Annotation of parser3/src/main/pa_string.C, revision 1.138
1.45 paf 1: /** @file
1.55 paf 2: Parser: string class. @see untasize_t.C.
1.46 paf 3:
1.137 paf 4: Copyright (c) 2001, 2002 ArtLebedev Group (http://www.artlebedev.com)
1.138 ! paf 5: Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru)
1.46 paf 6:
1.138 ! paf 7: $Id: pa_string.C,v 1.137 2002/02/08 07:27:49 paf Exp $
1.4 paf 8: */
9:
1.70 paf 10: #include "pcre.h"
11:
1.13 paf 12: #include "pa_pool.h"
1.12 paf 13: #include "pa_string.h"
1.5 paf 14: #include "pa_hash.h"
1.22 paf 15: #include "pa_exception.h"
1.53 paf 16: #include "pa_common.h"
1.60 paf 17: #include "pa_array.h"
18: #include "pa_globals.h"
1.61 paf 19: #include "pa_table.h"
1.101 parser 20: #include "pa_dictionary.h"
1.132 paf 21: #include "pa_charset.h"
1.60 paf 22:
1.75 paf 23: String::String(Pool& apool, const char *src, size_t src_size, bool tainted) :
1.120 paf 24: Pooled(apool) {
1.28 paf 25: last_chunk=&head;
26: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 27: append_here=head.rows;
1.123 paf 28: head_link=0;
1.28 paf 29: link_row=&head.rows[head.count];
1.41 paf 30:
31: if(src)
1.75 paf 32: if(tainted)
33: APPEND_TAINTED(src, src_size, 0, 0);
1.41 paf 34: else
1.75 paf 35: APPEND_CLEAN(src, src_size, 0, 0);
1.1 paf 36: }
37:
1.94 parser 38: String::String(const String& src) :
1.120 paf 39: Pooled(src.pool()) {
1.8 paf 40: head.count=CR_PREALLOCATED_COUNT;
41:
1.116 paf 42: uint src_used_rows=src.used_rows();
1.8 paf 43: if(src_used_rows<=head.count) {
1.55 paf 44: // all new rows fit size_to preallocated area
1.98 parser 45: last_chunk=&head;
1.116 paf 46: uint curr_chunk_rows=head.count;
1.8 paf 47: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
48: append_here=&head.rows[src_used_rows];
49: link_row=&head.rows[curr_chunk_rows];
50: } else {
51: // warning:
1.10 paf 52: // heavily relies on the fact
53: // "preallocated area is the same for all strings"
1.8 paf 54: //
55: // info:
56: // allocating only enough mem to fit src string rows
57: // next append would allocate a new chunk
58: //
1.55 paf 59: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 60: // preallocated chunk src to constructing head
61: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 paf 62: // remaining rows size_to new_chunk
1.116 paf 63: uint curr_chunk_rows=src_used_rows-head.count;
1.98 parser 64: last_chunk=static_cast<Chunk *>(
1.116 paf 65: malloc(sizeof(uint)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *), 9));
1.98 parser 66: last_chunk->count=curr_chunk_rows;
1.123 paf 67: head_link=last_chunk;
1.98 parser 68: append_here=link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 69:
1.123 paf 70: Chunk *old_chunk=src.head_link;
1.98 parser 71: Chunk::Row *new_rows=last_chunk->rows;
1.116 paf 72: uint rows_left_to_copy=last_chunk->count;
1.8 paf 73: while(true) {
1.116 paf 74: uint old_count=old_chunk->count;
1.8 paf 75: Chunk *next_chunk=old_chunk->rows[old_count].link;
76: if(next_chunk) {
77: // not last source chunk
78: // taking it all
79: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
80: new_rows+=old_count;
81: rows_left_to_copy-=old_count;
82:
83: old_chunk=next_chunk;
84: } else {
85: // the last source chunk
86: // taking only those rows of chunk that _left_to_copy
87: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
88: break;
89: }
90: }
1.5 paf 91: }
1.8 paf 92: link_row->link=0;
1.115 paf 93: src_used_rows;
1.120 paf 94: }
95:
96: size_t String::size() const {
97: size_t result=0;
1.123 paf 98: STRING_FOREACH_ROW(
1.120 paf 99: result+=row->item.size;
1.123 paf 100: );
1.120 paf 101: break2:
102: return result;
1.94 parser 103: }
104:
1.115 paf 105: /// @todo not very optimal
106: uint String::used_rows() const {
107: uint result=0;
1.123 paf 108: STRING_FOREACH_ROW(
109: result++;
110: );
1.115 paf 111: break2:
112: return result;
113: }
1.94 parser 114: void String::expand() {
1.116 paf 115: uint new_chunk_count=last_chunk->count+CR_GROW_COUNT;
1.122 paf 116: if(new_chunk_count>MAX_USHORT)
117: new_chunk_count=MAX_USHORT;
118:
1.94 parser 119: last_chunk=static_cast<Chunk *>(
1.116 paf 120: malloc(sizeof(uint)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *), 10));
1.94 parser 121: last_chunk->count=new_chunk_count;
122: link_row->link=last_chunk;
123: append_here=last_chunk->rows;
124: link_row=&last_chunk->rows[last_chunk->count];
125: link_row->link=0;
1.5 paf 126: }
1.28 paf 127:
1.13 paf 128: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 129: if(!src)
130: return *this;
1.26 paf 131: if(!size)
132: size=strlen(src);
133: if(!size)
1.9 paf 134: return *this;
1.122 paf 135:
136: while(size>MAX_USHORT) {
137: if(chunk_is_full())
138: expand();
139:
140: append_here->item.ptr=src;
141: append_here->item.size=MAX_USHORT;
142: append_here->item.lang=lang;
143: #ifndef NO_STRING_ORIGIN
144: append_here->item.origin.file=file;
145: append_here->item.origin.line=line;
146: #endif
147: append_here++;
148:
149: src+=MAX_USHORT;
150: size-=MAX_USHORT;
151: }
1.9 paf 152:
1.1 paf 153: if(chunk_is_full())
154: expand();
155:
156: append_here->item.ptr=src;
1.121 paf 157: append_here->item.size=size;
1.52 paf 158: append_here->item.lang=lang;
1.13 paf 159: #ifndef NO_STRING_ORIGIN
1.14 paf 160: append_here->item.origin.file=file;
161: append_here->item.origin.line=line;
1.13 paf 162: #endif
1.115 paf 163: append_here++;
1.1 paf 164:
165: return *this;
1.97 parser 166: }
167:
168: char String::first_char() const {
1.120 paf 169: if(!used_rows())
1.112 parser 170: throw Exception(0, 0,
1.97 parser 171: this,
172: "getting first char of empty string");
173:
174: return *head.rows[0].item.ptr;
1.1 paf 175: }
176:
1.16 paf 177: uint String::hash_code() const {
1.7 paf 178: uint result=0;
1.123 paf 179: STRING_FOREACH_ROW(
1.6 paf 180: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.123 paf 181: );
1.5 paf 182: break2:
183: return result;
184: }
185:
1.60 paf 186: /// @todo move 'lang' skipping to pos
187: int String::cmp(int& partial, const String& src,
188: size_t this_offset, Untaint_lang lang) const {
1.59 paf 189: partial=-1;
1.125 paf 190: size_t a_size=size();
191: this_offset=min(this_offset, a_size-1);
1.55 paf 192:
1.16 paf 193: const Chunk *a_chunk=&head;
194: const Chunk *b_chunk=&src.head;
195: const Chunk::Row *a_row=a_chunk->rows;
196: const Chunk::Row *b_row=b_chunk->rows;
1.55 paf 197: size_t a_offset=this_offset;
198: size_t b_offset=0;
1.9 paf 199: Chunk::Row *a_end=append_here;
200: Chunk::Row *b_end=src.append_here;
1.116 paf 201: uint a_countdown=a_chunk->count;
202: uint b_countdown=b_chunk->count;
203: int result;
1.60 paf 204: size_t pos=0;
1.33 paf 205:
1.125 paf 206: bool a_break=a_size==0;
1.91 parser 207: bool b_break=src.size()==0;
1.83 parser 208: if(!(a_break || b_break)) while(true) {
1.55 paf 209: if(pos+a_row->item.size > this_offset) {
1.136 paf 210: if(lang!=UL_UNSPECIFIED && a_row->item.lang>lang)
1.60 paf 211: return -1; // wrong lang -- bail out
212:
1.55 paf 213: int size_diff=
214: (a_row->item.size-a_offset)-
215: (b_row->item.size-b_offset);
216:
217: if(size_diff==0) { // a has same size as b
1.60 paf 218: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
219: a_row->item.size-a_offset);
1.55 paf 220: if(result)
221: return result;
1.60 paf 222: pos+=a_row->item.size;
1.55 paf 223: a_row++; a_countdown--; a_offset=0;
224: b_row++; b_countdown--; b_offset=0;
225: } else if (size_diff>0) { // a longer
1.60 paf 226: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
227: b_row->item.size-b_offset);
1.55 paf 228: if(result)
229: return result;
230: a_offset+=b_row->item.size-b_offset;
231: b_row++; b_countdown--; b_offset=0;
232: } else { // b longer
1.60 paf 233: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset,
234: a_row->item.size-a_offset);
1.55 paf 235: if(result)
236: return result;
237: b_offset+=a_row->item.size-a_offset;
1.60 paf 238: pos+=a_row->item.size;
1.55 paf 239: a_row++; a_countdown--; a_offset=0;
240: }
1.83 parser 241: if(b_break=b_row==b_end) {
242: a_break=a_row==a_end;
243: break;
244: }
1.55 paf 245: if(!b_countdown) {
246: b_chunk=b_row->link;
247: b_row=b_chunk->rows;
248: b_countdown=b_chunk->count;
249: }
250: } else {
1.60 paf 251: a_offset-=a_row->item.size;
252: pos+=a_row->item.size;
253: a_row++; a_countdown--;
1.9 paf 254: }
255:
1.83 parser 256: if(a_break=a_row==a_end) {
257: b_break=b_row==b_end;
258: break;
259: }
1.11 paf 260: if(!a_countdown) {
1.9 paf 261: a_chunk=a_row->link;
262: a_row=a_chunk->rows;
1.11 paf 263: a_countdown=a_chunk->count;
1.9 paf 264: }
1.27 paf 265: }
1.55 paf 266: if(a_break==b_break) { // ended simultaneously
267: partial=0; return 0;
268: } else if(a_break) { // first bytes equal, but a ended before b
269: partial=1; return -1;
270: } else {
271: partial=2; return +1;
272: }
1.27 paf 273: }
274:
1.60 paf 275: /// @todo move 'lang' skipping to pos
1.59 paf 276: int String::cmp(int& partial, const char* b_ptr, size_t src_size,
1.60 paf 277: size_t this_offset, Untaint_lang lang) const {
1.59 paf 278: partial=-1;
1.125 paf 279: size_t a_size=size();
1.50 paf 280: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.125 paf 281: this_offset=min(this_offset, a_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:
1.125 paf 291: bool a_break=a_size==0;
1.83 parser 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) {
1.136 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 {
1.125 paf 398: size_t self_size=size();
1.131 paf 399: for(; size_t(result)<self_size; result++) {
1.60 paf 400: int partial; cmp(partial, substr, result, lang);
1.58 paf 401: if(
402: partial==0 || // full match
403: partial==2) // 'substr' starts 'this'+'result'
404: return result;
405: }
406:
407: return -1;
408: }
409:
1.60 paf 410: int String::pos(const char *substr, size_t substr_size,
1.116 paf 411: int result, Untaint_lang lang) const {
1.125 paf 412: size_t self_size=size();
1.131 paf 413: for(; size_t(result)<self_size; result++) {
1.60 paf 414: int partial; cmp(partial, substr, substr_size, result, lang);
1.55 paf 415: if(
416: partial==0 || // full match
417: partial==2) // 'substr' starts 'this'+'result'
418: return result;
419: }
420:
421: return -1;
1.60 paf 422: }
423:
424: void String::split(Array& result,
425: size_t* pos_after_ref,
426: const char *delim, size_t delim_size,
427: Untaint_lang lang, int limit) const {
1.125 paf 428: size_t self_size=size();
1.60 paf 429: if(delim_size) {
430: size_t pos_after=pos_after_ref?*pos_after_ref:0;
431: int pos_before;
432: // while we have 'delim'...
433: for(; (pos_before=pos(delim, delim_size, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 434: result+=&mid(pos_after, pos_before);
1.60 paf 435: pos_after=pos_before+delim_size;
436: }
437: // last piece
1.124 paf 438: if(pos_after<self_size && limit) {
439: result+=&mid(pos_after, self_size);
440: pos_after=self_size;
1.60 paf 441: }
442: if(pos_after_ref)
443: *pos_after_ref=pos_after;
444: } else { // empty delim
445: result+=this;
446: if(pos_after_ref)
1.124 paf 447: *pos_after_ref+=self_size;
1.60 paf 448: }
449: }
450:
451: void String::split(Array& result,
452: size_t* pos_after_ref,
453: const String& delim, Untaint_lang lang,
454: int limit) const {
455: if(delim.size()) {
456: size_t pos_after=pos_after_ref?*pos_after_ref:0;
457: int pos_before;
458: // while we have 'delim'...
459: for(; (pos_before=pos(delim, pos_after, lang))>=0 && limit; limit--) {
1.69 paf 460: result+=&mid(pos_after, pos_before);
1.60 paf 461: pos_after=pos_before+delim.size();
462: }
463: // last piece
464: if(pos_after<size() && limit) {
1.69 paf 465: result+=&mid(pos_after, size());
1.60 paf 466: pos_after=size();
467: }
468: if(pos_after_ref)
469: *pos_after_ref=pos_after;
470: } else { // empty delim
471: result+=this;
472: if(pos_after_ref)
473: *pos_after_ref+=size();
474: }
1.61 paf 475: }
476:
1.63 paf 477: static void regex_options(char *options, int *result){
478: struct Regex_option {
479: char key;
480: int clear, set;
481: int *result;
482: } regex_option[]={
483: {'i', 0, PCRE_CASELESS, result}, // a=A
1.79 paf 484: {'s', 0, PCRE_DOTALL, result}, // \n\n$ [default]
1.63 paf 485: {'x', 0, PCRE_EXTENDED, result}, // whitespace in regex ignored
486: {'m', PCRE_DOTALL, PCRE_MULTILINE, result}, // ^aaa\n$^bbb\n$
487: {'g', 0, true, result+1}, // many rows
488: {0},
489: };
490: result[0]=PCRE_EXTRA | PCRE_DOTALL;
491: result[1]=0;
492:
493: if(options)
494: for(Regex_option *o=regex_option; o->key; o++)
495: if(
496: strchr(options, o->key) ||
497: strchr(options, toupper(o->key))) {
498: *(o->result)&=~o->clear;
499: *(o->result)|=o->set;
500: }
501: }
502:
1.88 parser 503: /// @todo maybe need speedup: some option to remove pre/match/post string generation
1.132 paf 504: bool String::match(
1.77 paf 505: const String *aorigin,
1.62 paf 506: const String& regexp,
1.63 paf 507: const String *options,
1.64 paf 508: Table **table,
1.95 parser 509: Row_action row_action, void *info,
510: bool *was_global) const {
1.64 paf 511:
1.73 paf 512: if(!regexp.size())
1.112 parser 513: throw Exception(0, 0,
1.73 paf 514: aorigin,
515: "regexp is empty");
1.118 paf 516: const char *pattern=regexp.cstr();
1.62 paf 517: const char *errptr;
518: int erroffset;
1.63 paf 519: int option_bits[2]; regex_options(options?options->cstr():0, option_bits);
1.95 parser 520: if(was_global)
521: *was_global=option_bits[1]!=0;
1.63 paf 522: pcre *code=pcre_compile(pattern, option_bits[0],
1.62 paf 523: &errptr, &erroffset,
1.132 paf 524: pool().get_source_charset().pcre_tables);
1.62 paf 525:
1.67 paf 526: if(!code)
1.112 parser 527: throw Exception(0, 0,
1.69 paf 528: ®exp.mid(erroffset, regexp.size()),
1.74 paf 529: "regular expression syntax error - %s", errptr);
1.62 paf 530:
1.63 paf 531: int info_substrings=pcre_info(code, 0, 0);
532: if(info_substrings<0) {
1.100 parser 533: pcre_free(code);
1.112 parser 534: throw Exception(0, 0,
1.73 paf 535: aorigin,
1.76 paf 536: "pcre_info error (%d)",
1.73 paf 537: info_substrings);
1.63 paf 538: }
539:
540: int startoffset=0;
1.118 paf 541: const char *subject=cstr();
1.62 paf 542: int length=strlen(subject);
1.63 paf 543: int ovecsize;
544: int *ovector=(int *)malloc(sizeof(int)*
1.114 paf 545: (ovecsize=(1/*match*/+info_substrings)*3), 11);
1.62 paf 546:
1.64 paf 547: { // create table
548: Array& columns=*NEW Array(pool());
549: columns+=string_pre_match_name;
550: columns+=string_match_name;
551: columns+=string_post_match_name;
552: for(int i=1; i<=info_substrings; i++) {
553: char *column=(char *)malloc(MAX_NUMBER);
554: snprintf(column, MAX_NUMBER, "%d", i);
555: columns+=NEW String(pool(), column); // .i column name
556: }
557: *table=NEW Table(pool(), aorigin, &columns);
1.62 paf 558: }
1.63 paf 559:
1.64 paf 560: int exec_option_bits=0;
1.63 paf 561: while(true) {
562: int exec_substrings=pcre_exec(code, 0,
563: subject, length, startoffset,
1.64 paf 564: exec_option_bits, ovector, ovecsize);
1.63 paf 565:
566: if(exec_substrings==PCRE_ERROR_NOMATCH) {
1.100 parser 567: pcre_free(code);
568: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 569: return option_bits[1]!=0; // global=true+table, not global=false
570: }
571:
572: if(exec_substrings<0) {
1.100 parser 573: pcre_free(code);
1.112 parser 574: throw Exception(0, 0,
1.63 paf 575: aorigin,
1.76 paf 576: "regular expression execute error (%d)",
1.63 paf 577: exec_substrings);
578: }
579:
580: Array& row=*NEW Array(pool());
1.81 paf 581: row+=&mid(0, ovector[0]); // .prematch column value
1.69 paf 582: row+=&mid(ovector[0], ovector[1]); // .match
1.81 paf 583: row+=&mid(ovector[1], size()); // .postmatch
1.63 paf 584:
585: for(int i=1; i<exec_substrings; i++) {
1.69 paf 586: // -1:-1 case handled peacefully by mid() itself
587: row+=&mid(ovector[i*2+0], ovector[i*2+1]); // .i column value
1.63 paf 588: }
589:
1.100 parser 590: row_action(**table, &row, startoffset, ovector[0], info);
1.63 paf 591:
1.100 parser 592: if(!option_bits[1] || startoffset==ovector[1]) { // not global | going to hang
593: pcre_free(code);
594: row_action(**table, 0/*last time, no row*/, 0, 0, info);
1.63 paf 595: return true;
596: }
1.100 parser 597: startoffset=ovector[1];
1.63 paf 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:
1.132 paf 606: String& String::change_case(Pool& pool,
1.82 parser 607: Change_case_kind kind) const {
1.132 paf 608: const unsigned char *tables=pool.get_source_charset().pcre_tables;
1.82 parser 609: String& result=*new(pool) String(pool);
610:
611: const unsigned char *a;
612: const unsigned char *b;
613: switch(kind) {
614: case CC_UPPER:
615: a=tables+lcc_offset;
616: b=tables+fcc_offset;
617: break;
618: case CC_LOWER:
619: a=tables+lcc_offset;
620: b=0;
621: break;
622: default:
1.112 parser 623: throw Exception(0, 0,
1.82 parser 624: this,
625: "unknown change case kind #%d",
626: static_cast<int>(kind)); // never
627: a=b=0; // calm, compiler
628: break; // never
629: }
630:
631: const Chunk *chunk=&head;
632: do {
633: const Chunk::Row *row=chunk->rows;
1.116 paf 634: for(uint i=0; i<chunk->count; i++, row++) {
1.82 parser 635: if(row==append_here)
636: goto break2;
637:
1.114 paf 638: char *new_cstr=(char *)pool.malloc(row->item.size, 12);
1.82 parser 639: char *dest=new_cstr;
640: const char *src=row->item.ptr;
641: for(int size=row->item.size; size--; src++) {
642: unsigned char c=a[(unsigned char)*src];
643: if(b)
644: c=b[c];
645:
646: *dest++=(char)c;
647: }
648:
649: result.APPEND(new_cstr, row->item.size,
1.123 paf 650: row->item.lang,
1.82 parser 651: row->item.origin.file, row->item.origin.line);
652: }
653: chunk=row->link;
654: } while(chunk);
655: break2:
1.89 parser 656:
1.101 parser 657: return result;
658: }
659:
1.108 parser 660: void String::join_chain(Pool& pool,
1.116 paf 661: uint& ai, const Chunk*& achunk, const Chunk::Row*& arow,
1.123 paf 662: uchar& joined_lang, const char *& joined_ptr, size_t& joined_size) const {
663: joined_lang=arow->item.lang;
1.108 parser 664:
665: // calc size
666: joined_size=0;
667: {
1.116 paf 668: uint start_i=ai;
1.108 parser 669: const Chunk::Row *start_row=arow;
670: const Chunk *chunk=achunk;
671: do {
672: const Chunk::Row *row=start_row;
1.116 paf 673: for(uint i=start_i; i<chunk->count; i++, row++) {
1.108 parser 674: if(row==append_here)
675: goto break21;
676:
677: if(row->item.lang==joined_lang)
678: joined_size+=row->item.size;
679: else
1.129 paf 680: goto break21;
1.108 parser 681: }
682: if(chunk=row->link) {
683: start_i=0;
684: start_row=chunk->rows;
685: } else
686: break;
687: } while(true);
688: break21:;
689: }
690:
691: // if one row, return simply itself
692: if(joined_size==arow->item.size) {
693: joined_ptr=arow->item.ptr;
694: ai++; arow++;
1.133 paf 695: if(ai==achunk->count) {
696: if(achunk=arow->link) {
697: ai=0;
1.134 paf 698: arow=achunk->rows;
1.133 paf 699: }
700: }
1.108 parser 701: } else {
702: // join adjacent rows
1.114 paf 703: char *ptr=(char *)pool.malloc(joined_size,13);
1.108 parser 704: joined_ptr=ptr;
1.116 paf 705: uint start_i=ai;
1.108 parser 706: const Chunk::Row *start_row=arow;
707: const Chunk *chunk=achunk;
1.116 paf 708: uint i;
1.108 parser 709: const Chunk::Row *row;
710: do {
711: row=start_row;
712: for(i=start_i; i<chunk->count; i++, row++) {
713: if(row==append_here)
714: goto break22;
715:
716: if(row->item.lang==joined_lang) {
717: memcpy(ptr, row->item.ptr, row->item.size);
718: ptr+=row->item.size;
719: } else
1.129 paf 720: goto break22;
1.108 parser 721: }
722: if(chunk=row->link) {
723: start_i=0;
724: start_row=chunk->rows;
725: } else
726: break;
727: } while(true);
728: break22:;
729:
730: // return joined rows
731: ai=i;
732: arow=row;
733: achunk=chunk;
734: }
735: }
736:
737: String& String::reconstruct(Pool& pool) const {
738: //_asm int 3;
739: String& result=*new(pool) String(pool);
740: const Chunk *chunk=&head;
1.133 paf 741: const Chunk::Row *row=chunk->rows;
742: for(uint i=0; i<chunk->count; ) {
743: if(row==append_here)
744: break;
1.108 parser 745:
1.133 paf 746: uchar joined_lang;
747: const char *joined_ptr;
748: size_t joined_size;
1.130 paf 749: #ifndef NO_STRING_ORIGIN
1.133 paf 750: const char *joined_origin_file=row->item.origin.file;
751: const size_t joined_origin_line=row->item.origin.line;
1.130 paf 752: #endif
1.133 paf 753: join_chain(pool, i, chunk, row,
754: joined_lang, joined_ptr, joined_size);
755:
756: result.APPEND(joined_ptr, joined_size, joined_lang,
757: joined_origin_file, joined_origin_line);
1.108 parser 758:
1.133 paf 759: if(!chunk)
760: break;
1.130 paf 761: }
1.108 parser 762:
763: return result;
764: };
765:
766: String& String::replace_in_reconstructed(Pool& pool, Dictionary& dict) const {
1.106 parser 767: //_asm int 3;
1.101 parser 768: String& result=*new(pool) String(pool);
1.123 paf 769: STRING_FOREACH_ROW(
770: const char *src=row->item.ptr;
771: size_t src_size=row->item.size;
772: char *new_cstr=(char *)pool.malloc((size_t)ceil(src_size*dict.max_ratio()), 14);
773: char *dest=new_cstr;
774: while(src_size) {
775: // there is a row where first column starts 'src'
776: if(Table::Item *item=dict.first_that_starts(src, src_size)) {
777: // get a=>b values
778: const String& a=*static_cast<Array *>(item)->get_string(0);
779: const String& b=*static_cast<Array *>(item)->get_string(1);
780: // skip 'a' in 'src' && reduce work size
781: src+=a.size(); src_size-=a.size();
782: // write 'b' to 'dest' && skip 'b' in 'dest'
783: b.store_to(dest); dest+=b.size();
784: } else {
785: // write a char to b && reduce work size
786: *dest++=*src++; src_size--;
1.101 parser 787: }
788: }
789:
1.123 paf 790: result.APPEND(new_cstr, dest-new_cstr,
791: row->item.lang,
792: row->item.origin.file, row->item.origin.line);
793: );
1.101 parser 794: break2:
1.89 parser 795: return result;
1.108 parser 796: }
797:
798: String& String::replace(Pool& pool, Dictionary& dict) const {
799: return reconstruct(pool).replace_in_reconstructed(pool, dict);
1.89 parser 800: }
801:
1.90 parser 802: double String::as_double() const {
1.89 parser 803: double result;
1.114 paf 804: const char *cstr;
805: char buf[MAX_NUMBER];
1.115 paf 806: if(head.rows+1==append_here) {
1.114 paf 807: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
808: memcpy(buf, head.rows[0].item.ptr, size);
809: buf[size]=0;
810: cstr=buf;
811: } else
812: cstr=this->cstr();
1.102 parser 813: char *error_pos;
1.89 parser 814: // 0xABC
1.99 parser 815: if(cstr[0]=='0')
816: if(cstr[1]=='x' || cstr[1]=='X')
817: result=(double)(unsigned long)strtol(cstr, &error_pos, 0);
818: else
1.102 parser 819: result=(double)strtod(cstr+1/*skip leading 0*/, &error_pos);
1.89 parser 820: else
1.99 parser 821: result=(double)strtod(cstr, &error_pos);
1.89 parser 822:
1.103 parser 823: if(*error_pos/*not EOS*/)
1.112 parser 824: throw Exception(0, 0,
1.89 parser 825: this,
826: "invalid number (double)");
827:
828: return result;
829: }
1.90 parser 830: int String::as_int() const {
1.89 parser 831: int result;
1.114 paf 832: const char *cstr;
833: char buf[MAX_NUMBER];
1.115 paf 834: if(head.rows+1==append_here) {
1.114 paf 835: int size=min(head.rows[0].item.size, MAX_NUMBER-1);
836: memcpy(buf, head.rows[0].item.ptr, size);
837: buf[size]=0;
838: cstr=buf;
839: } else
840: cstr=this->cstr();
1.102 parser 841: char *error_pos;
1.89 parser 842: // 0xABC
1.99 parser 843: if(cstr[0]=='0')
844: if(cstr[1]=='x' || cstr[1]=='X')
845: result=(int)(unsigned long)strtol(cstr, &error_pos, 0);
846: else
1.102 parser 847: result=(int)strtol(cstr+1/*skip leading 0*/, &error_pos, 0);
1.89 parser 848: else
849: result=(int)strtol(cstr, &error_pos, 0);
850:
1.103 parser 851: if(*error_pos/*not EOS*/)
1.112 parser 852: throw Exception(0, 0,
1.89 parser 853: this,
854: "invalid number (int)");
1.82 parser 855:
856: return result;
1.61 paf 857: }
1.113 parser 858:
1.128 paf 859: inline void ushort2uchars(ushort word, uchar& byte1, uchar& byte2) {
860: byte1=word&0xFF;
861: byte2=word>>8;
862: }
863: inline ushort uchars2ushort(uchar byte1, uchar byte2) {
864: return (byte2<<8) | byte1;
865: }
1.113 parser 866: /* @todo maybe network order worth spending some effort?
867: don't bothering myself with network byte order,
868: am not planning to be able to move resulting file across platforms
869: for now
870: */
871: void String::serialize(size_t prolog_size, void *& buf, size_t& buf_size) const {
872: buf_size=
873: prolog_size
1.126 paf 874: +used_rows()*(sizeof(uchar)+sizeof(ushort))
1.113 parser 875: +size();
1.114 paf 876: buf=malloc(buf_size,15);
1.113 parser 877: char *cur=(char *)buf+prolog_size;
878:
1.123 paf 879: STRING_FOREACH_ROW(
880: // lang
1.126 paf 881: memcpy(cur, &row->item.lang, sizeof(uchar));
882: cur+=sizeof(uchar);
1.123 paf 883: // size
1.128 paf 884: uchar byte1; uchar byte2;
885: ushort2uchars(row->item.size, byte1, byte2);
886: memcpy(cur, &byte1, sizeof(uchar)); cur+=sizeof(uchar);
887: memcpy(cur, &byte2, sizeof(uchar)); cur+=sizeof(uchar);
1.123 paf 888: // bytes
889: memcpy(cur, row->item.ptr, row->item.size);
890: cur+=row->item.size;
891: );
1.113 parser 892: break2:
893: ;
894: }
895: void String::deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file) {
1.135 paf 896: if(buf_size<=prolog_size)
897: return;
898:
1.126 paf 899: char *cur=(char *)buf+prolog_size;
1.113 parser 900: buf_size-=prolog_size;
901:
902: while(buf_size) {
1.128 paf 903: uchar lang=*(uchar *)(cur);
904:
905: ushort size=uchars2ushort(
906: *(uchar*)(cur+sizeof(uchar)*1),
907: *(uchar*)(cur+sizeof(uchar)*2)
908: );
909:
910: const char *ptr=(const char*)(cur+sizeof(uchar)*3);
1.126 paf 911: APPEND(ptr, size, lang, file, 0);
1.113 parser 912:
1.126 paf 913: size_t piece_size=sizeof(uchar)+sizeof(ushort)+size;
1.113 parser 914: cur+=piece_size;
915: buf_size-=piece_size;
916: }
917: }
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