src/string.c

Sun, 23 Oct 2022 13:32:16 +0200

author
Mike Becker <universe@uap-core.de>
date
Sun, 23 Oct 2022 13:32:16 +0200
changeset 592
bb69ef3ad1f3
parent 591
7df0bcaecffa
child 593
ea9b41b5ebbc
permissions
-rw-r--r--

enclose macro arguments in parenthesis

universe@576 1 /*
universe@576 2 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER.
universe@576 3 *
universe@576 4 * Copyright 2021 Mike Becker, Olaf Wintermann All rights reserved.
universe@576 5 *
universe@576 6 * Redistribution and use in source and binary forms, with or without
universe@576 7 * modification, are permitted provided that the following conditions are met:
universe@576 8 *
universe@576 9 * 1. Redistributions of source code must retain the above copyright
universe@576 10 * notice, this list of conditions and the following disclaimer.
universe@576 11 *
universe@576 12 * 2. Redistributions in binary form must reproduce the above copyright
universe@576 13 * notice, this list of conditions and the following disclaimer in the
universe@576 14 * documentation and/or other materials provided with the distribution.
universe@576 15 *
universe@576 16 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
universe@576 17 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
universe@576 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
universe@576 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
universe@576 20 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
universe@576 21 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
universe@576 22 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
universe@576 23 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
universe@576 24 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
universe@576 25 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
universe@576 26 * POSSIBILITY OF SUCH DAMAGE.
universe@576 27 */
universe@576 28
universe@576 29 #include "cx/string.h"
universe@579 30 #include "cx/utils.h"
universe@579 31
universe@579 32 #include <string.h>
universe@579 33 #include <stdarg.h>
universe@579 34 #include <stdint.h>
universe@581 35 #include <ctype.h>
universe@581 36
universe@581 37 #ifndef _WIN32
universe@581 38
universe@581 39 #include <strings.h> /* for strncasecmp() */
universe@581 40
universe@581 41 #endif /* _WIN32 */
universe@579 42
universe@579 43 cxmutstr cx_mutstr(char *cstring) {
universe@579 44 return (cxmutstr) {cstring, strlen(cstring)};
universe@579 45 }
universe@579 46
universe@579 47 cxmutstr cx_mutstrn(
universe@579 48 char *cstring,
universe@579 49 size_t length
universe@579 50 ) {
universe@579 51 return (cxmutstr) {cstring, length};
universe@579 52 }
universe@579 53
universe@579 54 cxstring cx_str(const char *cstring) {
universe@579 55 return (cxstring) {cstring, strlen(cstring)};
universe@579 56 }
universe@579 57
universe@579 58 cxstring cx_strn(
universe@579 59 const char *cstring,
universe@579 60 size_t length
universe@579 61 ) {
universe@579 62 return (cxstring) {cstring, length};
universe@579 63 }
universe@579 64
universe@579 65 cxstring cx_strcast(cxmutstr str) {
universe@579 66 return (cxstring) {str.ptr, str.length};
universe@579 67 }
universe@579 68
universe@579 69 void cx_strfree(cxmutstr *str) {
universe@579 70 free(str->ptr);
universe@579 71 str->ptr = NULL;
universe@579 72 str->length = 0;
universe@579 73 }
universe@579 74
universe@583 75 void cx_strfree_a(
universe@583 76 CxAllocator *alloc,
universe@583 77 cxmutstr *str
universe@583 78 ) {
universe@583 79 cxFree(alloc, str->ptr);
universe@583 80 str->ptr = NULL;
universe@583 81 str->length = 0;
universe@583 82 }
universe@583 83
universe@579 84 size_t cx_strlen(
universe@579 85 size_t count,
universe@579 86 ...
universe@579 87 ) {
universe@579 88 if (count == 0) return 0;
universe@579 89
universe@579 90 va_list ap;
universe@579 91 va_start(ap, count);
universe@579 92 size_t size = 0;
universe@579 93 cx_for_n(i, count) {
universe@579 94 cxstring str = va_arg(ap, cxstring);
universe@579 95 size += str.length;
universe@579 96 }
universe@579 97 va_end(ap);
universe@579 98
universe@579 99 return size;
universe@579 100 }
universe@579 101
universe@579 102 cxmutstr cx_strcat_a(
universe@579 103 CxAllocator *alloc,
universe@579 104 size_t count,
universe@579 105 ...
universe@579 106 ) {
universe@579 107 cxstring *strings = calloc(count, sizeof(cxstring));
universe@579 108 if (!strings) abort();
universe@579 109
universe@579 110 va_list ap;
universe@579 111 va_start(ap, count);
universe@579 112
universe@579 113 // get all args and overall length
universe@579 114 size_t slen = 0;
universe@579 115 cx_for_n(i, count) {
universe@579 116 cxstring s = va_arg (ap, cxstring);
universe@579 117 strings[i] = s;
universe@579 118 slen += s.length;
universe@579 119 }
universe@579 120
universe@579 121 // create new string
universe@579 122 cxmutstr result;
universe@579 123 result.ptr = cxMalloc(alloc, slen + 1);
universe@579 124 result.length = slen;
universe@579 125 if (result.ptr == NULL) abort();
universe@579 126
universe@579 127 // concatenate strings
universe@579 128 size_t pos = 0;
universe@579 129 cx_for_n(i, count) {
universe@579 130 cxstring s = strings[i];
universe@579 131 memcpy(result.ptr + pos, s.ptr, s.length);
universe@579 132 pos += s.length;
universe@579 133 }
universe@579 134
universe@579 135 // terminate string
universe@579 136 result.ptr[result.length] = '\0';
universe@579 137
universe@579 138 // free temporary array
universe@579 139 free(strings);
universe@579 140
universe@579 141 return result;
universe@579 142 }
universe@579 143
universe@580 144 cxstring cx_strsubs(
universe@580 145 cxstring string,
universe@580 146 size_t start
universe@580 147 ) {
universe@580 148 return cx_strsubsl(string, start, string.length - start);
universe@580 149 }
universe@579 150
universe@580 151 cxmutstr cx_strsubs_m(
universe@580 152 cxmutstr string,
universe@580 153 size_t start
universe@580 154 ) {
universe@580 155 return cx_strsubsl_m(string, start, string.length - start);
universe@580 156 }
universe@579 157
universe@580 158 cxstring cx_strsubsl(
universe@580 159 cxstring string,
universe@580 160 size_t start,
universe@580 161 size_t length
universe@580 162 ) {
universe@580 163 if (start > string.length) {
universe@580 164 return (cxstring) {NULL, 0};
universe@580 165 }
universe@580 166
universe@580 167 size_t rem_len = string.length - start;
universe@580 168 if (length > rem_len) {
universe@580 169 length = rem_len;
universe@580 170 }
universe@580 171
universe@580 172 return (cxstring) {string.ptr + start, length};
universe@580 173 }
universe@580 174
universe@580 175 cxmutstr cx_strsubsl_m(
universe@580 176 cxmutstr string,
universe@580 177 size_t start,
universe@580 178 size_t length
universe@580 179 ) {
universe@580 180 cxstring result = cx_strsubsl(cx_strcast(string), start, length);
universe@580 181 return (cxmutstr) {(char *) result.ptr, result.length};
universe@580 182 }
universe@580 183
universe@580 184 cxstring cx_strchr(
universe@580 185 cxstring string,
universe@580 186 int chr
universe@580 187 ) {
universe@580 188 chr = 0xFF & chr;
universe@580 189 // TODO: improve by comparing multiple bytes at once
universe@580 190 cx_for_n(i, string.length) {
universe@580 191 if (string.ptr[i] == chr) {
universe@580 192 return cx_strsubs(string, i);
universe@580 193 }
universe@580 194 }
universe@580 195 return (cxstring) {NULL, 0};
universe@580 196 }
universe@580 197
universe@580 198 cxmutstr cx_strchr_m(
universe@580 199 cxmutstr string,
universe@580 200 int chr
universe@580 201 ) {
universe@580 202 cxstring result = cx_strchr(cx_strcast(string), chr);
universe@580 203 return (cxmutstr) {(char *) result.ptr, result.length};
universe@580 204 }
universe@580 205
universe@580 206 cxstring cx_strrchr(
universe@580 207 cxstring string,
universe@580 208 int chr
universe@580 209 ) {
universe@580 210 chr = 0xFF & chr;
universe@580 211 size_t i = string.length;
universe@580 212 while (i > 0) {
universe@580 213 i--;
universe@580 214 // TODO: improve by comparing multiple bytes at once
universe@580 215 if (string.ptr[i] == chr) {
universe@580 216 return cx_strsubs(string, i);
universe@580 217 }
universe@580 218 }
universe@580 219 return (cxstring) {NULL, 0};
universe@580 220 }
universe@580 221
universe@580 222 cxmutstr cx_strrchr_m(
universe@580 223 cxmutstr string,
universe@580 224 int chr
universe@580 225 ) {
universe@580 226 cxstring result = cx_strrchr(cx_strcast(string), chr);
universe@580 227 return (cxmutstr) {(char *) result.ptr, result.length};
universe@580 228 }
universe@580 229
universe@591 230 #define STRSTR_SBO_BUFLEN 512
universe@580 231
universe@580 232 cxstring cx_strstr(
universe@580 233 cxstring haystack,
universe@580 234 cxstring needle
universe@580 235 ) {
universe@580 236 if (needle.length == 0) {
universe@580 237 return haystack;
universe@580 238 }
universe@580 239
universe@583 240 /* optimize for single-char needles */
universe@583 241 if (needle.length == 1) {
universe@583 242 return cx_strchr(haystack, *needle.ptr);
universe@583 243 }
universe@583 244
universe@580 245 /*
universe@580 246 * IMPORTANT:
universe@580 247 * Our prefix table contains the prefix length PLUS ONE
universe@580 248 * this is our decision, because we want to use the full range of size_t.
universe@580 249 * The original algorithm needs a (-1) at one single place,
universe@580 250 * and we want to avoid that.
universe@580 251 */
universe@580 252
universe@591 253 /* local prefix table */
universe@591 254 size_t s_prefix_table[STRSTR_SBO_BUFLEN];
universe@580 255
universe@591 256 /* check needle length and use appropriate prefix table */
universe@580 257 /* if the pattern exceeds static prefix table, allocate on the heap */
universe@591 258 bool useheap = needle.length >= STRSTR_SBO_BUFLEN;
universe@591 259 register size_t *ptable = useheap ? calloc(needle.length + 1,
universe@591 260 sizeof(size_t)) : s_prefix_table;
universe@580 261
universe@580 262 /* keep counter in registers */
universe@580 263 register size_t i, j;
universe@580 264
universe@580 265 /* fill prefix table */
universe@580 266 i = 0;
universe@580 267 j = 0;
universe@591 268 ptable[i] = j;
universe@580 269 while (i < needle.length) {
universe@580 270 while (j >= 1 && needle.ptr[j - 1] != needle.ptr[i]) {
universe@591 271 j = ptable[j - 1];
universe@580 272 }
universe@580 273 i++;
universe@580 274 j++;
universe@591 275 ptable[i] = j;
universe@580 276 }
universe@580 277
universe@580 278 /* search */
universe@580 279 cxstring result = {NULL, 0};
universe@580 280 i = 0;
universe@580 281 j = 1;
universe@580 282 while (i < haystack.length) {
universe@580 283 while (j >= 1 && haystack.ptr[i] != needle.ptr[j - 1]) {
universe@591 284 j = ptable[j - 1];
universe@580 285 }
universe@580 286 i++;
universe@580 287 j++;
universe@580 288 if (j - 1 == needle.length) {
universe@580 289 size_t start = i - needle.length;
universe@580 290 result.ptr = haystack.ptr + start;
universe@580 291 result.length = haystack.length - start;
universe@580 292 break;
universe@580 293 }
universe@580 294 }
universe@580 295
universe@580 296 /* if prefix table was allocated on the heap, free it */
universe@580 297 if (ptable != s_prefix_table) {
universe@580 298 free(ptable);
universe@580 299 }
universe@580 300
universe@580 301 return result;
universe@580 302 }
universe@580 303
universe@580 304 cxmutstr cx_strstr_m(
universe@580 305 cxmutstr haystack,
universe@580 306 cxstring needle
universe@580 307 ) {
universe@580 308 cxstring result = cx_strstr(cx_strcast(haystack), needle);
universe@580 309 return (cxmutstr) {(char *) result.ptr, result.length};
universe@580 310 }
universe@580 311
universe@580 312 size_t cx_strsplit(
universe@580 313 cxstring string,
universe@580 314 cxstring delim,
universe@580 315 size_t limit,
universe@580 316 cxstring *output
universe@580 317 ) {
universe@583 318 /* special case: output limit is zero */
universe@583 319 if (limit == 0) return 0;
universe@583 320
universe@583 321 /* special case: delimiter is empty */
universe@583 322 if (delim.length == 0) {
universe@583 323 output[0] = string;
universe@583 324 return 1;
universe@583 325 }
universe@583 326
universe@583 327 /* special cases: delimiter is at least as large as the string */
universe@583 328 if (delim.length >= string.length) {
universe@583 329 /* exact match */
universe@583 330 if (cx_strcmp(string, delim) == 0) {
universe@583 331 output[0] = cx_strn(string.ptr, 0);
universe@583 332 output[1] = cx_strn(string.ptr + string.length, 0);
universe@583 333 return 2;
universe@583 334 } else /* no match possible */ {
universe@583 335 output[0] = string;
universe@583 336 return 1;
universe@583 337 }
universe@583 338 }
universe@583 339
universe@583 340 size_t n = 0;
universe@583 341 cxstring curpos = string;
universe@583 342 while (1) {
universe@583 343 ++n;
universe@583 344 cxstring match = cx_strstr(curpos, delim);
universe@583 345 if (match.length > 0) {
universe@583 346 /* is the limit reached? */
universe@583 347 if (n < limit) {
universe@583 348 /* copy the current string to the array */
universe@583 349 cxstring item = cx_strn(curpos.ptr, match.ptr - curpos.ptr);
universe@583 350 output[n - 1] = item;
universe@583 351 size_t processed = item.length + delim.length;
universe@583 352 curpos.ptr += processed;
universe@583 353 curpos.length -= processed;
universe@583 354 } else {
universe@583 355 /* limit reached, copy the _full_ remaining string */
universe@583 356 output[n - 1] = curpos;
universe@583 357 break;
universe@583 358 }
universe@583 359 } else {
universe@583 360 /* no more matches, copy last string */
universe@583 361 output[n - 1] = curpos;
universe@583 362 break;
universe@583 363 }
universe@583 364 }
universe@583 365
universe@583 366 return n;
universe@580 367 }
universe@580 368
universe@580 369 size_t cx_strsplit_a(
universe@580 370 CxAllocator *allocator,
universe@580 371 cxstring string,
universe@580 372 cxstring delim,
universe@580 373 size_t limit,
universe@580 374 cxstring **output
universe@580 375 ) {
universe@583 376 /* find out how many splits we're going to make and allocate memory */
universe@583 377 size_t n = 0;
universe@583 378 cxstring curpos = string;
universe@583 379 while (1) {
universe@583 380 ++n;
universe@583 381 cxstring match = cx_strstr(curpos, delim);
universe@583 382 if (match.length > 0) {
universe@583 383 /* is the limit reached? */
universe@583 384 if (n < limit) {
universe@583 385 size_t processed = match.ptr - curpos.ptr + delim.length;
universe@583 386 curpos.ptr += processed;
universe@583 387 curpos.length -= processed;
universe@583 388 } else {
universe@583 389 /* limit reached */
universe@583 390 break;
universe@583 391 }
universe@583 392 } else {
universe@583 393 /* no more matches */
universe@583 394 break;
universe@583 395 }
universe@583 396 }
universe@583 397 *output = cxCalloc(allocator, n, sizeof(cxstring));
universe@583 398 return cx_strsplit(string, delim, n, *output);
universe@580 399 }
universe@580 400
universe@580 401 size_t cx_strsplit_m(
universe@580 402 cxmutstr string,
universe@580 403 cxstring delim,
universe@580 404 size_t limit,
universe@580 405 cxmutstr *output
universe@580 406 ) {
universe@580 407 return cx_strsplit(cx_strcast(string),
universe@580 408 delim, limit, (cxstring *) output);
universe@580 409 }
universe@580 410
universe@580 411 size_t cx_strsplit_ma(
universe@580 412 CxAllocator *allocator,
universe@580 413 cxmutstr string,
universe@580 414 cxstring delim,
universe@580 415 size_t limit,
universe@580 416 cxmutstr **output
universe@580 417 ) {
universe@580 418 return cx_strsplit_a(allocator, cx_strcast(string),
universe@580 419 delim, limit, (cxstring **) output);
universe@580 420 }
universe@581 421
universe@583 422 int cx_strcmp(
universe@583 423 cxstring s1,
universe@583 424 cxstring s2
universe@583 425 ) {
universe@581 426 if (s1.length == s2.length) {
universe@581 427 return memcmp(s1.ptr, s2.ptr, s1.length);
universe@581 428 } else if (s1.length > s2.length) {
universe@581 429 return 1;
universe@581 430 } else {
universe@581 431 return -1;
universe@581 432 }
universe@581 433 }
universe@581 434
universe@583 435 int cx_strcasecmp(
universe@583 436 cxstring s1,
universe@583 437 cxstring s2
universe@583 438 ) {
universe@581 439 if (s1.length == s2.length) {
universe@581 440 #ifdef _WIN32
universe@581 441 return _strnicmp(s1.ptr, s2.ptr, s1.length);
universe@581 442 #else
universe@581 443 return strncasecmp(s1.ptr, s2.ptr, s1.length);
universe@581 444 #endif
universe@581 445 } else if (s1.length > s2.length) {
universe@581 446 return 1;
universe@581 447 } else {
universe@581 448 return -1;
universe@581 449 }
universe@581 450 }
universe@581 451
universe@583 452 cxmutstr cx_strdup_a(
universe@583 453 CxAllocator *allocator,
universe@583 454 cxstring string
universe@583 455 ) {
universe@581 456 cxmutstr result = {
universe@581 457 cxMalloc(allocator, string.length + 1),
universe@581 458 string.length
universe@581 459 };
universe@581 460 if (result.ptr == NULL) {
universe@581 461 result.length = 0;
universe@581 462 return result;
universe@581 463 }
universe@581 464 memcpy(result.ptr, string.ptr, string.length);
universe@581 465 result.ptr[string.length] = '\0';
universe@581 466 return result;
universe@581 467 }
universe@581 468
universe@581 469 cxstring cx_strtrim(cxstring string) {
universe@581 470 cxstring result = string;
universe@581 471 // TODO: optimize by comparing multiple bytes at once
universe@581 472 while (result.length > 0 && isspace(*result.ptr)) {
universe@581 473 result.ptr++;
universe@581 474 result.length--;
universe@581 475 }
universe@581 476 while (result.length > 0 && isspace(result.ptr[result.length - 1])) {
universe@581 477 result.length--;
universe@581 478 }
universe@581 479 return result;
universe@581 480 }
universe@581 481
universe@581 482 cxmutstr cx_strtrim_m(cxmutstr string) {
universe@581 483 cxstring result = cx_strtrim(cx_strcast(string));
universe@581 484 return (cxmutstr) {(char *) result.ptr, result.length};
universe@581 485 }
universe@581 486
universe@583 487 bool cx_strprefix(
universe@583 488 cxstring string,
universe@583 489 cxstring prefix
universe@583 490 ) {
universe@581 491 if (string.length < prefix.length) return false;
universe@581 492 return memcmp(string.ptr, prefix.ptr, prefix.length) == 0;
universe@581 493 }
universe@581 494
universe@583 495 bool cx_strsuffix(
universe@583 496 cxstring string,
universe@583 497 cxstring suffix
universe@583 498 ) {
universe@581 499 if (string.length < suffix.length) return false;
universe@581 500 return memcmp(string.ptr + string.length - suffix.length,
universe@581 501 suffix.ptr, suffix.length) == 0;
universe@581 502 }
universe@581 503
universe@583 504 bool cx_strcaseprefix(
universe@583 505 cxstring string,
universe@583 506 cxstring prefix
universe@583 507 ) {
universe@581 508 if (string.length < prefix.length) return false;
universe@581 509 #ifdef _WIN32
universe@581 510 return _strnicmp(string.ptr, prefix.ptr, prefix.length) == 0;
universe@581 511 #else
universe@581 512 return strncasecmp(string.ptr, prefix.ptr, prefix.length) == 0;
universe@581 513 #endif
universe@581 514 }
universe@581 515
universe@583 516 bool cx_strcasesuffix(
universe@583 517 cxstring string,
universe@583 518 cxstring suffix
universe@583 519 ) {
universe@581 520 if (string.length < suffix.length) return false;
universe@581 521 #ifdef _WIN32
universe@581 522 return _strnicmp(string.ptr+string.length-suffix.length,
universe@581 523 suffix.ptr, suffix.length) == 0;
universe@581 524 #else
universe@581 525 return strncasecmp(string.ptr + string.length - suffix.length,
universe@581 526 suffix.ptr, suffix.length) == 0;
universe@581 527 #endif
universe@581 528 }
universe@582 529
universe@582 530 void cx_strlower(cxmutstr string) {
universe@582 531 cx_for_n(i, string.length) {
universe@582 532 string.ptr[i] = tolower(string.ptr[i]);
universe@582 533 }
universe@582 534 }
universe@582 535
universe@582 536 void cx_strupper(cxmutstr string) {
universe@582 537 cx_for_n(i, string.length) {
universe@582 538 string.ptr[i] = toupper(string.ptr[i]);
universe@582 539 }
universe@582 540 }
universe@583 541
universe@583 542 #define REPLACE_INDEX_BUFFER_MAX 100
universe@583 543
universe@583 544 struct cx_strreplace_ibuf {
universe@583 545 size_t *buf;
universe@583 546 struct cx_strreplace_ibuf *next;
universe@590 547 unsigned int len;
universe@583 548 };
universe@583 549
universe@583 550 static void cx_strrepl_free_ibuf(struct cx_strreplace_ibuf *buf) {
universe@583 551 while (buf) {
universe@583 552 struct cx_strreplace_ibuf *next = buf->next;
universe@583 553 free(buf->buf);
universe@583 554 free(buf);
universe@583 555 buf = next;
universe@583 556 }
universe@583 557 }
universe@583 558
universe@583 559 cxmutstr cx_strreplacen_a(
universe@583 560 CxAllocator *allocator,
universe@583 561 cxstring str,
universe@583 562 cxstring pattern,
universe@583 563 cxstring replacement,
universe@583 564 size_t replmax
universe@583 565 ) {
universe@583 566
universe@583 567 if (pattern.length == 0 || pattern.length > str.length || replmax == 0)
universe@583 568 return cx_strdup_a(allocator, str);
universe@583 569
universe@583 570 /* Compute expected buffer length */
universe@583 571 size_t ibufmax = str.length / pattern.length;
universe@583 572 size_t ibuflen = replmax < ibufmax ? replmax : ibufmax;
universe@583 573 if (ibuflen > REPLACE_INDEX_BUFFER_MAX) {
universe@583 574 ibuflen = REPLACE_INDEX_BUFFER_MAX;
universe@583 575 }
universe@583 576
universe@583 577 /* Allocate first index buffer */
universe@583 578 struct cx_strreplace_ibuf *firstbuf, *curbuf;
universe@583 579 firstbuf = curbuf = calloc(1, sizeof(struct cx_strreplace_ibuf));
universe@583 580 if (!firstbuf) return cx_mutstrn(NULL, 0);
universe@583 581 firstbuf->buf = calloc(ibuflen, sizeof(size_t));
universe@583 582 if (!firstbuf->buf) {
universe@583 583 free(firstbuf);
universe@583 584 return cx_mutstrn(NULL, 0);
universe@583 585 }
universe@583 586
universe@583 587 /* Search occurrences */
universe@583 588 cxstring searchstr = str;
universe@583 589 size_t found = 0;
universe@583 590 do {
universe@583 591 cxstring match = cx_strstr(searchstr, pattern);
universe@583 592 if (match.length > 0) {
universe@583 593 /* Allocate next buffer in chain, if required */
universe@583 594 if (curbuf->len == ibuflen) {
universe@583 595 struct cx_strreplace_ibuf *nextbuf =
universe@583 596 calloc(1, sizeof(struct cx_strreplace_ibuf));
universe@583 597 if (!nextbuf) {
universe@583 598 cx_strrepl_free_ibuf(firstbuf);
universe@583 599 return cx_mutstrn(NULL, 0);
universe@583 600 }
universe@583 601 nextbuf->buf = calloc(ibuflen, sizeof(size_t));
universe@583 602 if (!nextbuf->buf) {
universe@583 603 free(nextbuf);
universe@583 604 cx_strrepl_free_ibuf(firstbuf);
universe@583 605 return cx_mutstrn(NULL, 0);
universe@583 606 }
universe@583 607 curbuf->next = nextbuf;
universe@583 608 curbuf = nextbuf;
universe@583 609 }
universe@583 610
universe@583 611 /* Record match index */
universe@583 612 found++;
universe@583 613 size_t idx = match.ptr - str.ptr;
universe@583 614 curbuf->buf[curbuf->len++] = idx;
universe@583 615 searchstr.ptr = match.ptr + pattern.length;
universe@583 616 searchstr.length = str.length - idx - pattern.length;
universe@583 617 } else {
universe@583 618 break;
universe@583 619 }
universe@583 620 } while (searchstr.length > 0 && found < replmax);
universe@583 621
universe@583 622 /* Allocate result string */
universe@583 623 cxmutstr result;
universe@583 624 {
universe@583 625 ssize_t adjlen = (ssize_t) replacement.length - (ssize_t) pattern.length;
universe@583 626 size_t rcount = 0;
universe@583 627 curbuf = firstbuf;
universe@583 628 do {
universe@583 629 rcount += curbuf->len;
universe@583 630 curbuf = curbuf->next;
universe@583 631 } while (curbuf);
universe@583 632 result.length = str.length + rcount * adjlen;
universe@590 633 result.ptr = cxMalloc(allocator, result.length + 1);
universe@583 634 if (!result.ptr) {
universe@583 635 cx_strrepl_free_ibuf(firstbuf);
universe@583 636 return cx_mutstrn(NULL, 0);
universe@583 637 }
universe@583 638 }
universe@583 639
universe@583 640 /* Build result string */
universe@583 641 curbuf = firstbuf;
universe@583 642 size_t srcidx = 0;
universe@583 643 char *destptr = result.ptr;
universe@583 644 do {
universe@583 645 for (size_t i = 0; i < curbuf->len; i++) {
universe@583 646 /* Copy source part up to next match*/
universe@583 647 size_t idx = curbuf->buf[i];
universe@583 648 size_t srclen = idx - srcidx;
universe@583 649 if (srclen > 0) {
universe@583 650 memcpy(destptr, str.ptr + srcidx, srclen);
universe@583 651 destptr += srclen;
universe@583 652 srcidx += srclen;
universe@583 653 }
universe@583 654
universe@583 655 /* Copy the replacement and skip the source pattern */
universe@583 656 srcidx += pattern.length;
universe@583 657 memcpy(destptr, replacement.ptr, replacement.length);
universe@583 658 destptr += replacement.length;
universe@583 659 }
universe@583 660 curbuf = curbuf->next;
universe@583 661 } while (curbuf);
universe@583 662 memcpy(destptr, str.ptr + srcidx, str.length - srcidx);
universe@583 663
universe@590 664 /* Result is guaranteed to be zero-terminated */
universe@590 665 result.ptr[result.length] = '\0';
universe@590 666
universe@583 667 /* Free index buffer */
universe@583 668 cx_strrepl_free_ibuf(firstbuf);
universe@583 669
universe@583 670 return result;
universe@583 671 }
universe@583 672
universe@583 673

mercurial