Annotation of parser3/src/main/pa_string.C, revision 1.91
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:
1.91 ! parser 8: $Id: pa_string.C,v 1.90 2001/05/21 16:39:32 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"
1.60 paf 24:
1.75 paf 25: String::String(Pool& apool, const char *src, size_t src_size, bool tainted) :
1.84 parser 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];
1.8 paf 32: fused_rows=fsize=0;
1.41 paf 33:
34: if(src)
1.75 paf 35: if(tainted)
36: APPEND_TAINTED(src, src_size, 0, 0);
1.41 paf 37: else
1.75 paf 38: APPEND_CLEAN(src, src_size, 0, 0);
1.1 paf 39: }
40:
41: void String::expand() {
1.85 parser 42: size_t new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.28 paf 43: last_chunk=static_cast<Chunk *>(
1.55 paf 44: malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *)));
1.28 paf 45: last_chunk->count=new_chunk_count;
46: link_row->link=last_chunk;
47: append_here=last_chunk->rows;
48: link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 49: link_row->link=0;
1.1 paf 50: }
51:
1.40 paf 52: String::String(const String& src) : Pooled(src.pool()) {
1.8 paf 53: head.count=CR_PREALLOCATED_COUNT;
54:
1.55 paf 55: size_t src_used_rows=src.fused_rows;
1.8 paf 56: if(src_used_rows<=head.count) {
1.55 paf 57: // all new rows fit size_to preallocated area
58: size_t curr_chunk_rows=head.count;
1.8 paf 59: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
60: append_here=&head.rows[src_used_rows];
61: link_row=&head.rows[curr_chunk_rows];
62: } else {
63: // warning:
1.10 paf 64: // heavily relies on the fact
65: // "preallocated area is the same for all strings"
1.8 paf 66: //
67: // info:
68: // allocating only enough mem to fit src string rows
69: // next append would allocate a new chunk
70: //
1.55 paf 71: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 72: // preallocated chunk src to constructing head
73: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 paf 74: // remaining rows size_to new_chunk
75: size_t curr_chunk_rows=src_used_rows-head.count;
1.8 paf 76: Chunk *new_chunk=static_cast<Chunk *>(
1.55 paf 77: malloc(sizeof(size_t)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *)));
1.8 paf 78: new_chunk->count=curr_chunk_rows;
79: head.preallocated_link=new_chunk;
1.28 paf 80: append_here=link_row=&new_chunk->rows[new_chunk->count];
1.8 paf 81:
82: Chunk *old_chunk=src.head.preallocated_link;
83: Chunk::Row *new_rows=new_chunk->rows;
1.55 paf 84: size_t rows_left_to_copy=new_chunk->count;
1.8 paf 85: while(true) {
1.55 paf 86: size_t old_count=old_chunk->count;
1.8 paf 87: Chunk *next_chunk=old_chunk->rows[old_count].link;
88: if(next_chunk) {
89: // not last source chunk
90: // taking it all
91: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
92: new_rows+=old_count;
93: rows_left_to_copy-=old_count;
94:
95: old_chunk=next_chunk;
96: } else {
97: // the last source chunk
98: // taking only those rows of chunk that _left_to_copy
99: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
100: break;
101: }
102: }
1.5 paf 103: }
1.8 paf 104: link_row->link=0;
105: fused_rows=src_used_rows;
106: fsize=src.fsize;
1.5 paf 107: }
1.28 paf 108:
1.42 paf 109: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.60 paf 110: const Chunk *chunk=&src.head;
1.40 paf 111: do {
1.60 paf 112: const Chunk::Row *row=chunk->rows;
113: for(size_t i=0; i<chunk->count; i++, row++) {
114: if(row==src.append_here)
1.40 paf 115: goto break2;
1.60 paf 116:
117: APPEND(row->item.ptr, row->item.size,
118: (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang,
119: row->item.origin.file, row->item.origin.line);
1.40 paf 120: }
121: chunk=row->link;
122: } while(chunk);
123: break2:
1.60 paf 124: return *this;
1.34 paf 125: }
1.60 paf 126:
1.13 paf 127: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 128: if(!src)
129: return *this;
1.26 paf 130: if(!size)
131: size=strlen(src);
132: if(!size)
1.9 paf 133: return *this;
134:
1.1 paf 135: if(chunk_is_full())
136: expand();
137:
138: append_here->item.ptr=src;
1.26 paf 139: fsize+=append_here->item.size=size;
1.52 paf 140: append_here->item.lang=lang;
1.13 paf 141: #ifndef NO_STRING_ORIGIN
1.14 paf 142: append_here->item.origin.file=file;
143: append_here->item.origin.line=line;
1.13 paf 144: #endif
1.8 paf 145: append_here++; fused_rows++;
1.1 paf 146:
147: return *this;
148: }
149:
1.16 paf 150: uint String::hash_code() const {
1.7 paf 151: uint result=0;
1.5 paf 152:
1.16 paf 153: const Chunk *chunk=&head;
1.5 paf 154: do {
1.16 paf 155: const Chunk::Row *row=chunk->rows;
1.55 paf 156: for(size_t i=0; i<chunk->count; i++) {
1.5 paf 157: if(row==append_here)
158: goto break2;
159:
1.6 paf 160: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 161: row++;
162: }
163: chunk=row->link;
164: } while(chunk);
165: break2:
166: return result;
167: }
168:
1.60 paf 169: /// @todo move 'lang' skipping to pos
170: int String::cmp(int& partial, const String& src,
171: size_t this_offset, Untaint_lang lang) const {
1.59 paf 172: partial=-1;
1.55 paf 173: this_offset=min(this_offset, size()-1);
174:
1.16 paf 175: const Chunk *a_chunk=&head;
176: const Chunk *b_chunk=&src.head;
177: const Chunk::Row *a_row=a_chunk->rows;
178: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 179: size_t a_offset=this_offset;
180: size_t b_offset=0;
1.9 paf 181: Chunk::Row *a_end=append_here;
182: Chunk::Row *b_end=src.append_here;
1.55 paf 183: size_t a_countdown=a_chunk->count;
184: size_t b_countdown=b_chunk->count;
185: size_t result;
1.60 paf 186: size_t pos=0;
1.33 paf 187:
1.83 parser 188: bool a_break=size()==0;
1.91 ! parser 189: bool b_break=src.size()==0;
1.83 parser 190: if(!(a_break || b_break)) while(true) {
1.55 paf 191: if(pos+a_row->item.size > this_offset) {
1.71 paf 192: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 193: return -1; // wrong lang -- bail out
194:
1.55 paf 195: int size_diff=
196: (a_row->item.size-a_offset)-
197: (b_row->item.size-b_offset);
198:
199: if(size_diff==0) { // a has same size as b
1.60 paf 200: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
201: a_row->item.size-a_offset);
1.55 paf 202: if(result)
203: return result;
1.60 paf 204: pos+=a_row->item.size;
1.55 paf 205: a_row++; a_countdown--; a_offset=0;
206: b_row++; b_countdown--; b_offset=0;
207: } else if (size_diff>0) { // a longer
1.60 paf 208: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
209: b_row->item.size-b_offset);
1.55 paf 210: if(result)
211: return result;
212: a_offset+=b_row->item.size-b_offset;
213: b_row++; b_countdown--; b_offset=0;
214: } else { // b longer
1.60 paf 215: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
216: a_row->item.size-a_offset);
1.55 paf 217: if(result)
218: return result;
219: b_offset+=a_row->item.size-a_offset;
1.60 paf 220: pos+=a_row->item.size;
1.55 paf 221: a_row++; a_countdown--; a_offset=0;
222: }
1.83 parser 223: if(b_break=b_row==b_end) {
224: a_break=a_row==a_end;
225: break;
226: }
1.55 paf 227: if(!b_countdown) {
228: b_chunk=b_row->link;
229: b_row=b_chunk->rows;
230: b_countdown=b_chunk->count;
231: }
232: } else {
1.60 paf 233: a_offset-=a_row->item.size;
234: pos+=a_row->item.size;
235: a_row++; a_countdown--;
1.9 paf 236: }
237:
1.83 parser 238: if(a_break=a_row==a_end) {
239: b_break=b_row==b_end;
240: break;
241: }
1.11 paf 242: if(!a_countdown) {
1.9 paf 243: a_chunk=a_row->link;
244: a_row=a_chunk->rows;
1.11 paf 245: a_countdown=a_chunk->count;
1.9 paf 246: }
1.27 paf 247: }
1.55 paf 248: if(a_break==b_break) { // ended simultaneously
249: partial=0; return 0;
250: } else if(a_break) { // first bytes equal, but a ended before b
251: partial=1; return -1;
252: } else {
253: partial=2; return +1;
254: }
1.27 paf 255: }
256:
1.60 paf 257: /// @todo move 'lang' skipping to pos
1.59 paf 258: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 259: size_t this_offset, Untaint_lang lang) const {
1.59 paf 260: partial=-1;
1.50 paf 261: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 262: this_offset=min(this_offset, size()-1);
1.27 paf 263:
264: const Chunk *a_chunk=&head;
265: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 266: size_t a_offset=this_offset;
1.55 paf 267: size_t b_offset=0;
1.27 paf 268: Chunk::Row *a_end=append_here;
1.55 paf 269: size_t a_countdown=a_chunk->count;
1.60 paf 270: size_t pos=0;
1.52 paf 271:
1.83 parser 272: bool a_break=size()==0;
273: bool b_break=b_size==0;
274: if(!(a_break || b_break)) while(true) {
1.59 paf 275: if(pos+a_row->item.size > this_offset) {
1.71 paf 276: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 277: return -1; // wrong lang -- bail out
278:
1.59 paf 279: int size_diff=
280: (a_row->item.size-a_offset)-
281: (b_size-b_offset);
282:
283: if(size_diff==0) { // a has same size as b
284: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
285: a_row->item.size-a_offset)!=0)
286: return result;
1.60 paf 287: pos+=a_row->item.size;
1.59 paf 288: a_row++; a_countdown--; a_offset=0;
289: b_break=true;
290: } else if (size_diff>0) { // a longer
291: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
292: b_size-b_offset)!=0)
293: return result;
294: a_offset+=b_size-b_offset;
295: b_break=true;
296: } else { // b longer
297: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
298: a_row->item.size-a_offset)!=0)
299: return result;
300: b_offset+=a_row->item.size-a_offset;
1.60 paf 301: pos+=a_row->item.size;
1.59 paf 302: a_row++; a_countdown--; a_offset=0;
303: }
304: } else {
1.60 paf 305: a_offset-=a_row->item.size;
306: pos+=a_row->item.size;
307: a_row++; a_countdown--;
1.27 paf 308: }
309:
1.86 parser 310: a_break=a_row==a_end;
311: if(a_break || b_break)
1.83 parser 312: break;
1.27 paf 313: if(!a_countdown) {
314: a_chunk=a_row->link;
315: a_row=a_chunk->rows;
316: a_countdown=a_chunk->count;
1.9 paf 317: }
318: }
1.55 paf 319: if(a_break==b_break) { // ended simultaneously
320: partial=0; return 0;
321: } else if(a_break) { // first bytes equal, but a ended before b
322: partial=1; return -1;
323: } else {
324: partial=2; return +1;
325: }
1.5 paf 326: }
1.46 paf 327:
328: #ifndef NO_STRING_ORIGIN
329: const Origin& String::origin() const {
330: if(!fused_rows)
331: THROW(0, 0,
1.50 paf 332: 0,
333: "String::origin() of empty string called");
1.46 paf 334:
1.49 paf 335: // determining origin by last appended piece
1.50 paf 336: // because first one frequently constant.
337: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 338: // when last peice is constant,
339: // ex: parser_root_auto_path{dynamic} / auto.p{const}
340: // using first piece
341: Origin& last_origin=append_here[-1].item.origin;
342: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 343: }
344: #endif
1.53 paf 345:
1.69 paf 346: String& String::mid(size_t start, size_t finish) const {
1.53 paf 347: start=max(0, start);
348: finish=min(size(), finish);
1.60 paf 349: if(start==finish)
350: return *empty_string;
1.53 paf 351:
352: String& result=*NEW String(pool());
353:
354: size_t pos=0;
355: const Chunk *chunk=&head;
356: do {
357: const Chunk::Row *row=chunk->rows;
1.55 paf 358: for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 359: if(row==append_here)
360: goto break2;
361:
1.60 paf 362: size_t item_finish=pos+row->item.size;
363: if(item_finish > start) { // started now or already?
364: bool started=result.size()==0; // started now?
365: bool finished=finish <= item_finish; // finished now?
1.53 paf 366: size_t offset=started?start-pos:0;
367: size_t size=finished?finish-pos:row->item.size;
368: result.APPEND(
369: row->item.ptr+offset, size-offset,
370: row->item.lang,
371: row->item.origin.file, row->item.origin.line);
372: if(finished)
373: goto break2;
374: }
375: }
376: chunk=row->link;
377: } while(chunk);
378: break2:
1.60 paf 379: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
380: //cstr(), start, finish, result.cstr());
1.53 paf 381: return result;
1.54 paf 382: }
383:
1.60 paf 384: int String::pos(const String& substr,
385: size_t result, Untaint_lang lang) const {
1.58 paf 386: for(; result<size(); result++) {
1.60 paf 387: int partial; cmp(partial, substr, result, lang);
1.58 paf 388: if(
389: partial==0 || // full match
390: partial==2) // 'substr' starts 'this'+'result'
391: return result;
392: }
393:
394: return -1;
395: }
396:
1.60 paf 397: int String::pos(const char *substr, size_t substr_size,
398: size_t result, Untaint_lang lang) const {
1.57 paf 399: for(; result<size(); result++) {
1.60 paf 400: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 401: if(
402: partial==0 || // full match
403: partial==2) // 'substr' starts 'this'+'result'
404: return result;
405: }
406:
407: return -1;
1.60 paf 408: }
409:
410: void String::split(Array& result,
411: size_t* pos_after_ref,
412: const char *delim, size_t delim_size,
413: Untaint_lang lang, int limit) const {
414: if(delim_size) {
415: size_t pos_after=pos_after_ref?*pos_after_ref:0;
416: int pos_before;
417: // while we have 'delim'...
418: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 419: result+=&mid(pos_after, pos_before);
1.60 paf 420: pos_after=pos_before+delim_size;
421: }
422: // last piece
423: if(pos_after<size() && limit) {
1.69 paf 424: result+=&mid(pos_after, size());
1.60 paf 425: pos_after=size();
426: }
427: if(pos_after_ref)
428: *pos_after_ref=pos_after;
429: } else { // empty delim
430: result+=this;
431: if(pos_after_ref)
432: *pos_after_ref+=size();
433: }
434: }
435:
436: void String::split(Array& result,
437: size_t* pos_after_ref,
438: const String& delim, Untaint_lang lang,
439: int limit) const {
440: if(delim.size()) {
441: size_t pos_after=pos_after_ref?*pos_after_ref:0;
442: int pos_before;
443: // while we have 'delim'...
444: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 445: result+=&mid(pos_after, pos_before);
1.60 paf 446: pos_after=pos_before+delim.size();
447: }
448: // last piece
449: if(pos_after<size() && limit) {
1.69 paf 450: result+=&mid(pos_after, size());
1.60 paf 451: pos_after=size();
452: }
453: if(pos_after_ref)
454: *pos_after_ref=pos_after;
455: } else { // empty delim
456: result+=this;
457: if(pos_after_ref)
458: *pos_after_ref+=size();
459: }
1.61 paf 460: }
461:
1.63 paf 462: static void regex_options(char *options, int *result){
463: struct Regex_option {
464: char key;
465: int clear, set;
466: int *result;
467: } regex_option[]={
468: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 469: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 470: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
471: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
472: {'g', 0, true, result+1}, // many rows
473: {0},
474: };
475: result[0]=PCRE_EXTRA | PCRE_DOTALL;
476: result[1]=0;
477:
478: if(options)
479: for(Regex_option *o=regex_option; o->key; o++)
480: if(
481: strchr(options, o->key) ||
482: strchr(options, toupper(o->key))) {
483: *(o->result)&=~o->clear;
484: *(o->result)|=o->set;
485: }
486: }
487:
1.88 parser 488: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.77 paf 489: bool String::match(const unsigned char *pcre_tables,
490: const String *aorigin,
1.62 paf 491: const String& regexp,
1.63 paf 492: const String *options,
1.64 paf 493: Table **table,
494: Row_action row_action, void *info) const {
495:
1.73 paf 496: if(!regexp.size())
497: THROW(0, 0,
498: aorigin,
499: "regexp is empty");
1.68 paf 500: const char *pattern=regexp.cstr(UL_AS_IS);
1.62 paf 501: const char *errptr;
502: int erroffset;
1.63 paf 503: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
504: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 505: &errptr, &erroffset,
1.74 paf 506: pcre_tables);
1.62 paf 507:
1.67 paf 508: if(!code)
1.62 paf 509: THROW(0, 0,
1.69 paf 510: ®exp.mid(erroffset, regexp.size()),
1.74 paf 511: "regular expression syntax error - %s", errptr);
1.62 paf 512:
1.63 paf 513: int info_substrings=pcre_info(code, 0, 0);
514: if(info_substrings<0) {
515: (*pcre_free)(code);
516: THROW(0, 0,
1.73 paf 517: aorigin,
1.76 paf 518: "pcre_info error (%d)",
1.73 paf 519: info_substrings);
1.63 paf 520: }
521:
522: int startoffset=0;
1.68 paf 523: const char *subject=cstr(UL_AS_IS);
1.62 paf 524: int length=strlen(subject);
1.63 paf 525: int ovecsize;
526: int *ovector=(int *)malloc(sizeof(int)*
1.65 paf 527: (ovecsize=(1/*match*/+info_substrings)*3));
1.62 paf 528:
1.64 paf 529: { // create table
530: Array& columns=*NEW Array(pool());
531: columns+=string_pre_match_name;
532: columns+=string_match_name;
533: columns+=string_post_match_name;
534: for(int i=1; i<=info_substrings; i++) {
535: char *column=(char *)malloc(MAX_NUMBER);
536: snprintf(column, MAX_NUMBER, "%d", i);
537: columns+=NEW String(pool(), column); // .i column name
538: }
539: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 540: }
1.63 paf 541:
1.64 paf 542: int exec_option_bits=0;
1.63 paf 543: while(true) {
544: int exec_substrings=pcre_exec(code, 0,
545: subject, length, startoffset,
1.64 paf 546: exec_option_bits, ovector, ovecsize);
1.63 paf 547:
548: if(exec_substrings==PCRE_ERROR_NOMATCH) {
549: (*pcre_free)(code);
1.67 paf 550: (*row_action)(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 551: return option_bits[1]!=0; // global=true+table, not global=false
552: }
553:
554: if(exec_substrings<0) {
555: (*pcre_free)(code);
556: THROW(0, 0,
557: aorigin,
1.76 paf 558: "regular expression execute error (%d)",
1.63 paf 559: exec_substrings);
560: }
561:
562: Array& row=*NEW Array(pool());
1.81 paf 563: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 564: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 565: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 566:
567: for(int i=1; i<exec_substrings; i++) {
1.69 paf 568: // -1:-1 case handled peacefully by mid() itself
569: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 570: }
571:
1.67 paf 572: (*row_action)(**table, &row, startoffset, ovector[0], info);
1.63 paf 573:
1.67 paf 574: if(!option_bits[1] || !(startoffset=ovector[1])) { // not global | going to hang
1.63 paf 575: (*pcre_free)(code);
1.67 paf 576: (*row_action)(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 577: return true;
578: }
579:
580: /*
581: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 582: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 583: */
584: }
1.82 parser 585: }
586:
587: String& String::change_case(Pool& pool, const unsigned char *tables,
588: Change_case_kind kind) const {
589: String& result=*new(pool) String(pool);
590:
591: const unsigned char *a;
592: const unsigned char *b;
593: switch(kind) {
594: case CC_UPPER:
595: a=tables+lcc_offset;
596: b=tables+fcc_offset;
597: break;
598: case CC_LOWER:
599: a=tables+lcc_offset;
600: b=0;
601: break;
602: default:
603: PTHROW(0, 0,
604: this,
605: "unknown change case kind #%d",
606: static_cast<int>(kind)); // never
607: a=b=0; // calm, compiler
608: break; // never
609: }
610:
611: const Chunk *chunk=&head;
612: do {
613: const Chunk::Row *row=chunk->rows;
614: for(size_t i=0; i<chunk->count; i++, row++) {
615: if(row==append_here)
616: goto break2;
617:
618: char *new_cstr=(char *)pool.malloc(row->item.size);
619: char *dest=new_cstr;
620: const char *src=row->item.ptr;
621: for(int size=row->item.size; size--; src++) {
622: unsigned char c=a[(unsigned char)*src];
623: if(b)
624: c=b[c];
625:
626: *dest++=(char)c;
627: }
628:
629: result.APPEND(new_cstr, row->item.size,
630: row->item.lang,
631: row->item.origin.file, row->item.origin.line);
632: }
633: chunk=row->link;
634: } while(chunk);
635: break2:
1.89 parser 636:
637: return result;
638: }
639:
1.90 parser 640: double String::as_double() const {
1.89 parser 641: double result;
642: const char *cstr=this->cstr();
643: char *error_pos=0;
644: // 0xABC
645: if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X'))
646: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
647: else
648: result=strtod(cstr, &error_pos);
649:
650: if(error_pos && *error_pos)
651: THROW(0, 0,
652: this,
653: "invalid number (double)");
654:
655: return result;
656: }
1.90 parser 657: int String::as_int() const {
1.89 parser 658: int result;
659: const char *cstr=this->cstr();
660: char *error_pos=0;
661: // 0xABC
662: if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X'))
663: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
664: else
665: result=(int)strtol(cstr, &error_pos, 0);
666:
667: if(error_pos && *error_pos)
668: THROW(0, 0,
669: this,
670: "invalid number (int)");
1.82 parser 671:
672: return result;
1.61 paf 673: }
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