Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 1 | /*********************************************************** |
Guido van Rossum | 9bfef44 | 1993-03-29 10:43:31 +0000 | [diff] [blame] | 2 | Copyright 1991, 1992, 1993 by Stichting Mathematisch Centrum, |
| 3 | Amsterdam, The Netherlands. |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 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 | |
| 25 | /* strop module */ |
| 26 | |
| 27 | #include "allobjects.h" |
| 28 | #include "modsupport.h" |
| 29 | |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 30 | #include <ctype.h> |
Guido van Rossum | e22e644 | 1993-07-09 10:51:31 +0000 | [diff] [blame] | 31 | /* XXX This file assumes that the <ctype.h> is*() functions |
| 32 | XXX are defined for all 8-bit characters! */ |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 33 | |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 34 | |
| 35 | static object * |
| 36 | strop_split(self, args) |
| 37 | object *self; /* Not used */ |
| 38 | object *args; |
| 39 | { |
| 40 | int len, i, j; |
| 41 | char *s; |
| 42 | char c; |
| 43 | object *list, *item; |
| 44 | |
| 45 | if (!getargs(args, "s#", &s, &len)) |
| 46 | return NULL; |
| 47 | list = newlistobject(0); |
| 48 | if (list == NULL) |
| 49 | return NULL; |
| 50 | |
| 51 | i = 0; |
| 52 | while (i < len) { |
| 53 | while (i < len && |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 54 | ((c = s[i]), isspace(c))) { |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 55 | i = i+1; |
| 56 | } |
| 57 | j = i; |
| 58 | while (i < len && |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 59 | !((c = s[i]), isspace(c))) { |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 60 | i = i+1; |
| 61 | } |
| 62 | if (j < i) { |
| 63 | item = newsizedstringobject(s+j, (int)(i-j)); |
| 64 | if (item == NULL || addlistitem(list, item) < 0) { |
| 65 | DECREF(list); |
| 66 | return NULL; |
| 67 | } |
| 68 | } |
| 69 | } |
| 70 | |
| 71 | return list; |
| 72 | } |
| 73 | |
| 74 | |
| 75 | static object * |
| 76 | strop_splitfields(self, args) |
| 77 | object *self; /* Not used */ |
| 78 | object *args; |
| 79 | { |
| 80 | int len, n, i, j; |
| 81 | char *s, *sub; |
| 82 | char c; |
| 83 | object *list, *item; |
| 84 | |
| 85 | if (!getargs(args, "(s#s#)", &s, &len, &sub, &n)) |
| 86 | return NULL; |
| 87 | if (n == 0) { |
| 88 | err_setstr(ValueError, "empty separator"); |
| 89 | return NULL; |
| 90 | } |
| 91 | |
| 92 | list = newlistobject(0); |
| 93 | if (list == NULL) |
| 94 | return NULL; |
| 95 | |
| 96 | i = j = 0; |
| 97 | while (i+n <= len) { |
| 98 | if (s[i] == sub[0] && (n == 1 || strncmp(s+i, sub, n) == 0)) { |
| 99 | item = newsizedstringobject(s+j, (int)(i-j)); |
| 100 | if (item == NULL || addlistitem(list, item) < 0) { |
| 101 | DECREF(list); |
| 102 | return NULL; |
| 103 | } |
| 104 | i = j = i + n; |
| 105 | } |
| 106 | else |
| 107 | i++; |
| 108 | } |
| 109 | item = newsizedstringobject(s+j, (int)(len-j)); |
| 110 | if (item == NULL || addlistitem(list, item) < 0) { |
| 111 | DECREF(list); |
| 112 | return NULL; |
| 113 | } |
| 114 | |
| 115 | return list; |
| 116 | } |
| 117 | |
| 118 | |
| 119 | static object * |
Guido van Rossum | c89705d | 1992-11-26 08:54:07 +0000 | [diff] [blame] | 120 | strop_joinfields(self, args) |
| 121 | object *self; /* Not used */ |
| 122 | object *args; |
| 123 | { |
| 124 | object *seq, *item, *res; |
| 125 | object * (*getitem) FPROTO((object *, int)); |
| 126 | char *sep, *p; |
| 127 | int seplen, seqlen, reslen, itemlen, i; |
| 128 | |
| 129 | if (!getargs(args, "(Os#)", &seq, &sep, &seplen)) |
| 130 | return NULL; |
| 131 | if (is_listobject(seq)) { |
| 132 | getitem = getlistitem; |
| 133 | seqlen = getlistsize(seq); |
| 134 | } |
| 135 | else if (is_tupleobject(seq)) { |
| 136 | getitem = gettupleitem; |
| 137 | seqlen = gettuplesize(seq); |
| 138 | } |
| 139 | else { |
| 140 | err_setstr(TypeError, "first argument must be list/tuple"); |
| 141 | return NULL; |
| 142 | } |
| 143 | reslen = 0; |
| 144 | for (i = 0; i < seqlen; i++) { |
| 145 | item = getitem(seq, i); |
| 146 | if (!is_stringobject(item)) { |
| 147 | err_setstr(TypeError, |
| 148 | "first argument must be list/tuple of strings"); |
| 149 | return NULL; |
| 150 | } |
| 151 | if (i > 0) |
| 152 | reslen = reslen + seplen; |
| 153 | reslen = reslen + getstringsize(item); |
| 154 | } |
| 155 | if (seqlen == 1) { |
| 156 | /* Optimization if there's only one item */ |
| 157 | item = getitem(seq, 0); |
| 158 | INCREF(item); |
| 159 | return item; |
| 160 | } |
| 161 | res = newsizedstringobject((char *)NULL, reslen); |
| 162 | if (res == NULL) |
| 163 | return NULL; |
| 164 | p = getstringvalue(res); |
| 165 | for (i = 0; i < seqlen; i++) { |
| 166 | item = getitem(seq, i); |
| 167 | if (i > 0) { |
| 168 | memcpy(p, sep, seplen); |
| 169 | p += seplen; |
| 170 | } |
| 171 | itemlen = getstringsize(item); |
| 172 | memcpy(p, getstringvalue(item), itemlen); |
| 173 | p += itemlen; |
| 174 | } |
| 175 | if (p != getstringvalue(res) + reslen) { |
| 176 | err_setstr(SystemError, "strop.joinfields: assertion failed"); |
| 177 | return NULL; |
| 178 | } |
| 179 | return res; |
| 180 | } |
| 181 | |
| 182 | |
| 183 | static object * |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 184 | strop_index(self, args) |
| 185 | object *self; /* Not used */ |
| 186 | object *args; |
| 187 | { |
| 188 | char *s, *sub; |
| 189 | int len, n, i; |
| 190 | |
| 191 | if (getargs(args, "(s#s#i)", &s, &len, &sub, &n, &i)) { |
| 192 | if (i < 0 || i+n > len) { |
| 193 | err_setstr(ValueError, "start offset out of range"); |
| 194 | return NULL; |
| 195 | } |
| 196 | } |
| 197 | else { |
| 198 | err_clear(); |
| 199 | if (!getargs(args, "(s#s#)", &s, &len, &sub, &n)) |
| 200 | return NULL; |
| 201 | i = 0; |
| 202 | } |
| 203 | |
| 204 | if (n == 0) |
| 205 | return newintobject((long)i); |
| 206 | |
| 207 | len -= n; |
| 208 | for (; i <= len; i++) { |
| 209 | if (s[i] == sub[0]) { |
| 210 | if (n == 1 || strncmp(s+i, sub, n) == 0) |
| 211 | return newintobject((long)i); |
| 212 | } |
| 213 | } |
| 214 | |
| 215 | err_setstr(ValueError, "substring not found"); |
| 216 | return NULL; |
| 217 | } |
| 218 | |
| 219 | |
| 220 | static object * |
| 221 | strop_strip(self, args) |
| 222 | object *self; /* Not used */ |
| 223 | object *args; |
| 224 | { |
| 225 | char *s; |
| 226 | int len, i, j; |
| 227 | char c; |
| 228 | |
| 229 | if (!getargs(args, "s#", &s, &len)) |
| 230 | return NULL; |
| 231 | |
| 232 | i = 0; |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 233 | while (i < len && ((c = s[i]), isspace(c))) { |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 234 | i++; |
| 235 | } |
| 236 | |
| 237 | j = len; |
| 238 | do { |
| 239 | j--; |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 240 | } while (j >= i && ((c = s[j]), isspace(c))); |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 241 | j++; |
| 242 | |
| 243 | if (i == 0 && j == len) { |
| 244 | INCREF(args); |
| 245 | return args; |
| 246 | } |
| 247 | else |
| 248 | return newsizedstringobject(s+i, j-i); |
| 249 | } |
| 250 | |
| 251 | |
Guido van Rossum | 5c85062 | 1992-09-11 23:55:51 +0000 | [diff] [blame] | 252 | static object * |
| 253 | strop_lower(self, args) |
| 254 | object *self; /* Not used */ |
| 255 | object *args; |
| 256 | { |
| 257 | char *s; |
| 258 | int i, n; |
| 259 | object *new; |
| 260 | int changed; |
| 261 | |
| 262 | if (!getargs(args, "s#", &s, &n)) |
| 263 | return NULL; |
| 264 | new = newsizedstringobject(s, n); |
| 265 | if (new == NULL) |
| 266 | return NULL; |
| 267 | s = getstringvalue(new); |
| 268 | changed = 0; |
| 269 | for (i = 0; i < n; i++) { |
| 270 | char c = s[i]; |
| 271 | if (isupper(c)) { |
| 272 | changed = 1; |
| 273 | s[i] = tolower(c); |
| 274 | } |
| 275 | } |
| 276 | if (!changed) { |
| 277 | DECREF(new); |
| 278 | INCREF(args); |
| 279 | return args; |
| 280 | } |
| 281 | return new; |
| 282 | } |
| 283 | |
| 284 | |
| 285 | static object * |
| 286 | strop_upper(self, args) |
| 287 | object *self; /* Not used */ |
| 288 | object *args; |
| 289 | { |
| 290 | char *s; |
| 291 | int i, n; |
| 292 | object *new; |
| 293 | int changed; |
| 294 | |
| 295 | if (!getargs(args, "s#", &s, &n)) |
| 296 | return NULL; |
| 297 | new = newsizedstringobject(s, n); |
| 298 | if (new == NULL) |
| 299 | return NULL; |
| 300 | s = getstringvalue(new); |
| 301 | changed = 0; |
| 302 | for (i = 0; i < n; i++) { |
| 303 | char c = s[i]; |
| 304 | if (islower(c)) { |
| 305 | changed = 1; |
| 306 | s[i] = toupper(c); |
| 307 | } |
| 308 | } |
| 309 | if (!changed) { |
| 310 | DECREF(new); |
| 311 | INCREF(args); |
| 312 | return args; |
| 313 | } |
| 314 | return new; |
| 315 | } |
| 316 | |
| 317 | |
| 318 | static object * |
| 319 | strop_swapcase(self, args) |
| 320 | object *self; /* Not used */ |
| 321 | object *args; |
| 322 | { |
| 323 | char *s; |
| 324 | int i, n; |
| 325 | object *new; |
| 326 | int changed; |
| 327 | |
| 328 | if (!getargs(args, "s#", &s, &n)) |
| 329 | return NULL; |
| 330 | new = newsizedstringobject(s, n); |
| 331 | if (new == NULL) |
| 332 | return NULL; |
| 333 | s = getstringvalue(new); |
| 334 | changed = 0; |
| 335 | for (i = 0; i < n; i++) { |
| 336 | char c = s[i]; |
| 337 | if (islower(c)) { |
| 338 | changed = 1; |
| 339 | s[i] = toupper(c); |
| 340 | } |
| 341 | else if (isupper(c)) { |
| 342 | changed = 1; |
| 343 | s[i] = tolower(c); |
| 344 | } |
| 345 | } |
| 346 | if (!changed) { |
| 347 | DECREF(new); |
| 348 | INCREF(args); |
| 349 | return args; |
| 350 | } |
| 351 | return new; |
| 352 | } |
| 353 | |
| 354 | |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 355 | /* List of functions defined in the module */ |
| 356 | |
| 357 | static struct methodlist strop_methods[] = { |
| 358 | {"index", strop_index}, |
Guido van Rossum | c89705d | 1992-11-26 08:54:07 +0000 | [diff] [blame] | 359 | {"joinfields", strop_joinfields}, |
Guido van Rossum | 5c85062 | 1992-09-11 23:55:51 +0000 | [diff] [blame] | 360 | {"lower", strop_lower}, |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 361 | {"split", strop_split}, |
| 362 | {"splitfields", strop_splitfields}, |
| 363 | {"strip", strop_strip}, |
Guido van Rossum | 5c85062 | 1992-09-11 23:55:51 +0000 | [diff] [blame] | 364 | {"swapcase", strop_swapcase}, |
| 365 | {"upper", strop_upper}, |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 366 | {NULL, NULL} /* sentinel */ |
| 367 | }; |
| 368 | |
| 369 | |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 370 | void |
| 371 | initstrop() |
| 372 | { |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 373 | object *m, *d, *s; |
| 374 | char buf[256]; |
| 375 | int c, n; |
| 376 | m = initmodule("strop", strop_methods); |
| 377 | d = getmoduledict(m); |
Guido van Rossum | e22e644 | 1993-07-09 10:51:31 +0000 | [diff] [blame] | 378 | |
| 379 | /* Create 'whitespace' object */ |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 380 | n = 0; |
| 381 | for (c = 1; c < 256; c++) { |
| 382 | if (isspace(c)) |
| 383 | buf[n++] = c; |
| 384 | } |
Sjoerd Mullender | d96ec44 | 1993-09-14 08:37:39 +0000 | [diff] [blame] | 385 | s = newsizedstringobject(buf, n); |
Guido van Rossum | e22e644 | 1993-07-09 10:51:31 +0000 | [diff] [blame] | 386 | if (s) { |
Guido van Rossum | e22e644 | 1993-07-09 10:51:31 +0000 | [diff] [blame] | 387 | dictinsert(d, "whitespace", s); |
| 388 | DECREF(s); |
| 389 | } |
| 390 | /* Create 'lowercase' object */ |
| 391 | n = 0; |
| 392 | for (c = 1; c < 256; c++) { |
| 393 | if (islower(c)) |
| 394 | buf[n++] = c; |
| 395 | } |
Guido van Rossum | d05eb8b | 1993-07-08 11:12:36 +0000 | [diff] [blame] | 396 | s = newsizedstringobject(buf, n); |
Guido van Rossum | e22e644 | 1993-07-09 10:51:31 +0000 | [diff] [blame] | 397 | if (s) { |
| 398 | dictinsert(d, "lowercase", s); |
| 399 | DECREF(s); |
| 400 | } |
| 401 | |
| 402 | /* Create 'uppercase' object */ |
| 403 | n = 0; |
| 404 | for (c = 1; c < 256; c++) { |
| 405 | if (isupper(c)) |
| 406 | buf[n++] = c; |
| 407 | } |
| 408 | s = newsizedstringobject(buf, n); |
| 409 | if (s) { |
| 410 | dictinsert(d, "uppercase", s); |
| 411 | DECREF(s); |
| 412 | } |
| 413 | |
| 414 | if (err_occurred()) |
| 415 | fatal("can't initialize module strop"); |
Guido van Rossum | e270b43 | 1992-09-03 20:21:07 +0000 | [diff] [blame] | 416 | } |