Annotation of parser3/src/main/pa_string.C, revision 1.93
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)
1.4 paf 7: */
1.93 ! parser 8: static const char *RCSId="$Id: double.C,v 1.31 2001/06/28 07:41:59 parser Exp $";
1.4 paf 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"
1.60 paf 23:
1.75 paf 24: String::String(Pool& apool, const char *src, size_t src_size, bool tainted) :
1.84 parser 25: Pooled(apool) {
1.28 paf 26: last_chunk=&head;
27: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 28: append_here=head.rows;
1.2 paf 29: head.preallocated_link=0;
1.28 paf 30: link_row=&head.rows[head.count];
1.8 paf 31: fused_rows=fsize=0;
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:
40: void String::expand() {
1.85 parser 41: size_t new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.28 paf 42: last_chunk=static_cast<Chunk *>(
1.55 paf 43: malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *)));
1.28 paf 44: last_chunk->count=new_chunk_count;
45: link_row->link=last_chunk;
46: append_here=last_chunk->rows;
47: link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 48: link_row->link=0;
1.1 paf 49: }
50:
1.40 paf 51: String::String(const String& src) : Pooled(src.pool()) {
1.8 paf 52: head.count=CR_PREALLOCATED_COUNT;
53:
1.55 paf 54: size_t src_used_rows=src.fused_rows;
1.8 paf 55: if(src_used_rows<=head.count) {
1.55 paf 56: // all new rows fit size_to preallocated area
57: size_t curr_chunk_rows=head.count;
1.8 paf 58: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
59: append_here=&head.rows[src_used_rows];
60: link_row=&head.rows[curr_chunk_rows];
61: } else {
62: // warning:
1.10 paf 63: // heavily relies on the fact
64: // "preallocated area is the same for all strings"
1.8 paf 65: //
66: // info:
67: // allocating only enough mem to fit src string rows
68: // next append would allocate a new chunk
69: //
1.55 paf 70: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 71: // preallocated chunk src to constructing head
72: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 paf 73: // remaining rows size_to new_chunk
74: size_t curr_chunk_rows=src_used_rows-head.count;
1.8 paf 75: Chunk *new_chunk=static_cast<Chunk *>(
1.55 paf 76: malloc(sizeof(size_t)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *)));
1.8 paf 77: new_chunk->count=curr_chunk_rows;
78: head.preallocated_link=new_chunk;
1.28 paf 79: append_here=link_row=&new_chunk->rows[new_chunk->count];
1.8 paf 80:
81: Chunk *old_chunk=src.head.preallocated_link;
82: Chunk::Row *new_rows=new_chunk->rows;
1.55 paf 83: size_t rows_left_to_copy=new_chunk->count;
1.8 paf 84: while(true) {
1.55 paf 85: size_t old_count=old_chunk->count;
1.8 paf 86: Chunk *next_chunk=old_chunk->rows[old_count].link;
87: if(next_chunk) {
88: // not last source chunk
89: // taking it all
90: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
91: new_rows+=old_count;
92: rows_left_to_copy-=old_count;
93:
94: old_chunk=next_chunk;
95: } else {
96: // the last source chunk
97: // taking only those rows of chunk that _left_to_copy
98: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
99: break;
100: }
101: }
1.5 paf 102: }
1.8 paf 103: link_row->link=0;
104: fused_rows=src_used_rows;
105: fsize=src.fsize;
1.5 paf 106: }
1.28 paf 107:
1.42 paf 108: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.60 paf 109: const Chunk *chunk=&src.head;
1.40 paf 110: do {
1.60 paf 111: const Chunk::Row *row=chunk->rows;
112: for(size_t i=0; i<chunk->count; i++, row++) {
113: if(row==src.append_here)
1.40 paf 114: goto break2;
1.60 paf 115:
116: APPEND(row->item.ptr, row->item.size,
117: (lang!=UL_PASS_APPENDED && (row->item.lang==UL_TAINTED || forced))?lang:row->item.lang,
118: row->item.origin.file, row->item.origin.line);
1.40 paf 119: }
120: chunk=row->link;
121: } while(chunk);
122: break2:
1.60 paf 123: return *this;
1.34 paf 124: }
1.60 paf 125:
1.13 paf 126: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 127: if(!src)
128: return *this;
1.26 paf 129: if(!size)
130: size=strlen(src);
131: if(!size)
1.9 paf 132: return *this;
133:
1.1 paf 134: if(chunk_is_full())
135: expand();
136:
137: append_here->item.ptr=src;
1.26 paf 138: fsize+=append_here->item.size=size;
1.52 paf 139: append_here->item.lang=lang;
1.13 paf 140: #ifndef NO_STRING_ORIGIN
1.14 paf 141: append_here->item.origin.file=file;
142: append_here->item.origin.line=line;
1.13 paf 143: #endif
1.8 paf 144: append_here++; fused_rows++;
1.1 paf 145:
146: return *this;
147: }
148:
1.16 paf 149: uint String::hash_code() const {
1.7 paf 150: uint result=0;
1.5 paf 151:
1.16 paf 152: const Chunk *chunk=&head;
1.5 paf 153: do {
1.16 paf 154: const Chunk::Row *row=chunk->rows;
1.55 paf 155: for(size_t i=0; i<chunk->count; i++) {
1.5 paf 156: if(row==append_here)
157: goto break2;
158:
1.6 paf 159: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 160: row++;
161: }
162: chunk=row->link;
163: } while(chunk);
164: break2:
165: return result;
166: }
167:
1.60 paf 168: /// @todo move 'lang' skipping to pos
169: int String::cmp(int& partial, const String& src,
170: size_t this_offset, Untaint_lang lang) const {
1.59 paf 171: partial=-1;
1.55 paf 172: this_offset=min(this_offset, size()-1);
173:
1.16 paf 174: const Chunk *a_chunk=&head;
175: const Chunk *b_chunk=&src.head;
176: const Chunk::Row *a_row=a_chunk->rows;
177: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 178: size_t a_offset=this_offset;
179: size_t b_offset=0;
1.9 paf 180: Chunk::Row *a_end=append_here;
181: Chunk::Row *b_end=src.append_here;
1.55 paf 182: size_t a_countdown=a_chunk->count;
183: size_t b_countdown=b_chunk->count;
184: size_t result;
1.60 paf 185: size_t pos=0;
1.33 paf 186:
1.83 parser 187: bool a_break=size()==0;
1.91 parser 188: bool b_break=src.size()==0;
1.83 parser 189: if(!(a_break || b_break)) while(true) {
1.55 paf 190: if(pos+a_row->item.size > this_offset) {
1.71 paf 191: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 192: return -1; // wrong lang -- bail out
193:
1.55 paf 194: int size_diff=
195: (a_row->item.size-a_offset)-
196: (b_row->item.size-b_offset);
197:
198: if(size_diff==0) { // a has same size as b
1.60 paf 199: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
200: a_row->item.size-a_offset);
1.55 paf 201: if(result)
202: return result;
1.60 paf 203: pos+=a_row->item.size;
1.55 paf 204: a_row++; a_countdown--; a_offset=0;
205: b_row++; b_countdown--; b_offset=0;
206: } else if (size_diff>0) { // a longer
1.60 paf 207: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
208: b_row->item.size-b_offset);
1.55 paf 209: if(result)
210: return result;
211: a_offset+=b_row->item.size-b_offset;
212: b_row++; b_countdown--; b_offset=0;
213: } else { // b longer
1.60 paf 214: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
215: a_row->item.size-a_offset);
1.55 paf 216: if(result)
217: return result;
218: b_offset+=a_row->item.size-a_offset;
1.60 paf 219: pos+=a_row->item.size;
1.55 paf 220: a_row++; a_countdown--; a_offset=0;
221: }
1.83 parser 222: if(b_break=b_row==b_end) {
223: a_break=a_row==a_end;
224: break;
225: }
1.55 paf 226: if(!b_countdown) {
227: b_chunk=b_row->link;
228: b_row=b_chunk->rows;
229: b_countdown=b_chunk->count;
230: }
231: } else {
1.60 paf 232: a_offset-=a_row->item.size;
233: pos+=a_row->item.size;
234: a_row++; a_countdown--;
1.9 paf 235: }
236:
1.83 parser 237: if(a_break=a_row==a_end) {
238: b_break=b_row==b_end;
239: break;
240: }
1.11 paf 241: if(!a_countdown) {
1.9 paf 242: a_chunk=a_row->link;
243: a_row=a_chunk->rows;
1.11 paf 244: a_countdown=a_chunk->count;
1.9 paf 245: }
1.27 paf 246: }
1.55 paf 247: if(a_break==b_break) { // ended simultaneously
248: partial=0; return 0;
249: } else if(a_break) { // first bytes equal, but a ended before b
250: partial=1; return -1;
251: } else {
252: partial=2; return +1;
253: }
1.27 paf 254: }
255:
1.60 paf 256: /// @todo move 'lang' skipping to pos
1.59 paf 257: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 258: size_t this_offset, Untaint_lang lang) const {
1.59 paf 259: partial=-1;
1.50 paf 260: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.59 paf 261: this_offset=min(this_offset, size()-1);
1.27 paf 262:
263: const Chunk *a_chunk=&head;
264: const Chunk::Row *a_row=a_chunk->rows;
1.59 paf 265: size_t a_offset=this_offset;
1.55 paf 266: size_t b_offset=0;
1.27 paf 267: Chunk::Row *a_end=append_here;
1.55 paf 268: size_t a_countdown=a_chunk->count;
1.60 paf 269: size_t pos=0;
1.52 paf 270:
1.83 parser 271: bool a_break=size()==0;
272: bool b_break=b_size==0;
273: if(!(a_break || b_break)) while(true) {
1.59 paf 274: if(pos+a_row->item.size > this_offset) {
1.71 paf 275: if(lang!=UL_UNSPECIFIED && a_row->item.lang!=lang)
1.60 paf 276: return -1; // wrong lang -- bail out
277:
1.59 paf 278: int size_diff=
279: (a_row->item.size-a_offset)-
280: (b_size-b_offset);
281:
282: if(size_diff==0) { // a has same size as b
283: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
284: a_row->item.size-a_offset)!=0)
285: return result;
1.60 paf 286: pos+=a_row->item.size;
1.59 paf 287: a_row++; a_countdown--; a_offset=0;
288: b_break=true;
289: } else if (size_diff>0) { // a longer
290: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
291: b_size-b_offset)!=0)
292: return result;
293: a_offset+=b_size-b_offset;
294: b_break=true;
295: } else { // b longer
296: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset,
297: a_row->item.size-a_offset)!=0)
298: return result;
299: b_offset+=a_row->item.size-a_offset;
1.60 paf 300: pos+=a_row->item.size;
1.59 paf 301: a_row++; a_countdown--; a_offset=0;
302: }
303: } else {
1.60 paf 304: a_offset-=a_row->item.size;
305: pos+=a_row->item.size;
306: a_row++; a_countdown--;
1.27 paf 307: }
308:
1.86 parser 309: a_break=a_row==a_end;
310: if(a_break || b_break)
1.83 parser 311: break;
1.27 paf 312: if(!a_countdown) {
313: a_chunk=a_row->link;
314: a_row=a_chunk->rows;
315: a_countdown=a_chunk->count;
1.9 paf 316: }
317: }
1.55 paf 318: if(a_break==b_break) { // ended simultaneously
319: partial=0; return 0;
320: } else if(a_break) { // first bytes equal, but a ended before b
321: partial=1; return -1;
322: } else {
323: partial=2; return +1;
324: }
1.5 paf 325: }
1.46 paf 326:
327: #ifndef NO_STRING_ORIGIN
328: const Origin& String::origin() const {
329: if(!fused_rows)
330: THROW(0, 0,
1.50 paf 331: 0,
332: "String::origin() of empty string called");
1.46 paf 333:
1.49 paf 334: // determining origin by last appended piece
1.50 paf 335: // because first one frequently constant.
336: // ex: ^load[/file] "document_root" + "/file"
1.80 paf 337: // when last peice is constant,
338: // ex: parser_root_auto_path{dynamic} / auto.p{const}
339: // using first piece
340: Origin& last_origin=append_here[-1].item.origin;
341: return last_origin.file ? last_origin : head.rows[0].item.origin;
1.46 paf 342: }
343: #endif
1.53 paf 344:
1.69 paf 345: String& String::mid(size_t start, size_t finish) const {
1.53 paf 346: start=max(0, start);
347: finish=min(size(), finish);
1.60 paf 348: if(start==finish)
349: return *empty_string;
1.53 paf 350:
351: String& result=*NEW String(pool());
352:
353: size_t pos=0;
354: const Chunk *chunk=&head;
355: do {
356: const Chunk::Row *row=chunk->rows;
1.55 paf 357: for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 358: if(row==append_here)
359: goto break2;
360:
1.60 paf 361: size_t item_finish=pos+row->item.size;
362: if(item_finish > start) { // started now or already?
363: bool started=result.size()==0; // started now?
364: bool finished=finish <= item_finish; // finished now?
1.53 paf 365: size_t offset=started?start-pos:0;
366: size_t size=finished?finish-pos:row->item.size;
367: result.APPEND(
368: row->item.ptr+offset, size-offset,
369: row->item.lang,
370: row->item.origin.file, row->item.origin.line);
371: if(finished)
372: goto break2;
373: }
374: }
375: chunk=row->link;
376: } while(chunk);
377: break2:
1.60 paf 378: // SAPI::log(pool(), "piece of '%s' from %d to %d is '%s'",
379: //cstr(), start, finish, result.cstr());
1.53 paf 380: return result;
1.54 paf 381: }
382:
1.60 paf 383: int String::pos(const String& substr,
384: size_t result, Untaint_lang lang) const {
1.58 paf 385: for(; result<size(); result++) {
1.60 paf 386: int partial; cmp(partial, substr, result, lang);
1.58 paf 387: if(
388: partial==0 || // full match
389: partial==2) // 'substr' starts 'this'+'result'
390: return result;
391: }
392:
393: return -1;
394: }
395:
1.60 paf 396: int String::pos(const char *substr, size_t substr_size,
397: size_t result, Untaint_lang lang) const {
1.57 paf 398: for(; result<size(); result++) {
1.60 paf 399: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 400: if(
401: partial==0 || // full match
402: partial==2) // 'substr' starts 'this'+'result'
403: return result;
404: }
405:
406: return -1;
1.60 paf 407: }
408:
409: void String::split(Array& result,
410: size_t* pos_after_ref,
411: const char *delim, size_t delim_size,
412: Untaint_lang lang, int limit) const {
413: if(delim_size) {
414: size_t pos_after=pos_after_ref?*pos_after_ref:0;
415: int pos_before;
416: // while we have 'delim'...
417: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 418: result+=&mid(pos_after, pos_before);
1.60 paf 419: pos_after=pos_before+delim_size;
420: }
421: // last piece
422: if(pos_after<size() && limit) {
1.69 paf 423: result+=&mid(pos_after, size());
1.60 paf 424: pos_after=size();
425: }
426: if(pos_after_ref)
427: *pos_after_ref=pos_after;
428: } else { // empty delim
429: result+=this;
430: if(pos_after_ref)
431: *pos_after_ref+=size();
432: }
433: }
434:
435: void String::split(Array& result,
436: size_t* pos_after_ref,
437: const String& delim, Untaint_lang lang,
438: int limit) const {
439: if(delim.size()) {
440: size_t pos_after=pos_after_ref?*pos_after_ref:0;
441: int pos_before;
442: // while we have 'delim'...
443: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 444: result+=&mid(pos_after, pos_before);
1.60 paf 445: pos_after=pos_before+delim.size();
446: }
447: // last piece
448: if(pos_after<size() && limit) {
1.69 paf 449: result+=&mid(pos_after, size());
1.60 paf 450: pos_after=size();
451: }
452: if(pos_after_ref)
453: *pos_after_ref=pos_after;
454: } else { // empty delim
455: result+=this;
456: if(pos_after_ref)
457: *pos_after_ref+=size();
458: }
1.61 paf 459: }
460:
1.63 paf 461: static void regex_options(char *options, int *result){
462: struct Regex_option {
463: char key;
464: int clear, set;
465: int *result;
466: } regex_option[]={
467: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 468: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 469: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
470: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
471: {'g', 0, true, result+1}, // many rows
472: {0},
473: };
474: result[0]=PCRE_EXTRA | PCRE_DOTALL;
475: result[1]=0;
476:
477: if(options)
478: for(Regex_option *o=regex_option; o->key; o++)
479: if(
480: strchr(options, o->key) ||
481: strchr(options, toupper(o->key))) {
482: *(o->result)&=~o->clear;
483: *(o->result)|=o->set;
484: }
485: }
486:
1.88 parser 487: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.77 paf 488: bool String::match(const unsigned char *pcre_tables,
489: const String *aorigin,
1.62 paf 490: const String& regexp,
1.63 paf 491: const String *options,
1.64 paf 492: Table **table,
493: Row_action row_action, void *info) const {
494:
1.73 paf 495: if(!regexp.size())
496: THROW(0, 0,
497: aorigin,
498: "regexp is empty");
1.68 paf 499: const char *pattern=regexp.cstr(UL_AS_IS);
1.62 paf 500: const char *errptr;
501: int erroffset;
1.63 paf 502: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
503: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 504: &errptr, &erroffset,
1.74 paf 505: pcre_tables);
1.62 paf 506:
1.67 paf 507: if(!code)
1.62 paf 508: THROW(0, 0,
1.69 paf 509: ®exp.mid(erroffset, regexp.size()),
1.74 paf 510: "regular expression syntax error - %s", errptr);
1.62 paf 511:
1.63 paf 512: int info_substrings=pcre_info(code, 0, 0);
513: if(info_substrings<0) {
514: (*pcre_free)(code);
515: THROW(0, 0,
1.73 paf 516: aorigin,
1.76 paf 517: "pcre_info error (%d)",
1.73 paf 518: info_substrings);
1.63 paf 519: }
520:
521: int startoffset=0;
1.68 paf 522: const char *subject=cstr(UL_AS_IS);
1.62 paf 523: int length=strlen(subject);
1.63 paf 524: int ovecsize;
525: int *ovector=(int *)malloc(sizeof(int)*
1.65 paf 526: (ovecsize=(1/*match*/+info_substrings)*3));
1.62 paf 527:
1.64 paf 528: { // create table
529: Array& columns=*NEW Array(pool());
530: columns+=string_pre_match_name;
531: columns+=string_match_name;
532: columns+=string_post_match_name;
533: for(int i=1; i<=info_substrings; i++) {
534: char *column=(char *)malloc(MAX_NUMBER);
535: snprintf(column, MAX_NUMBER, "%d", i);
536: columns+=NEW String(pool(), column); // .i column name
537: }
538: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 539: }
1.63 paf 540:
1.64 paf 541: int exec_option_bits=0;
1.63 paf 542: while(true) {
543: int exec_substrings=pcre_exec(code, 0,
544: subject, length, startoffset,
1.64 paf 545: exec_option_bits, ovector, ovecsize);
1.63 paf 546:
547: if(exec_substrings==PCRE_ERROR_NOMATCH) {
548: (*pcre_free)(code);
1.67 paf 549: (*row_action)(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 550: return option_bits[1]!=0; // global=true+table, not global=false
551: }
552:
553: if(exec_substrings<0) {
554: (*pcre_free)(code);
555: THROW(0, 0,
556: aorigin,
1.76 paf 557: "regular expression execute error (%d)",
1.63 paf 558: exec_substrings);
559: }
560:
561: Array& row=*NEW Array(pool());
1.81 paf 562: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 563: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 564: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 565:
566: for(int i=1; i<exec_substrings; i++) {
1.69 paf 567: // -1:-1 case handled peacefully by mid() itself
568: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 569: }
570:
1.67 paf 571: (*row_action)(**table, &row, startoffset, ovector[0], info);
1.63 paf 572:
1.67 paf 573: if(!option_bits[1] || !(startoffset=ovector[1])) { // not global | going to hang
1.63 paf 574: (*pcre_free)(code);
1.67 paf 575: (*row_action)(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 576: return true;
577: }
578:
579: /*
580: if(option_bits[0] & PCRE_MULTILINE)
1.64 paf 581: exec_option_bits|=PCRE_NOTBOL; // start of subject+startoffset not BOL
1.63 paf 582: */
583: }
1.82 parser 584: }
585:
586: String& String::change_case(Pool& pool, const unsigned char *tables,
587: Change_case_kind kind) const {
588: String& result=*new(pool) String(pool);
589:
590: const unsigned char *a;
591: const unsigned char *b;
592: switch(kind) {
593: case CC_UPPER:
594: a=tables+lcc_offset;
595: b=tables+fcc_offset;
596: break;
597: case CC_LOWER:
598: a=tables+lcc_offset;
599: b=0;
600: break;
601: default:
602: PTHROW(0, 0,
603: this,
604: "unknown change case kind #%d",
605: static_cast<int>(kind)); // never
606: a=b=0; // calm, compiler
607: break; // never
608: }
609:
610: const Chunk *chunk=&head;
611: do {
612: const Chunk::Row *row=chunk->rows;
613: for(size_t i=0; i<chunk->count; i++, row++) {
614: if(row==append_here)
615: goto break2;
616:
617: char *new_cstr=(char *)pool.malloc(row->item.size);
618: char *dest=new_cstr;
619: const char *src=row->item.ptr;
620: for(int size=row->item.size; size--; src++) {
621: unsigned char c=a[(unsigned char)*src];
622: if(b)
623: c=b[c];
624:
625: *dest++=(char)c;
626: }
627:
628: result.APPEND(new_cstr, row->item.size,
629: row->item.lang,
630: row->item.origin.file, row->item.origin.line);
631: }
632: chunk=row->link;
633: } while(chunk);
634: break2:
1.89 parser 635:
636: return result;
637: }
638:
1.90 parser 639: double String::as_double() const {
1.89 parser 640: double result;
641: const char *cstr=this->cstr();
642: char *error_pos=0;
643: // 0xABC
644: if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X'))
645: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
646: else
647: result=strtod(cstr, &error_pos);
648:
649: if(error_pos && *error_pos)
650: THROW(0, 0,
651: this,
652: "invalid number (double)");
653:
654: return result;
655: }
1.90 parser 656: int String::as_int() const {
1.89 parser 657: int result;
658: const char *cstr=this->cstr();
659: char *error_pos=0;
660: // 0xABC
661: if(cstr[0]=='0' && (cstr[1]=='x' || cstr[1]=='X'))
662: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
663: else
664: result=(int)strtol(cstr, &error_pos, 0);
665:
666: if(error_pos && *error_pos)
667: THROW(0, 0,
668: this,
669: "invalid number (int)");
1.82 parser 670:
671: return result;
1.61 paf 672: }
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