Annotation of parser3/src/main/pa_string.C, revision 1.55
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:
1.55 ! paf 8: $Id: pa_string.C,v 1.54 2001/03/29 15:36:16 paf Exp $
1.4 paf 9: */
10:
1.48 paf 11: #include "pa_config_includes.h"
1.1 paf 12:
1.13 paf 13: #include "pa_pool.h"
1.12 paf 14: #include "pa_string.h"
1.5 paf 15: #include "pa_hash.h"
1.22 paf 16: #include "pa_exception.h"
1.53 paf 17: #include "pa_common.h"
1.1 paf 18:
1.18 paf 19: // String
20:
1.55 ! paf 21: String::String(Pool& apool, const char *src, bool tasize_ted) :
1.17 paf 22: Pooled(apool) {
1.28 paf 23: last_chunk=&head;
24: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 25: append_here=head.rows;
1.2 paf 26: head.preallocated_link=0;
1.28 paf 27: link_row=&head.rows[head.count];
1.8 paf 28: fused_rows=fsize=0;
1.41 paf 29:
30: if(src)
1.55 ! paf 31: if(tasize_ted)
1.41 paf 32: APPEND_TAINTED(src, 0, 0, 0);
33: else
1.53 paf 34: APPEND_CONST(src);
1.1 paf 35: }
36:
37: void String::expand() {
1.55 ! paf 38: size_t new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100;
1.28 paf 39: last_chunk=static_cast<Chunk *>(
1.55 ! paf 40: malloc(sizeof(size_t)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *)));
1.28 paf 41: last_chunk->count=new_chunk_count;
42: link_row->link=last_chunk;
43: append_here=last_chunk->rows;
44: link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 45: link_row->link=0;
1.1 paf 46: }
47:
1.40 paf 48: String::String(const String& src) : Pooled(src.pool()) {
1.8 paf 49: head.count=CR_PREALLOCATED_COUNT;
50:
1.55 ! paf 51: size_t src_used_rows=src.fused_rows;
1.8 paf 52: if(src_used_rows<=head.count) {
1.55 ! paf 53: // all new rows fit size_to preallocated area
! 54: size_t curr_chunk_rows=head.count;
1.8 paf 55: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
56: append_here=&head.rows[src_used_rows];
57: link_row=&head.rows[curr_chunk_rows];
58: } else {
59: // warning:
1.10 paf 60: // heavily relies on the fact
61: // "preallocated area is the same for all strings"
1.8 paf 62: //
63: // info:
64: // allocating only enough mem to fit src string rows
65: // next append would allocate a new chunk
66: //
1.55 ! paf 67: // new rows don't fit size_to preallocated area: splitting size_to two chunks
1.8 paf 68: // preallocated chunk src to constructing head
69: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
1.55 ! paf 70: // remaining rows size_to new_chunk
! 71: size_t curr_chunk_rows=src_used_rows-head.count;
1.8 paf 72: Chunk *new_chunk=static_cast<Chunk *>(
1.55 ! paf 73: malloc(sizeof(size_t)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *)));
1.8 paf 74: new_chunk->count=curr_chunk_rows;
75: head.preallocated_link=new_chunk;
1.28 paf 76: append_here=link_row=&new_chunk->rows[new_chunk->count];
1.8 paf 77:
78: Chunk *old_chunk=src.head.preallocated_link;
79: Chunk::Row *new_rows=new_chunk->rows;
1.55 ! paf 80: size_t rows_left_to_copy=new_chunk->count;
1.8 paf 81: while(true) {
1.55 ! paf 82: size_t old_count=old_chunk->count;
1.8 paf 83: Chunk *next_chunk=old_chunk->rows[old_count].link;
84: if(next_chunk) {
85: // not last source chunk
86: // taking it all
87: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
88: new_rows+=old_count;
89: rows_left_to_copy-=old_count;
90:
91: old_chunk=next_chunk;
92: } else {
93: // the last source chunk
94: // taking only those rows of chunk that _left_to_copy
95: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
96: break;
97: }
98: }
1.5 paf 99: }
1.8 paf 100: link_row->link=0;
101: fused_rows=src_used_rows;
102: fsize=src.fsize;
1.5 paf 103: }
1.28 paf 104:
1.42 paf 105: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.55 ! paf 106: size_t src_used_rows=src.fused_rows;
! 107: size_t dst_free_rows=link_row-append_here;
1.28 paf 108:
109: if(src_used_rows<=dst_free_rows) {
1.55 ! paf 110: // all new rows fit size_to last chunk
1.28 paf 111: memcpy(append_here, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
1.42 paf 112: set_lang(append_here, lang, forced, src_used_rows);
1.28 paf 113: append_here+=src_used_rows;
114: } else {
1.55 ! paf 115: // not all new rows fit size_to last chunk: shrinking it to used part,
! 116: size_t used_rows=last_chunk->count-dst_free_rows;
! 117: //size_t *countp=append_here
1.28 paf 118: link_row=&last_chunk->rows[last_chunk->count=used_rows];
119: // allocating only enough mem to fit src string rows
120: // next append would allocate a new chunk
121: last_chunk=static_cast<Chunk *>(
1.55 ! paf 122: malloc(sizeof(size_t)+sizeof(Chunk::Row)*src_used_rows+sizeof(Chunk *)));
1.28 paf 123: last_chunk->count=src_used_rows;
124: link_row->link=last_chunk;
125: append_here=link_row=&last_chunk->rows[src_used_rows];
126:
1.31 paf 127: const Chunk *old_chunk=&src.head;
1.28 paf 128: Chunk::Row *new_rows=last_chunk->rows;
1.55 ! paf 129: size_t rows_left_to_copy=src_used_rows;
1.28 paf 130: while(true) {
1.55 ! paf 131: size_t old_count=old_chunk->count;
1.28 paf 132: Chunk *next_chunk=old_chunk->rows[old_count].link;
133: if(next_chunk) {
134: // not last source chunk
135: // taking it all
136: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
1.42 paf 137: set_lang(new_rows, lang, forced, old_count);
1.28 paf 138: new_rows+=old_count;
139: rows_left_to_copy-=old_count;
140:
141: old_chunk=next_chunk;
142: } else {
143: // the last source chunk
144: // taking only those rows of chunk that _left_to_copy
145: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
1.42 paf 146: set_lang(new_rows, lang, forced, rows_left_to_copy);
1.28 paf 147: break;
148: }
149: }
1.29 paf 150: link_row->link=0;
1.28 paf 151: }
152: fused_rows+=src_used_rows;
153: fsize+=src.fsize;
154:
155: return *this;
1.23 paf 156: }
1.42 paf 157: void String::set_lang(Chunk::Row *row, Untaint_lang lang, bool forced, size_t size) {
1.47 paf 158: if(lang==UL_PASS_APPENDED)
1.34 paf 159: return;
160:
161: while(size--) {
162: Untaint_lang& item_lang=(row++)->item.lang;
1.55 ! paf 163: if(item_lang==UL_YES || forced) // tasize_ted? need untasize_t language assignment
! 164: item_lang=lang; // assign untasize_t language
1.34 paf 165: }
1.40 paf 166: }
167:
1.42 paf 168: /*void String::change_lang(Untaint_lang lang) {
1.40 paf 169: Chunk *chunk=&head;
170: do {
171: Chunk::Row *row=chunk->rows;
1.55 ! paf 172: for(size_t i=0; i<chunk->count; i++) {
1.40 paf 173: if(row==append_here)
174: goto break2;
175:
176: row->item.lang=lang;
177: row++;
178: }
179: chunk=row->link;
180: } while(chunk);
181: break2:
182: return;
1.34 paf 183: }
1.42 paf 184: */
1.13 paf 185: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 186: if(!src)
187: return *this;
1.26 paf 188: if(!size)
189: size=strlen(src);
190: if(!size)
1.9 paf 191: return *this;
192:
1.1 paf 193: if(chunk_is_full())
194: expand();
195:
196: append_here->item.ptr=src;
1.26 paf 197: fsize+=append_here->item.size=size;
1.52 paf 198: append_here->item.lang=lang;
1.13 paf 199: #ifndef NO_STRING_ORIGIN
1.14 paf 200: append_here->item.origin.file=file;
201: append_here->item.origin.line=line;
1.13 paf 202: #endif
1.8 paf 203: append_here++; fused_rows++;
1.1 paf 204:
205: return *this;
206: }
207:
1.16 paf 208: uint String::hash_code() const {
1.7 paf 209: uint result=0;
1.5 paf 210:
1.16 paf 211: const Chunk *chunk=&head;
1.5 paf 212: do {
1.16 paf 213: const Chunk::Row *row=chunk->rows;
1.55 ! paf 214: for(size_t i=0; i<chunk->count; i++) {
1.5 paf 215: if(row==append_here)
216: goto break2;
217:
1.6 paf 218: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 219: row++;
220: }
221: chunk=row->link;
222: } while(chunk);
223: break2:
224: return result;
225: }
226:
1.55 ! paf 227: int String::cmp(int& partial, const String& src, size_t this_offset) const {
! 228: this_offset=min(this_offset, size()-1);
! 229:
1.16 paf 230: const Chunk *a_chunk=&head;
231: const Chunk *b_chunk=&src.head;
232: const Chunk::Row *a_row=a_chunk->rows;
233: const Chunk::Row *b_row=b_chunk->rows;
1.55 ! paf 234: size_t a_offset=this_offset;
! 235: size_t b_offset=0;
1.9 paf 236: Chunk::Row *a_end=append_here;
237: Chunk::Row *b_end=src.append_here;
1.55 ! paf 238: size_t a_countdown=a_chunk->count;
! 239: size_t b_countdown=b_chunk->count;
1.9 paf 240: bool a_break=false;
241: bool b_break=false;
1.55 ! paf 242: size_t result;
! 243: for(size_t pos=0; true; pos+=a_row->item.size) {
1.33 paf 244: a_break=a_row==a_end;
245: b_break=b_row==b_end;
246: if(a_break || b_break)
247: break;
248:
1.55 ! paf 249: if(pos+a_row->item.size > this_offset) {
! 250: int size_diff=
! 251: (a_row->item.size-a_offset)-
! 252: (b_row->item.size-b_offset);
! 253:
! 254: if(size_diff==0) { // a has same size as b
! 255: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
! 256: if(result)
! 257: return result;
! 258: a_row++; a_countdown--; a_offset=0;
! 259: b_row++; b_countdown--; b_offset=0;
! 260: } else if (size_diff>0) { // a longer
! 261: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset);
! 262: if(result)
! 263: return result;
! 264: a_offset+=b_row->item.size-b_offset;
! 265: b_row++; b_countdown--; b_offset=0;
! 266: } else { // b longer
! 267: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
! 268: if(result)
! 269: return result;
! 270: b_offset+=a_row->item.size-a_offset;
! 271: a_row++; a_countdown--; a_offset=0;
! 272: }
1.9 paf 273:
1.55 ! paf 274: if(!b_countdown) {
! 275: b_chunk=b_row->link;
! 276: b_row=b_chunk->rows;
! 277: b_countdown=b_chunk->count;
! 278: }
! 279: } else {
! 280: a_row++; a_countdown--; a_offset-=a_row->item.size;
1.9 paf 281: }
282:
1.11 paf 283: if(!a_countdown) {
1.9 paf 284: a_chunk=a_row->link;
285: a_row=a_chunk->rows;
1.11 paf 286: a_countdown=a_chunk->count;
1.9 paf 287: }
1.27 paf 288: }
1.55 ! paf 289: if(a_break==b_break) { // ended simultaneously
! 290: partial=0; return 0;
! 291: } else if(a_break) { // first bytes equal, but a ended before b
! 292: partial=1; return -1;
! 293: } else {
! 294: partial=2; return +1;
! 295: }
1.27 paf 296: }
297:
1.55 ! paf 298: int String::cmp(int& partial, const char* b_ptr, size_t src_size) const {
1.50 paf 299: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.27 paf 300:
1.51 paf 301: partial=-1;
1.27 paf 302: const Chunk *a_chunk=&head;
303: const Chunk::Row *a_row=a_chunk->rows;
1.55 ! paf 304: size_t a_offset=0;
! 305: size_t b_offset=0;
1.27 paf 306: Chunk::Row *a_end=append_here;
1.55 ! paf 307: size_t a_countdown=a_chunk->count;
1.27 paf 308: bool a_break=false;
309: bool b_break=false;
310: while(true) {
1.52 paf 311: a_break=a_row==a_end;
312: if(a_break || b_break)
313: break;
314:
1.27 paf 315: int size_diff=
316: (a_row->item.size-a_offset)-
317: (b_size-b_offset);
318:
319: if(size_diff==0) { // a has same size as b
1.55 ! paf 320: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
1.50 paf 321: return result;
1.27 paf 322: a_row++; a_countdown--; a_offset=0;
323: b_break=true;
324: } else if (size_diff>0) { // a longer
1.55 ! paf 325: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, b_size-b_offset)!=0)
1.50 paf 326: return result;
1.27 paf 327: a_offset+=b_size-b_offset;
328: b_break=true;
329: } else { // b longer
1.55 ! paf 330: if(size_t result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
1.50 paf 331: return result;
1.27 paf 332: b_offset+=a_row->item.size-a_offset;
333: a_row++; a_countdown--; a_offset=0;
334: }
335:
336: if(!a_countdown) {
337: a_chunk=a_row->link;
338: a_row=a_chunk->rows;
339: a_countdown=a_chunk->count;
1.9 paf 340: }
341: }
1.55 ! paf 342: if(a_break==b_break) { // ended simultaneously
! 343: partial=0; return 0;
! 344: } else if(a_break) { // first bytes equal, but a ended before b
! 345: partial=1; return -1;
! 346: } else {
! 347: partial=2; return +1;
! 348: }
1.5 paf 349: }
1.46 paf 350:
351: #ifndef NO_STRING_ORIGIN
352: const Origin& String::origin() const {
353: if(!fused_rows)
354: THROW(0, 0,
1.50 paf 355: 0,
356: "String::origin() of empty string called");
1.46 paf 357:
1.49 paf 358: // determining origin by last appended piece
1.50 paf 359: // because first one frequently constant.
360: // ex: ^load[/file] "document_root" + "/file"
1.49 paf 361: return append_here[-1].item.origin;
1.46 paf 362: }
363: #endif
1.53 paf 364:
365: String& String::piece(size_t start, size_t finish) const {
366: start=max(0, start);
367: finish=min(size(), finish);
368:
369: String& result=*NEW String(pool());
370:
371: size_t pos=0;
372: const Chunk *chunk=&head;
373: do {
374: const Chunk::Row *row=chunk->rows;
1.55 ! paf 375: for(size_t i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
1.53 paf 376: if(row==append_here)
377: goto break2;
378:
379: if(start>=pos) { // started now or already?
380: size_t item_finish=pos+row->item.size;
381: bool started=start < item_finish; // started now?
382: bool finished=finish < item_finish; // finished now?
383: size_t offset=started?start-pos:0;
384: size_t size=finished?finish-pos:row->item.size;
385: result.APPEND(
386: row->item.ptr+offset, size-offset,
387: row->item.lang,
388: row->item.origin.file, row->item.origin.line);
389: if(finished)
390: goto break2;
391: }
392: }
393: chunk=row->link;
394: } while(chunk);
395: break2:
396: return result;
1.54 paf 397: }
398:
399: size_t String::pos(const String& substr) const {
1.55 ! paf 400: for(size_t result=0; result<size(); result++) {
! 401: int partial; cmp(partial, substr, result);
! 402: if(
! 403: partial==0 || // full match
! 404: partial==2) // 'substr' starts 'this'+'result'
! 405: return result;
! 406: }
! 407:
! 408: return -1;
1.53 paf 409: }
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