Guido van Rossum | f70e43a | 1991-02-19 12:39:46 +0000 | [diff] [blame] | 1 | /*********************************************************** |
| 2 | Copyright 1991 by Stichting Mathematisch Centrum, Amsterdam, The |
| 3 | Netherlands. |
| 4 | |
| 5 | All Rights Reserved |
| 6 | |
| 7 | Permission to use, copy, modify, and distribute this software and its |
| 8 | documentation for any purpose and without fee is hereby granted, |
| 9 | provided that the above copyright notice appear in all copies and that |
| 10 | both that copyright notice and this permission notice appear in |
| 11 | supporting documentation, and that the names of Stichting Mathematisch |
| 12 | Centrum or CWI not be used in advertising or publicity pertaining to |
| 13 | distribution of the software without specific, written prior permission. |
| 14 | |
| 15 | STICHTING MATHEMATISCH CENTRUM DISCLAIMS ALL WARRANTIES WITH REGARD TO |
| 16 | THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND |
| 17 | FITNESS, IN NO EVENT SHALL STICHTING MATHEMATISCH CENTRUM BE LIABLE |
| 18 | FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
| 19 | WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
| 20 | ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT |
| 21 | OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
| 22 | |
| 23 | ******************************************************************/ |
| 24 | |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 25 | /* String object implementation */ |
| 26 | |
Guido van Rossum | 3f5da24 | 1990-12-20 15:06:42 +0000 | [diff] [blame] | 27 | #include "allobjects.h" |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 28 | |
| 29 | object * |
| 30 | newsizedstringobject(str, size) |
| 31 | char *str; |
| 32 | int size; |
| 33 | { |
| 34 | register stringobject *op = (stringobject *) |
| 35 | malloc(sizeof(stringobject) + size * sizeof(char)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 36 | if (op == NULL) |
| 37 | return err_nomem(); |
| 38 | NEWREF(op); |
| 39 | op->ob_type = &Stringtype; |
| 40 | op->ob_size = size; |
| 41 | if (str != NULL) |
| 42 | memcpy(op->ob_sval, str, size); |
| 43 | op->ob_sval[size] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 44 | return (object *) op; |
| 45 | } |
| 46 | |
| 47 | object * |
| 48 | newstringobject(str) |
| 49 | char *str; |
| 50 | { |
| 51 | register unsigned int size = strlen(str); |
| 52 | register stringobject *op = (stringobject *) |
| 53 | malloc(sizeof(stringobject) + size * sizeof(char)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 54 | if (op == NULL) |
| 55 | return err_nomem(); |
| 56 | NEWREF(op); |
| 57 | op->ob_type = &Stringtype; |
| 58 | op->ob_size = size; |
| 59 | strcpy(op->ob_sval, str); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 60 | return (object *) op; |
| 61 | } |
| 62 | |
| 63 | unsigned int |
| 64 | getstringsize(op) |
| 65 | register object *op; |
| 66 | { |
| 67 | if (!is_stringobject(op)) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 68 | err_badcall(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 69 | return -1; |
| 70 | } |
| 71 | return ((stringobject *)op) -> ob_size; |
| 72 | } |
| 73 | |
| 74 | /*const*/ char * |
| 75 | getstringvalue(op) |
| 76 | register object *op; |
| 77 | { |
| 78 | if (!is_stringobject(op)) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 79 | err_badcall(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 80 | return NULL; |
| 81 | } |
| 82 | return ((stringobject *)op) -> ob_sval; |
| 83 | } |
| 84 | |
| 85 | /* Methods */ |
| 86 | |
| 87 | static void |
| 88 | stringprint(op, fp, flags) |
| 89 | stringobject *op; |
| 90 | FILE *fp; |
| 91 | int flags; |
| 92 | { |
| 93 | int i; |
| 94 | char c; |
| 95 | if (flags & PRINT_RAW) { |
| 96 | fwrite(op->ob_sval, 1, (int) op->ob_size, fp); |
| 97 | return; |
| 98 | } |
| 99 | fprintf(fp, "'"); |
| 100 | for (i = 0; i < op->ob_size; i++) { |
| 101 | c = op->ob_sval[i]; |
| 102 | if (c == '\'' || c == '\\') |
| 103 | fprintf(fp, "\\%c", c); |
| 104 | else if (c < ' ' || c >= 0177) |
| 105 | fprintf(fp, "\\%03o", c&0377); |
| 106 | else |
| 107 | putc(c, fp); |
| 108 | } |
| 109 | fprintf(fp, "'"); |
| 110 | } |
| 111 | |
| 112 | static object * |
| 113 | stringrepr(op) |
| 114 | register stringobject *op; |
| 115 | { |
| 116 | /* XXX overflow? */ |
| 117 | int newsize = 2 + 4 * op->ob_size * sizeof(char); |
| 118 | object *v = newsizedstringobject((char *)NULL, newsize); |
| 119 | if (v == NULL) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 120 | return err_nomem(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 121 | } |
| 122 | else { |
| 123 | register int i; |
| 124 | register char c; |
| 125 | register char *p; |
| 126 | NEWREF(v); |
| 127 | v->ob_type = &Stringtype; |
| 128 | ((stringobject *)v)->ob_size = newsize; |
| 129 | p = ((stringobject *)v)->ob_sval; |
| 130 | *p++ = '\''; |
| 131 | for (i = 0; i < op->ob_size; i++) { |
| 132 | c = op->ob_sval[i]; |
| 133 | if (c == '\'' || c == '\\') |
| 134 | *p++ = '\\', *p++ = c; |
| 135 | else if (c < ' ' || c >= 0177) { |
| 136 | sprintf(p, "\\%03o", c&0377); |
| 137 | while (*p != '\0') |
| 138 | p++; |
| 139 | |
| 140 | } |
| 141 | else |
| 142 | *p++ = c; |
| 143 | } |
| 144 | *p++ = '\''; |
| 145 | *p = '\0'; |
| 146 | resizestring(&v, (int) (p - ((stringobject *)v)->ob_sval)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 147 | return v; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 148 | } |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 149 | } |
| 150 | |
| 151 | static int |
| 152 | stringlength(a) |
| 153 | stringobject *a; |
| 154 | { |
| 155 | return a->ob_size; |
| 156 | } |
| 157 | |
| 158 | static object * |
| 159 | stringconcat(a, bb) |
| 160 | register stringobject *a; |
| 161 | register object *bb; |
| 162 | { |
| 163 | register unsigned int size; |
| 164 | register stringobject *op; |
| 165 | if (!is_stringobject(bb)) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 166 | err_badarg(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 167 | return NULL; |
| 168 | } |
| 169 | #define b ((stringobject *)bb) |
| 170 | /* Optimize cases with empty left or right operand */ |
| 171 | if (a->ob_size == 0) { |
| 172 | INCREF(bb); |
| 173 | return bb; |
| 174 | } |
| 175 | if (b->ob_size == 0) { |
| 176 | INCREF(a); |
| 177 | return (object *)a; |
| 178 | } |
| 179 | size = a->ob_size + b->ob_size; |
| 180 | op = (stringobject *) |
| 181 | malloc(sizeof(stringobject) + size * sizeof(char)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 182 | if (op == NULL) |
| 183 | return err_nomem(); |
| 184 | NEWREF(op); |
| 185 | op->ob_type = &Stringtype; |
| 186 | op->ob_size = size; |
| 187 | memcpy(op->ob_sval, a->ob_sval, (int) a->ob_size); |
| 188 | memcpy(op->ob_sval + a->ob_size, b->ob_sval, (int) b->ob_size); |
| 189 | op->ob_sval[size] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 190 | return (object *) op; |
| 191 | #undef b |
| 192 | } |
| 193 | |
| 194 | static object * |
| 195 | stringrepeat(a, n) |
| 196 | register stringobject *a; |
| 197 | register int n; |
| 198 | { |
| 199 | register int i; |
| 200 | register unsigned int size; |
| 201 | register stringobject *op; |
| 202 | if (n < 0) |
| 203 | n = 0; |
| 204 | size = a->ob_size * n; |
| 205 | if (size == a->ob_size) { |
| 206 | INCREF(a); |
| 207 | return (object *)a; |
| 208 | } |
| 209 | op = (stringobject *) |
| 210 | malloc(sizeof(stringobject) + size * sizeof(char)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 211 | if (op == NULL) |
| 212 | return err_nomem(); |
| 213 | NEWREF(op); |
| 214 | op->ob_type = &Stringtype; |
| 215 | op->ob_size = size; |
| 216 | for (i = 0; i < size; i += a->ob_size) |
| 217 | memcpy(op->ob_sval+i, a->ob_sval, (int) a->ob_size); |
| 218 | op->ob_sval[size] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 219 | return (object *) op; |
| 220 | } |
| 221 | |
| 222 | /* String slice a[i:j] consists of characters a[i] ... a[j-1] */ |
| 223 | |
| 224 | static object * |
| 225 | stringslice(a, i, j) |
| 226 | register stringobject *a; |
| 227 | register int i, j; /* May be negative! */ |
| 228 | { |
| 229 | if (i < 0) |
| 230 | i = 0; |
| 231 | if (j < 0) |
| 232 | j = 0; /* Avoid signed/unsigned bug in next line */ |
| 233 | if (j > a->ob_size) |
| 234 | j = a->ob_size; |
| 235 | if (i == 0 && j == a->ob_size) { /* It's the same as a */ |
| 236 | INCREF(a); |
| 237 | return (object *)a; |
| 238 | } |
| 239 | if (j < i) |
| 240 | j = i; |
| 241 | return newsizedstringobject(a->ob_sval + i, (int) (j-i)); |
| 242 | } |
| 243 | |
| 244 | static object * |
| 245 | stringitem(a, i) |
| 246 | stringobject *a; |
| 247 | register int i; |
| 248 | { |
| 249 | if (i < 0 || i >= a->ob_size) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 250 | err_setstr(IndexError, "string index out of range"); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 251 | return NULL; |
| 252 | } |
| 253 | return stringslice(a, i, i+1); |
| 254 | } |
| 255 | |
| 256 | static int |
| 257 | stringcompare(a, b) |
| 258 | stringobject *a, *b; |
| 259 | { |
Guido van Rossum | 253919f | 1991-02-13 23:18:39 +0000 | [diff] [blame] | 260 | int len_a = a->ob_size, len_b = b->ob_size; |
| 261 | int min_len = (len_a < len_b) ? len_a : len_b; |
| 262 | int cmp = memcmp(a->ob_sval, b->ob_sval, min_len); |
| 263 | if (cmp != 0) |
| 264 | return cmp; |
| 265 | return (len_a < len_b) ? -1 : (len_a > len_b) ? 1 : 0; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 266 | } |
| 267 | |
| 268 | static sequence_methods string_as_sequence = { |
| 269 | stringlength, /*tp_length*/ |
| 270 | stringconcat, /*tp_concat*/ |
| 271 | stringrepeat, /*tp_repeat*/ |
| 272 | stringitem, /*tp_item*/ |
| 273 | stringslice, /*tp_slice*/ |
| 274 | 0, /*tp_ass_item*/ |
| 275 | 0, /*tp_ass_slice*/ |
| 276 | }; |
| 277 | |
| 278 | typeobject Stringtype = { |
| 279 | OB_HEAD_INIT(&Typetype) |
| 280 | 0, |
| 281 | "string", |
| 282 | sizeof(stringobject), |
| 283 | sizeof(char), |
| 284 | free, /*tp_dealloc*/ |
| 285 | stringprint, /*tp_print*/ |
| 286 | 0, /*tp_getattr*/ |
| 287 | 0, /*tp_setattr*/ |
| 288 | stringcompare, /*tp_compare*/ |
| 289 | stringrepr, /*tp_repr*/ |
| 290 | 0, /*tp_as_number*/ |
| 291 | &string_as_sequence, /*tp_as_sequence*/ |
| 292 | 0, /*tp_as_mapping*/ |
| 293 | }; |
| 294 | |
| 295 | void |
| 296 | joinstring(pv, w) |
| 297 | register object **pv; |
| 298 | register object *w; |
| 299 | { |
| 300 | register object *v; |
| 301 | if (*pv == NULL || w == NULL || !is_stringobject(*pv)) |
| 302 | return; |
| 303 | v = stringconcat((stringobject *) *pv, w); |
| 304 | DECREF(*pv); |
| 305 | *pv = v; |
| 306 | } |
| 307 | |
| 308 | /* The following function breaks the notion that strings are immutable: |
| 309 | it changes the size of a string. We get away with this only if there |
| 310 | is only one module referencing the object. You can also think of it |
| 311 | as creating a new string object and destroying the old one, only |
| 312 | more efficiently. In any case, don't use this if the string may |
| 313 | already be known to some other part of the code... */ |
| 314 | |
| 315 | int |
| 316 | resizestring(pv, newsize) |
| 317 | object **pv; |
| 318 | int newsize; |
| 319 | { |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 320 | register object *v; |
| 321 | register stringobject *sv; |
| 322 | v = *pv; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 323 | if (!is_stringobject(v) || v->ob_refcnt != 1) { |
| 324 | *pv = 0; |
| 325 | DECREF(v); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 326 | err_badcall(); |
| 327 | return -1; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 328 | } |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 329 | /* XXX UNREF/NEWREF interface should be more symmetrical */ |
Guido van Rossum | 392ab32 | 1990-11-18 17:41:19 +0000 | [diff] [blame] | 330 | #ifdef REF_DEBUG |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 331 | --ref_total; |
| 332 | #endif |
| 333 | UNREF(v); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 334 | *pv = (object *) |
| 335 | realloc((char *)v, |
| 336 | sizeof(stringobject) + newsize * sizeof(char)); |
| 337 | if (*pv == NULL) { |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 338 | DEL(v); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 339 | err_nomem(); |
| 340 | return -1; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 341 | } |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 342 | NEWREF(*pv); |
| 343 | sv = (stringobject *) *pv; |
| 344 | sv->ob_size = newsize; |
| 345 | sv->ob_sval[newsize] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 346 | return 0; |
| 347 | } |