Annotation of parser3/src/main/pa_string.C, revision 1.48
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:
1.48 ! paf 8: $Id: pa_string.C,v 1.47 2001/03/21 14:06:46 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"
1.1 paf 17:
1.18 paf 18: // String
19:
1.41 paf 20: String::String(Pool& apool, const char *src, bool tainted) :
1.17 paf 21: Pooled(apool) {
1.28 paf 22: last_chunk=&head;
23: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 24: append_here=head.rows;
1.2 paf 25: head.preallocated_link=0;
1.28 paf 26: link_row=&head.rows[head.count];
1.8 paf 27: fused_rows=fsize=0;
1.41 paf 28:
29: if(src)
30: if(tainted)
31: APPEND_TAINTED(src, 0, 0, 0);
32: else
33: APPEND(src, 0, 0, 0);
1.1 paf 34: }
35:
36: void String::expand() {
1.28 paf 37: int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100;
38: last_chunk=static_cast<Chunk *>(
1.30 paf 39: malloc(sizeof(int)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *)));
1.28 paf 40: last_chunk->count=new_chunk_count;
41: link_row->link=last_chunk;
42: append_here=last_chunk->rows;
43: link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 44: link_row->link=0;
1.1 paf 45: }
46:
1.40 paf 47: String::String(const String& src) : Pooled(src.pool()) {
1.8 paf 48: head.count=CR_PREALLOCATED_COUNT;
49:
1.44 paf 50: int src_used_rows=src.fused_rows;
1.8 paf 51: if(src_used_rows<=head.count) {
1.10 paf 52: // all new rows fit into preallocated area
1.28 paf 53: int curr_chunk_rows=head.count;
1.8 paf 54: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
55: append_here=&head.rows[src_used_rows];
56: link_row=&head.rows[curr_chunk_rows];
57: } else {
58: // warning:
1.10 paf 59: // heavily relies on the fact
60: // "preallocated area is the same for all strings"
1.8 paf 61: //
62: // info:
63: // allocating only enough mem to fit src string rows
64: // next append would allocate a new chunk
65: //
66: // new rows don't fit into preallocated area: splitting into two chunks
67: // preallocated chunk src to constructing head
68: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
69: // remaining rows into new_chunk
1.28 paf 70: int curr_chunk_rows=src_used_rows-head.count;
1.8 paf 71: Chunk *new_chunk=static_cast<Chunk *>(
1.30 paf 72: malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *)));
1.8 paf 73: new_chunk->count=curr_chunk_rows;
74: head.preallocated_link=new_chunk;
1.28 paf 75: append_here=link_row=&new_chunk->rows[new_chunk->count];
1.8 paf 76:
77: Chunk *old_chunk=src.head.preallocated_link;
78: Chunk::Row *new_rows=new_chunk->rows;
1.28 paf 79: int rows_left_to_copy=new_chunk->count;
1.8 paf 80: while(true) {
81: int old_count=old_chunk->count;
82: Chunk *next_chunk=old_chunk->rows[old_count].link;
83: if(next_chunk) {
84: // not last source chunk
85: // taking it all
86: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
87: new_rows+=old_count;
88: rows_left_to_copy-=old_count;
89:
90: old_chunk=next_chunk;
91: } else {
92: // the last source chunk
93: // taking only those rows of chunk that _left_to_copy
94: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
95: break;
96: }
97: }
1.5 paf 98: }
1.8 paf 99: link_row->link=0;
100: fused_rows=src_used_rows;
101: fsize=src.fsize;
1.5 paf 102: }
1.28 paf 103:
1.42 paf 104: String& String::append(const String& src, Untaint_lang lang, bool forced) {
1.44 paf 105: int src_used_rows=src.fused_rows;
1.28 paf 106: int dst_free_rows=link_row-append_here;
107:
108: if(src_used_rows<=dst_free_rows) {
109: // all new rows fit into last chunk
110: memcpy(append_here, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
1.42 paf 111: set_lang(append_here, lang, forced, src_used_rows);
1.28 paf 112: append_here+=src_used_rows;
113: } else {
1.31 paf 114: // not all new rows fit into last chunk: shrinking it to used part,
1.28 paf 115: int used_rows=last_chunk->count-dst_free_rows;
116: //int *countp=append_here
117: link_row=&last_chunk->rows[last_chunk->count=used_rows];
118: // allocating only enough mem to fit src string rows
119: // next append would allocate a new chunk
120: last_chunk=static_cast<Chunk *>(
1.30 paf 121: malloc(sizeof(int)+sizeof(Chunk::Row)*src_used_rows+sizeof(Chunk *)));
1.28 paf 122: last_chunk->count=src_used_rows;
123: link_row->link=last_chunk;
124: append_here=link_row=&last_chunk->rows[src_used_rows];
125:
1.31 paf 126: const Chunk *old_chunk=&src.head;
1.28 paf 127: Chunk::Row *new_rows=last_chunk->rows;
128: int rows_left_to_copy=src_used_rows;
129: while(true) {
130: int old_count=old_chunk->count;
131: Chunk *next_chunk=old_chunk->rows[old_count].link;
132: if(next_chunk) {
133: // not last source chunk
134: // taking it all
135: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
1.42 paf 136: set_lang(new_rows, lang, forced, old_count);
1.28 paf 137: new_rows+=old_count;
138: rows_left_to_copy-=old_count;
139:
140: old_chunk=next_chunk;
141: } else {
142: // the last source chunk
143: // taking only those rows of chunk that _left_to_copy
144: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
1.42 paf 145: set_lang(new_rows, lang, forced, rows_left_to_copy);
1.28 paf 146: break;
147: }
148: }
1.29 paf 149: link_row->link=0;
1.28 paf 150: }
151: fused_rows+=src_used_rows;
152: fsize+=src.fsize;
153:
154: return *this;
1.23 paf 155: }
1.42 paf 156: void String::set_lang(Chunk::Row *row, Untaint_lang lang, bool forced, size_t size) {
1.47 paf 157: if(lang==UL_PASS_APPENDED)
1.34 paf 158: return;
159:
160: while(size--) {
161: Untaint_lang& item_lang=(row++)->item.lang;
1.47 paf 162: if(item_lang==UL_YES || forced) // tainted? need untaint language assignment
1.34 paf 163: item_lang=lang; // assign untaint language
164: }
1.40 paf 165: }
166:
1.42 paf 167: /*void String::change_lang(Untaint_lang lang) {
1.40 paf 168: Chunk *chunk=&head;
169: do {
170: Chunk::Row *row=chunk->rows;
171: for(int i=0; i<chunk->count; i++) {
172: if(row==append_here)
173: goto break2;
174:
175: row->item.lang=lang;
176: row++;
177: }
178: chunk=row->link;
179: } while(chunk);
180: break2:
181: return;
1.34 paf 182: }
1.42 paf 183: */
1.13 paf 184: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 185: if(!src)
186: return *this;
1.26 paf 187: if(!size)
188: size=strlen(src);
189: if(!size)
1.9 paf 190: return *this;
191:
1.1 paf 192: if(chunk_is_full())
193: expand();
194:
195: append_here->item.ptr=src;
1.26 paf 196: fsize+=append_here->item.size=size;
1.47 paf 197: append_here->item.lang=tainted?UL_YES:UL_NO;
1.13 paf 198: #ifndef NO_STRING_ORIGIN
1.14 paf 199: append_here->item.origin.file=file;
200: append_here->item.origin.line=line;
1.13 paf 201: #endif
1.8 paf 202: append_here++; fused_rows++;
1.1 paf 203:
204: return *this;
205: }
206:
1.16 paf 207: uint String::hash_code() const {
1.7 paf 208: uint result=0;
1.5 paf 209:
1.16 paf 210: const Chunk *chunk=&head;
1.5 paf 211: do {
1.16 paf 212: const Chunk::Row *row=chunk->rows;
1.5 paf 213: for(int i=0; i<chunk->count; i++) {
214: if(row==append_here)
215: goto break2;
216:
1.6 paf 217: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 218: row++;
219: }
220: chunk=row->link;
221: } while(chunk);
222: break2:
223: return result;
224: }
225:
1.32 paf 226: int String::cmp(const String& src) const {
1.16 paf 227: const Chunk *a_chunk=&head;
228: const Chunk *b_chunk=&src.head;
229: const Chunk::Row *a_row=a_chunk->rows;
230: const Chunk::Row *b_row=b_chunk->rows;
1.9 paf 231: int a_offset=0;
232: int b_offset=0;
233: Chunk::Row *a_end=append_here;
234: Chunk::Row *b_end=src.append_here;
1.11 paf 235: int a_countdown=a_chunk->count;
236: int b_countdown=b_chunk->count;
1.9 paf 237: bool a_break=false;
238: bool b_break=false;
1.32 paf 239: int result;
1.9 paf 240: while(true) {
1.33 paf 241: a_break=a_row==a_end;
242: b_break=b_row==b_end;
243: if(a_break || b_break)
244: break;
245:
1.9 paf 246: int size_diff=
247: (a_row->item.size-a_offset)-
248: (b_row->item.size-b_offset);
249:
250: if(size_diff==0) { // a has same size as b
1.32 paf 251: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
252: if(result)
253: return result;
1.11 paf 254: a_row++; a_countdown--; a_offset=0;
255: b_row++; b_countdown--; b_offset=0;
1.9 paf 256: } else if (size_diff>0) { // a longer
1.32 paf 257: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset);
258: if(result)
259: return result;
1.9 paf 260: a_offset+=b_row->item.size-b_offset;
1.11 paf 261: b_row++; b_countdown--; b_offset=0;
1.9 paf 262: } else { // b longer
1.32 paf 263: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
264: if(result)
265: return result;
1.9 paf 266: b_offset+=a_row->item.size-a_offset;
1.11 paf 267: a_row++; a_countdown--; a_offset=0;
1.9 paf 268: }
269:
1.11 paf 270: if(!a_countdown) {
1.9 paf 271: a_chunk=a_row->link;
272: a_row=a_chunk->rows;
1.11 paf 273: a_countdown=a_chunk->count;
1.9 paf 274: }
1.11 paf 275: if(!b_countdown) {
1.9 paf 276: b_chunk=b_row->link;
277: b_row=b_chunk->rows;
1.11 paf 278: b_countdown=b_chunk->count;
1.27 paf 279: }
280: }
1.32 paf 281: if(a_break==b_break) // ended simultaneously
282: result=0;
283: else if(a_break) // first bytes equal, but a ended before b
284: result=-1;
285: else
286: result=+1;
287: return result;
1.27 paf 288: }
289:
1.39 paf 290: bool String::operator == (const char* b_ptr) const {
1.27 paf 291: size_t b_size=b_ptr?strlen(b_ptr):0;
292: if(size() != b_size)
293: return false;
294:
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) {
304: int size_diff=
305: (a_row->item.size-a_offset)-
306: (b_size-b_offset);
307:
308: if(size_diff==0) { // a has same size as b
309: if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
310: return false;
311: a_row++; a_countdown--; a_offset=0;
312: b_break=true;
313: } else if (size_diff>0) { // a longer
314: if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, b_size-b_offset)!=0)
315: return false;
316: a_offset+=b_size-b_offset;
317: b_break=true;
318: } else { // b longer
319: if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
320: return false;
321: b_offset+=a_row->item.size-a_offset;
322: a_row++; a_countdown--; a_offset=0;
323: }
324:
325: a_break=a_row==a_end;
326: if(a_break || b_break)
327: break;
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: }
335: return a_break==b_break;
1.5 paf 336: }
1.46 paf 337:
338: #ifndef NO_STRING_ORIGIN
339: const Origin& String::origin() const {
340: if(!fused_rows)
341: THROW(0, 0,
342: 0,
343: "String::origin() of empty string called");
344:
345: return head.rows[0].item.origin;
346: }
347: #endif
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