Annotation of parser3/src/main/pa_string.C, revision 1.53
1.45 paf 1: /** @file
1.46 paf 2: Parser: string class. @see untaint.C.
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:
1.53 ! paf 8: $Id: pa_string.C,v 1.52 2001/03/25 08:52:36 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:
1.41 paf 21: String::String(Pool& apool, const char *src, bool tainted) :
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)
31: if(tainted)
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() {
1.28 paf 38: int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100;
39: last_chunk=static_cast<Chunk *>(
1.30 paf 40: malloc(sizeof(int)+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:
1.44 paf 51: int src_used_rows=src.fused_rows;
1.8 paf 52: if(src_used_rows<=head.count) {
1.10 paf 53: // all new rows fit into preallocated area
1.28 paf 54: int 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: //
67: // new rows don't fit into preallocated area: splitting into two chunks
68: // preallocated chunk src to constructing head
69: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
70: // remaining rows into new_chunk
1.28 paf 71: int curr_chunk_rows=src_used_rows-head.count;
1.8 paf 72: Chunk *new_chunk=static_cast<Chunk *>(
1.30 paf 73: malloc(sizeof(int)+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;
1.28 paf 80: int rows_left_to_copy=new_chunk->count;
1.8 paf 81: while(true) {
82: int old_count=old_chunk->count;
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) {
1.44 paf 106: int src_used_rows=src.fused_rows;
1.28 paf 107: int dst_free_rows=link_row-append_here;
108:
109: if(src_used_rows<=dst_free_rows) {
110: // all new rows fit into last chunk
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 {
1.31 paf 115: // not all new rows fit into last chunk: shrinking it to used part,
1.28 paf 116: int used_rows=last_chunk->count-dst_free_rows;
117: //int *countp=append_here
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 *>(
1.30 paf 122: malloc(sizeof(int)+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;
129: int rows_left_to_copy=src_used_rows;
130: while(true) {
131: int old_count=old_chunk->count;
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;
1.47 paf 163: if(item_lang==UL_YES || forced) // tainted? need untaint language assignment
1.34 paf 164: item_lang=lang; // assign untaint language
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;
172: for(int i=0; i<chunk->count; i++) {
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;
1.5 paf 214: for(int i=0; i<chunk->count; i++) {
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:
1.32 paf 227: int String::cmp(const String& src) const {
1.16 paf 228: const Chunk *a_chunk=&head;
229: const Chunk *b_chunk=&src.head;
230: const Chunk::Row *a_row=a_chunk->rows;
231: const Chunk::Row *b_row=b_chunk->rows;
1.9 paf 232: int a_offset=0;
233: int b_offset=0;
234: Chunk::Row *a_end=append_here;
235: Chunk::Row *b_end=src.append_here;
1.11 paf 236: int a_countdown=a_chunk->count;
237: int b_countdown=b_chunk->count;
1.9 paf 238: bool a_break=false;
239: bool b_break=false;
1.32 paf 240: int result;
1.9 paf 241: while(true) {
1.33 paf 242: a_break=a_row==a_end;
243: b_break=b_row==b_end;
244: if(a_break || b_break)
245: break;
246:
1.9 paf 247: int size_diff=
248: (a_row->item.size-a_offset)-
249: (b_row->item.size-b_offset);
250:
251: if(size_diff==0) { // a has same size as b
1.32 paf 252: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
253: if(result)
254: return result;
1.11 paf 255: a_row++; a_countdown--; a_offset=0;
256: b_row++; b_countdown--; b_offset=0;
1.9 paf 257: } else if (size_diff>0) { // a longer
1.32 paf 258: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset);
259: if(result)
260: return result;
1.9 paf 261: a_offset+=b_row->item.size-b_offset;
1.11 paf 262: b_row++; b_countdown--; b_offset=0;
1.9 paf 263: } else { // b longer
1.32 paf 264: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
265: if(result)
266: return result;
1.9 paf 267: b_offset+=a_row->item.size-a_offset;
1.11 paf 268: a_row++; a_countdown--; a_offset=0;
1.9 paf 269: }
270:
1.11 paf 271: if(!a_countdown) {
1.9 paf 272: a_chunk=a_row->link;
273: a_row=a_chunk->rows;
1.11 paf 274: a_countdown=a_chunk->count;
1.9 paf 275: }
1.11 paf 276: if(!b_countdown) {
1.9 paf 277: b_chunk=b_row->link;
278: b_row=b_chunk->rows;
1.11 paf 279: b_countdown=b_chunk->count;
1.27 paf 280: }
281: }
1.32 paf 282: if(a_break==b_break) // ended simultaneously
283: result=0;
284: else if(a_break) // first bytes equal, but a ended before b
285: result=-1;
286: else
287: result=+1;
288: return result;
1.27 paf 289: }
290:
1.50 paf 291: int String::cmp(const char* b_ptr, int& partial, size_t src_size) const {
292: size_t b_size=src_size?src_size:b_ptr?strlen(b_ptr):0;
1.27 paf 293:
1.51 paf 294: partial=-1;
1.27 paf 295: const Chunk *a_chunk=&head;
296: const Chunk::Row *a_row=a_chunk->rows;
297: int a_offset=0;
298: int b_offset=0;
299: Chunk::Row *a_end=append_here;
300: int a_countdown=a_chunk->count;
301: bool a_break=false;
302: bool b_break=false;
303: while(true) {
1.52 paf 304: a_break=a_row==a_end;
305: if(a_break || b_break)
306: break;
307:
1.27 paf 308: int size_diff=
309: (a_row->item.size-a_offset)-
310: (b_size-b_offset);
311:
312: if(size_diff==0) { // a has same size as b
1.50 paf 313: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
314: return result;
1.27 paf 315: a_row++; a_countdown--; a_offset=0;
316: b_break=true;
317: } else if (size_diff>0) { // a longer
1.50 paf 318: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, b_size-b_offset)!=0)
319: return result;
1.27 paf 320: a_offset+=b_size-b_offset;
321: b_break=true;
322: } else { // b longer
1.50 paf 323: if(int result=memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
324: return result;
1.27 paf 325: b_offset+=a_row->item.size-a_offset;
326: a_row++; a_countdown--; a_offset=0;
327: }
328:
329: if(!a_countdown) {
330: a_chunk=a_row->link;
331: a_row=a_chunk->rows;
332: a_countdown=a_chunk->count;
1.9 paf 333: }
334: }
1.50 paf 335: if(a_break==b_break) // ended simultaneously
1.51 paf 336: return partial=0;
1.50 paf 337: else if(a_break) // first bytes equal, but a ended before b
1.51 paf 338: return partial=1;
1.50 paf 339: else
1.51 paf 340: return partial=2;
1.5 paf 341: }
1.46 paf 342:
343: #ifndef NO_STRING_ORIGIN
344: const Origin& String::origin() const {
345: if(!fused_rows)
346: THROW(0, 0,
1.50 paf 347: 0,
348: "String::origin() of empty string called");
1.46 paf 349:
1.49 paf 350: // determining origin by last appended piece
1.50 paf 351: // because first one frequently constant.
352: // ex: ^load[/file] "document_root" + "/file"
1.49 paf 353: return append_here[-1].item.origin;
1.46 paf 354: }
355: #endif
1.53 ! paf 356:
! 357: String& String::piece(size_t start, size_t finish) const {
! 358: start=max(0, start);
! 359: finish=min(size(), finish);
! 360:
! 361: String& result=*NEW String(pool());
! 362:
! 363: size_t pos=0;
! 364: const Chunk *chunk=&head;
! 365: do {
! 366: const Chunk::Row *row=chunk->rows;
! 367: for(int i=0; i<chunk->count; pos+=row->item.size, i++, row++) {
! 368: if(row==append_here)
! 369: goto break2;
! 370:
! 371: if(start>=pos) { // started now or already?
! 372: size_t item_finish=pos+row->item.size;
! 373: bool started=start < item_finish; // started now?
! 374: bool finished=finish < item_finish; // finished now?
! 375: size_t offset=started?start-pos:0;
! 376: size_t size=finished?finish-pos:row->item.size;
! 377: result.APPEND(
! 378: row->item.ptr+offset, size-offset,
! 379: row->item.lang,
! 380: row->item.origin.file, row->item.origin.line);
! 381: if(finished)
! 382: goto break2;
! 383: }
! 384: }
! 385: chunk=row->link;
! 386: } while(chunk);
! 387: break2:
! 388: return result;
! 389: }
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