Guido van Rossum | f70e43a | 1991-02-19 12:39:46 +0000 | [diff] [blame] | 1 | /*********************************************************** |
Guido van Rossum | bab9d03 | 1992-04-05 14:26:55 +0000 | [diff] [blame] | 2 | Copyright 1991, 1992 by Stichting Mathematisch Centrum, Amsterdam, The |
Guido van Rossum | f70e43a | 1991-02-19 12:39:46 +0000 | [diff] [blame] | 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 | |
Guido van Rossum | 719f5fa | 1992-03-27 17:31:02 +0000 | [diff] [blame] | 63 | void |
| 64 | stringdealloc(op) |
| 65 | object *op; |
| 66 | { |
| 67 | DEL(op); |
| 68 | } |
| 69 | |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 70 | unsigned int |
| 71 | getstringsize(op) |
| 72 | register object *op; |
| 73 | { |
| 74 | if (!is_stringobject(op)) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 75 | err_badcall(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 76 | return -1; |
| 77 | } |
| 78 | return ((stringobject *)op) -> ob_size; |
| 79 | } |
| 80 | |
| 81 | /*const*/ char * |
| 82 | getstringvalue(op) |
| 83 | register object *op; |
| 84 | { |
| 85 | if (!is_stringobject(op)) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 86 | err_badcall(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 87 | return NULL; |
| 88 | } |
| 89 | return ((stringobject *)op) -> ob_sval; |
| 90 | } |
| 91 | |
| 92 | /* Methods */ |
| 93 | |
Guido van Rossum | bcaa31c | 1991-06-07 22:58:57 +0000 | [diff] [blame] | 94 | static int |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 95 | stringprint(op, fp, flags) |
| 96 | stringobject *op; |
| 97 | FILE *fp; |
| 98 | int flags; |
| 99 | { |
| 100 | int i; |
| 101 | char c; |
Guido van Rossum | bcaa31c | 1991-06-07 22:58:57 +0000 | [diff] [blame] | 102 | /* XXX Ought to check for interrupts when writing long strings */ |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 103 | if (flags & PRINT_RAW) { |
| 104 | fwrite(op->ob_sval, 1, (int) op->ob_size, fp); |
Guido van Rossum | bcaa31c | 1991-06-07 22:58:57 +0000 | [diff] [blame] | 105 | return 0; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 106 | } |
| 107 | fprintf(fp, "'"); |
| 108 | for (i = 0; i < op->ob_size; i++) { |
| 109 | c = op->ob_sval[i]; |
| 110 | if (c == '\'' || c == '\\') |
| 111 | fprintf(fp, "\\%c", c); |
| 112 | else if (c < ' ' || c >= 0177) |
| 113 | fprintf(fp, "\\%03o", c&0377); |
| 114 | else |
| 115 | putc(c, fp); |
| 116 | } |
| 117 | fprintf(fp, "'"); |
Guido van Rossum | bcaa31c | 1991-06-07 22:58:57 +0000 | [diff] [blame] | 118 | return 0; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 119 | } |
| 120 | |
| 121 | static object * |
| 122 | stringrepr(op) |
| 123 | register stringobject *op; |
| 124 | { |
| 125 | /* XXX overflow? */ |
| 126 | int newsize = 2 + 4 * op->ob_size * sizeof(char); |
| 127 | object *v = newsizedstringobject((char *)NULL, newsize); |
| 128 | if (v == NULL) { |
Guido van Rossum | bcaa31c | 1991-06-07 22:58:57 +0000 | [diff] [blame] | 129 | return NULL; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 130 | } |
| 131 | else { |
| 132 | register int i; |
| 133 | register char c; |
| 134 | register char *p; |
| 135 | NEWREF(v); |
| 136 | v->ob_type = &Stringtype; |
| 137 | ((stringobject *)v)->ob_size = newsize; |
| 138 | p = ((stringobject *)v)->ob_sval; |
| 139 | *p++ = '\''; |
| 140 | for (i = 0; i < op->ob_size; i++) { |
| 141 | c = op->ob_sval[i]; |
| 142 | if (c == '\'' || c == '\\') |
| 143 | *p++ = '\\', *p++ = c; |
| 144 | else if (c < ' ' || c >= 0177) { |
| 145 | sprintf(p, "\\%03o", c&0377); |
| 146 | while (*p != '\0') |
| 147 | p++; |
| 148 | |
| 149 | } |
| 150 | else |
| 151 | *p++ = c; |
| 152 | } |
| 153 | *p++ = '\''; |
| 154 | *p = '\0'; |
| 155 | resizestring(&v, (int) (p - ((stringobject *)v)->ob_sval)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 156 | return v; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 157 | } |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 158 | } |
| 159 | |
| 160 | static int |
| 161 | stringlength(a) |
| 162 | stringobject *a; |
| 163 | { |
| 164 | return a->ob_size; |
| 165 | } |
| 166 | |
| 167 | static object * |
| 168 | stringconcat(a, bb) |
| 169 | register stringobject *a; |
| 170 | register object *bb; |
| 171 | { |
| 172 | register unsigned int size; |
| 173 | register stringobject *op; |
| 174 | if (!is_stringobject(bb)) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 175 | err_badarg(); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 176 | return NULL; |
| 177 | } |
| 178 | #define b ((stringobject *)bb) |
| 179 | /* Optimize cases with empty left or right operand */ |
| 180 | if (a->ob_size == 0) { |
| 181 | INCREF(bb); |
| 182 | return bb; |
| 183 | } |
| 184 | if (b->ob_size == 0) { |
| 185 | INCREF(a); |
| 186 | return (object *)a; |
| 187 | } |
| 188 | size = a->ob_size + b->ob_size; |
| 189 | op = (stringobject *) |
| 190 | malloc(sizeof(stringobject) + size * sizeof(char)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 191 | if (op == NULL) |
| 192 | return err_nomem(); |
| 193 | NEWREF(op); |
| 194 | op->ob_type = &Stringtype; |
| 195 | op->ob_size = size; |
| 196 | memcpy(op->ob_sval, a->ob_sval, (int) a->ob_size); |
| 197 | memcpy(op->ob_sval + a->ob_size, b->ob_sval, (int) b->ob_size); |
| 198 | op->ob_sval[size] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 199 | return (object *) op; |
| 200 | #undef b |
| 201 | } |
| 202 | |
| 203 | static object * |
| 204 | stringrepeat(a, n) |
| 205 | register stringobject *a; |
| 206 | register int n; |
| 207 | { |
| 208 | register int i; |
| 209 | register unsigned int size; |
| 210 | register stringobject *op; |
| 211 | if (n < 0) |
| 212 | n = 0; |
| 213 | size = a->ob_size * n; |
| 214 | if (size == a->ob_size) { |
| 215 | INCREF(a); |
| 216 | return (object *)a; |
| 217 | } |
| 218 | op = (stringobject *) |
| 219 | malloc(sizeof(stringobject) + size * sizeof(char)); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 220 | if (op == NULL) |
| 221 | return err_nomem(); |
| 222 | NEWREF(op); |
| 223 | op->ob_type = &Stringtype; |
| 224 | op->ob_size = size; |
| 225 | for (i = 0; i < size; i += a->ob_size) |
| 226 | memcpy(op->ob_sval+i, a->ob_sval, (int) a->ob_size); |
| 227 | op->ob_sval[size] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 228 | return (object *) op; |
| 229 | } |
| 230 | |
| 231 | /* String slice a[i:j] consists of characters a[i] ... a[j-1] */ |
| 232 | |
| 233 | static object * |
| 234 | stringslice(a, i, j) |
| 235 | register stringobject *a; |
| 236 | register int i, j; /* May be negative! */ |
| 237 | { |
| 238 | if (i < 0) |
| 239 | i = 0; |
| 240 | if (j < 0) |
| 241 | j = 0; /* Avoid signed/unsigned bug in next line */ |
| 242 | if (j > a->ob_size) |
| 243 | j = a->ob_size; |
| 244 | if (i == 0 && j == a->ob_size) { /* It's the same as a */ |
| 245 | INCREF(a); |
| 246 | return (object *)a; |
| 247 | } |
| 248 | if (j < i) |
| 249 | j = i; |
| 250 | return newsizedstringobject(a->ob_sval + i, (int) (j-i)); |
| 251 | } |
| 252 | |
Guido van Rossum | daa8bb3 | 1991-04-04 10:48:33 +0000 | [diff] [blame] | 253 | #ifdef __STDC__ |
| 254 | #include <limits.h> |
| 255 | #else |
| 256 | #ifndef UCHAR_MAX |
| 257 | #define UCHAR_MAX 255 |
| 258 | #endif |
| 259 | #endif |
| 260 | |
| 261 | static object *characters[UCHAR_MAX + 1]; |
| 262 | |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 263 | static object * |
| 264 | stringitem(a, i) |
| 265 | stringobject *a; |
| 266 | register int i; |
| 267 | { |
Guido van Rossum | daa8bb3 | 1991-04-04 10:48:33 +0000 | [diff] [blame] | 268 | int c; |
| 269 | object *v; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 270 | if (i < 0 || i >= a->ob_size) { |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 271 | err_setstr(IndexError, "string index out of range"); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 272 | return NULL; |
| 273 | } |
Guido van Rossum | daa8bb3 | 1991-04-04 10:48:33 +0000 | [diff] [blame] | 274 | c = a->ob_sval[i] & UCHAR_MAX; |
| 275 | v = characters[c]; |
| 276 | if (v == NULL) { |
| 277 | v = newsizedstringobject((char *)NULL, 1); |
| 278 | if (v == NULL) |
| 279 | return NULL; |
| 280 | characters[c] = v; |
| 281 | ((stringobject *)v)->ob_sval[0] = c; |
| 282 | } |
| 283 | INCREF(v); |
| 284 | return v; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 285 | } |
| 286 | |
| 287 | static int |
| 288 | stringcompare(a, b) |
| 289 | stringobject *a, *b; |
| 290 | { |
Guido van Rossum | 253919f | 1991-02-13 23:18:39 +0000 | [diff] [blame] | 291 | int len_a = a->ob_size, len_b = b->ob_size; |
| 292 | int min_len = (len_a < len_b) ? len_a : len_b; |
| 293 | int cmp = memcmp(a->ob_sval, b->ob_sval, min_len); |
| 294 | if (cmp != 0) |
| 295 | return cmp; |
| 296 | 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] | 297 | } |
| 298 | |
| 299 | static sequence_methods string_as_sequence = { |
Guido van Rossum | f380e66 | 1991-06-04 19:36:32 +0000 | [diff] [blame] | 300 | stringlength, /*sq_length*/ |
| 301 | stringconcat, /*sq_concat*/ |
| 302 | stringrepeat, /*sq_repeat*/ |
| 303 | stringitem, /*sq_item*/ |
| 304 | stringslice, /*sq_slice*/ |
| 305 | 0, /*sq_ass_item*/ |
| 306 | 0, /*sq_ass_slice*/ |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 307 | }; |
| 308 | |
| 309 | typeobject Stringtype = { |
| 310 | OB_HEAD_INIT(&Typetype) |
| 311 | 0, |
| 312 | "string", |
| 313 | sizeof(stringobject), |
| 314 | sizeof(char), |
Guido van Rossum | 719f5fa | 1992-03-27 17:31:02 +0000 | [diff] [blame] | 315 | stringdealloc, /*tp_dealloc*/ |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 316 | stringprint, /*tp_print*/ |
| 317 | 0, /*tp_getattr*/ |
| 318 | 0, /*tp_setattr*/ |
| 319 | stringcompare, /*tp_compare*/ |
| 320 | stringrepr, /*tp_repr*/ |
| 321 | 0, /*tp_as_number*/ |
| 322 | &string_as_sequence, /*tp_as_sequence*/ |
| 323 | 0, /*tp_as_mapping*/ |
| 324 | }; |
| 325 | |
| 326 | void |
| 327 | joinstring(pv, w) |
| 328 | register object **pv; |
| 329 | register object *w; |
| 330 | { |
| 331 | register object *v; |
| 332 | if (*pv == NULL || w == NULL || !is_stringobject(*pv)) |
| 333 | return; |
| 334 | v = stringconcat((stringobject *) *pv, w); |
| 335 | DECREF(*pv); |
| 336 | *pv = v; |
| 337 | } |
| 338 | |
| 339 | /* The following function breaks the notion that strings are immutable: |
| 340 | it changes the size of a string. We get away with this only if there |
| 341 | is only one module referencing the object. You can also think of it |
| 342 | as creating a new string object and destroying the old one, only |
| 343 | more efficiently. In any case, don't use this if the string may |
| 344 | already be known to some other part of the code... */ |
| 345 | |
| 346 | int |
| 347 | resizestring(pv, newsize) |
| 348 | object **pv; |
| 349 | int newsize; |
| 350 | { |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 351 | register object *v; |
| 352 | register stringobject *sv; |
| 353 | v = *pv; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 354 | if (!is_stringobject(v) || v->ob_refcnt != 1) { |
| 355 | *pv = 0; |
| 356 | DECREF(v); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 357 | err_badcall(); |
| 358 | return -1; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 359 | } |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 360 | /* XXX UNREF/NEWREF interface should be more symmetrical */ |
Guido van Rossum | 392ab32 | 1990-11-18 17:41:19 +0000 | [diff] [blame] | 361 | #ifdef REF_DEBUG |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 362 | --ref_total; |
| 363 | #endif |
| 364 | UNREF(v); |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 365 | *pv = (object *) |
| 366 | realloc((char *)v, |
| 367 | sizeof(stringobject) + newsize * sizeof(char)); |
| 368 | if (*pv == NULL) { |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 369 | DEL(v); |
Guido van Rossum | 2a9096b | 1990-10-21 22:15:08 +0000 | [diff] [blame] | 370 | err_nomem(); |
| 371 | return -1; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 372 | } |
Guido van Rossum | 921842f | 1990-11-18 17:30:23 +0000 | [diff] [blame] | 373 | NEWREF(*pv); |
| 374 | sv = (stringobject *) *pv; |
| 375 | sv->ob_size = newsize; |
| 376 | sv->ob_sval[newsize] = '\0'; |
Guido van Rossum | 85a5fbb | 1990-10-14 12:07:46 +0000 | [diff] [blame] | 377 | return 0; |
| 378 | } |