Annotation of parser3/src/main/pa_string.C, revision 1.41
1.4 paf 1: /*
1.36 paf 2: Parser
3: Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com)
1.37 paf 4: Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf)
1.36 paf 5:
1.41 ! paf 6: $Id: pa_string.C,v 1.40 2001/03/18 13:22:07 paf Exp $
1.4 paf 7: */
8:
1.1 paf 9: #include <string.h>
10:
1.13 paf 11: #include "pa_pool.h"
1.12 paf 12: #include "pa_string.h"
1.5 paf 13: #include "pa_hash.h"
1.22 paf 14: #include "pa_exception.h"
1.1 paf 15:
1.18 paf 16: // String
17:
1.41 ! paf 18: String::String(Pool& apool, const char *src, bool tainted) :
1.17 paf 19: Pooled(apool) {
1.28 paf 20: last_chunk=&head;
21: head.count=CR_PREALLOCATED_COUNT;
1.5 paf 22: append_here=head.rows;
1.2 paf 23: head.preallocated_link=0;
1.28 paf 24: link_row=&head.rows[head.count];
1.8 paf 25: fused_rows=fsize=0;
1.41 ! paf 26:
! 27: if(src)
! 28: if(tainted)
! 29: APPEND_TAINTED(src, 0, 0, 0);
! 30: else
! 31: APPEND(src, 0, 0, 0);
1.1 paf 32: }
33:
34: void String::expand() {
1.28 paf 35: int new_chunk_count=last_chunk->count+last_chunk->count*CR_GROW_PERCENT/100;
36: last_chunk=static_cast<Chunk *>(
1.30 paf 37: malloc(sizeof(int)+sizeof(Chunk::Row)*new_chunk_count+sizeof(Chunk *)));
1.28 paf 38: last_chunk->count=new_chunk_count;
39: link_row->link=last_chunk;
40: append_here=last_chunk->rows;
41: link_row=&last_chunk->rows[last_chunk->count];
1.8 paf 42: link_row->link=0;
1.1 paf 43: }
44:
1.40 paf 45: String::String(const String& src) : Pooled(src.pool()) {
1.8 paf 46: head.count=CR_PREALLOCATED_COUNT;
47:
48: int src_used_rows=src.used_rows();
49: if(src_used_rows<=head.count) {
1.10 paf 50: // all new rows fit into preallocated area
1.28 paf 51: int curr_chunk_rows=head.count;
1.8 paf 52: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
53: append_here=&head.rows[src_used_rows];
54: link_row=&head.rows[curr_chunk_rows];
55: } else {
56: // warning:
1.10 paf 57: // heavily relies on the fact
58: // "preallocated area is the same for all strings"
1.8 paf 59: //
60: // info:
61: // allocating only enough mem to fit src string rows
62: // next append would allocate a new chunk
63: //
64: // new rows don't fit into preallocated area: splitting into two chunks
65: // preallocated chunk src to constructing head
66: memcpy(head.rows, src.head.rows, sizeof(Chunk::Row)*head.count);
67: // remaining rows into new_chunk
1.28 paf 68: int curr_chunk_rows=src_used_rows-head.count;
1.8 paf 69: Chunk *new_chunk=static_cast<Chunk *>(
1.30 paf 70: malloc(sizeof(int)+sizeof(Chunk::Row)*curr_chunk_rows+sizeof(Chunk *)));
1.8 paf 71: new_chunk->count=curr_chunk_rows;
72: head.preallocated_link=new_chunk;
1.28 paf 73: append_here=link_row=&new_chunk->rows[new_chunk->count];
1.8 paf 74:
75: Chunk *old_chunk=src.head.preallocated_link;
76: Chunk::Row *new_rows=new_chunk->rows;
1.28 paf 77: int rows_left_to_copy=new_chunk->count;
1.8 paf 78: while(true) {
79: int old_count=old_chunk->count;
80: Chunk *next_chunk=old_chunk->rows[old_count].link;
81: if(next_chunk) {
82: // not last source chunk
83: // taking it all
84: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
85: new_rows+=old_count;
86: rows_left_to_copy-=old_count;
87:
88: old_chunk=next_chunk;
89: } else {
90: // the last source chunk
91: // taking only those rows of chunk that _left_to_copy
92: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
93: break;
94: }
95: }
1.5 paf 96: }
1.8 paf 97: link_row->link=0;
98: fused_rows=src_used_rows;
99: fsize=src.fsize;
1.5 paf 100: }
1.28 paf 101:
1.34 paf 102: String& String::append(const String& src, Untaint_lang lang) {
1.28 paf 103: int src_used_rows=src.used_rows();
104: int dst_free_rows=link_row-append_here;
105:
106: if(src_used_rows<=dst_free_rows) {
107: // all new rows fit into last chunk
108: memcpy(append_here, src.head.rows, sizeof(Chunk::Row)*src_used_rows);
1.34 paf 109: set_lang(append_here, lang, src_used_rows);
1.28 paf 110: append_here+=src_used_rows;
111: } else {
1.31 paf 112: // not all new rows fit into last chunk: shrinking it to used part,
1.28 paf 113: int used_rows=last_chunk->count-dst_free_rows;
114: //int *countp=append_here
115: link_row=&last_chunk->rows[last_chunk->count=used_rows];
116: // allocating only enough mem to fit src string rows
117: // next append would allocate a new chunk
118: last_chunk=static_cast<Chunk *>(
1.30 paf 119: malloc(sizeof(int)+sizeof(Chunk::Row)*src_used_rows+sizeof(Chunk *)));
1.28 paf 120: last_chunk->count=src_used_rows;
121: link_row->link=last_chunk;
122: append_here=link_row=&last_chunk->rows[src_used_rows];
123:
1.31 paf 124: const Chunk *old_chunk=&src.head;
1.28 paf 125: Chunk::Row *new_rows=last_chunk->rows;
126: int rows_left_to_copy=src_used_rows;
127: while(true) {
128: int old_count=old_chunk->count;
129: Chunk *next_chunk=old_chunk->rows[old_count].link;
130: if(next_chunk) {
131: // not last source chunk
132: // taking it all
133: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*old_count);
1.34 paf 134: set_lang(new_rows, lang, old_count);
1.28 paf 135: new_rows+=old_count;
136: rows_left_to_copy-=old_count;
137:
138: old_chunk=next_chunk;
139: } else {
140: // the last source chunk
141: // taking only those rows of chunk that _left_to_copy
142: memcpy(new_rows, old_chunk->rows, sizeof(Chunk::Row)*rows_left_to_copy);
1.34 paf 143: set_lang(new_rows, lang, rows_left_to_copy);
1.28 paf 144: break;
145: }
146: }
1.29 paf 147: link_row->link=0;
1.28 paf 148: }
149: fused_rows+=src_used_rows;
150: fsize+=src.fsize;
151:
152: return *this;
1.23 paf 153: }
1.34 paf 154: void String::set_lang(Chunk::Row *row, Untaint_lang lang, size_t size) {
1.40 paf 155: if(lang==PASS_APPEND)
1.34 paf 156: return;
157:
158: while(size--) {
159: Untaint_lang& item_lang=(row++)->item.lang;
160: if(item_lang==YES) // tainted? need untaint language assignment
161: item_lang=lang; // assign untaint language
162: }
1.40 paf 163: }
164:
165: void String::change_lang(Untaint_lang lang) {
166: Chunk *chunk=&head;
167: do {
168: Chunk::Row *row=chunk->rows;
169: for(int i=0; i<chunk->count; i++) {
170: if(row==append_here)
171: goto break2;
172:
173: row->item.lang=lang;
174: row++;
175: }
176: chunk=row->link;
177: } while(chunk);
178: break2:
179: return;
1.34 paf 180: }
1.5 paf 181:
1.13 paf 182: String& String::real_append(STRING_APPEND_PARAMS) {
1.9 paf 183: if(!src)
184: return *this;
1.26 paf 185: if(!size)
186: size=strlen(src);
187: if(!size)
1.9 paf 188: return *this;
189:
1.1 paf 190: if(chunk_is_full())
191: expand();
192:
193: append_here->item.ptr=src;
1.26 paf 194: fsize+=append_here->item.size=size;
1.38 paf 195: append_here->item.lang=tainted?YES:NO;
1.13 paf 196: #ifndef NO_STRING_ORIGIN
1.14 paf 197: append_here->item.origin.file=file;
198: append_here->item.origin.line=line;
1.13 paf 199: #endif
1.8 paf 200: append_here++; fused_rows++;
1.1 paf 201:
202: return *this;
203: }
204:
1.16 paf 205: uint String::hash_code() const {
1.7 paf 206: uint result=0;
1.5 paf 207:
1.16 paf 208: const Chunk *chunk=&head;
1.5 paf 209: do {
1.16 paf 210: const Chunk::Row *row=chunk->rows;
1.5 paf 211: for(int i=0; i<chunk->count; i++) {
212: if(row==append_here)
213: goto break2;
214:
1.6 paf 215: result=Hash::generic_code(result, row->item.ptr, row->item.size);
1.5 paf 216: row++;
217: }
218: chunk=row->link;
219: } while(chunk);
220: break2:
221: return result;
222: }
223:
1.32 paf 224: int String::cmp(const String& src) const {
1.16 paf 225: const Chunk *a_chunk=&head;
226: const Chunk *b_chunk=&src.head;
227: const Chunk::Row *a_row=a_chunk->rows;
228: const Chunk::Row *b_row=b_chunk->rows;
1.9 paf 229: int a_offset=0;
230: int b_offset=0;
231: Chunk::Row *a_end=append_here;
232: Chunk::Row *b_end=src.append_here;
1.11 paf 233: int a_countdown=a_chunk->count;
234: int b_countdown=b_chunk->count;
1.9 paf 235: bool a_break=false;
236: bool b_break=false;
1.32 paf 237: int result;
1.9 paf 238: while(true) {
1.33 paf 239: a_break=a_row==a_end;
240: b_break=b_row==b_end;
241: if(a_break || b_break)
242: break;
243:
1.9 paf 244: int size_diff=
245: (a_row->item.size-a_offset)-
246: (b_row->item.size-b_offset);
247:
248: if(size_diff==0) { // a has same size as b
1.32 paf 249: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
250: if(result)
251: return result;
1.11 paf 252: a_row++; a_countdown--; a_offset=0;
253: b_row++; b_countdown--; b_offset=0;
1.9 paf 254: } else if (size_diff>0) { // a longer
1.32 paf 255: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, b_row->item.size-b_offset);
256: if(result)
257: return result;
1.9 paf 258: a_offset+=b_row->item.size-b_offset;
1.11 paf 259: b_row++; b_countdown--; b_offset=0;
1.9 paf 260: } else { // b longer
1.32 paf 261: result=memcmp(a_row->item.ptr+a_offset, b_row->item.ptr+b_offset, a_row->item.size-a_offset);
262: if(result)
263: return result;
1.9 paf 264: b_offset+=a_row->item.size-a_offset;
1.11 paf 265: a_row++; a_countdown--; a_offset=0;
1.9 paf 266: }
267:
1.11 paf 268: if(!a_countdown) {
1.9 paf 269: a_chunk=a_row->link;
270: a_row=a_chunk->rows;
1.11 paf 271: a_countdown=a_chunk->count;
1.9 paf 272: }
1.11 paf 273: if(!b_countdown) {
1.9 paf 274: b_chunk=b_row->link;
275: b_row=b_chunk->rows;
1.11 paf 276: b_countdown=b_chunk->count;
1.27 paf 277: }
278: }
1.32 paf 279: if(a_break==b_break) // ended simultaneously
280: result=0;
281: else if(a_break) // first bytes equal, but a ended before b
282: result=-1;
283: else
284: result=+1;
285: return result;
1.27 paf 286: }
287:
1.39 paf 288: bool String::operator == (const char* b_ptr) const {
1.27 paf 289: size_t b_size=b_ptr?strlen(b_ptr):0;
290: if(size() != b_size)
291: return false;
292:
293: const Chunk *a_chunk=&head;
294: const Chunk::Row *a_row=a_chunk->rows;
295: int a_offset=0;
296: int b_offset=0;
297: Chunk::Row *a_end=append_here;
298: int a_countdown=a_chunk->count;
299: bool a_break=false;
300: bool b_break=false;
301: while(true) {
302: int size_diff=
303: (a_row->item.size-a_offset)-
304: (b_size-b_offset);
305:
306: if(size_diff==0) { // a has same size as b
307: if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
308: return false;
309: a_row++; a_countdown--; a_offset=0;
310: b_break=true;
311: } else if (size_diff>0) { // a longer
312: if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, b_size-b_offset)!=0)
313: return false;
314: a_offset+=b_size-b_offset;
315: b_break=true;
316: } else { // b longer
317: if(memcmp(a_row->item.ptr+a_offset, b_ptr+b_offset, a_row->item.size-a_offset)!=0)
318: return false;
319: b_offset+=a_row->item.size-a_offset;
320: a_row++; a_countdown--; a_offset=0;
321: }
322:
323: a_break=a_row==a_end;
324: if(a_break || b_break)
325: break;
326:
327: if(!a_countdown) {
328: a_chunk=a_row->link;
329: a_row=a_chunk->rows;
330: a_countdown=a_chunk->count;
1.9 paf 331: }
332: }
333: return a_break==b_break;
1.5 paf 334: }
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