| 1 |
#if STRINGLIB_IS_UNICODE |
| 2 |
# error "transmogrify.h only compatible with byte-wise strings" |
| 3 |
#endif |
| 4 |
|
| 5 |
/* the more complicated methods. parts of these should be pulled out into the |
| 6 |
shared code in bytes_methods.c to cut down on duplicate code bloat. */ |
| 7 |
|
| 8 |
/*[clinic input] |
| 9 |
class B "PyObject *" "&PyType_Type" |
| 10 |
[clinic start generated code]*/ |
| 11 |
/*[clinic end generated code: output=da39a3ee5e6b4b0d input=2935558188d97c76]*/ |
| 12 |
|
| 13 |
#include "clinic/transmogrify.h.h" |
| 14 |
|
| 15 |
static inline PyObject * |
| 16 |
return_self(PyObject *self) |
| 17 |
{ |
| 18 |
#if !STRINGLIB_MUTABLE |
| 19 |
if (STRINGLIB_CHECK_EXACT(self)) { |
| 20 |
return Py_NewRef(self); |
| 21 |
} |
| 22 |
#endif |
| 23 |
return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); |
| 24 |
} |
| 25 |
|
| 26 |
/*[clinic input] |
| 27 |
B.expandtabs as stringlib_expandtabs |
| 28 |
|
| 29 |
tabsize: int = 8 |
| 30 |
|
| 31 |
Return a copy where all tab characters are expanded using spaces. |
| 32 |
|
| 33 |
If tabsize is not given, a tab size of 8 characters is assumed. |
| 34 |
[clinic start generated code]*/ |
| 35 |
|
| 36 |
static PyObject * |
| 37 |
stringlib_expandtabs_impl(PyObject *self, int tabsize) |
| 38 |
/*[clinic end generated code: output=069cb7fae72e4c2b input=3c6d3b12aa3ccbea]*/ |
| 39 |
{ |
| 40 |
const char *e, *p; |
| 41 |
char *q; |
| 42 |
Py_ssize_t i, j; |
| 43 |
PyObject *u; |
| 44 |
|
| 45 |
/* First pass: determine size of output string */ |
| 46 |
i = j = 0; |
| 47 |
e = STRINGLIB_STR(self) + STRINGLIB_LEN(self); |
| 48 |
for (p = STRINGLIB_STR(self); p < e; p++) { |
| 49 |
if (*p == '\t') { |
| 50 |
if (tabsize > 0) { |
| 51 |
Py_ssize_t incr = tabsize - (j % tabsize); |
| 52 |
if (j > PY_SSIZE_T_MAX - incr) |
| 53 |
goto overflow; |
| 54 |
j += incr; |
| 55 |
} |
| 56 |
} |
| 57 |
else { |
| 58 |
if (j > PY_SSIZE_T_MAX - 1) |
| 59 |
goto overflow; |
| 60 |
j++; |
| 61 |
if (*p == '\n' || *p == '\r') { |
| 62 |
if (i > PY_SSIZE_T_MAX - j) |
| 63 |
goto overflow; |
| 64 |
i += j; |
| 65 |
j = 0; |
| 66 |
} |
| 67 |
} |
| 68 |
} |
| 69 |
|
| 70 |
if (i > PY_SSIZE_T_MAX - j) |
| 71 |
goto overflow; |
| 72 |
|
| 73 |
/* Second pass: create output string and fill it */ |
| 74 |
u = STRINGLIB_NEW(NULL, i + j); |
| 75 |
if (!u) |
| 76 |
return NULL; |
| 77 |
|
| 78 |
j = 0; |
| 79 |
q = STRINGLIB_STR(u); |
| 80 |
|
| 81 |
for (p = STRINGLIB_STR(self); p < e; p++) { |
| 82 |
if (*p == '\t') { |
| 83 |
if (tabsize > 0) { |
| 84 |
i = tabsize - (j % tabsize); |
| 85 |
j += i; |
| 86 |
while (i--) |
| 87 |
*q++ = ' '; |
| 88 |
} |
| 89 |
} |
| 90 |
else { |
| 91 |
j++; |
| 92 |
*q++ = *p; |
| 93 |
if (*p == '\n' || *p == '\r') |
| 94 |
j = 0; |
| 95 |
} |
| 96 |
} |
| 97 |
|
| 98 |
return u; |
| 99 |
overflow: |
| 100 |
PyErr_SetString(PyExc_OverflowError, "result too long"); |
| 101 |
return NULL; |
| 102 |
} |
| 103 |
|
| 104 |
static inline PyObject * |
| 105 |
pad(PyObject *self, Py_ssize_t left, Py_ssize_t right, char fill) |
| 106 |
{ |
| 107 |
PyObject *u; |
| 108 |
|
| 109 |
if (left < 0) |
| 110 |
left = 0; |
| 111 |
if (right < 0) |
| 112 |
right = 0; |
| 113 |
|
| 114 |
if (left == 0 && right == 0) { |
| 115 |
return return_self(self); |
| 116 |
} |
| 117 |
|
| 118 |
u = STRINGLIB_NEW(NULL, left + STRINGLIB_LEN(self) + right); |
| 119 |
if (u) { |
| 120 |
if (left) |
| 121 |
memset(STRINGLIB_STR(u), fill, left); |
| 122 |
memcpy(STRINGLIB_STR(u) + left, |
| 123 |
STRINGLIB_STR(self), |
| 124 |
STRINGLIB_LEN(self)); |
| 125 |
if (right) |
| 126 |
memset(STRINGLIB_STR(u) + left + STRINGLIB_LEN(self), |
| 127 |
fill, right); |
| 128 |
} |
| 129 |
|
| 130 |
return u; |
| 131 |
} |
| 132 |
|
| 133 |
/*[clinic input] |
| 134 |
B.ljust as stringlib_ljust |
| 135 |
|
| 136 |
width: Py_ssize_t |
| 137 |
fillchar: char = b' ' |
| 138 |
/ |
| 139 |
|
| 140 |
Return a left-justified string of length width. |
| 141 |
|
| 142 |
Padding is done using the specified fill character. |
| 143 |
[clinic start generated code]*/ |
| 144 |
|
| 145 |
static PyObject * |
| 146 |
stringlib_ljust_impl(PyObject *self, Py_ssize_t width, char fillchar) |
| 147 |
/*[clinic end generated code: output=c79ca173c5ff8337 input=eff2d014bc7d80df]*/ |
| 148 |
{ |
| 149 |
if (STRINGLIB_LEN(self) >= width) { |
| 150 |
return return_self(self); |
| 151 |
} |
| 152 |
|
| 153 |
return pad(self, 0, width - STRINGLIB_LEN(self), fillchar); |
| 154 |
} |
| 155 |
|
| 156 |
|
| 157 |
/*[clinic input] |
| 158 |
B.rjust as stringlib_rjust |
| 159 |
|
| 160 |
width: Py_ssize_t |
| 161 |
fillchar: char = b' ' |
| 162 |
/ |
| 163 |
|
| 164 |
Return a right-justified string of length width. |
| 165 |
|
| 166 |
Padding is done using the specified fill character. |
| 167 |
[clinic start generated code]*/ |
| 168 |
|
| 169 |
static PyObject * |
| 170 |
stringlib_rjust_impl(PyObject *self, Py_ssize_t width, char fillchar) |
| 171 |
/*[clinic end generated code: output=7df5d728a5439570 input=218b0bd31308955d]*/ |
| 172 |
{ |
| 173 |
if (STRINGLIB_LEN(self) >= width) { |
| 174 |
return return_self(self); |
| 175 |
} |
| 176 |
|
| 177 |
return pad(self, width - STRINGLIB_LEN(self), 0, fillchar); |
| 178 |
} |
| 179 |
|
| 180 |
|
| 181 |
/*[clinic input] |
| 182 |
B.center as stringlib_center |
| 183 |
|
| 184 |
width: Py_ssize_t |
| 185 |
fillchar: char = b' ' |
| 186 |
/ |
| 187 |
|
| 188 |
Return a centered string of length width. |
| 189 |
|
| 190 |
Padding is done using the specified fill character. |
| 191 |
[clinic start generated code]*/ |
| 192 |
|
| 193 |
static PyObject * |
| 194 |
stringlib_center_impl(PyObject *self, Py_ssize_t width, char fillchar) |
| 195 |
/*[clinic end generated code: output=d8da2e055288b4c2 input=3776fd278765d89b]*/ |
| 196 |
{ |
| 197 |
Py_ssize_t marg, left; |
| 198 |
|
| 199 |
if (STRINGLIB_LEN(self) >= width) { |
| 200 |
return return_self(self); |
| 201 |
} |
| 202 |
|
| 203 |
marg = width - STRINGLIB_LEN(self); |
| 204 |
left = marg / 2 + (marg & width & 1); |
| 205 |
|
| 206 |
return pad(self, left, marg - left, fillchar); |
| 207 |
} |
| 208 |
|
| 209 |
/*[clinic input] |
| 210 |
B.zfill as stringlib_zfill |
| 211 |
|
| 212 |
width: Py_ssize_t |
| 213 |
/ |
| 214 |
|
| 215 |
Pad a numeric string with zeros on the left, to fill a field of the given width. |
| 216 |
|
| 217 |
The original string is never truncated. |
| 218 |
[clinic start generated code]*/ |
| 219 |
|
| 220 |
static PyObject * |
| 221 |
stringlib_zfill_impl(PyObject *self, Py_ssize_t width) |
| 222 |
/*[clinic end generated code: output=0b3c684a7f1b2319 input=2da6d7b8e9bcb19a]*/ |
| 223 |
{ |
| 224 |
Py_ssize_t fill; |
| 225 |
PyObject *s; |
| 226 |
char *p; |
| 227 |
|
| 228 |
if (STRINGLIB_LEN(self) >= width) { |
| 229 |
return return_self(self); |
| 230 |
} |
| 231 |
|
| 232 |
fill = width - STRINGLIB_LEN(self); |
| 233 |
|
| 234 |
s = pad(self, fill, 0, '0'); |
| 235 |
|
| 236 |
if (s == NULL) |
| 237 |
return NULL; |
| 238 |
|
| 239 |
p = STRINGLIB_STR(s); |
| 240 |
if (p[fill] == '+' || p[fill] == '-') { |
| 241 |
/* move sign to beginning of string */ |
| 242 |
p[0] = p[fill]; |
| 243 |
p[fill] = '0'; |
| 244 |
} |
| 245 |
|
| 246 |
return s; |
| 247 |
} |
| 248 |
|
| 249 |
|
| 250 |
/* find and count characters and substrings */ |
| 251 |
|
| 252 |
#define findchar(target, target_len, c) \ |
| 253 |
((char *)memchr((const void *)(target), c, target_len)) |
| 254 |
|
| 255 |
|
| 256 |
static Py_ssize_t |
| 257 |
countchar(const char *target, Py_ssize_t target_len, char c, |
| 258 |
Py_ssize_t maxcount) |
| 259 |
{ |
| 260 |
Py_ssize_t count = 0; |
| 261 |
const char *start = target; |
| 262 |
const char *end = target + target_len; |
| 263 |
|
| 264 |
while ((start = findchar(start, end - start, c)) != NULL) { |
| 265 |
count++; |
| 266 |
if (count >= maxcount) |
| 267 |
break; |
| 268 |
start += 1; |
| 269 |
} |
| 270 |
return count; |
| 271 |
} |
| 272 |
|
| 273 |
|
| 274 |
/* Algorithms for different cases of string replacement */ |
| 275 |
|
| 276 |
/* len(self)>=1, from="", len(to)>=1, maxcount>=1 */ |
| 277 |
static PyObject * |
| 278 |
stringlib_replace_interleave(PyObject *self, |
| 279 |
const char *to_s, Py_ssize_t to_len, |
| 280 |
Py_ssize_t maxcount) |
| 281 |
{ |
| 282 |
const char *self_s; |
| 283 |
char *result_s; |
| 284 |
Py_ssize_t self_len, result_len; |
| 285 |
Py_ssize_t count, i; |
| 286 |
PyObject *result; |
| 287 |
|
| 288 |
self_len = STRINGLIB_LEN(self); |
| 289 |
|
| 290 |
/* 1 at the end plus 1 after every character; |
| 291 |
count = min(maxcount, self_len + 1) */ |
| 292 |
if (maxcount <= self_len) { |
| 293 |
count = maxcount; |
| 294 |
} |
| 295 |
else { |
| 296 |
/* Can't overflow: self_len + 1 <= maxcount <= PY_SSIZE_T_MAX. */ |
| 297 |
count = self_len + 1; |
| 298 |
} |
| 299 |
|
| 300 |
/* Check for overflow */ |
| 301 |
/* result_len = count * to_len + self_len; */ |
| 302 |
assert(count > 0); |
| 303 |
if (to_len > (PY_SSIZE_T_MAX - self_len) / count) { |
| 304 |
PyErr_SetString(PyExc_OverflowError, |
| 305 |
"replace bytes is too long"); |
| 306 |
return NULL; |
| 307 |
} |
| 308 |
result_len = count * to_len + self_len; |
| 309 |
result = STRINGLIB_NEW(NULL, result_len); |
| 310 |
if (result == NULL) { |
| 311 |
return NULL; |
| 312 |
} |
| 313 |
|
| 314 |
self_s = STRINGLIB_STR(self); |
| 315 |
result_s = STRINGLIB_STR(result); |
| 316 |
|
| 317 |
if (to_len > 1) { |
| 318 |
/* Lay the first one down (guaranteed this will occur) */ |
| 319 |
memcpy(result_s, to_s, to_len); |
| 320 |
result_s += to_len; |
| 321 |
count -= 1; |
| 322 |
|
| 323 |
for (i = 0; i < count; i++) { |
| 324 |
*result_s++ = *self_s++; |
| 325 |
memcpy(result_s, to_s, to_len); |
| 326 |
result_s += to_len; |
| 327 |
} |
| 328 |
} |
| 329 |
else { |
| 330 |
result_s[0] = to_s[0]; |
| 331 |
result_s += to_len; |
| 332 |
count -= 1; |
| 333 |
for (i = 0; i < count; i++) { |
| 334 |
*result_s++ = *self_s++; |
| 335 |
result_s[0] = to_s[0]; |
| 336 |
result_s += to_len; |
| 337 |
} |
| 338 |
} |
| 339 |
|
| 340 |
/* Copy the rest of the original string */ |
| 341 |
memcpy(result_s, self_s, self_len - i); |
| 342 |
|
| 343 |
return result; |
| 344 |
} |
| 345 |
|
| 346 |
/* Special case for deleting a single character */ |
| 347 |
/* len(self)>=1, len(from)==1, to="", maxcount>=1 */ |
| 348 |
static PyObject * |
| 349 |
stringlib_replace_delete_single_character(PyObject *self, |
| 350 |
char from_c, Py_ssize_t maxcount) |
| 351 |
{ |
| 352 |
const char *self_s, *start, *next, *end; |
| 353 |
char *result_s; |
| 354 |
Py_ssize_t self_len, result_len; |
| 355 |
Py_ssize_t count; |
| 356 |
PyObject *result; |
| 357 |
|
| 358 |
self_len = STRINGLIB_LEN(self); |
| 359 |
self_s = STRINGLIB_STR(self); |
| 360 |
|
| 361 |
count = countchar(self_s, self_len, from_c, maxcount); |
| 362 |
if (count == 0) { |
| 363 |
return return_self(self); |
| 364 |
} |
| 365 |
|
| 366 |
result_len = self_len - count; /* from_len == 1 */ |
| 367 |
assert(result_len>=0); |
| 368 |
|
| 369 |
result = STRINGLIB_NEW(NULL, result_len); |
| 370 |
if (result == NULL) { |
| 371 |
return NULL; |
| 372 |
} |
| 373 |
result_s = STRINGLIB_STR(result); |
| 374 |
|
| 375 |
start = self_s; |
| 376 |
end = self_s + self_len; |
| 377 |
while (count-- > 0) { |
| 378 |
next = findchar(start, end - start, from_c); |
| 379 |
if (next == NULL) |
| 380 |
break; |
| 381 |
memcpy(result_s, start, next - start); |
| 382 |
result_s += (next - start); |
| 383 |
start = next + 1; |
| 384 |
} |
| 385 |
memcpy(result_s, start, end - start); |
| 386 |
|
| 387 |
return result; |
| 388 |
} |
| 389 |
|
| 390 |
/* len(self)>=1, len(from)>=2, to="", maxcount>=1 */ |
| 391 |
|
| 392 |
static PyObject * |
| 393 |
stringlib_replace_delete_substring(PyObject *self, |
| 394 |
const char *from_s, Py_ssize_t from_len, |
| 395 |
Py_ssize_t maxcount) |
| 396 |
{ |
| 397 |
const char *self_s, *start, *next, *end; |
| 398 |
char *result_s; |
| 399 |
Py_ssize_t self_len, result_len; |
| 400 |
Py_ssize_t count, offset; |
| 401 |
PyObject *result; |
| 402 |
|
| 403 |
self_len = STRINGLIB_LEN(self); |
| 404 |
self_s = STRINGLIB_STR(self); |
| 405 |
|
| 406 |
count = stringlib_count(self_s, self_len, |
| 407 |
from_s, from_len, |
| 408 |
maxcount); |
| 409 |
|
| 410 |
if (count == 0) { |
| 411 |
/* no matches */ |
| 412 |
return return_self(self); |
| 413 |
} |
| 414 |
|
| 415 |
result_len = self_len - (count * from_len); |
| 416 |
assert (result_len>=0); |
| 417 |
|
| 418 |
result = STRINGLIB_NEW(NULL, result_len); |
| 419 |
if (result == NULL) { |
| 420 |
return NULL; |
| 421 |
} |
| 422 |
result_s = STRINGLIB_STR(result); |
| 423 |
|
| 424 |
start = self_s; |
| 425 |
end = self_s + self_len; |
| 426 |
while (count-- > 0) { |
| 427 |
offset = stringlib_find(start, end - start, |
| 428 |
from_s, from_len, |
| 429 |
0); |
| 430 |
if (offset == -1) |
| 431 |
break; |
| 432 |
next = start + offset; |
| 433 |
|
| 434 |
memcpy(result_s, start, next - start); |
| 435 |
|
| 436 |
result_s += (next - start); |
| 437 |
start = next + from_len; |
| 438 |
} |
| 439 |
memcpy(result_s, start, end - start); |
| 440 |
return result; |
| 441 |
} |
| 442 |
|
| 443 |
/* len(self)>=1, len(from)==len(to)==1, maxcount>=1 */ |
| 444 |
static PyObject * |
| 445 |
stringlib_replace_single_character_in_place(PyObject *self, |
| 446 |
char from_c, char to_c, |
| 447 |
Py_ssize_t maxcount) |
| 448 |
{ |
| 449 |
const char *self_s, *end; |
| 450 |
char *result_s, *start, *next; |
| 451 |
Py_ssize_t self_len; |
| 452 |
PyObject *result; |
| 453 |
|
| 454 |
/* The result string will be the same size */ |
| 455 |
self_s = STRINGLIB_STR(self); |
| 456 |
self_len = STRINGLIB_LEN(self); |
| 457 |
|
| 458 |
next = findchar(self_s, self_len, from_c); |
| 459 |
|
| 460 |
if (next == NULL) { |
| 461 |
/* No matches; return the original bytes */ |
| 462 |
return return_self(self); |
| 463 |
} |
| 464 |
|
| 465 |
/* Need to make a new bytes */ |
| 466 |
result = STRINGLIB_NEW(NULL, self_len); |
| 467 |
if (result == NULL) { |
| 468 |
return NULL; |
| 469 |
} |
| 470 |
result_s = STRINGLIB_STR(result); |
| 471 |
memcpy(result_s, self_s, self_len); |
| 472 |
|
| 473 |
/* change everything in-place, starting with this one */ |
| 474 |
start = result_s + (next - self_s); |
| 475 |
*start = to_c; |
| 476 |
start++; |
| 477 |
end = result_s + self_len; |
| 478 |
|
| 479 |
while (--maxcount > 0) { |
| 480 |
next = findchar(start, end - start, from_c); |
| 481 |
if (next == NULL) |
| 482 |
break; |
| 483 |
*next = to_c; |
| 484 |
start = next + 1; |
| 485 |
} |
| 486 |
|
| 487 |
return result; |
| 488 |
} |
| 489 |
|
| 490 |
/* len(self)>=1, len(from)==len(to)>=2, maxcount>=1 */ |
| 491 |
static PyObject * |
| 492 |
stringlib_replace_substring_in_place(PyObject *self, |
| 493 |
const char *from_s, Py_ssize_t from_len, |
| 494 |
const char *to_s, Py_ssize_t to_len, |
| 495 |
Py_ssize_t maxcount) |
| 496 |
{ |
| 497 |
const char *self_s, *end; |
| 498 |
char *result_s, *start; |
| 499 |
Py_ssize_t self_len, offset; |
| 500 |
PyObject *result; |
| 501 |
|
| 502 |
/* The result bytes will be the same size */ |
| 503 |
|
| 504 |
self_s = STRINGLIB_STR(self); |
| 505 |
self_len = STRINGLIB_LEN(self); |
| 506 |
|
| 507 |
offset = stringlib_find(self_s, self_len, |
| 508 |
from_s, from_len, |
| 509 |
0); |
| 510 |
if (offset == -1) { |
| 511 |
/* No matches; return the original bytes */ |
| 512 |
return return_self(self); |
| 513 |
} |
| 514 |
|
| 515 |
/* Need to make a new bytes */ |
| 516 |
result = STRINGLIB_NEW(NULL, self_len); |
| 517 |
if (result == NULL) { |
| 518 |
return NULL; |
| 519 |
} |
| 520 |
result_s = STRINGLIB_STR(result); |
| 521 |
memcpy(result_s, self_s, self_len); |
| 522 |
|
| 523 |
/* change everything in-place, starting with this one */ |
| 524 |
start = result_s + offset; |
| 525 |
memcpy(start, to_s, from_len); |
| 526 |
start += from_len; |
| 527 |
end = result_s + self_len; |
| 528 |
|
| 529 |
while ( --maxcount > 0) { |
| 530 |
offset = stringlib_find(start, end - start, |
| 531 |
from_s, from_len, |
| 532 |
0); |
| 533 |
if (offset == -1) |
| 534 |
break; |
| 535 |
memcpy(start + offset, to_s, from_len); |
| 536 |
start += offset + from_len; |
| 537 |
} |
| 538 |
|
| 539 |
return result; |
| 540 |
} |
| 541 |
|
| 542 |
/* len(self)>=1, len(from)==1, len(to)>=2, maxcount>=1 */ |
| 543 |
static PyObject * |
| 544 |
stringlib_replace_single_character(PyObject *self, |
| 545 |
char from_c, |
| 546 |
const char *to_s, Py_ssize_t to_len, |
| 547 |
Py_ssize_t maxcount) |
| 548 |
{ |
| 549 |
const char *self_s, *start, *next, *end; |
| 550 |
char *result_s; |
| 551 |
Py_ssize_t self_len, result_len; |
| 552 |
Py_ssize_t count; |
| 553 |
PyObject *result; |
| 554 |
|
| 555 |
self_s = STRINGLIB_STR(self); |
| 556 |
self_len = STRINGLIB_LEN(self); |
| 557 |
|
| 558 |
count = countchar(self_s, self_len, from_c, maxcount); |
| 559 |
if (count == 0) { |
| 560 |
/* no matches, return unchanged */ |
| 561 |
return return_self(self); |
| 562 |
} |
| 563 |
|
| 564 |
/* use the difference between current and new, hence the "-1" */ |
| 565 |
/* result_len = self_len + count * (to_len-1) */ |
| 566 |
assert(count > 0); |
| 567 |
if (to_len - 1 > (PY_SSIZE_T_MAX - self_len) / count) { |
| 568 |
PyErr_SetString(PyExc_OverflowError, "replace bytes is too long"); |
| 569 |
return NULL; |
| 570 |
} |
| 571 |
result_len = self_len + count * (to_len - 1); |
| 572 |
|
| 573 |
result = STRINGLIB_NEW(NULL, result_len); |
| 574 |
if (result == NULL) { |
| 575 |
return NULL; |
| 576 |
} |
| 577 |
result_s = STRINGLIB_STR(result); |
| 578 |
|
| 579 |
start = self_s; |
| 580 |
end = self_s + self_len; |
| 581 |
while (count-- > 0) { |
| 582 |
next = findchar(start, end - start, from_c); |
| 583 |
if (next == NULL) |
| 584 |
break; |
| 585 |
|
| 586 |
if (next == start) { |
| 587 |
/* replace with the 'to' */ |
| 588 |
memcpy(result_s, to_s, to_len); |
| 589 |
result_s += to_len; |
| 590 |
start += 1; |
| 591 |
} else { |
| 592 |
/* copy the unchanged old then the 'to' */ |
| 593 |
memcpy(result_s, start, next - start); |
| 594 |
result_s += (next - start); |
| 595 |
memcpy(result_s, to_s, to_len); |
| 596 |
result_s += to_len; |
| 597 |
start = next + 1; |
| 598 |
} |
| 599 |
} |
| 600 |
/* Copy the remainder of the remaining bytes */ |
| 601 |
memcpy(result_s, start, end - start); |
| 602 |
|
| 603 |
return result; |
| 604 |
} |
| 605 |
|
| 606 |
/* len(self)>=1, len(from)>=2, len(to)>=2, maxcount>=1 */ |
| 607 |
static PyObject * |
| 608 |
stringlib_replace_substring(PyObject *self, |
| 609 |
const char *from_s, Py_ssize_t from_len, |
| 610 |
const char *to_s, Py_ssize_t to_len, |
| 611 |
Py_ssize_t maxcount) |
| 612 |
{ |
| 613 |
const char *self_s, *start, *next, *end; |
| 614 |
char *result_s; |
| 615 |
Py_ssize_t self_len, result_len; |
| 616 |
Py_ssize_t count, offset; |
| 617 |
PyObject *result; |
| 618 |
|
| 619 |
self_s = STRINGLIB_STR(self); |
| 620 |
self_len = STRINGLIB_LEN(self); |
| 621 |
|
| 622 |
count = stringlib_count(self_s, self_len, |
| 623 |
from_s, from_len, |
| 624 |
maxcount); |
| 625 |
|
| 626 |
if (count == 0) { |
| 627 |
/* no matches, return unchanged */ |
| 628 |
return return_self(self); |
| 629 |
} |
| 630 |
|
| 631 |
/* Check for overflow */ |
| 632 |
/* result_len = self_len + count * (to_len-from_len) */ |
| 633 |
assert(count > 0); |
| 634 |
if (to_len - from_len > (PY_SSIZE_T_MAX - self_len) / count) { |
| 635 |
PyErr_SetString(PyExc_OverflowError, "replace bytes is too long"); |
| 636 |
return NULL; |
| 637 |
} |
| 638 |
result_len = self_len + count * (to_len - from_len); |
| 639 |
|
| 640 |
result = STRINGLIB_NEW(NULL, result_len); |
| 641 |
if (result == NULL) { |
| 642 |
return NULL; |
| 643 |
} |
| 644 |
result_s = STRINGLIB_STR(result); |
| 645 |
|
| 646 |
start = self_s; |
| 647 |
end = self_s + self_len; |
| 648 |
while (count-- > 0) { |
| 649 |
offset = stringlib_find(start, end - start, |
| 650 |
from_s, from_len, |
| 651 |
0); |
| 652 |
if (offset == -1) |
| 653 |
break; |
| 654 |
next = start + offset; |
| 655 |
if (next == start) { |
| 656 |
/* replace with the 'to' */ |
| 657 |
memcpy(result_s, to_s, to_len); |
| 658 |
result_s += to_len; |
| 659 |
start += from_len; |
| 660 |
} else { |
| 661 |
/* copy the unchanged old then the 'to' */ |
| 662 |
memcpy(result_s, start, next - start); |
| 663 |
result_s += (next - start); |
| 664 |
memcpy(result_s, to_s, to_len); |
| 665 |
result_s += to_len; |
| 666 |
start = next + from_len; |
| 667 |
} |
| 668 |
} |
| 669 |
/* Copy the remainder of the remaining bytes */ |
| 670 |
memcpy(result_s, start, end - start); |
| 671 |
|
| 672 |
return result; |
| 673 |
} |
| 674 |
|
| 675 |
|
| 676 |
static PyObject * |
| 677 |
stringlib_replace(PyObject *self, |
| 678 |
const char *from_s, Py_ssize_t from_len, |
| 679 |
const char *to_s, Py_ssize_t to_len, |
| 680 |
Py_ssize_t maxcount) |
| 681 |
{ |
| 682 |
if (STRINGLIB_LEN(self) < from_len) { |
| 683 |
/* nothing to do; return the original bytes */ |
| 684 |
return return_self(self); |
| 685 |
} |
| 686 |
if (maxcount < 0) { |
| 687 |
maxcount = PY_SSIZE_T_MAX; |
| 688 |
} else if (maxcount == 0) { |
| 689 |
/* nothing to do; return the original bytes */ |
| 690 |
return return_self(self); |
| 691 |
} |
| 692 |
|
| 693 |
/* Handle zero-length special cases */ |
| 694 |
if (from_len == 0) { |
| 695 |
if (to_len == 0) { |
| 696 |
/* nothing to do; return the original bytes */ |
| 697 |
return return_self(self); |
| 698 |
} |
| 699 |
/* insert the 'to' bytes everywhere. */ |
| 700 |
/* >>> b"Python".replace(b"", b".") */ |
| 701 |
/* b'.P.y.t.h.o.n.' */ |
| 702 |
return stringlib_replace_interleave(self, to_s, to_len, maxcount); |
| 703 |
} |
| 704 |
|
| 705 |
if (to_len == 0) { |
| 706 |
/* delete all occurrences of 'from' bytes */ |
| 707 |
if (from_len == 1) { |
| 708 |
return stringlib_replace_delete_single_character( |
| 709 |
self, from_s[0], maxcount); |
| 710 |
} else { |
| 711 |
return stringlib_replace_delete_substring( |
| 712 |
self, from_s, from_len, maxcount); |
| 713 |
} |
| 714 |
} |
| 715 |
|
| 716 |
/* Handle special case where both bytes have the same length */ |
| 717 |
|
| 718 |
if (from_len == to_len) { |
| 719 |
if (from_len == 1) { |
| 720 |
return stringlib_replace_single_character_in_place( |
| 721 |
self, from_s[0], to_s[0], maxcount); |
| 722 |
} else { |
| 723 |
return stringlib_replace_substring_in_place( |
| 724 |
self, from_s, from_len, to_s, to_len, maxcount); |
| 725 |
} |
| 726 |
} |
| 727 |
|
| 728 |
/* Otherwise use the more generic algorithms */ |
| 729 |
if (from_len == 1) { |
| 730 |
return stringlib_replace_single_character( |
| 731 |
self, from_s[0], to_s, to_len, maxcount); |
| 732 |
} else { |
| 733 |
/* len('from')>=2, len('to')>=1 */ |
| 734 |
return stringlib_replace_substring( |
| 735 |
self, from_s, from_len, to_s, to_len, maxcount); |
| 736 |
} |
| 737 |
} |
| 738 |
|
| 739 |
#undef findchar |