Annotation of parser3/src/main/pa_string.C, revision 1.123
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:
1.123 ! paf 7: $Id: pa_string.C,v 1.122 2001/11/16 14:25:02 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;
1.123 ! paf 30: head_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;
1.123 ! paf 69: head_link=last_chunk;
1.98 parser 70: append_here=link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 71:
1.123 ! paf 72: Chunk *old_chunk=src.head_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;
1.123 ! paf 100: STRING_FOREACH_ROW(
1.120 paf 101: result+=row->item.size;
1.123 ! paf 102: );
1.120 paf 103: break2:
104: return result;
1.94 parser 105: }
106:
1.115 paf 107: /// @todo not very optimal
108: uint String::used_rows() const {
109: uint result=0;
1.123 ! paf 110: STRING_FOREACH_ROW(
! 111: result++;
! 112: );
1.115 paf 113: break2:
114: return result;
115: }
1.94 parser 116: void String::expand() {
1.116 paf 117: uint new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.122 paf 118: if(new_chunk_count>MAX_USHORT)
119: new_chunk_count=MAX_USHORT;
120:
1.94 parser 121: last_chunk=static_cast<Chunk *>(
1.116 paf 122: malloc(sizeof(uint)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *), 10));
1.94 parser 123: last_chunk->count=new_chunk_count;
124: link_row->link=last_chunk;
125: append_here=last_chunk->rows;
126: link_row=&last_chunk->rows[last_chunk->count];
127: link_row->link=0;
1.5 paf 128: }
1.28 paf 129:
1.13 paf 130: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 131: if(!src)
132: return *this;
1.26 paf 133: if(!size)
134: size=strlen(src);
135: if(!size)
1.9 paf 136: return *this;
1.122 paf 137:
138: while(size>MAX_USHORT) {
139: if(chunk_is_full())
140: expand();
141:
142: append_here->item.ptr=src;
143: append_here->item.size=MAX_USHORT;
144: append_here->item.lang=lang;
145: #ifndef NO_STRING_ORIGIN
146: append_here->item.origin.file=file;
147: append_here->item.origin.line=line;
148: #endif
149: append_here++;
150:
151: src+=MAX_USHORT;
152: size-=MAX_USHORT;
153: }
1.9 paf 154:
1.1 paf 155: if(chunk_is_full())
156: expand();
157:
158: append_here->item.ptr=src;
1.121 paf 159: append_here->item.size=size;
1.52 paf 160: append_here->item.lang=lang;
1.13 paf 161: #ifndef NO_STRING_ORIGIN
1.14 paf 162: append_here->item.origin.file=file;
163: append_here->item.origin.line=line;
1.13 paf 164: #endif
1.115 paf 165: append_here++;
1.1 paf 166:
167: return *this;
1.97 parser 168: }
169:
170: char String::first_char() const {
1.120 paf 171: if(!used_rows())
1.112 parser 172: throw Exception(0, 0,
1.97 parser 173: this,
174: "getting first char of empty string");
175:
176: return *head.rows[0].item.ptr;
1.1 paf 177: }
178:
1.16 paf 179: uint String::hash_code() const {
1.7 paf 180: uint result=0;
1.123 ! paf 181: STRING_FOREACH_ROW(
1.6 paf 182: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.123 ! paf 183: );
1.5 paf 184: break2:
185: return result;
186: }
187:
1.60 paf 188: /// @todo move 'lang' skipping to pos
189: int String::cmp(int& partial, const String& src,
190: size_t this_offset, Untaint_lang lang) const {
1.59 paf 191: partial=-1;
1.55 paf 192: this_offset=min(this_offset, size()-1);
193:
1.16 paf 194: const Chunk *a_chunk=&head;
195: const Chunk *b_chunk=&src.head;
196: const Chunk::Row *a_row=a_chunk->rows;
197: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 198: size_t a_offset=this_offset;
199: size_t b_offset=0;
1.9 paf 200: Chunk::Row *a_end=append_here;
201: Chunk::Row *b_end=src.append_here;
1.116 paf 202: uint a_countdown=a_chunk->count;
203: uint b_countdown=b_chunk->count;
204: int result;
1.60 paf 205: size_t pos=0;
1.33 paf 206:
1.83 parser 207: bool a_break=size()==0;
1.91 parser 208: bool b_break=src.size()==0;
1.83 parser 209: if(!(a_break || b_break)) while(true) {
1.55 paf 210: if(pos+a_row->item.size > this_offset) {
1.71 paf 211: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 212: return -1; // wrong lang -- bail out
213:
1.55 paf 214: int size_diff=
215: (a_row->item.size-a_offset)-
216: (b_row->item.size-b_offset);
217:
218: if(size_diff==0) { // a has same size as b
1.60 paf 219: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
220: a_row->item.size-a_offset);
1.55 paf 221: if(result)
222: return result;
1.60 paf 223: pos+=a_row->item.size;
1.55 paf 224: a_row++; a_countdown--; a_offset=0;
225: b_row++; b_countdown--; b_offset=0;
226: } else if (size_diff>0) { // a longer
1.60 paf 227: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
228: b_row->item.size-b_offset);
1.55 paf 229: if(result)
230: return result;
231: a_offset+=b_row->item.size-b_offset;
232: b_row++; b_countdown--; b_offset=0;
233: } else { // b longer
1.60 paf 234: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
235: a_row->item.size-a_offset);
1.55 paf 236: if(result)
237: return result;
238: b_offset+=a_row->item.size-a_offset;
1.60 paf 239: pos+=a_row->item.size;
1.55 paf 240: a_row++; a_countdown--; a_offset=0;
241: }
1.83 parser 242: if(b_break=b_row==b_end) {
243: a_break=a_row==a_end;
244: break;
245: }
1.55 paf 246: if(!b_countdown) {
247: b_chunk=b_row->link;
248: b_row=b_chunk->rows;
249: b_countdown=b_chunk->count;
250: }
251: } else {
1.60 paf 252: a_offset-=a_row->item.size;
253: pos+=a_row->item.size;
254: a_row++; a_countdown--;
1.9 paf 255: }
256:
1.83 parser 257: if(a_break=a_row==a_end) {
258: b_break=b_row==b_end;
259: break;
260: }
1.11 paf 261: if(!a_countdown) {
1.9 paf 262: a_chunk=a_row->link;
263: a_row=a_chunk->rows;
1.11 paf 264: a_countdown=a_chunk->count;
1.9 paf 265: }
1.27 paf 266: }
1.55 paf 267: if(a_break==b_break) { // ended simultaneously
268: partial=0; return 0;
269: } else if(a_break) { // first bytes equal, but a ended before b
270: partial=1; return -1;
271: } else {
272: partial=2; return +1;
273: }
1.27 paf 274: }
275:
1.60 paf 276: /// @todo move 'lang' skipping to pos
1.59 paf 277: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 278: size_t this_offset, Untaint_lang lang) const {
1.59 paf 279: partial=-1;
1.50 paf 280: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 281: this_offset=min(this_offset, size()-1);
1.27 paf 282:
283: const Chunk *a_chunk=&head;
284: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 285: size_t a_offset=this_offset;
1.55 paf 286: size_t b_offset=0;
1.27 paf 287: Chunk::Row *a_end=append_here;
1.116 paf 288: uint a_countdown=a_chunk->count;
1.60 paf 289: size_t pos=0;
1.52 paf 290:
1.83 parser 291: bool a_break=size()==0;
292: bool b_break=b_size==0;
293: if(!(a_break || b_break)) while(true) {
1.59 paf 294: if(pos+a_row->item.size > this_offset) {
1.71 paf 295: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 296: return -1; // wrong lang -- bail out
297:
1.59 paf 298: int size_diff=
299: (a_row->item.size-a_offset)-
300: (b_size-b_offset);
301:
302: if(size_diff==0) { // a has same size as b
1.116 paf 303: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 304: a_row->item.size-a_offset)!=0)
305: return result;
1.60 paf 306: pos+=a_row->item.size;
1.59 paf 307: a_row++; a_countdown--; a_offset=0;
308: b_break=true;
309: } else if (size_diff>0) { // a longer
1.116 paf 310: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 311: b_size-b_offset)!=0)
312: return result;
313: a_offset+=b_size-b_offset;
314: b_break=true;
315: } else { // b longer
1.116 paf 316: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
1.59 paf 317: a_row->item.size-a_offset)!=0)
318: return result;
319: b_offset+=a_row->item.size-a_offset;
1.60 paf 320: pos+=a_row->item.size;
1.59 paf 321: a_row++; a_countdown--; a_offset=0;
322: }
323: } else {
1.60 paf 324: a_offset-=a_row->item.size;
325: pos+=a_row->item.size;
326: a_row++; a_countdown--;
1.27 paf 327: }
328:
1.86 parser 329: a_break=a_row==a_end;
330: if(a_break || b_break)
1.83 parser 331: break;
1.27 paf 332: if(!a_countdown) {
333: a_chunk=a_row->link;
334: a_row=a_chunk->rows;
335: a_countdown=a_chunk->count;
1.9 paf 336: }
337: }
1.55 paf 338: if(a_break==b_break) { // ended simultaneously
339: partial=0; return 0;
340: } else if(a_break) { // first bytes equal, but a ended before b
341: partial=1; return -1;
342: } else {
343: partial=2; return +1;
344: }
1.5 paf 345: }
1.46 paf 346:
347: #ifndef NO_STRING_ORIGIN
348: const Origin& String::origin() const {
1.120 paf 349: if(!used_rows()) {
1.96 parser 350: static const Origin empty_origin={"empty string"};
351: return empty_origin;
352: }
1.46 paf 353:
1.49 paf 354: // determining origin by last appended piece
1.50 paf 355: // because first one frequently constant.
356: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 357: // when last peice is constant,
358: // ex: parser_root_auto_path{dynamic} / auto.p{const}
359: // using first piece
360: Origin& last_origin=append_here[-1].item.origin;
361: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 362: }
363: #endif
1.53 paf 364:
1.69 paf 365: String& String::mid(size_t start, size_t finish) const {
1.107 parser 366: String& result=*NEW String(pool());
367:
1.53 paf 368: start=max(0, start);
1.111 parser 369: finish=min(size(), finish);
1.60 paf 370: if(start==finish)
1.107 parser 371: return result;
1.53 paf 372:
373: size_t pos=0;
1.123 ! paf 374: STRING_FOREACH_ROW(
! 375: size_t item_finish=pos+row->item.size;
! 376: if(item_finish > start) { // started now or already?
! 377: bool started=result.size()==0; // started now?
! 378: bool finished=finish <= item_finish; // finished now?
! 379: size_t offset=started?start-pos:0;
! 380: size_t size=finished?finish-pos:row->item.size;
! 381: result.APPEND(
! 382: row->item.ptr+offset, size-offset,
! 383: row->item.lang,
! 384: row->item.origin.file, row->item.origin.line);
! 385: if(finished)
1.53 paf 386: goto break2;
387: }
1.123 ! paf 388: pos+=row->item.size;
! 389: );
1.53 paf 390: break2:
1.60 paf 391: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
392: //cstr(), start, finish, result.cstr());
1.53 paf 393: return result;
1.54 paf 394: }
395:
1.60 paf 396: int String::pos(const String& substr,
1.116 paf 397: int result, Untaint_lang lang) const {
1.58 paf 398: for(; result<size(); result++) {
1.60 paf 399: int partial; cmp(partial, substr, result, lang);
1.58 paf 400: if(
401: partial==0 || // full match
402: partial==2) // 'substr' starts 'this'+'result'
403: return result;
404: }
405:
406: return -1;
407: }
408:
1.60 paf 409: int String::pos(const char *substr, size_t substr_size,
1.116 paf 410: int result, Untaint_lang lang) const {
1.57 paf 411: for(; result<size(); result++) {
1.60 paf 412: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 413: if(
414: partial==0 || // full match
415: partial==2) // 'substr' starts 'this'+'result'
416: return result;
417: }
418:
419: return -1;
1.60 paf 420: }
421:
422: void String::split(Array& result,
423: size_t* pos_after_ref,
424: const char *delim, size_t delim_size,
425: Untaint_lang lang, int limit) const {
426: if(delim_size) {
427: size_t pos_after=pos_after_ref?*pos_after_ref:0;
428: int pos_before;
429: // while we have 'delim'...
430: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 431: result+=&mid(pos_after, pos_before);
1.60 paf 432: pos_after=pos_before+delim_size;
433: }
434: // last piece
435: if(pos_after<size() && limit) {
1.69 paf 436: result+=&mid(pos_after, size());
1.60 paf 437: pos_after=size();
438: }
439: if(pos_after_ref)
440: *pos_after_ref=pos_after;
441: } else { // empty delim
442: result+=this;
443: if(pos_after_ref)
444: *pos_after_ref+=size();
445: }
446: }
447:
448: void String::split(Array& result,
449: size_t* pos_after_ref,
450: const String& delim, Untaint_lang lang,
451: int limit) const {
452: if(delim.size()) {
453: size_t pos_after=pos_after_ref?*pos_after_ref:0;
454: int pos_before;
455: // while we have 'delim'...
456: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 457: result+=&mid(pos_after, pos_before);
1.60 paf 458: pos_after=pos_before+delim.size();
459: }
460: // last piece
461: if(pos_after<size() && limit) {
1.69 paf 462: result+=&mid(pos_after, size());
1.60 paf 463: pos_after=size();
464: }
465: if(pos_after_ref)
466: *pos_after_ref=pos_after;
467: } else { // empty delim
468: result+=this;
469: if(pos_after_ref)
470: *pos_after_ref+=size();
471: }
1.61 paf 472: }
473:
1.63 paf 474: static void regex_options(char *options, int *result){
475: struct Regex_option {
476: char key;
477: int clear, set;
478: int *result;
479: } regex_option[]={
480: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 481: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 482: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
483: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
484: {'g', 0, true, result+1}, // many rows
485: {0},
486: };
487: result[0]=PCRE_EXTRA | PCRE_DOTALL;
488: result[1]=0;
489:
490: if(options)
491: for(Regex_option *o=regex_option; o->key; o++)
492: if(
493: strchr(options, o->key) ||
494: strchr(options, toupper(o->key))) {
495: *(o->result)&=~o->clear;
496: *(o->result)|=o->set;
497: }
498: }
499:
1.88 parser 500: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.77 paf 501: bool String::match(const unsigned char *pcre_tables,
502: const String *aorigin,
1.62 paf 503: const String& regexp,
1.63 paf 504: const String *options,
1.64 paf 505: Table **table,
1.95 parser 506: Row_action row_action, void *info,
507: bool *was_global) const {
1.64 paf 508:
1.73 paf 509: if(!regexp.size())
1.112 parser 510: throw Exception(0, 0,
1.73 paf 511: aorigin,
512: "regexp is empty");
1.118 paf 513: const char *pattern=regexp.cstr();
1.62 paf 514: const char *errptr;
515: int erroffset;
1.63 paf 516: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
1.95 parser 517: if(was_global)
518: *was_global=option_bits[1]!=0;
1.63 paf 519: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 520: &errptr, &erroffset,
1.74 paf 521: pcre_tables);
1.62 paf 522:
1.67 paf 523: if(!code)
1.112 parser 524: throw Exception(0, 0,
1.69 paf 525: ®exp.mid(erroffset, regexp.size()),
1.74 paf 526: "regular expression syntax error - %s", errptr);
1.62 paf 527:
1.63 paf 528: int info_substrings=pcre_info(code, 0, 0);
529: if(info_substrings<0) {
1.100 parser 530: pcre_free(code);
1.112 parser 531: throw Exception(0, 0,
1.73 paf 532: aorigin,
1.76 paf 533: "pcre_info error (%d)",
1.73 paf 534: info_substrings);
1.63 paf 535: }
536:
537: int startoffset=0;
1.118 paf 538: const char *subject=cstr();
1.62 paf 539: int length=strlen(subject);
1.63 paf 540: int ovecsize;
541: int *ovector=(int *)malloc(sizeof(int)*
1.114 paf 542: (ovecsize=(1/*match*/+info_substrings)*3), 11);
1.62 paf 543:
1.64 paf 544: { // create table
545: Array& columns=*NEW Array(pool());
546: columns+=string_pre_match_name;
547: columns+=string_match_name;
548: columns+=string_post_match_name;
549: for(int i=1; i<=info_substrings; i++) {
550: char *column=(char *)malloc(MAX_NUMBER);
551: snprintf(column, MAX_NUMBER, "%d", i);
552: columns+=NEW String(pool(), column); // .i column name
553: }
554: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 555: }
1.63 paf 556:
1.64 paf 557: int exec_option_bits=0;
1.63 paf 558: while(true) {
559: int exec_substrings=pcre_exec(code, 0,
560: subject, length, startoffset,
1.64 paf 561: exec_option_bits, ovector, ovecsize);
1.63 paf 562:
563: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 564: pcre_free(code);
565: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 566: return option_bits[1]!=0; // global=true+table, not global=false
567: }
568:
569: if(exec_substrings<0) {
1.100 parser 570: pcre_free(code);
1.112 parser 571: throw Exception(0, 0,
1.63 paf 572: aorigin,
1.76 paf 573: "regular expression execute error (%d)",
1.63 paf 574: exec_substrings);
575: }
576:
577: Array& row=*NEW Array(pool());
1.81 paf 578: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 579: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 580: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 581:
582: for(int i=1; i<exec_substrings; i++) {
1.69 paf 583: // -1:-1 case handled peacefully by mid() itself
584: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 585: }
586:
1.100 parser 587: row_action(**table, &row, startoffset, ovector[0], info);
1.63 paf 588:
1.100 parser 589: if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang
590: pcre_free(code);
591: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 592: return true;
593: }
1.100 parser 594: startoffset=ovector[1];
1.63 paf 595:
596: /*
597: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 598: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 599: */
600: }
1.82 parser 601: }
602:
603: String& String::change_case(Pool& pool, const unsigned char *tables,
604: Change_case_kind kind) const {
605: String& result=*new(pool) String(pool);
606:
607: const unsigned char *a;
608: const unsigned char *b;
609: switch(kind) {
610: case CC_UPPER:
611: a=tables+lcc_offset;
612: b=tables+fcc_offset;
613: break;
614: case CC_LOWER:
615: a=tables+lcc_offset;
616: b=0;
617: break;
618: default:
1.112 parser 619: throw Exception(0, 0,
1.82 parser 620: this,
621: "unknown change case kind #%d",
622: static_cast<int>(kind)); // never
623: a=b=0; // calm, compiler
624: break; // never
625: }
626:
627: const Chunk *chunk=&head;
628: do {
629: const Chunk::Row *row=chunk->rows;
1.116 paf 630: for(uint i=0; i<chunk->count; i++, row++) {
1.82 parser 631: if(row==append_here)
632: goto break2;
633:
1.114 paf 634: char *new_cstr=(char *)pool.malloc(row->item.size, 12);
1.82 parser 635: char *dest=new_cstr;
636: const char *src=row->item.ptr;
637: for(int size=row->item.size; size--; src++) {
638: unsigned char c=a[(unsigned char)*src];
639: if(b)
640: c=b[c];
641:
642: *dest++=(char)c;
643: }
644:
645: result.APPEND(new_cstr, row->item.size,
1.123 ! paf 646: row->item.lang,
1.82 parser 647: row->item.origin.file, row->item.origin.line);
648: }
649: chunk=row->link;
650: } while(chunk);
651: break2:
1.89 parser 652:
1.101 parser 653: return result;
654: }
655:
1.108 parser 656: void String::join_chain(Pool& pool,
1.116 paf 657: uint& ai, const Chunk*& achunk, const Chunk::Row*& arow,
1.123 ! paf 658: uchar& joined_lang, const char *& joined_ptr, size_t& joined_size) const {
! 659: joined_lang=arow->item.lang;
1.108 parser 660:
661: // calc size
662: joined_size=0;
663: {
1.116 paf 664: uint start_i=ai;
1.108 parser 665: const Chunk::Row *start_row=arow;
666: const Chunk *chunk=achunk;
667: do {
668: const Chunk::Row *row=start_row;
1.116 paf 669: for(uint i=start_i; i<chunk->count; i++, row++) {
1.108 parser 670: if(row==append_here)
671: goto break21;
672:
673: if(row->item.lang==joined_lang)
674: joined_size+=row->item.size;
675: else
676: break;
677: }
678: if(chunk=row->link) {
679: start_i=0;
680: start_row=chunk->rows;
681: } else
682: break;
683: } while(true);
684: break21:;
685: }
686:
687: // if one row, return simply itself
688: if(joined_size==arow->item.size) {
689: joined_ptr=arow->item.ptr;
690: ai++; arow++;
691: if(ai==achunk->count)
692: achunk=arow->link;
693: } else {
694: // join adjacent rows
1.114 paf 695: char *ptr=(char *)pool.malloc(joined_size,13);
1.108 parser 696: joined_ptr=ptr;
1.116 paf 697: uint start_i=ai;
1.108 parser 698: const Chunk::Row *start_row=arow;
699: const Chunk *chunk=achunk;
1.116 paf 700: uint i;
1.108 parser 701: const Chunk::Row *row;
702: do {
703: row=start_row;
704: for(i=start_i; i<chunk->count; i++, row++) {
705: if(row==append_here)
706: goto break22;
707:
708: if(row->item.lang==joined_lang) {
709: memcpy(ptr, row->item.ptr, row->item.size);
710: ptr+=row->item.size;
711: } else
712: break;
713: }
714: if(chunk=row->link) {
715: start_i=0;
716: start_row=chunk->rows;
717: } else
718: break;
719: } while(true);
720: break22:;
721:
722: // return joined rows
723: ai=i;
724: arow=row;
725: achunk=chunk;
726: }
727: }
728:
729: String& String::reconstruct(Pool& pool) const {
730: //_asm int 3;
731: String& result=*new(pool) String(pool);
732: const Chunk *chunk=&head;
733: do {
734: const Chunk::Row *row=chunk->rows;
1.116 paf 735: for(uint i=0; i<chunk->count; ) {
1.108 parser 736: if(row==append_here)
737: goto break2;
738:
1.123 ! paf 739: uchar joined_lang;
1.109 parser 740: const char *joined_ptr;
1.108 parser 741: size_t joined_size;
742: join_chain(pool, i, chunk, row,
743: joined_lang, joined_ptr, joined_size);
744:
745: result.APPEND(joined_ptr, joined_size,
746: joined_lang,
747: row->item.origin.file, row->item.origin.line);
748: if(!chunk)
749: goto break2;
750: }
751: } while(true);
752: break2:
753:
754: return result;
755: };
756:
757: String& String::replace_in_reconstructed(Pool& pool, Dictionary& dict) const {
1.106 parser 758: //_asm int 3;
1.101 parser 759: String& result=*new(pool) String(pool);
1.123 ! paf 760: STRING_FOREACH_ROW(
! 761: const char *src=row->item.ptr;
! 762: size_t src_size=row->item.size;
! 763: char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14);
! 764: char *dest=new_cstr;
! 765: while(src_size) {
! 766: // there is a row where first column starts 'src'
! 767: if(Table::Item *item=dict.first_that_starts(src, src_size)) {
! 768: // get a=>b values
! 769: const String& a=*static_cast<Array *>(item)->get_string(0);
! 770: const String& b=*static_cast<Array *>(item)->get_string(1);
! 771: // skip 'a' in 'src' && reduce work size
! 772: src+=a.size(); src_size-=a.size();
! 773: // write 'b' to 'dest' && skip 'b' in 'dest'
! 774: b.store_to(dest); dest+=b.size();
! 775: } else {
! 776: // write a char to b && reduce work size
! 777: *dest++=*src++; src_size--;
1.101 parser 778: }
779: }
780:
1.123 ! paf 781: result.APPEND(new_cstr, dest-new_cstr,
! 782: row->item.lang,
! 783: row->item.origin.file, row->item.origin.line);
! 784: );
1.101 parser 785: break2:
1.89 parser 786: return result;
1.108 parser 787: }
788:
789: String& String::replace(Pool& pool, Dictionary& dict) const {
790: return reconstruct(pool).replace_in_reconstructed(pool, dict);
1.89 parser 791: }
792:
1.90 parser 793: double String::as_double() const {
1.89 parser 794: double result;
1.114 paf 795: const char *cstr;
796: char buf[MAX_NUMBER];
1.115 paf 797: if(head.rows+1==append_here) {
1.114 paf 798: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
799: memcpy(buf, head.rows[0].item.ptr, size);
800: buf[size]=0;
801: cstr=buf;
802: } else
803: cstr=this->cstr();
1.102 parser 804: char *error_pos;
1.89 parser 805: // 0xABC
1.99 parser 806: if(cstr[0]=='0')
807: if(cstr[1]=='x' || cstr[1]=='X')
808: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
809: else
1.102 parser 810: result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos);
1.89 parser 811: else
1.99 parser 812: result=(double)strtod(cstr, &error_pos);
1.89 parser 813:
1.103 parser 814: if(*error_pos/*not EOS*/)
1.112 parser 815: throw Exception(0, 0,
1.89 parser 816: this,
817: "invalid number (double)");
818:
819: return result;
820: }
1.90 parser 821: int String::as_int() const {
1.89 parser 822: int 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=(int)(unsigned long)strtol(cstr, &error_pos, 0);
837: else
1.102 parser 838: result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 839: else
840: result=(int)strtol(cstr, &error_pos, 0);
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 (int)");
1.82 parser 846:
847: return result;
1.61 paf 848: }
1.113 parser 849:
850: /* @todo maybe network order worth spending some effort?
851: don't bothering myself with network byte order,
852: am not planning to be able to move resulting file across platforms
853: for now
854: */
855: void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const {
856: buf_size=
857: prolog_size
1.115 paf 858: +used_rows()*(sizeof(Untaint_lang)+sizeof(size_t))
1.113 parser 859: +size();
1.114 paf 860: buf=malloc(buf_size,15);
1.113 parser 861: char *cur=(char *)buf+prolog_size;
862:
1.123 ! paf 863: STRING_FOREACH_ROW(
! 864: // lang
! 865: memcpy(cur, &row->item.lang, sizeof(Untaint_lang));
! 866: cur+=sizeof(Untaint_lang);
! 867: // size
! 868: memcpy(cur, &row->item.size, sizeof(size_t));
! 869: cur+=sizeof(size_t);
! 870: // bytes
! 871: memcpy(cur, row->item.ptr, row->item.size);
! 872: cur+=row->item.size;
! 873: );
1.113 parser 874: break2:
875: ;
876: }
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: #ifndef DOXYGEN
884: struct Serialized_piece {
885: String::Untaint_lang lang;
886: size_t size;
887: char ptr[1];
888: };
889: #endif
890:
891: void String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) {
892: char *cur=((char *)buf)+prolog_size;
893: buf_size-=prolog_size;
894:
895: while(buf_size) {
896: Serialized_piece& p=*(Serialized_piece *)cur;
897: APPEND(p.ptr, p.size, p.lang, file, 0);
898:
899: size_t piece_size=sizeof(p.lang)+sizeof(p.size)+p.size;
900: cur+=piece_size;
901: buf_size-=piece_size;
902: }
903: }
904:
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