Annotation of parser3/src/main/pa_string.C, revision 1.83
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)
1.46 paf 5:
1.37 paf 6: Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf)
1.36 paf 7:
1.83 ! parser 8: $Id: pa_string.C,v 1.82 2001/05/14 13:18:07 parser Exp $
1.4 paf 9: */
10:
1.48 paf 11: #include "pa_config_includes.h"
1.1 paf 12:
1.70 paf 13: #include "pcre.h"
1.82 parser 14: #include "internal.h"
1.70 paf 15:
1.13 paf 16: #include "pa_pool.h"
1.12 paf 17: #include "pa_string.h"
1.5 paf 18: #include "pa_hash.h"
1.22 paf 19: #include "pa_exception.h"
1.53 paf 20: #include "pa_common.h"
1.60 paf 21: #include "pa_array.h"
22: #include "pa_globals.h"
1.61 paf 23: #include "pa_table.h"
1.62 paf 24: #include "pa_threads.h"
1.60 paf 25:
1.83 ! parser 26: #include "pa_sapi.h"
! 27: #define STRING_STAT_MAX_PIECES 1000
! 28: int string_stat_pieces[STRING_STAT_MAX_PIECES];
! 29: void log_string_stats(Pool& pool) {
! 30: for(int i=0; i<STRING_STAT_MAX_PIECES; i++)
! 31: if(int v=string_stat_pieces[i])
! 32: SAPI::log(pool, "%i: %10d",
! 33: i, v);
! 34: }
1.1 paf 35:
1.18 paf 36:
1.75 paf 37: String::String(Pool& apool, const char *src, size_t src_size, bool tainted) :
1.83 ! parser 38: Pooled(apool),expand_times(0) {
! 39: string_stat_pieces[0]++;
1.28 paf 40: last_chunk=&head;
41: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 42: append_here=head.rows;
1.2 paf 43: head.preallocated_link=0;
1.28 paf 44: link_row=&head.rows[head.count];
1.8 paf 45: fused_rows=fsize=0;
1.41 paf 46:
47: if(src)
1.75 paf 48: if(tainted)
49: APPEND_TAINTED(src, src_size, 0, 0);
1.41 paf 50: else
1.75 paf 51: APPEND_CLEAN(src, src_size, 0, 0);
1.1 paf 52: }
53:
54: void String::expand() {
1.83 ! parser 55: {
! 56: int index=min(++expand_times, STRING_STAT_MAX_PIECES-1);
! 57: if(index)
! 58: string_stat_pieces[index-1]++;
! 59: string_stat_pieces[index]++;
! 60: }
! 61:
1.55 paf 62: size_t new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100;
1.28 paf 63: last_chunk=static_cast<Chunk *>(
1.55 paf 64: malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *)));
1.28 paf 65: last_chunk->count=new_chunk_count;
66: link_row->link=last_chunk;
67: append_here=last_chunk->rows;
68: link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 69: link_row->link=0;
1.1 paf 70: }
71:
1.40 paf 72: String::String(const String& src) : Pooled(src.pool()) {
1.83 ! parser 73: string_stat_pieces[0]++;
1.8 paf 74: head.count=CR_PREALLOCATED_COUNT;
75:
1.55 paf 76: size_t src_used_rows=src.fused_rows;
1.8 paf 77: if(src_used_rows<=head.count) {
1.55 paf 78: // all new rows fit size_to preallocated area
79: size_t curr_chunk_rows=head.count;
1.8 paf 80: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
81: append_here=&head.rows[src_used_rows];
82: link_row=&head.rows[curr_chunk_rows];
83: } else {
84: // warning:
1.10 paf 85: // heavily relies on the fact
86: // "preallocated area is the same for all strings"
1.8 paf 87: //
88: // info:
89: // allocating only enough mem to fit src string rows
90: // next append would allocate a new chunk
91: //
1.55 paf 92: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 93: // preallocated chunk src to constructing head
94: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 paf 95: // remaining rows size_to new_chunk
96: size_t curr_chunk_rows=src_used_rows-head.count;
1.8 paf 97: Chunk *new_chunk=static_cast<Chunk *>(
1.55 paf 98: malloc(sizeof(size_t)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *)));
1.8 paf 99: new_chunk->count=curr_chunk_rows;
100: head.preallocated_link=new_chunk;
1.28 paf 101: append_here=link_row=&new_chunk->rows[new_chunk->count];
1.8 paf 102:
103: Chunk *old_chunk=src.head.preallocated_link;
104: Chunk::Row *new_rows=new_chunk->rows;
1.55 paf 105: size_t rows_left_to_copy=new_chunk->count;
1.8 paf 106: while(true) {
1.55 paf 107: size_t old_count=old_chunk->count;
1.8 paf 108: Chunk *next_chunk=old_chunk->rows[old_count].link;
109: if(next_chunk) {
110: // not last source chunk
111: // taking it all
112: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
113: new_rows+=old_count;
114: rows_left_to_copy-=old_count;
115:
116: old_chunk=next_chunk;
117: } else {
118: // the last source chunk
119: // taking only those rows of chunk that _left_to_copy
120: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
121: break;
122: }
123: }
1.5 paf 124: }
1.8 paf 125: link_row->link=0;
126: fused_rows=src_used_rows;
127: fsize=src.fsize;
1.5 paf 128: }
1.28 paf 129:
1.42 paf 130: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.60 paf 131: const Chunk *chunk=&src.head;
1.40 paf 132: do {
1.60 paf 133: const Chunk::Row *row=chunk->rows;
134: for(size_t i=0; i<chunk->count; i++, row++) {
135: if(row==src.append_here)
1.40 paf 136: goto break2;
1.60 paf 137:
138: APPEND(row->item.ptr, row->item.size,
139: (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang,
140: row->item.origin.file, row->item.origin.line);
1.40 paf 141: }
142: chunk=row->link;
143: } while(chunk);
144: break2:
1.60 paf 145: return *this;
1.34 paf 146: }
1.60 paf 147:
1.13 paf 148: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 149: if(!src)
150: return *this;
1.26 paf 151: if(!size)
152: size=strlen(src);
153: if(!size)
1.9 paf 154: return *this;
155:
1.1 paf 156: if(chunk_is_full())
157: expand();
158:
159: append_here->item.ptr=src;
1.26 paf 160: fsize+=append_here->item.size=size;
1.52 paf 161: append_here->item.lang=lang;
1.13 paf 162: #ifndef NO_STRING_ORIGIN
1.14 paf 163: append_here->item.origin.file=file;
164: append_here->item.origin.line=line;
1.13 paf 165: #endif
1.8 paf 166: append_here++; fused_rows++;
1.1 paf 167:
168: return *this;
169: }
170:
1.16 paf 171: uint String::hash_code() const {
1.7 paf 172: uint result=0;
1.5 paf 173:
1.16 paf 174: const Chunk *chunk=&head;
1.5 paf 175: do {
1.16 paf 176: const Chunk::Row *row=chunk->rows;
1.55 paf 177: for(size_t i=0; i<chunk->count; i++) {
1.5 paf 178: if(row==append_here)
179: goto break2;
180:
1.6 paf 181: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 182: row++;
183: }
184: chunk=row->link;
185: } while(chunk);
186: break2:
187: return result;
188: }
189:
1.60 paf 190: /// @todo move 'lang' skipping to pos
191: int String::cmp(int& partial, const String& src,
192: size_t this_offset, Untaint_lang lang) const {
1.59 paf 193: partial=-1;
1.55 paf 194: this_offset=min(this_offset, size()-1);
195:
1.16 paf 196: const Chunk *a_chunk=&head;
197: const Chunk *b_chunk=&src.head;
198: const Chunk::Row *a_row=a_chunk->rows;
199: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 200: size_t a_offset=this_offset;
201: size_t b_offset=0;
1.9 paf 202: Chunk::Row *a_end=append_here;
203: Chunk::Row *b_end=src.append_here;
1.55 paf 204: size_t a_countdown=a_chunk->count;
205: size_t b_countdown=b_chunk->count;
206: size_t result;
1.60 paf 207: size_t pos=0;
1.33 paf 208:
1.83 ! parser 209: bool a_break=size()==0;
! 210: bool b_break=size()==0;
! 211: if(!(a_break || b_break)) while(true) {
1.55 paf 212: if(pos+a_row->item.size > this_offset) {
1.71 paf 213: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 214: return -1; // wrong lang -- bail out
215:
1.55 paf 216: int size_diff=
217: (a_row->item.size-a_offset)-
218: (b_row->item.size-b_offset);
219:
220: if(size_diff==0) { // a has same size as b
1.60 paf 221: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
222: a_row->item.size-a_offset);
1.55 paf 223: if(result)
224: return result;
1.60 paf 225: pos+=a_row->item.size;
1.55 paf 226: a_row++; a_countdown--; a_offset=0;
227: b_row++; b_countdown--; b_offset=0;
228: } else if (size_diff>0) { // a longer
1.60 paf 229: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
230: b_row->item.size-b_offset);
1.55 paf 231: if(result)
232: return result;
233: a_offset+=b_row->item.size-b_offset;
234: b_row++; b_countdown--; b_offset=0;
235: } else { // b longer
1.60 paf 236: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
237: a_row->item.size-a_offset);
1.55 paf 238: if(result)
239: return result;
240: b_offset+=a_row->item.size-a_offset;
1.60 paf 241: pos+=a_row->item.size;
1.55 paf 242: a_row++; a_countdown--; a_offset=0;
243: }
1.83 ! parser 244: if(b_break=b_row==b_end) {
! 245: a_break=a_row==a_end;
! 246: break;
! 247: }
1.55 paf 248: if(!b_countdown) {
249: b_chunk=b_row->link;
250: b_row=b_chunk->rows;
251: b_countdown=b_chunk->count;
252: }
253: } else {
1.60 paf 254: a_offset-=a_row->item.size;
255: pos+=a_row->item.size;
256: a_row++; a_countdown--;
1.9 paf 257: }
258:
1.83 ! parser 259: if(a_break=a_row==a_end) {
! 260: b_break=b_row==b_end;
! 261: break;
! 262: }
1.11 paf 263: if(!a_countdown) {
1.9 paf 264: a_chunk=a_row->link;
265: a_row=a_chunk->rows;
1.11 paf 266: a_countdown=a_chunk->count;
1.9 paf 267: }
1.27 paf 268: }
1.55 paf 269: if(a_break==b_break) { // ended simultaneously
270: partial=0; return 0;
271: } else if(a_break) { // first bytes equal, but a ended before b
272: partial=1; return -1;
273: } else {
274: partial=2; return +1;
275: }
1.27 paf 276: }
277:
1.60 paf 278: /// @todo move 'lang' skipping to pos
1.59 paf 279: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 280: size_t this_offset, Untaint_lang lang) const {
1.59 paf 281: partial=-1;
1.50 paf 282: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 283: this_offset=min(this_offset, size()-1);
1.27 paf 284:
285: const Chunk *a_chunk=&head;
286: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 287: size_t a_offset=this_offset;
1.55 paf 288: size_t b_offset=0;
1.27 paf 289: Chunk::Row *a_end=append_here;
1.55 paf 290: size_t a_countdown=a_chunk->count;
1.60 paf 291: size_t pos=0;
1.52 paf 292:
1.83 ! parser 293: bool a_break=size()==0;
! 294: bool b_break=b_size==0;
! 295: if(!(a_break || b_break)) while(true) {
1.59 paf 296: if(pos+a_row->item.size > this_offset) {
1.71 paf 297: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 298: return -1; // wrong lang -- bail out
299:
1.59 paf 300: int size_diff=
301: (a_row->item.size-a_offset)-
302: (b_size-b_offset);
303:
304: if(size_diff==0) { // a has same size as b
305: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
306: a_row->item.size-a_offset)!=0)
307: return result;
1.60 paf 308: pos+=a_row->item.size;
1.59 paf 309: a_row++; a_countdown--; a_offset=0;
310: b_break=true;
311: } else if (size_diff>0) { // a longer
312: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
313: b_size-b_offset)!=0)
314: return result;
315: a_offset+=b_size-b_offset;
316: b_break=true;
317: } else { // b longer
318: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
319: a_row->item.size-a_offset)!=0)
320: return result;
321: b_offset+=a_row->item.size-a_offset;
1.60 paf 322: pos+=a_row->item.size;
1.59 paf 323: a_row++; a_countdown--; a_offset=0;
324: }
325: } else {
1.60 paf 326: a_offset-=a_row->item.size;
327: pos+=a_row->item.size;
328: a_row++; a_countdown--;
1.27 paf 329: }
330:
1.83 ! parser 331: if(a_break=a_row==a_end)
! 332: break;
1.27 paf 333: if(!a_countdown) {
334: a_chunk=a_row->link;
335: a_row=a_chunk->rows;
336: a_countdown=a_chunk->count;
1.9 paf 337: }
338: }
1.55 paf 339: if(a_break==b_break) { // ended simultaneously
340: partial=0; return 0;
341: } else if(a_break) { // first bytes equal, but a ended before b
342: partial=1; return -1;
343: } else {
344: partial=2; return +1;
345: }
1.5 paf 346: }
1.46 paf 347:
348: #ifndef NO_STRING_ORIGIN
349: const Origin& String::origin() const {
350: if(!fused_rows)
351: THROW(0, 0,
1.50 paf 352: 0,
353: "String::origin() of empty string called");
1.46 paf 354:
1.49 paf 355: // determining origin by last appended piece
1.50 paf 356: // because first one frequently constant.
357: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 358: // when last peice is constant,
359: // ex: parser_root_auto_path{dynamic} / auto.p{const}
360: // using first piece
361: Origin& last_origin=append_here[-1].item.origin;
362: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 363: }
364: #endif
1.53 paf 365:
1.69 paf 366: String& String::mid(size_t start, size_t finish) const {
1.53 paf 367: start=max(0, start);
368: finish=min(size(), finish);
1.60 paf 369: if(start==finish)
370: return *empty_string;
1.53 paf 371:
372: String& result=*NEW String(pool());
373:
374: size_t pos=0;
375: const Chunk *chunk=&head;
376: do {
377: const Chunk::Row *row=chunk->rows;
1.55 paf 378: for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 379: if(row==append_here)
380: goto break2;
381:
1.60 paf 382: size_t item_finish=pos+row->item.size;
383: if(item_finish > start) { // started now or already?
384: bool started=result.size()==0; // started now?
385: bool finished=finish <= item_finish; // finished now?
1.53 paf 386: size_t offset=started?start-pos:0;
387: size_t size=finished?finish-pos:row->item.size;
388: result.APPEND(
389: row->item.ptr+offset, size-offset,
390: row->item.lang,
391: row->item.origin.file, row->item.origin.line);
392: if(finished)
393: goto break2;
394: }
395: }
396: chunk=row->link;
397: } while(chunk);
398: break2:
1.60 paf 399: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
400: //cstr(), start, finish, result.cstr());
1.53 paf 401: return result;
1.54 paf 402: }
403:
1.60 paf 404: int String::pos(const String& substr,
405: size_t result, Untaint_lang lang) const {
1.58 paf 406: for(; result<size(); result++) {
1.60 paf 407: int partial; cmp(partial, substr, result, lang);
1.58 paf 408: if(
409: partial==0 || // full match
410: partial==2) // 'substr' starts 'this'+'result'
411: return result;
412: }
413:
414: return -1;
415: }
416:
1.60 paf 417: int String::pos(const char *substr, size_t substr_size,
418: size_t result, Untaint_lang lang) const {
1.57 paf 419: for(; result<size(); result++) {
1.60 paf 420: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 421: if(
422: partial==0 || // full match
423: partial==2) // 'substr' starts 'this'+'result'
424: return result;
425: }
426:
427: return -1;
1.60 paf 428: }
429:
430: void String::split(Array& result,
431: size_t* pos_after_ref,
432: const char *delim, size_t delim_size,
433: Untaint_lang lang, int limit) const {
434: if(delim_size) {
435: size_t pos_after=pos_after_ref?*pos_after_ref:0;
436: int pos_before;
437: // while we have 'delim'...
438: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 439: result+=&mid(pos_after, pos_before);
1.60 paf 440: pos_after=pos_before+delim_size;
441: }
442: // last piece
443: if(pos_after<size() && limit) {
1.69 paf 444: result+=&mid(pos_after, size());
1.60 paf 445: pos_after=size();
446: }
447: if(pos_after_ref)
448: *pos_after_ref=pos_after;
449: } else { // empty delim
450: result+=this;
451: if(pos_after_ref)
452: *pos_after_ref+=size();
453: }
454: }
455:
456: void String::split(Array& result,
457: size_t* pos_after_ref,
458: const String& delim, Untaint_lang lang,
459: int limit) const {
460: if(delim.size()) {
461: size_t pos_after=pos_after_ref?*pos_after_ref:0;
462: int pos_before;
463: // while we have 'delim'...
464: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 465: result+=&mid(pos_after, pos_before);
1.60 paf 466: pos_after=pos_before+delim.size();
467: }
468: // last piece
469: if(pos_after<size() && limit) {
1.69 paf 470: result+=&mid(pos_after, size());
1.60 paf 471: pos_after=size();
472: }
473: if(pos_after_ref)
474: *pos_after_ref=pos_after;
475: } else { // empty delim
476: result+=this;
477: if(pos_after_ref)
478: *pos_after_ref+=size();
479: }
1.61 paf 480: }
481:
1.63 paf 482: static void regex_options(char *options, int *result){
483: struct Regex_option {
484: char key;
485: int clear, set;
486: int *result;
487: } regex_option[]={
488: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 489: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 490: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
491: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
492: {'g', 0, true, result+1}, // many rows
493: {0},
494: };
495: result[0]=PCRE_EXTRA | PCRE_DOTALL;
496: result[1]=0;
497:
498: if(options)
499: for(Regex_option *o=regex_option; o->key; o++)
500: if(
501: strchr(options, o->key) ||
502: strchr(options, toupper(o->key))) {
503: *(o->result)&=~o->clear;
504: *(o->result)|=o->set;
505: }
506: }
507:
1.77 paf 508: bool String::match(const unsigned char *pcre_tables,
509: const String *aorigin,
1.62 paf 510: const String& regexp,
1.63 paf 511: const String *options,
1.64 paf 512: Table **table,
513: Row_action row_action, void *info) const {
514:
1.73 paf 515: if(!regexp.size())
516: THROW(0, 0,
517: aorigin,
518: "regexp is empty");
1.68 paf 519: const char *pattern=regexp.cstr(UL_AS_IS);
1.62 paf 520: const char *errptr;
521: int erroffset;
1.63 paf 522: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
523: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 524: &errptr, &erroffset,
1.74 paf 525: pcre_tables);
1.62 paf 526:
1.67 paf 527: if(!code)
1.62 paf 528: THROW(0, 0,
1.69 paf 529: ®exp.mid(erroffset, regexp.size()),
1.74 paf 530: "regular expression syntax error - %s", errptr);
1.62 paf 531:
1.63 paf 532: int info_substrings=pcre_info(code, 0, 0);
533: if(info_substrings<0) {
534: (*pcre_free)(code);
535: THROW(0, 0,
1.73 paf 536: aorigin,
1.76 paf 537: "pcre_info error (%d)",
1.73 paf 538: info_substrings);
1.63 paf 539: }
540:
541: int startoffset=0;
1.68 paf 542: const char *subject=cstr(UL_AS_IS);
1.62 paf 543: int length=strlen(subject);
1.63 paf 544: int ovecsize;
545: int *ovector=(int *)malloc(sizeof(int)*
1.65 paf 546: (ovecsize=(1/*match*/+info_substrings)*3));
1.62 paf 547:
1.64 paf 548: { // create table
549: Array& columns=*NEW Array(pool());
550: columns+=string_pre_match_name;
551: columns+=string_match_name;
552: columns+=string_post_match_name;
553: for(int i=1; i<=info_substrings; i++) {
554: char *column=(char *)malloc(MAX_NUMBER);
555: snprintf(column, MAX_NUMBER, "%d", i);
556: columns+=NEW String(pool(), column); // .i column name
557: }
558: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 559: }
1.63 paf 560:
1.64 paf 561: int exec_option_bits=0;
1.63 paf 562: while(true) {
563: int exec_substrings=pcre_exec(code, 0,
564: subject, length, startoffset,
1.64 paf 565: exec_option_bits, ovector, ovecsize);
1.63 paf 566:
567: if(exec_substrings==PCRE_ERROR_NOMATCH) {
568: (*pcre_free)(code);
1.67 paf 569: (*row_action)(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 570: return option_bits[1]!=0; // global=true+table, not global=false
571: }
572:
573: if(exec_substrings<0) {
574: (*pcre_free)(code);
575: THROW(0, 0,
576: aorigin,
1.76 paf 577: "regular expression execute error (%d)",
1.63 paf 578: exec_substrings);
579: }
580:
581: Array& row=*NEW Array(pool());
1.81 paf 582: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 583: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 584: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 585:
586: for(int i=1; i<exec_substrings; i++) {
1.69 paf 587: // -1:-1 case handled peacefully by mid() itself
588: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 589: }
590:
1.67 paf 591: (*row_action)(**table, &row, startoffset, ovector[0], info);
1.63 paf 592:
1.67 paf 593: if(!option_bits[1] || !(startoffset=ovector[1])) { // not global | going to hang
1.63 paf 594: (*pcre_free)(code);
1.67 paf 595: (*row_action)(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 596: return true;
597: }
598:
599: /*
600: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 601: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 602: */
603: }
1.82 parser 604: }
605:
606: String& String::change_case(Pool& pool, const unsigned char *tables,
607: Change_case_kind kind) const {
608: String& result=*new(pool) String(pool);
609:
610: const unsigned char *a;
611: const unsigned char *b;
612: switch(kind) {
613: case CC_UPPER:
614: a=tables+lcc_offset;
615: b=tables+fcc_offset;
616: break;
617: case CC_LOWER:
618: a=tables+lcc_offset;
619: b=0;
620: break;
621: default:
622: PTHROW(0, 0,
623: this,
624: "unknown change case kind #%d",
625: static_cast<int>(kind)); // never
626: a=b=0; // calm, compiler
627: break; // never
628: }
629:
630: const Chunk *chunk=&head;
631: do {
632: const Chunk::Row *row=chunk->rows;
633: for(size_t i=0; i<chunk->count; i++, row++) {
634: if(row==append_here)
635: goto break2;
636:
637: char *new_cstr=(char *)pool.malloc(row->item.size);
638: char *dest=new_cstr;
639: const char *src=row->item.ptr;
640: for(int size=row->item.size; size--; src++) {
641: unsigned char c=a[(unsigned char)*src];
642: if(b)
643: c=b[c];
644:
645: *dest++=(char)c;
646: }
647:
648: result.APPEND(new_cstr, row->item.size,
649: row->item.lang,
650: row->item.origin.file, row->item.origin.line);
651: }
652: chunk=row->link;
653: } while(chunk);
654: break2:
655:
656: return result;
1.61 paf 657: }
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