Annotation of parser3/src/main/pa_string.C, revision 1.121
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:
1.121 ! paf 7: $Id: pa_string.C,v 1.120 2001/11/16 13:51:14 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;
1.121 ! paf 175: append_here->item.size=size;
1.52 paf 176: append_here->item.lang=lang;
1.13 paf 177: #ifndef NO_STRING_ORIGIN
1.14 paf 178: append_here->item.origin.file=file;
179: append_here->item.origin.line=line;
1.13 paf 180: #endif
1.115 paf 181: append_here++;
1.1 paf 182:
183: return *this;
1.97 parser 184: }
185:
186: char String::first_char() const {
1.120 paf 187: if(!used_rows())
1.112 parser 188: throw Exception(0, 0,
1.97 parser 189: this,
190: "getting first char of empty string");
191:
192: return *head.rows[0].item.ptr;
1.1 paf 193: }
194:
1.16 paf 195: uint String::hash_code() const {
1.7 paf 196: uint result=0;
1.5 paf 197:
1.16 paf 198: const Chunk *chunk=&head;
1.5 paf 199: do {
1.16 paf 200: const Chunk::Row *row=chunk->rows;
1.116 paf 201: for(uint i=0; i<chunk->count; i++) {
1.5 paf 202: if(row==append_here)
203: goto break2;
204:
1.6 paf 205: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 206: row++;
207: }
208: chunk=row->link;
209: } while(chunk);
210: break2:
211: return result;
212: }
213:
1.60 paf 214: /// @todo move 'lang' skipping to pos
215: int String::cmp(int& partial, const String& src,
216: size_t this_offset, Untaint_lang lang) const {
1.59 paf 217: partial=-1;
1.55 paf 218: this_offset=min(this_offset, size()-1);
219:
1.16 paf 220: const Chunk *a_chunk=&head;
221: const Chunk *b_chunk=&src.head;
222: const Chunk::Row *a_row=a_chunk->rows;
223: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 224: size_t a_offset=this_offset;
225: size_t b_offset=0;
1.9 paf 226: Chunk::Row *a_end=append_here;
227: Chunk::Row *b_end=src.append_here;
1.116 paf 228: uint a_countdown=a_chunk->count;
229: uint b_countdown=b_chunk->count;
230: int result;
1.60 paf 231: size_t pos=0;
1.33 paf 232:
1.83 parser 233: bool a_break=size()==0;
1.91 parser 234: bool b_break=src.size()==0;
1.83 parser 235: if(!(a_break || b_break)) while(true) {
1.55 paf 236: if(pos+a_row->item.size > this_offset) {
1.71 paf 237: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 238: return -1; // wrong lang -- bail out
239:
1.55 paf 240: int size_diff=
241: (a_row->item.size-a_offset)-
242: (b_row->item.size-b_offset);
243:
244: if(size_diff==0) { // a has same size as b
1.60 paf 245: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
246: a_row->item.size-a_offset);
1.55 paf 247: if(result)
248: return result;
1.60 paf 249: pos+=a_row->item.size;
1.55 paf 250: a_row++; a_countdown--; a_offset=0;
251: b_row++; b_countdown--; b_offset=0;
252: } else if (size_diff>0) { // a longer
1.60 paf 253: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
254: b_row->item.size-b_offset);
1.55 paf 255: if(result)
256: return result;
257: a_offset+=b_row->item.size-b_offset;
258: b_row++; b_countdown--; b_offset=0;
259: } else { // b longer
1.60 paf 260: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
261: a_row->item.size-a_offset);
1.55 paf 262: if(result)
263: return result;
264: b_offset+=a_row->item.size-a_offset;
1.60 paf 265: pos+=a_row->item.size;
1.55 paf 266: a_row++; a_countdown--; a_offset=0;
267: }
1.83 parser 268: if(b_break=b_row==b_end) {
269: a_break=a_row==a_end;
270: break;
271: }
1.55 paf 272: if(!b_countdown) {
273: b_chunk=b_row->link;
274: b_row=b_chunk->rows;
275: b_countdown=b_chunk->count;
276: }
277: } else {
1.60 paf 278: a_offset-=a_row->item.size;
279: pos+=a_row->item.size;
280: a_row++; a_countdown--;
1.9 paf 281: }
282:
1.83 parser 283: if(a_break=a_row==a_end) {
284: b_break=b_row==b_end;
285: break;
286: }
1.11 paf 287: if(!a_countdown) {
1.9 paf 288: a_chunk=a_row->link;
289: a_row=a_chunk->rows;
1.11 paf 290: a_countdown=a_chunk->count;
1.9 paf 291: }
1.27 paf 292: }
1.55 paf 293: if(a_break==b_break) { // ended simultaneously
294: partial=0; return 0;
295: } else if(a_break) { // first bytes equal, but a ended before b
296: partial=1; return -1;
297: } else {
298: partial=2; return +1;
299: }
1.27 paf 300: }
301:
1.60 paf 302: /// @todo move 'lang' skipping to pos
1.59 paf 303: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 304: size_t this_offset, Untaint_lang lang) const {
1.59 paf 305: partial=-1;
1.50 paf 306: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 307: this_offset=min(this_offset, size()-1);
1.27 paf 308:
309: const Chunk *a_chunk=&head;
310: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 311: size_t a_offset=this_offset;
1.55 paf 312: size_t b_offset=0;
1.27 paf 313: Chunk::Row *a_end=append_here;
1.116 paf 314: uint a_countdown=a_chunk->count;
1.60 paf 315: size_t pos=0;
1.52 paf 316:
1.83 parser 317: bool a_break=size()==0;
318: bool b_break=b_size==0;
319: if(!(a_break || b_break)) while(true) {
1.59 paf 320: if(pos+a_row->item.size > this_offset) {
1.71 paf 321: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 322: return -1; // wrong lang -- bail out
323:
1.59 paf 324: int size_diff=
325: (a_row->item.size-a_offset)-
326: (b_size-b_offset);
327:
328: if(size_diff==0) { // a has same size as b
1.116 paf 329: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 330: a_row->item.size-a_offset)!=0)
331: return result;
1.60 paf 332: pos+=a_row->item.size;
1.59 paf 333: a_row++; a_countdown--; a_offset=0;
334: b_break=true;
335: } else if (size_diff>0) { // a longer
1.116 paf 336: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 337: b_size-b_offset)!=0)
338: return result;
339: a_offset+=b_size-b_offset;
340: b_break=true;
341: } else { // b longer
1.116 paf 342: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 343: a_row->item.size-a_offset)!=0)
344: return result;
345: b_offset+=a_row->item.size-a_offset;
1.60 paf 346: pos+=a_row->item.size;
1.59 paf 347: a_row++; a_countdown--; a_offset=0;
348: }
349: } else {
1.60 paf 350: a_offset-=a_row->item.size;
351: pos+=a_row->item.size;
352: a_row++; a_countdown--;
1.27 paf 353: }
354:
1.86 parser 355: a_break=a_row==a_end;
356: if(a_break || b_break)
1.83 parser 357: break;
1.27 paf 358: if(!a_countdown) {
359: a_chunk=a_row->link;
360: a_row=a_chunk->rows;
361: a_countdown=a_chunk->count;
1.9 paf 362: }
363: }
1.55 paf 364: if(a_break==b_break) { // ended simultaneously
365: partial=0; return 0;
366: } else if(a_break) { // first bytes equal, but a ended before b
367: partial=1; return -1;
368: } else {
369: partial=2; return +1;
370: }
1.5 paf 371: }
1.46 paf 372:
373: #ifndef NO_STRING_ORIGIN
374: const Origin& String::origin() const {
1.120 paf 375: if(!used_rows()) {
1.96 parser 376: static const Origin empty_origin={"empty string"};
377: return empty_origin;
378: }
1.46 paf 379:
1.49 paf 380: // determining origin by last appended piece
1.50 paf 381: // because first one frequently constant.
382: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 383: // when last peice is constant,
384: // ex: parser_root_auto_path{dynamic} / auto.p{const}
385: // using first piece
386: Origin& last_origin=append_here[-1].item.origin;
387: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 388: }
389: #endif
1.53 paf 390:
1.69 paf 391: String& String::mid(size_t start, size_t finish) const {
1.107 parser 392: String& result=*NEW String(pool());
393:
1.53 paf 394: start=max(0, start);
1.111 parser 395: finish=min(size(), finish);
1.60 paf 396: if(start==finish)
1.107 parser 397: return result;
1.53 paf 398:
399: size_t pos=0;
400: const Chunk *chunk=&head;
401: do {
402: const Chunk::Row *row=chunk->rows;
1.116 paf 403: for(uint i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 404: if(row==append_here)
405: goto break2;
406:
1.60 paf 407: size_t item_finish=pos+row->item.size;
408: if(item_finish > start) { // started now or already?
409: bool started=result.size()==0; // started now?
410: bool finished=finish <= item_finish; // finished now?
1.53 paf 411: size_t offset=started?start-pos:0;
412: size_t size=finished?finish-pos:row->item.size;
413: result.APPEND(
414: row->item.ptr+offset, size-offset,
1.117 paf 415: (Untaint_lang)row->item.lang,
1.53 paf 416: row->item.origin.file, row->item.origin.line);
417: if(finished)
418: goto break2;
419: }
420: }
421: chunk=row->link;
422: } while(chunk);
423: break2:
1.60 paf 424: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
425: //cstr(), start, finish, result.cstr());
1.53 paf 426: return result;
1.54 paf 427: }
428:
1.60 paf 429: int String::pos(const String& substr,
1.116 paf 430: int result, Untaint_lang lang) const {
1.58 paf 431: for(; result<size(); result++) {
1.60 paf 432: int partial; cmp(partial, substr, result, lang);
1.58 paf 433: if(
434: partial==0 || // full match
435: partial==2) // 'substr' starts 'this'+'result'
436: return result;
437: }
438:
439: return -1;
440: }
441:
1.60 paf 442: int String::pos(const char *substr, size_t substr_size,
1.116 paf 443: int result, Untaint_lang lang) const {
1.57 paf 444: for(; result<size(); result++) {
1.60 paf 445: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 446: if(
447: partial==0 || // full match
448: partial==2) // 'substr' starts 'this'+'result'
449: return result;
450: }
451:
452: return -1;
1.60 paf 453: }
454:
455: void String::split(Array& result,
456: size_t* pos_after_ref,
457: const char *delim, size_t delim_size,
458: Untaint_lang lang, int limit) const {
459: if(delim_size) {
460: size_t pos_after=pos_after_ref?*pos_after_ref:0;
461: int pos_before;
462: // while we have 'delim'...
463: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 464: result+=&mid(pos_after, pos_before);
1.60 paf 465: pos_after=pos_before+delim_size;
466: }
467: // last piece
468: if(pos_after<size() && limit) {
1.69 paf 469: result+=&mid(pos_after, size());
1.60 paf 470: pos_after=size();
471: }
472: if(pos_after_ref)
473: *pos_after_ref=pos_after;
474: } else { // empty delim
475: result+=this;
476: if(pos_after_ref)
477: *pos_after_ref+=size();
478: }
479: }
480:
481: void String::split(Array& result,
482: size_t* pos_after_ref,
483: const String& delim, Untaint_lang lang,
484: int limit) const {
485: if(delim.size()) {
486: size_t pos_after=pos_after_ref?*pos_after_ref:0;
487: int pos_before;
488: // while we have 'delim'...
489: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 490: result+=&mid(pos_after, pos_before);
1.60 paf 491: pos_after=pos_before+delim.size();
492: }
493: // last piece
494: if(pos_after<size() && limit) {
1.69 paf 495: result+=&mid(pos_after, size());
1.60 paf 496: pos_after=size();
497: }
498: if(pos_after_ref)
499: *pos_after_ref=pos_after;
500: } else { // empty delim
501: result+=this;
502: if(pos_after_ref)
503: *pos_after_ref+=size();
504: }
1.61 paf 505: }
506:
1.63 paf 507: static void regex_options(char *options, int *result){
508: struct Regex_option {
509: char key;
510: int clear, set;
511: int *result;
512: } regex_option[]={
513: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 514: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 515: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
516: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
517: {'g', 0, true, result+1}, // many rows
518: {0},
519: };
520: result[0]=PCRE_EXTRA | PCRE_DOTALL;
521: result[1]=0;
522:
523: if(options)
524: for(Regex_option *o=regex_option; o->key; o++)
525: if(
526: strchr(options, o->key) ||
527: strchr(options, toupper(o->key))) {
528: *(o->result)&=~o->clear;
529: *(o->result)|=o->set;
530: }
531: }
532:
1.88 parser 533: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.77 paf 534: bool String::match(const unsigned char *pcre_tables,
535: const String *aorigin,
1.62 paf 536: const String& regexp,
1.63 paf 537: const String *options,
1.64 paf 538: Table **table,
1.95 parser 539: Row_action row_action, void *info,
540: bool *was_global) const {
1.64 paf 541:
1.73 paf 542: if(!regexp.size())
1.112 parser 543: throw Exception(0, 0,
1.73 paf 544: aorigin,
545: "regexp is empty");
1.118 paf 546: const char *pattern=regexp.cstr();
1.62 paf 547: const char *errptr;
548: int erroffset;
1.63 paf 549: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
1.95 parser 550: if(was_global)
551: *was_global=option_bits[1]!=0;
1.63 paf 552: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 553: &errptr, &erroffset,
1.74 paf 554: pcre_tables);
1.62 paf 555:
1.67 paf 556: if(!code)
1.112 parser 557: throw Exception(0, 0,
1.69 paf 558: ®exp.mid(erroffset, regexp.size()),
1.74 paf 559: "regular expression syntax error - %s", errptr);
1.62 paf 560:
1.63 paf 561: int info_substrings=pcre_info(code, 0, 0);
562: if(info_substrings<0) {
1.100 parser 563: pcre_free(code);
1.112 parser 564: throw Exception(0, 0,
1.73 paf 565: aorigin,
1.76 paf 566: "pcre_info error (%d)",
1.73 paf 567: info_substrings);
1.63 paf 568: }
569:
570: int startoffset=0;
1.118 paf 571: const char *subject=cstr();
1.62 paf 572: int length=strlen(subject);
1.63 paf 573: int ovecsize;
574: int *ovector=(int *)malloc(sizeof(int)*
1.114 paf 575: (ovecsize=(1/*match*/+info_substrings)*3), 11);
1.62 paf 576:
1.64 paf 577: { // create table
578: Array& columns=*NEW Array(pool());
579: columns+=string_pre_match_name;
580: columns+=string_match_name;
581: columns+=string_post_match_name;
582: for(int i=1; i<=info_substrings; i++) {
583: char *column=(char *)malloc(MAX_NUMBER);
584: snprintf(column, MAX_NUMBER, "%d", i);
585: columns+=NEW String(pool(), column); // .i column name
586: }
587: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 588: }
1.63 paf 589:
1.64 paf 590: int exec_option_bits=0;
1.63 paf 591: while(true) {
592: int exec_substrings=pcre_exec(code, 0,
593: subject, length, startoffset,
1.64 paf 594: exec_option_bits, ovector, ovecsize);
1.63 paf 595:
596: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 597: pcre_free(code);
598: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 599: return option_bits[1]!=0; // global=true+table, not global=false
600: }
601:
602: if(exec_substrings<0) {
1.100 parser 603: pcre_free(code);
1.112 parser 604: throw Exception(0, 0,
1.63 paf 605: aorigin,
1.76 paf 606: "regular expression execute error (%d)",
1.63 paf 607: exec_substrings);
608: }
609:
610: Array& row=*NEW Array(pool());
1.81 paf 611: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 612: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 613: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 614:
615: for(int i=1; i<exec_substrings; i++) {
1.69 paf 616: // -1:-1 case handled peacefully by mid() itself
617: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 618: }
619:
1.100 parser 620: row_action(**table, &row, startoffset, ovector[0], info);
1.63 paf 621:
1.100 parser 622: if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang
623: pcre_free(code);
624: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 625: return true;
626: }
1.100 parser 627: startoffset=ovector[1];
1.63 paf 628:
629: /*
630: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 631: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 632: */
633: }
1.82 parser 634: }
635:
636: String& String::change_case(Pool& pool, const unsigned char *tables,
637: Change_case_kind kind) const {
638: String& result=*new(pool) String(pool);
639:
640: const unsigned char *a;
641: const unsigned char *b;
642: switch(kind) {
643: case CC_UPPER:
644: a=tables+lcc_offset;
645: b=tables+fcc_offset;
646: break;
647: case CC_LOWER:
648: a=tables+lcc_offset;
649: b=0;
650: break;
651: default:
1.112 parser 652: throw Exception(0, 0,
1.82 parser 653: this,
654: "unknown change case kind #%d",
655: static_cast<int>(kind)); // never
656: a=b=0; // calm, compiler
657: break; // never
658: }
659:
660: const Chunk *chunk=&head;
661: do {
662: const Chunk::Row *row=chunk->rows;
1.116 paf 663: for(uint i=0; i<chunk->count; i++, row++) {
1.82 parser 664: if(row==append_here)
665: goto break2;
666:
1.114 paf 667: char *new_cstr=(char *)pool.malloc(row->item.size, 12);
1.82 parser 668: char *dest=new_cstr;
669: const char *src=row->item.ptr;
670: for(int size=row->item.size; size--; src++) {
671: unsigned char c=a[(unsigned char)*src];
672: if(b)
673: c=b[c];
674:
675: *dest++=(char)c;
676: }
677:
678: result.APPEND(new_cstr, row->item.size,
1.117 paf 679: (Untaint_lang)row->item.lang,
1.82 parser 680: row->item.origin.file, row->item.origin.line);
681: }
682: chunk=row->link;
683: } while(chunk);
684: break2:
1.89 parser 685:
1.101 parser 686: return result;
687: }
688:
1.108 parser 689: void String::join_chain(Pool& pool,
1.116 paf 690: uint& ai, const Chunk*& achunk, const Chunk::Row*& arow,
1.108 parser 691: Untaint_lang& joined_lang, const char *& joined_ptr, size_t& joined_size) const {
1.117 paf 692: joined_lang=(Untaint_lang)arow->item.lang;
1.108 parser 693:
694: // calc size
695: joined_size=0;
696: {
1.116 paf 697: uint start_i=ai;
1.108 parser 698: const Chunk::Row *start_row=arow;
699: const Chunk *chunk=achunk;
700: do {
701: const Chunk::Row *row=start_row;
1.116 paf 702: for(uint i=start_i; i<chunk->count; i++, row++) {
1.108 parser 703: if(row==append_here)
704: goto break21;
705:
706: if(row->item.lang==joined_lang)
707: joined_size+=row->item.size;
708: else
709: break;
710: }
711: if(chunk=row->link) {
712: start_i=0;
713: start_row=chunk->rows;
714: } else
715: break;
716: } while(true);
717: break21:;
718: }
719:
720: // if one row, return simply itself
721: if(joined_size==arow->item.size) {
722: joined_ptr=arow->item.ptr;
723: ai++; arow++;
724: if(ai==achunk->count)
725: achunk=arow->link;
726: } else {
727: // join adjacent rows
1.114 paf 728: char *ptr=(char *)pool.malloc(joined_size,13);
1.108 parser 729: joined_ptr=ptr;
1.116 paf 730: uint start_i=ai;
1.108 parser 731: const Chunk::Row *start_row=arow;
732: const Chunk *chunk=achunk;
1.116 paf 733: uint i;
1.108 parser 734: const Chunk::Row *row;
735: do {
736: row=start_row;
737: for(i=start_i; i<chunk->count; i++, row++) {
738: if(row==append_here)
739: goto break22;
740:
741: if(row->item.lang==joined_lang) {
742: memcpy(ptr, row->item.ptr, row->item.size);
743: ptr+=row->item.size;
744: } else
745: break;
746: }
747: if(chunk=row->link) {
748: start_i=0;
749: start_row=chunk->rows;
750: } else
751: break;
752: } while(true);
753: break22:;
754:
755: // return joined rows
756: ai=i;
757: arow=row;
758: achunk=chunk;
759: }
760: }
761:
762: String& String::reconstruct(Pool& pool) const {
763: //_asm int 3;
764: String& result=*new(pool) String(pool);
765: const Chunk *chunk=&head;
766: do {
767: const Chunk::Row *row=chunk->rows;
1.116 paf 768: for(uint i=0; i<chunk->count; ) {
1.108 parser 769: if(row==append_here)
770: goto break2;
771:
772: Untaint_lang joined_lang;
1.109 parser 773: const char *joined_ptr;
1.108 parser 774: size_t joined_size;
775: join_chain(pool, i, chunk, row,
776: joined_lang, joined_ptr, joined_size);
777:
778: result.APPEND(joined_ptr, joined_size,
779: joined_lang,
780: row->item.origin.file, row->item.origin.line);
781: if(!chunk)
782: goto break2;
783: }
784: } while(true);
785: break2:
786:
787: return result;
788: };
789:
790: String& String::replace_in_reconstructed(Pool& pool, Dictionary& dict) const {
1.106 parser 791: //_asm int 3;
1.101 parser 792: String& result=*new(pool) String(pool);
793: const Chunk *chunk=&head;
794: do {
795: const Chunk::Row *row=chunk->rows;
1.116 paf 796: for(uint i=0; i<chunk->count; i++, row++) {
1.101 parser 797: if(row==append_here)
798: goto break2;
799:
800: const char *src=row->item.ptr;
801: size_t src_size=row->item.size;
1.114 paf 802: char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14);
1.101 parser 803: char *dest=new_cstr;
804: while(src_size) {
805: // there is a row where first column starts 'src'
1.106 parser 806: if(Table::Item *item=dict.first_that_starts(src, src_size)) {
1.101 parser 807: // get a=>b values
808: const String& a=*static_cast<Array *>(item)->get_string(0);
809: const String& b=*static_cast<Array *>(item)->get_string(1);
1.105 parser 810: // skip 'a' in 'src' && reduce work size
811: src+=a.size(); src_size-=a.size();
812: // write 'b' to 'dest' && skip 'b' in 'dest'
813: b.store_to(dest); dest+=b.size();
1.101 parser 814: } else {
1.105 parser 815: // write a char to b && reduce work size
816: *dest++=*src++; src_size--;
1.101 parser 817: }
818: }
819:
820: result.APPEND(new_cstr, dest-new_cstr,
1.117 paf 821: (Untaint_lang)row->item.lang,
1.101 parser 822: row->item.origin.file, row->item.origin.line);
823: }
824: chunk=row->link;
825: } while(chunk);
826:
827: break2:
1.89 parser 828: return result;
1.108 parser 829: }
830:
831: String& String::replace(Pool& pool, Dictionary& dict) const {
832: return reconstruct(pool).replace_in_reconstructed(pool, dict);
1.89 parser 833: }
834:
1.90 parser 835: double String::as_double() const {
1.89 parser 836: double result;
1.114 paf 837: const char *cstr;
838: char buf[MAX_NUMBER];
1.115 paf 839: if(head.rows+1==append_here) {
1.114 paf 840: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
841: memcpy(buf, head.rows[0].item.ptr, size);
842: buf[size]=0;
843: cstr=buf;
844: } else
845: cstr=this->cstr();
1.102 parser 846: char *error_pos;
1.89 parser 847: // 0xABC
1.99 parser 848: if(cstr[0]=='0')
849: if(cstr[1]=='x' || cstr[1]=='X')
850: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
851: else
1.102 parser 852: result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos);
1.89 parser 853: else
1.99 parser 854: result=(double)strtod(cstr, &error_pos);
1.89 parser 855:
1.103 parser 856: if(*error_pos/*not EOS*/)
1.112 parser 857: throw Exception(0, 0,
1.89 parser 858: this,
859: "invalid number (double)");
860:
861: return result;
862: }
1.90 parser 863: int String::as_int() const {
1.89 parser 864: int result;
1.114 paf 865: const char *cstr;
866: char buf[MAX_NUMBER];
1.115 paf 867: if(head.rows+1==append_here) {
1.114 paf 868: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
869: memcpy(buf, head.rows[0].item.ptr, size);
870: buf[size]=0;
871: cstr=buf;
872: } else
873: cstr=this->cstr();
1.102 parser 874: char *error_pos;
1.89 parser 875: // 0xABC
1.99 parser 876: if(cstr[0]=='0')
877: if(cstr[1]=='x' || cstr[1]=='X')
878: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
879: else
1.102 parser 880: result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 881: else
882: result=(int)strtol(cstr, &error_pos, 0);
883:
1.103 parser 884: if(*error_pos/*not EOS*/)
1.112 parser 885: throw Exception(0, 0,
1.89 parser 886: this,
887: "invalid number (int)");
1.82 parser 888:
889: return result;
1.61 paf 890: }
1.113 parser 891:
892: /* @todo maybe network order worth spending some effort?
893: don't bothering myself with network byte order,
894: am not planning to be able to move resulting file across platforms
895: for now
896: */
897: void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const {
898: buf_size=
899: prolog_size
1.115 paf 900: +used_rows()*(sizeof(Untaint_lang)+sizeof(size_t))
1.113 parser 901: +size();
1.114 paf 902: buf=malloc(buf_size,15);
1.113 parser 903: char *cur=(char *)buf+prolog_size;
904:
905: const Chunk *chunk=&head;
906: do {
907: const Chunk::Row *row=chunk->rows;
1.116 paf 908: for(uint i=0; i<chunk->count; i++) {
1.113 parser 909: if(row==append_here)
910: goto break2;
911:
912: // lang
913: memcpy(cur, &row->item.lang, sizeof(Untaint_lang));
914: cur+=sizeof(Untaint_lang);
915: // size
916: memcpy(cur, &row->item.size, sizeof(size_t));
917: cur+=sizeof(size_t);
918: // bytes
919: memcpy(cur, row->item.ptr, row->item.size);
920: cur+=row->item.size;
921:
922: row++;
923: }
924: chunk=row->link;
925: } while(chunk);
926: break2:
927: ;
928: }
929:
930: /* @todo maybe network order worth spending some effort?
931: don't bothering myself with network byte order,
932: am not planning to be able to move resulting file across platforms
933: for now
934: */
935: #ifndef DOXYGEN
936: struct Serialized_piece {
937: String::Untaint_lang lang;
938: size_t size;
939: char ptr[1];
940: };
941: #endif
942:
943: void String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) {
944: char *cur=((char *)buf)+prolog_size;
945: buf_size-=prolog_size;
946:
947: while(buf_size) {
948: Serialized_piece& p=*(Serialized_piece *)cur;
949: APPEND(p.ptr, p.size, p.lang, file, 0);
950:
951: size_t piece_size=sizeof(p.lang)+sizeof(p.size)+p.size;
952: cur+=piece_size;
953: buf_size-=piece_size;
954: }
955: }
956:
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