Annotation of parser3/src/main/pa_string.C, revision 1.120
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)
1.119 paf 5: Author: Alexander Petrosyan <paf@design.ru> (http://paf.design.ru)
1.46 paf 6:
1.120 ! paf 7: $Id: pa_string.C,v 1.119 2001/11/05 11:46:28 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) :
1.120 ! paf 26: Pooled(apool) {
1.28 paf 27: last_chunk=&head;
28: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 29: append_here=head.rows;
1.2 paf 30: head.preallocated_link=0;
1.28 paf 31: link_row=&head.rows[head.count];
1.41 paf 32:
33: if(src)
1.75 paf 34: if(tainted)
35: APPEND_TAINTED(src, src_size, 0, 0);
1.41 paf 36: else
1.75 paf 37: APPEND_CLEAN(src, src_size, 0, 0);
1.1 paf 38: }
39:
1.94 parser 40: String::String(const String& src) :
1.120 ! paf 41: Pooled(src.pool()) {
1.8 paf 42: head.count=CR_PREALLOCATED_COUNT;
43:
1.116 paf 44: uint src_used_rows=src.used_rows();
1.8 paf 45: if(src_used_rows<=head.count) {
1.55 paf 46: // all new rows fit size_to preallocated area
1.98 parser 47: last_chunk=&head;
1.116 paf 48: uint curr_chunk_rows=head.count;
1.8 paf 49: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
50: append_here=&head.rows[src_used_rows];
51: link_row=&head.rows[curr_chunk_rows];
52: } else {
53: // warning:
1.10 paf 54: // heavily relies on the fact
55: // "preallocated area is the same for all strings"
1.8 paf 56: //
57: // info:
58: // allocating only enough mem to fit src string rows
59: // next append would allocate a new chunk
60: //
1.55 paf 61: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 62: // preallocated chunk src to constructing head
63: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 paf 64: // remaining rows size_to new_chunk
1.116 paf 65: uint curr_chunk_rows=src_used_rows-head.count;
1.98 parser 66: last_chunk=static_cast<Chunk *>(
1.116 paf 67: malloc(sizeof(uint)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *), 9));
1.98 parser 68: last_chunk->count=curr_chunk_rows;
69: head.preallocated_link=last_chunk;
70: append_here=link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 71:
72: Chunk *old_chunk=src.head.preallocated_link;
1.98 parser 73: Chunk::Row *new_rows=last_chunk->rows;
1.116 paf 74: uint rows_left_to_copy=last_chunk->count;
1.8 paf 75: while(true) {
1.116 paf 76: uint old_count=old_chunk->count;
1.8 paf 77: Chunk *next_chunk=old_chunk->rows[old_count].link;
78: if(next_chunk) {
79: // not last source chunk
80: // taking it all
81: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
82: new_rows+=old_count;
83: rows_left_to_copy-=old_count;
84:
85: old_chunk=next_chunk;
86: } else {
87: // the last source chunk
88: // taking only those rows of chunk that _left_to_copy
89: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
90: break;
91: }
92: }
1.5 paf 93: }
1.8 paf 94: link_row->link=0;
1.115 paf 95: src_used_rows;
1.120 ! paf 96: }
! 97:
! 98: size_t String::size() const {
! 99: size_t result=0;
! 100: const Chunk *chunk=&head;
! 101: do {
! 102: const Chunk::Row *row=chunk->rows;
! 103: for(uint i=0; i<chunk->count; i++, row++) {
! 104: if(row==append_here)
! 105: goto break2;
! 106:
! 107: result+=row->item.size;
! 108: }
! 109: chunk=row->link;
! 110: } while(chunk);
! 111:
! 112: break2:
! 113: return result;
1.94 parser 114: }
115:
1.115 paf 116: /// @todo not very optimal
117: uint String::used_rows() const {
118: uint result=0;
119: const Chunk *chunk=&head;
120: do {
121: const Chunk::Row *row=chunk->rows;
1.116 paf 122: for(uint i=0; i<chunk->count; i++, row++) {
1.115 paf 123: if(row==append_here)
124: goto break2;
125:
126: result++;
127: }
128: chunk=row->link;
129: } while(chunk);
130:
131: break2:
132: return result;
133: }
1.94 parser 134: void String::expand() {
1.116 paf 135: uint new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.94 parser 136: last_chunk=static_cast<Chunk *>(
1.116 paf 137: malloc(sizeof(uint)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *), 10));
1.94 parser 138: last_chunk->count=new_chunk_count;
139: link_row->link=last_chunk;
140: append_here=last_chunk->rows;
141: link_row=&last_chunk->rows[last_chunk->count];
142: link_row->link=0;
1.5 paf 143: }
1.28 paf 144:
1.42 paf 145: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.60 paf 146: const Chunk *chunk=&src.head;
1.40 paf 147: do {
1.60 paf 148: const Chunk::Row *row=chunk->rows;
1.116 paf 149: for(uint i=0; i<chunk->count; i++, row++) {
1.60 paf 150: if(row==src.append_here)
1.40 paf 151: goto break2;
1.60 paf 152:
153: APPEND(row->item.ptr, row->item.size,
1.117 paf 154: (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:(Untaint_lang)row->item.lang,
1.60 paf 155: row->item.origin.file, row->item.origin.line);
1.40 paf 156: }
157: chunk=row->link;
158: } while(chunk);
159: break2:
1.60 paf 160: return *this;
1.34 paf 161: }
1.60 paf 162:
1.13 paf 163: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 164: if(!src)
165: return *this;
1.26 paf 166: if(!size)
167: size=strlen(src);
168: if(!size)
1.9 paf 169: return *this;
170:
1.1 paf 171: if(chunk_is_full())
172: expand();
173:
174: append_here->item.ptr=src;
1.52 paf 175: append_here->item.lang=lang;
1.13 paf 176: #ifndef NO_STRING_ORIGIN
1.14 paf 177: append_here->item.origin.file=file;
178: append_here->item.origin.line=line;
1.13 paf 179: #endif
1.115 paf 180: append_here++;
1.1 paf 181:
182: return *this;
1.97 parser 183: }
184:
185: char String::first_char() const {
1.120 ! paf 186: if(!used_rows())
1.112 parser 187: throw Exception(0, 0,
1.97 parser 188: this,
189: "getting first char of empty string");
190:
191: return *head.rows[0].item.ptr;
1.1 paf 192: }
193:
1.16 paf 194: uint String::hash_code() const {
1.7 paf 195: uint result=0;
1.5 paf 196:
1.16 paf 197: const Chunk *chunk=&head;
1.5 paf 198: do {
1.16 paf 199: const Chunk::Row *row=chunk->rows;
1.116 paf 200: for(uint i=0; i<chunk->count; i++) {
1.5 paf 201: if(row==append_here)
202: goto break2;
203:
1.6 paf 204: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 205: row++;
206: }
207: chunk=row->link;
208: } while(chunk);
209: break2:
210: return result;
211: }
212:
1.60 paf 213: /// @todo move 'lang' skipping to pos
214: int String::cmp(int& partial, const String& src,
215: size_t this_offset, Untaint_lang lang) const {
1.59 paf 216: partial=-1;
1.55 paf 217: this_offset=min(this_offset, size()-1);
218:
1.16 paf 219: const Chunk *a_chunk=&head;
220: const Chunk *b_chunk=&src.head;
221: const Chunk::Row *a_row=a_chunk->rows;
222: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 223: size_t a_offset=this_offset;
224: size_t b_offset=0;
1.9 paf 225: Chunk::Row *a_end=append_here;
226: Chunk::Row *b_end=src.append_here;
1.116 paf 227: uint a_countdown=a_chunk->count;
228: uint b_countdown=b_chunk->count;
229: int result;
1.60 paf 230: size_t pos=0;
1.33 paf 231:
1.83 parser 232: bool a_break=size()==0;
1.91 parser 233: bool b_break=src.size()==0;
1.83 parser 234: if(!(a_break || b_break)) while(true) {
1.55 paf 235: if(pos+a_row->item.size > this_offset) {
1.71 paf 236: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 237: return -1; // wrong lang -- bail out
238:
1.55 paf 239: int size_diff=
240: (a_row->item.size-a_offset)-
241: (b_row->item.size-b_offset);
242:
243: if(size_diff==0) { // a has same size as b
1.60 paf 244: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
245: a_row->item.size-a_offset);
1.55 paf 246: if(result)
247: return result;
1.60 paf 248: pos+=a_row->item.size;
1.55 paf 249: a_row++; a_countdown--; a_offset=0;
250: b_row++; b_countdown--; b_offset=0;
251: } else if (size_diff>0) { // a longer
1.60 paf 252: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
253: b_row->item.size-b_offset);
1.55 paf 254: if(result)
255: return result;
256: a_offset+=b_row->item.size-b_offset;
257: b_row++; b_countdown--; b_offset=0;
258: } else { // b longer
1.60 paf 259: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
260: a_row->item.size-a_offset);
1.55 paf 261: if(result)
262: return result;
263: b_offset+=a_row->item.size-a_offset;
1.60 paf 264: pos+=a_row->item.size;
1.55 paf 265: a_row++; a_countdown--; a_offset=0;
266: }
1.83 parser 267: if(b_break=b_row==b_end) {
268: a_break=a_row==a_end;
269: break;
270: }
1.55 paf 271: if(!b_countdown) {
272: b_chunk=b_row->link;
273: b_row=b_chunk->rows;
274: b_countdown=b_chunk->count;
275: }
276: } else {
1.60 paf 277: a_offset-=a_row->item.size;
278: pos+=a_row->item.size;
279: a_row++; a_countdown--;
1.9 paf 280: }
281:
1.83 parser 282: if(a_break=a_row==a_end) {
283: b_break=b_row==b_end;
284: break;
285: }
1.11 paf 286: if(!a_countdown) {
1.9 paf 287: a_chunk=a_row->link;
288: a_row=a_chunk->rows;
1.11 paf 289: a_countdown=a_chunk->count;
1.9 paf 290: }
1.27 paf 291: }
1.55 paf 292: if(a_break==b_break) { // ended simultaneously
293: partial=0; return 0;
294: } else if(a_break) { // first bytes equal, but a ended before b
295: partial=1; return -1;
296: } else {
297: partial=2; return +1;
298: }
1.27 paf 299: }
300:
1.60 paf 301: /// @todo move 'lang' skipping to pos
1.59 paf 302: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 303: size_t this_offset, Untaint_lang lang) const {
1.59 paf 304: partial=-1;
1.50 paf 305: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 306: this_offset=min(this_offset, size()-1);
1.27 paf 307:
308: const Chunk *a_chunk=&head;
309: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 310: size_t a_offset=this_offset;
1.55 paf 311: size_t b_offset=0;
1.27 paf 312: Chunk::Row *a_end=append_here;
1.116 paf 313: uint a_countdown=a_chunk->count;
1.60 paf 314: size_t pos=0;
1.52 paf 315:
1.83 parser 316: bool a_break=size()==0;
317: bool b_break=b_size==0;
318: if(!(a_break || b_break)) while(true) {
1.59 paf 319: if(pos+a_row->item.size > this_offset) {
1.71 paf 320: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 321: return -1; // wrong lang -- bail out
322:
1.59 paf 323: int size_diff=
324: (a_row->item.size-a_offset)-
325: (b_size-b_offset);
326:
327: if(size_diff==0) { // a has same size as b
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;
1.60 paf 331: pos+=a_row->item.size;
1.59 paf 332: a_row++; a_countdown--; a_offset=0;
333: b_break=true;
334: } else if (size_diff>0) { // a longer
1.116 paf 335: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 336: b_size-b_offset)!=0)
337: return result;
338: a_offset+=b_size-b_offset;
339: b_break=true;
340: } else { // b longer
1.116 paf 341: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 342: a_row->item.size-a_offset)!=0)
343: return result;
344: b_offset+=a_row->item.size-a_offset;
1.60 paf 345: pos+=a_row->item.size;
1.59 paf 346: a_row++; a_countdown--; a_offset=0;
347: }
348: } else {
1.60 paf 349: a_offset-=a_row->item.size;
350: pos+=a_row->item.size;
351: a_row++; a_countdown--;
1.27 paf 352: }
353:
1.86 parser 354: a_break=a_row==a_end;
355: if(a_break || b_break)
1.83 parser 356: break;
1.27 paf 357: if(!a_countdown) {
358: a_chunk=a_row->link;
359: a_row=a_chunk->rows;
360: a_countdown=a_chunk->count;
1.9 paf 361: }
362: }
1.55 paf 363: if(a_break==b_break) { // ended simultaneously
364: partial=0; return 0;
365: } else if(a_break) { // first bytes equal, but a ended before b
366: partial=1; return -1;
367: } else {
368: partial=2; return +1;
369: }
1.5 paf 370: }
1.46 paf 371:
372: #ifndef NO_STRING_ORIGIN
373: const Origin& String::origin() const {
1.120 ! paf 374: if(!used_rows()) {
1.96 parser 375: static const Origin empty_origin={"empty string"};
376: return empty_origin;
377: }
1.46 paf 378:
1.49 paf 379: // determining origin by last appended piece
1.50 paf 380: // because first one frequently constant.
381: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 382: // when last peice is constant,
383: // ex: parser_root_auto_path{dynamic} / auto.p{const}
384: // using first piece
385: Origin& last_origin=append_here[-1].item.origin;
386: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 387: }
388: #endif
1.53 paf 389:
1.69 paf 390: String& String::mid(size_t start, size_t finish) const {
1.107 parser 391: String& result=*NEW String(pool());
392:
1.53 paf 393: start=max(0, start);
1.111 parser 394: finish=min(size(), finish);
1.60 paf 395: if(start==finish)
1.107 parser 396: return result;
1.53 paf 397:
398: size_t pos=0;
399: const Chunk *chunk=&head;
400: do {
401: const Chunk::Row *row=chunk->rows;
1.116 paf 402: for(uint i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 403: if(row==append_here)
404: goto break2;
405:
1.60 paf 406: size_t item_finish=pos+row->item.size;
407: if(item_finish > start) { // started now or already?
408: bool started=result.size()==0; // started now?
409: bool finished=finish <= item_finish; // finished now?
1.53 paf 410: size_t offset=started?start-pos:0;
411: size_t size=finished?finish-pos:row->item.size;
412: result.APPEND(
413: row->item.ptr+offset, size-offset,
1.117 paf 414: (Untaint_lang)row->item.lang,
1.53 paf 415: row->item.origin.file, row->item.origin.line);
416: if(finished)
417: goto break2;
418: }
419: }
420: chunk=row->link;
421: } while(chunk);
422: break2:
1.60 paf 423: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
424: //cstr(), start, finish, result.cstr());
1.53 paf 425: return result;
1.54 paf 426: }
427:
1.60 paf 428: int String::pos(const String& substr,
1.116 paf 429: int result, Untaint_lang lang) const {
1.58 paf 430: for(; result<size(); result++) {
1.60 paf 431: int partial; cmp(partial, substr, result, lang);
1.58 paf 432: if(
433: partial==0 || // full match
434: partial==2) // 'substr' starts 'this'+'result'
435: return result;
436: }
437:
438: return -1;
439: }
440:
1.60 paf 441: int String::pos(const char *substr, size_t substr_size,
1.116 paf 442: int result, Untaint_lang lang) const {
1.57 paf 443: for(; result<size(); result++) {
1.60 paf 444: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 445: if(
446: partial==0 || // full match
447: partial==2) // 'substr' starts 'this'+'result'
448: return result;
449: }
450:
451: return -1;
1.60 paf 452: }
453:
454: void String::split(Array& result,
455: size_t* pos_after_ref,
456: const char *delim, size_t delim_size,
457: Untaint_lang lang, int limit) const {
458: if(delim_size) {
459: size_t pos_after=pos_after_ref?*pos_after_ref:0;
460: int pos_before;
461: // while we have 'delim'...
462: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 463: result+=&mid(pos_after, pos_before);
1.60 paf 464: pos_after=pos_before+delim_size;
465: }
466: // last piece
467: if(pos_after<size() && limit) {
1.69 paf 468: result+=&mid(pos_after, size());
1.60 paf 469: pos_after=size();
470: }
471: if(pos_after_ref)
472: *pos_after_ref=pos_after;
473: } else { // empty delim
474: result+=this;
475: if(pos_after_ref)
476: *pos_after_ref+=size();
477: }
478: }
479:
480: void String::split(Array& result,
481: size_t* pos_after_ref,
482: const String& delim, Untaint_lang lang,
483: int limit) const {
484: if(delim.size()) {
485: size_t pos_after=pos_after_ref?*pos_after_ref:0;
486: int pos_before;
487: // while we have 'delim'...
488: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 489: result+=&mid(pos_after, pos_before);
1.60 paf 490: pos_after=pos_before+delim.size();
491: }
492: // last piece
493: if(pos_after<size() && limit) {
1.69 paf 494: result+=&mid(pos_after, size());
1.60 paf 495: pos_after=size();
496: }
497: if(pos_after_ref)
498: *pos_after_ref=pos_after;
499: } else { // empty delim
500: result+=this;
501: if(pos_after_ref)
502: *pos_after_ref+=size();
503: }
1.61 paf 504: }
505:
1.63 paf 506: static void regex_options(char *options, int *result){
507: struct Regex_option {
508: char key;
509: int clear, set;
510: int *result;
511: } regex_option[]={
512: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 513: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 514: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
515: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
516: {'g', 0, true, result+1}, // many rows
517: {0},
518: };
519: result[0]=PCRE_EXTRA | PCRE_DOTALL;
520: result[1]=0;
521:
522: if(options)
523: for(Regex_option *o=regex_option; o->key; o++)
524: if(
525: strchr(options, o->key) ||
526: strchr(options, toupper(o->key))) {
527: *(o->result)&=~o->clear;
528: *(o->result)|=o->set;
529: }
530: }
531:
1.88 parser 532: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.77 paf 533: bool String::match(const unsigned char *pcre_tables,
534: const String *aorigin,
1.62 paf 535: const String& regexp,
1.63 paf 536: const String *options,
1.64 paf 537: Table **table,
1.95 parser 538: Row_action row_action, void *info,
539: bool *was_global) const {
1.64 paf 540:
1.73 paf 541: if(!regexp.size())
1.112 parser 542: throw Exception(0, 0,
1.73 paf 543: aorigin,
544: "regexp is empty");
1.118 paf 545: const char *pattern=regexp.cstr();
1.62 paf 546: const char *errptr;
547: int erroffset;
1.63 paf 548: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
1.95 parser 549: if(was_global)
550: *was_global=option_bits[1]!=0;
1.63 paf 551: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 552: &errptr, &erroffset,
1.74 paf 553: pcre_tables);
1.62 paf 554:
1.67 paf 555: if(!code)
1.112 parser 556: throw Exception(0, 0,
1.69 paf 557: ®exp.mid(erroffset, regexp.size()),
1.74 paf 558: "regular expression syntax error - %s", errptr);
1.62 paf 559:
1.63 paf 560: int info_substrings=pcre_info(code, 0, 0);
561: if(info_substrings<0) {
1.100 parser 562: pcre_free(code);
1.112 parser 563: throw Exception(0, 0,
1.73 paf 564: aorigin,
1.76 paf 565: "pcre_info error (%d)",
1.73 paf 566: info_substrings);
1.63 paf 567: }
568:
569: int startoffset=0;
1.118 paf 570: const char *subject=cstr();
1.62 paf 571: int length=strlen(subject);
1.63 paf 572: int ovecsize;
573: int *ovector=(int *)malloc(sizeof(int)*
1.114 paf 574: (ovecsize=(1/*match*/+info_substrings)*3), 11);
1.62 paf 575:
1.64 paf 576: { // create table
577: Array& columns=*NEW Array(pool());
578: columns+=string_pre_match_name;
579: columns+=string_match_name;
580: columns+=string_post_match_name;
581: for(int i=1; i<=info_substrings; i++) {
582: char *column=(char *)malloc(MAX_NUMBER);
583: snprintf(column, MAX_NUMBER, "%d", i);
584: columns+=NEW String(pool(), column); // .i column name
585: }
586: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 587: }
1.63 paf 588:
1.64 paf 589: int exec_option_bits=0;
1.63 paf 590: while(true) {
591: int exec_substrings=pcre_exec(code, 0,
592: subject, length, startoffset,
1.64 paf 593: exec_option_bits, ovector, ovecsize);
1.63 paf 594:
595: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 596: pcre_free(code);
597: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 598: return option_bits[1]!=0; // global=true+table, not global=false
599: }
600:
601: if(exec_substrings<0) {
1.100 parser 602: pcre_free(code);
1.112 parser 603: throw Exception(0, 0,
1.63 paf 604: aorigin,
1.76 paf 605: "regular expression execute error (%d)",
1.63 paf 606: exec_substrings);
607: }
608:
609: Array& row=*NEW Array(pool());
1.81 paf 610: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 611: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 612: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 613:
614: for(int i=1; i<exec_substrings; i++) {
1.69 paf 615: // -1:-1 case handled peacefully by mid() itself
616: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 617: }
618:
1.100 parser 619: row_action(**table, &row, startoffset, ovector[0], info);
1.63 paf 620:
1.100 parser 621: if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang
622: pcre_free(code);
623: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 624: return true;
625: }
1.100 parser 626: startoffset=ovector[1];
1.63 paf 627:
628: /*
629: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 630: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 631: */
632: }
1.82 parser 633: }
634:
635: String& String::change_case(Pool& pool, const unsigned char *tables,
636: Change_case_kind kind) const {
637: String& result=*new(pool) String(pool);
638:
639: const unsigned char *a;
640: const unsigned char *b;
641: switch(kind) {
642: case CC_UPPER:
643: a=tables+lcc_offset;
644: b=tables+fcc_offset;
645: break;
646: case CC_LOWER:
647: a=tables+lcc_offset;
648: b=0;
649: break;
650: default:
1.112 parser 651: throw Exception(0, 0,
1.82 parser 652: this,
653: "unknown change case kind #%d",
654: static_cast<int>(kind)); // never
655: a=b=0; // calm, compiler
656: break; // never
657: }
658:
659: const Chunk *chunk=&head;
660: do {
661: const Chunk::Row *row=chunk->rows;
1.116 paf 662: for(uint i=0; i<chunk->count; i++, row++) {
1.82 parser 663: if(row==append_here)
664: goto break2;
665:
1.114 paf 666: char *new_cstr=(char *)pool.malloc(row->item.size, 12);
1.82 parser 667: char *dest=new_cstr;
668: const char *src=row->item.ptr;
669: for(int size=row->item.size; size--; src++) {
670: unsigned char c=a[(unsigned char)*src];
671: if(b)
672: c=b[c];
673:
674: *dest++=(char)c;
675: }
676:
677: result.APPEND(new_cstr, row->item.size,
1.117 paf 678: (Untaint_lang)row->item.lang,
1.82 parser 679: row->item.origin.file, row->item.origin.line);
680: }
681: chunk=row->link;
682: } while(chunk);
683: break2:
1.89 parser 684:
1.101 parser 685: return result;
686: }
687:
1.108 parser 688: void String::join_chain(Pool& pool,
1.116 paf 689: uint& ai, const Chunk*& achunk, const Chunk::Row*& arow,
1.108 parser 690: Untaint_lang& joined_lang, const char *& joined_ptr, size_t& joined_size) const {
1.117 paf 691: joined_lang=(Untaint_lang)arow->item.lang;
1.108 parser 692:
693: // calc size
694: joined_size=0;
695: {
1.116 paf 696: uint start_i=ai;
1.108 parser 697: const Chunk::Row *start_row=arow;
698: const Chunk *chunk=achunk;
699: do {
700: const Chunk::Row *row=start_row;
1.116 paf 701: for(uint i=start_i; i<chunk->count; i++, row++) {
1.108 parser 702: if(row==append_here)
703: goto break21;
704:
705: if(row->item.lang==joined_lang)
706: joined_size+=row->item.size;
707: else
708: break;
709: }
710: if(chunk=row->link) {
711: start_i=0;
712: start_row=chunk->rows;
713: } else
714: break;
715: } while(true);
716: break21:;
717: }
718:
719: // if one row, return simply itself
720: if(joined_size==arow->item.size) {
721: joined_ptr=arow->item.ptr;
722: ai++; arow++;
723: if(ai==achunk->count)
724: achunk=arow->link;
725: } else {
726: // join adjacent rows
1.114 paf 727: char *ptr=(char *)pool.malloc(joined_size,13);
1.108 parser 728: joined_ptr=ptr;
1.116 paf 729: uint start_i=ai;
1.108 parser 730: const Chunk::Row *start_row=arow;
731: const Chunk *chunk=achunk;
1.116 paf 732: uint i;
1.108 parser 733: const Chunk::Row *row;
734: do {
735: row=start_row;
736: for(i=start_i; i<chunk->count; i++, row++) {
737: if(row==append_here)
738: goto break22;
739:
740: if(row->item.lang==joined_lang) {
741: memcpy(ptr, row->item.ptr, row->item.size);
742: ptr+=row->item.size;
743: } else
744: break;
745: }
746: if(chunk=row->link) {
747: start_i=0;
748: start_row=chunk->rows;
749: } else
750: break;
751: } while(true);
752: break22:;
753:
754: // return joined rows
755: ai=i;
756: arow=row;
757: achunk=chunk;
758: }
759: }
760:
761: String& String::reconstruct(Pool& pool) const {
762: //_asm int 3;
763: String& result=*new(pool) String(pool);
764: const Chunk *chunk=&head;
765: do {
766: const Chunk::Row *row=chunk->rows;
1.116 paf 767: for(uint i=0; i<chunk->count; ) {
1.108 parser 768: if(row==append_here)
769: goto break2;
770:
771: Untaint_lang joined_lang;
1.109 parser 772: const char *joined_ptr;
1.108 parser 773: size_t joined_size;
774: join_chain(pool, i, chunk, row,
775: joined_lang, joined_ptr, joined_size);
776:
777: result.APPEND(joined_ptr, joined_size,
778: joined_lang,
779: row->item.origin.file, row->item.origin.line);
780: if(!chunk)
781: goto break2;
782: }
783: } while(true);
784: break2:
785:
786: return result;
787: };
788:
789: String& String::replace_in_reconstructed(Pool& pool, Dictionary& dict) const {
1.106 parser 790: //_asm int 3;
1.101 parser 791: String& result=*new(pool) String(pool);
792: const Chunk *chunk=&head;
793: do {
794: const Chunk::Row *row=chunk->rows;
1.116 paf 795: for(uint i=0; i<chunk->count; i++, row++) {
1.101 parser 796: if(row==append_here)
797: goto break2;
798:
799: const char *src=row->item.ptr;
800: size_t src_size=row->item.size;
1.114 paf 801: char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14);
1.101 parser 802: char *dest=new_cstr;
803: while(src_size) {
804: // there is a row where first column starts 'src'
1.106 parser 805: if(Table::Item *item=dict.first_that_starts(src, src_size)) {
1.101 parser 806: // get a=>b values
807: const String& a=*static_cast<Array *>(item)->get_string(0);
808: const String& b=*static_cast<Array *>(item)->get_string(1);
1.105 parser 809: // skip 'a' in 'src' && reduce work size
810: src+=a.size(); src_size-=a.size();
811: // write 'b' to 'dest' && skip 'b' in 'dest'
812: b.store_to(dest); dest+=b.size();
1.101 parser 813: } else {
1.105 parser 814: // write a char to b && reduce work size
815: *dest++=*src++; src_size--;
1.101 parser 816: }
817: }
818:
819: result.APPEND(new_cstr, dest-new_cstr,
1.117 paf 820: (Untaint_lang)row->item.lang,
1.101 parser 821: row->item.origin.file, row->item.origin.line);
822: }
823: chunk=row->link;
824: } while(chunk);
825:
826: break2:
1.89 parser 827: return result;
1.108 parser 828: }
829:
830: String& String::replace(Pool& pool, Dictionary& dict) const {
831: return reconstruct(pool).replace_in_reconstructed(pool, dict);
1.89 parser 832: }
833:
1.90 parser 834: double String::as_double() const {
1.89 parser 835: double result;
1.114 paf 836: const char *cstr;
837: char buf[MAX_NUMBER];
1.115 paf 838: if(head.rows+1==append_here) {
1.114 paf 839: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
840: memcpy(buf, head.rows[0].item.ptr, size);
841: buf[size]=0;
842: cstr=buf;
843: } else
844: cstr=this->cstr();
1.102 parser 845: char *error_pos;
1.89 parser 846: // 0xABC
1.99 parser 847: if(cstr[0]=='0')
848: if(cstr[1]=='x' || cstr[1]=='X')
849: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
850: else
1.102 parser 851: result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos);
1.89 parser 852: else
1.99 parser 853: result=(double)strtod(cstr, &error_pos);
1.89 parser 854:
1.103 parser 855: if(*error_pos/*not EOS*/)
1.112 parser 856: throw Exception(0, 0,
1.89 parser 857: this,
858: "invalid number (double)");
859:
860: return result;
861: }
1.90 parser 862: int String::as_int() const {
1.89 parser 863: int result;
1.114 paf 864: const char *cstr;
865: char buf[MAX_NUMBER];
1.115 paf 866: if(head.rows+1==append_here) {
1.114 paf 867: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
868: memcpy(buf, head.rows[0].item.ptr, size);
869: buf[size]=0;
870: cstr=buf;
871: } else
872: cstr=this->cstr();
1.102 parser 873: char *error_pos;
1.89 parser 874: // 0xABC
1.99 parser 875: if(cstr[0]=='0')
876: if(cstr[1]=='x' || cstr[1]=='X')
877: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
878: else
1.102 parser 879: result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 880: else
881: result=(int)strtol(cstr, &error_pos, 0);
882:
1.103 parser 883: if(*error_pos/*not EOS*/)
1.112 parser 884: throw Exception(0, 0,
1.89 parser 885: this,
886: "invalid number (int)");
1.82 parser 887:
888: return result;
1.61 paf 889: }
1.113 parser 890:
891: /* @todo maybe network order worth spending some effort?
892: don't bothering myself with network byte order,
893: am not planning to be able to move resulting file across platforms
894: for now
895: */
896: void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const {
897: buf_size=
898: prolog_size
1.115 paf 899: +used_rows()*(sizeof(Untaint_lang)+sizeof(size_t))
1.113 parser 900: +size();
1.114 paf 901: buf=malloc(buf_size,15);
1.113 parser 902: char *cur=(char *)buf+prolog_size;
903:
904: const Chunk *chunk=&head;
905: do {
906: const Chunk::Row *row=chunk->rows;
1.116 paf 907: for(uint i=0; i<chunk->count; i++) {
1.113 parser 908: if(row==append_here)
909: goto break2;
910:
911: // lang
912: memcpy(cur, &row->item.lang, sizeof(Untaint_lang));
913: cur+=sizeof(Untaint_lang);
914: // size
915: memcpy(cur, &row->item.size, sizeof(size_t));
916: cur+=sizeof(size_t);
917: // bytes
918: memcpy(cur, row->item.ptr, row->item.size);
919: cur+=row->item.size;
920:
921: row++;
922: }
923: chunk=row->link;
924: } while(chunk);
925: break2:
926: ;
927: }
928:
929: /* @todo maybe network order worth spending some effort?
930: don't bothering myself with network byte order,
931: am not planning to be able to move resulting file across platforms
932: for now
933: */
934: #ifndef DOXYGEN
935: struct Serialized_piece {
936: String::Untaint_lang lang;
937: size_t size;
938: char ptr[1];
939: };
940: #endif
941:
942: void String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) {
943: char *cur=((char *)buf)+prolog_size;
944: buf_size-=prolog_size;
945:
946: while(buf_size) {
947: Serialized_piece& p=*(Serialized_piece *)cur;
948: APPEND(p.ptr, p.size, p.lang, file, 0);
949:
950: size_t piece_size=sizeof(p.lang)+sizeof(p.size)+p.size;
951: cur+=piece_size;
952: buf_size-=piece_size;
953: }
954: }
955:
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