint i=0; i<chunk->count; i++) {
1.113 parser 895: if(row==append_here)
896: goto break2;
897:
898: // lang
899: memcpy(cur, &row->item.lang, sizeof(Untaint_lang));
900: cur+=sizeof(Untaint_lang);
901: // size
902: memcpy(cur, &row->item.size, sizeof(size_t));
903: cur+=sizeof(size_t);
904: // bytes
905: memcpy(cur, row->item.ptr, row->item.size);
906: cur+=row->item.size;
907:
908: row++;
909: }
910: chunk=row->link;
911: } while(chunk);
912: break2:
913: ;
914: }
915:
916: /* @todo maybe network order worth spending some effort?
917: don't bothering myself with network byte order,
918: am not planning to be able to move resulting file across platforms
919: for now
920: */
921: #ifndef DOXYGEN
922: struct Serialized_piece {
923: String::Untaint_lang lang;
924: size_t size;
925: char ptr[1];
926: };
927: #endif
928:
929: void String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) {
930: char *cur=((char *)buf)+prolog_size;
931: buf_size-=prolog_size;
932:
933: while(buf_size) {
934: Serialized_piece& p=*(Serialized_piece *)cur;
935: APPEND(p.ptr, p.size, p.lang, file, 0);
936:
937: size_t piece_size=sizeof(p.lang)+sizeof(p.size)+p.size;
938: cur+=piece_size;
939: buf_size-=piece_size;
940: }
941: }
942:
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