1 /*- 2 * SPDX-License-Identifier: BSD-3-Clause 3 * 4 * Copyright (c) 1991, 1993, 1994 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Steve Hayman of the Computer Science Department, Indiana University, 9 * Michiro Hikida and David Goodenough. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 3. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 */ 35 36 #ifndef lint 37 static const char copyright[] = 38 "@(#) Copyright (c) 1991, 1993, 1994\n\ 39 The Regents of the University of California. All rights reserved.\n"; 40 #endif /* not lint */ 41 42 #ifndef lint 43 #if 0 44 static char sccsid[] = "@(#)join.c 8.6 (Berkeley) 5/4/95"; 45 #endif 46 #endif /* not lint */ 47 #include <sys/cdefs.h> 48 __FBSDID("$FreeBSD$"); 49 50 #include <sys/param.h> 51 52 #include <err.h> 53 #include <errno.h> 54 #include <limits.h> 55 #include <locale.h> 56 #include <stdio.h> 57 #include <stdlib.h> 58 #include <string.h> 59 #include <unistd.h> 60 #include <wchar.h> 61 62 /* 63 * There's a structure per input file which encapsulates the state of the 64 * file. We repeatedly read lines from each file until we've read in all 65 * the consecutive lines from the file with a common join field. Then we 66 * compare the set of lines with an equivalent set from the other file. 67 */ 68 typedef struct { 69 char *line; /* line */ 70 u_long linealloc; /* line allocated count */ 71 char **fields; /* line field(s) */ 72 u_long fieldcnt; /* line field(s) count */ 73 u_long fieldalloc; /* line field(s) allocated count */ 74 } LINE; 75 76 typedef struct { 77 FILE *fp; /* file descriptor */ 78 u_long joinf; /* join field (-1, -2, -j) */ 79 int unpair; /* output unpairable lines (-a) */ 80 u_long number; /* 1 for file 1, 2 for file 2 */ 81 82 LINE *set; /* set of lines with same field */ 83 int pushbool; /* if pushback is set */ 84 u_long pushback; /* line on the stack */ 85 u_long setcnt; /* set count */ 86 u_long setalloc; /* set allocated count */ 87 } INPUT; 88 static INPUT input1 = { NULL, 0, 0, 1, NULL, 0, 0, 0, 0 }, 89 input2 = { NULL, 0, 0, 2, NULL, 0, 0, 0, 0 }; 90 91 typedef struct { 92 u_long filenum; /* file number */ 93 u_long fieldno; /* field number */ 94 } OLIST; 95 static OLIST *olist; /* output field list */ 96 static u_long olistcnt; /* output field list count */ 97 static u_long olistalloc; /* output field allocated count */ 98 99 static int joinout = 1; /* show lines with matched join fields (-v) */ 100 static int needsep; /* need separator character */ 101 static int spans = 1; /* span multiple delimiters (-t) */ 102 static char *empty; /* empty field replacement string (-e) */ 103 static wchar_t default_tabchar[] = L" \t"; 104 static wchar_t *tabchar = default_tabchar; /* delimiter characters (-t) */ 105 106 static int cmp(LINE *, u_long, LINE *, u_long); 107 static void fieldarg(char *); 108 static void joinlines(INPUT *, INPUT *); 109 static int mbscoll(const char *, const char *); 110 static char *mbssep(char **, const wchar_t *); 111 static void obsolete(char **); 112 static void outfield(LINE *, u_long, int); 113 static void outoneline(INPUT *, LINE *); 114 static void outtwoline(INPUT *, LINE *, INPUT *, LINE *); 115 static void slurp(INPUT *); 116 static wchar_t *towcs(const char *); 117 static void usage(void) __dead2; 118 119 int 120 main(int argc, char *argv[]) 121 { 122 INPUT *F1, *F2; 123 int aflag, ch, cval, vflag; 124 char *end; 125 126 setlocale(LC_ALL, ""); 127 128 F1 = &input1; 129 F2 = &input2; 130 131 aflag = vflag = 0; 132 obsolete(argv); 133 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 134 switch (ch) { 135 case '\01': /* See comment in obsolete(). */ 136 aflag = 1; 137 F1->unpair = F2->unpair = 1; 138 break; 139 case '1': 140 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) 141 errx(1, "-1 option field number less than 1"); 142 if (*end) 143 errx(1, "illegal field number -- %s", optarg); 144 --F1->joinf; 145 break; 146 case '2': 147 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) 148 errx(1, "-2 option field number less than 1"); 149 if (*end) 150 errx(1, "illegal field number -- %s", optarg); 151 --F2->joinf; 152 break; 153 case 'a': 154 aflag = 1; 155 switch(strtol(optarg, &end, 10)) { 156 case 1: 157 F1->unpair = 1; 158 break; 159 case 2: 160 F2->unpair = 1; 161 break; 162 default: 163 errx(1, "-a option file number not 1 or 2"); 164 break; 165 } 166 if (*end) 167 errx(1, "illegal file number -- %s", optarg); 168 break; 169 case 'e': 170 empty = optarg; 171 break; 172 case 'j': 173 if ((F1->joinf = F2->joinf = 174 strtol(optarg, &end, 10)) < 1) 175 errx(1, "-j option field number less than 1"); 176 if (*end) 177 errx(1, "illegal field number -- %s", optarg); 178 --F1->joinf; 179 --F2->joinf; 180 break; 181 case 'o': 182 fieldarg(optarg); 183 break; 184 case 't': 185 spans = 0; 186 if (mbrtowc(&tabchar[0], optarg, MB_LEN_MAX, NULL) != 187 strlen(optarg)) 188 errx(1, "illegal tab character specification"); 189 tabchar[1] = L'\0'; 190 break; 191 case 'v': 192 vflag = 1; 193 joinout = 0; 194 switch (strtol(optarg, &end, 10)) { 195 case 1: 196 F1->unpair = 1; 197 break; 198 case 2: 199 F2->unpair = 1; 200 break; 201 default: 202 errx(1, "-v option file number not 1 or 2"); 203 break; 204 } 205 if (*end) 206 errx(1, "illegal file number -- %s", optarg); 207 break; 208 case '?': 209 default: 210 usage(); 211 } 212 } 213 argc -= optind; 214 argv += optind; 215 216 if (aflag && vflag) 217 errx(1, "the -a and -v options are mutually exclusive"); 218 219 if (argc != 2) 220 usage(); 221 222 /* Open the files; "-" means stdin. */ 223 if (!strcmp(*argv, "-")) 224 F1->fp = stdin; 225 else if ((F1->fp = fopen(*argv, "r")) == NULL) 226 err(1, "%s", *argv); 227 ++argv; 228 if (!strcmp(*argv, "-")) 229 F2->fp = stdin; 230 else if ((F2->fp = fopen(*argv, "r")) == NULL) 231 err(1, "%s", *argv); 232 if (F1->fp == stdin && F2->fp == stdin) 233 errx(1, "only one input file may be stdin"); 234 235 slurp(F1); 236 slurp(F2); 237 while (F1->setcnt && F2->setcnt) { 238 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 239 if (cval == 0) { 240 /* Oh joy, oh rapture, oh beauty divine! */ 241 if (joinout) 242 joinlines(F1, F2); 243 slurp(F1); 244 slurp(F2); 245 } else if (cval < 0) { 246 /* File 1 takes the lead... */ 247 if (F1->unpair) 248 joinlines(F1, NULL); 249 slurp(F1); 250 } else { 251 /* File 2 takes the lead... */ 252 if (F2->unpair) 253 joinlines(F2, NULL); 254 slurp(F2); 255 } 256 } 257 258 /* 259 * Now that one of the files is used up, optionally output any 260 * remaining lines from the other file. 261 */ 262 if (F1->unpair) 263 while (F1->setcnt) { 264 joinlines(F1, NULL); 265 slurp(F1); 266 } 267 if (F2->unpair) 268 while (F2->setcnt) { 269 joinlines(F2, NULL); 270 slurp(F2); 271 } 272 exit(0); 273 } 274 275 static void 276 slurp(INPUT *F) 277 { 278 LINE *lp, *lastlp, tmp; 279 size_t len; 280 int cnt; 281 char *bp, *fieldp; 282 283 /* 284 * Read all of the lines from an input file that have the same 285 * join field. 286 */ 287 F->setcnt = 0; 288 for (lastlp = NULL;; ++F->setcnt) { 289 /* 290 * If we're out of space to hold line structures, allocate 291 * more. Initialize the structure so that we know that this 292 * is new space. 293 */ 294 if (F->setcnt == F->setalloc) { 295 cnt = F->setalloc; 296 F->setalloc += 50; 297 if ((F->set = realloc(F->set, 298 F->setalloc * sizeof(LINE))) == NULL) 299 err(1, NULL); 300 memset(F->set + cnt, 0, 50 * sizeof(LINE)); 301 302 /* re-set lastlp in case it moved */ 303 if (lastlp != NULL) 304 lastlp = &F->set[F->setcnt - 1]; 305 } 306 307 /* 308 * Get any pushed back line, else get the next line. Allocate 309 * space as necessary. If taking the line from the stack swap 310 * the two structures so that we don't lose space allocated to 311 * either structure. This could be avoided by doing another 312 * level of indirection, but it's probably okay as is. 313 */ 314 lp = &F->set[F->setcnt]; 315 if (F->setcnt) 316 lastlp = &F->set[F->setcnt - 1]; 317 if (F->pushbool) { 318 tmp = F->set[F->setcnt]; 319 F->set[F->setcnt] = F->set[F->pushback]; 320 F->set[F->pushback] = tmp; 321 F->pushbool = 0; 322 continue; 323 } 324 if ((bp = fgetln(F->fp, &len)) == NULL) 325 return; 326 if (lp->linealloc <= len + 1) { 327 lp->linealloc += MAX(100, len + 1 - lp->linealloc); 328 if ((lp->line = 329 realloc(lp->line, lp->linealloc)) == NULL) 330 err(1, NULL); 331 } 332 memmove(lp->line, bp, len); 333 334 /* Replace trailing newline, if it exists. */ 335 if (bp[len - 1] == '\n') 336 lp->line[len - 1] = '\0'; 337 else 338 lp->line[len] = '\0'; 339 bp = lp->line; 340 341 /* Split the line into fields, allocate space as necessary. */ 342 lp->fieldcnt = 0; 343 while ((fieldp = mbssep(&bp, tabchar)) != NULL) { 344 if (spans && *fieldp == '\0') 345 continue; 346 if (lp->fieldcnt == lp->fieldalloc) { 347 lp->fieldalloc += 50; 348 if ((lp->fields = realloc(lp->fields, 349 lp->fieldalloc * sizeof(char *))) == NULL) 350 err(1, NULL); 351 } 352 lp->fields[lp->fieldcnt++] = fieldp; 353 } 354 355 /* See if the join field value has changed. */ 356 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) { 357 F->pushbool = 1; 358 F->pushback = F->setcnt; 359 break; 360 } 361 } 362 } 363 364 static char * 365 mbssep(char **stringp, const wchar_t *delim) 366 { 367 char *s, *tok; 368 const wchar_t *spanp; 369 wchar_t c, sc; 370 size_t n; 371 372 if ((s = *stringp) == NULL) 373 return (NULL); 374 for (tok = s;;) { 375 n = mbrtowc(&c, s, MB_LEN_MAX, NULL); 376 if (n == (size_t)-1 || n == (size_t)-2) 377 errc(1, EILSEQ, NULL); /* XXX */ 378 s += n; 379 spanp = delim; 380 do { 381 if ((sc = *spanp++) == c) { 382 if (c == 0) 383 s = NULL; 384 else 385 s[-n] = '\0'; 386 *stringp = s; 387 return (tok); 388 } 389 } while (sc != 0); 390 } 391 } 392 393 static int 394 cmp(LINE *lp1, u_long fieldno1, LINE *lp2, u_long fieldno2) 395 { 396 if (lp1->fieldcnt <= fieldno1) 397 return (lp2->fieldcnt <= fieldno2 ? 0 : 1); 398 if (lp2->fieldcnt <= fieldno2) 399 return (-1); 400 return (mbscoll(lp1->fields[fieldno1], lp2->fields[fieldno2])); 401 } 402 403 static int 404 mbscoll(const char *s1, const char *s2) 405 { 406 wchar_t *w1, *w2; 407 int ret; 408 409 if (MB_CUR_MAX == 1) 410 return (strcoll(s1, s2)); 411 if ((w1 = towcs(s1)) == NULL || (w2 = towcs(s2)) == NULL) 412 err(1, NULL); /* XXX */ 413 ret = wcscoll(w1, w2); 414 free(w1); 415 free(w2); 416 return (ret); 417 } 418 419 static wchar_t * 420 towcs(const char *s) 421 { 422 wchar_t *wcs; 423 size_t n; 424 425 if ((n = mbsrtowcs(NULL, &s, 0, NULL)) == (size_t)-1) 426 return (NULL); 427 if ((wcs = malloc((n + 1) * sizeof(*wcs))) == NULL) 428 return (NULL); 429 mbsrtowcs(wcs, &s, n + 1, NULL); 430 return (wcs); 431 } 432 433 static void 434 joinlines(INPUT *F1, INPUT *F2) 435 { 436 u_long cnt1, cnt2; 437 438 /* 439 * Output the results of a join comparison. The output may be from 440 * either file 1 or file 2 (in which case the first argument is the 441 * file from which to output) or from both. 442 */ 443 if (F2 == NULL) { 444 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 445 outoneline(F1, &F1->set[cnt1]); 446 return; 447 } 448 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 449 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 450 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 451 } 452 453 static void 454 outoneline(INPUT *F, LINE *lp) 455 { 456 u_long cnt; 457 458 /* 459 * Output a single line from one of the files, according to the 460 * join rules. This happens when we are writing unmatched single 461 * lines. Output empty fields in the right places. 462 */ 463 if (olist) 464 for (cnt = 0; cnt < olistcnt; ++cnt) { 465 if (olist[cnt].filenum == (unsigned)F->number) 466 outfield(lp, olist[cnt].fieldno, 0); 467 else if (olist[cnt].filenum == 0) 468 outfield(lp, F->joinf, 0); 469 else 470 outfield(lp, 0, 1); 471 } 472 else { 473 /* 474 * Output the join field, then the remaining fields. 475 */ 476 outfield(lp, F->joinf, 0); 477 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 478 if (F->joinf != cnt) 479 outfield(lp, cnt, 0); 480 } 481 (void)printf("\n"); 482 if (ferror(stdout)) 483 err(1, "stdout"); 484 needsep = 0; 485 } 486 487 static void 488 outtwoline(INPUT *F1, LINE *lp1, INPUT *F2, LINE *lp2) 489 { 490 u_long cnt; 491 492 /* Output a pair of lines according to the join list (if any). */ 493 if (olist) 494 for (cnt = 0; cnt < olistcnt; ++cnt) 495 if (olist[cnt].filenum == 0) { 496 if (lp1->fieldcnt >= F1->joinf) 497 outfield(lp1, F1->joinf, 0); 498 else 499 outfield(lp2, F2->joinf, 0); 500 } else if (olist[cnt].filenum == 1) 501 outfield(lp1, olist[cnt].fieldno, 0); 502 else /* if (olist[cnt].filenum == 2) */ 503 outfield(lp2, olist[cnt].fieldno, 0); 504 else { 505 /* 506 * Output the join field, then the remaining fields from F1 507 * and F2. 508 */ 509 outfield(lp1, F1->joinf, 0); 510 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 511 if (F1->joinf != cnt) 512 outfield(lp1, cnt, 0); 513 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 514 if (F2->joinf != cnt) 515 outfield(lp2, cnt, 0); 516 } 517 (void)printf("\n"); 518 if (ferror(stdout)) 519 err(1, "stdout"); 520 needsep = 0; 521 } 522 523 static void 524 outfield(LINE *lp, u_long fieldno, int out_empty) 525 { 526 if (needsep++) 527 (void)printf("%lc", (wint_t)*tabchar); 528 if (!ferror(stdout)) { 529 if (lp->fieldcnt <= fieldno || out_empty) { 530 if (empty != NULL) 531 (void)printf("%s", empty); 532 } else { 533 if (*lp->fields[fieldno] == '\0') 534 return; 535 (void)printf("%s", lp->fields[fieldno]); 536 } 537 } 538 if (ferror(stdout)) 539 err(1, "stdout"); 540 } 541 542 /* 543 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 544 * fields. 545 */ 546 static void 547 fieldarg(char *option) 548 { 549 u_long fieldno, filenum; 550 char *end, *token; 551 552 while ((token = strsep(&option, ", \t")) != NULL) { 553 if (*token == '\0') 554 continue; 555 if (token[0] == '0') 556 filenum = fieldno = 0; 557 else if ((token[0] == '1' || token[0] == '2') && 558 token[1] == '.') { 559 filenum = token[0] - '0'; 560 fieldno = strtol(token + 2, &end, 10); 561 if (*end) 562 errx(1, "malformed -o option field"); 563 if (fieldno == 0) 564 errx(1, "field numbers are 1 based"); 565 --fieldno; 566 } else 567 errx(1, "malformed -o option field"); 568 if (olistcnt == olistalloc) { 569 olistalloc += 50; 570 if ((olist = realloc(olist, 571 olistalloc * sizeof(OLIST))) == NULL) 572 err(1, NULL); 573 } 574 olist[olistcnt].filenum = filenum; 575 olist[olistcnt].fieldno = fieldno; 576 ++olistcnt; 577 } 578 } 579 580 static void 581 obsolete(char **argv) 582 { 583 size_t len; 584 char **p, *ap, *t; 585 586 while ((ap = *++argv) != NULL) { 587 /* Return if "--". */ 588 if (ap[0] == '-' && ap[1] == '-') 589 return; 590 /* skip if not an option */ 591 if (ap[0] != '-') 592 continue; 593 switch (ap[1]) { 594 case 'a': 595 /* 596 * The original join allowed "-a", which meant the 597 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 598 * only specifies this as "-a 1" and "a -2", so we 599 * have to use another option flag, one that is 600 * unlikely to ever be used or accidentally entered 601 * on the command line. (Well, we could reallocate 602 * the argv array, but that hardly seems worthwhile.) 603 */ 604 if (ap[2] == '\0' && (argv[1] == NULL || 605 (strcmp(argv[1], "1") != 0 && 606 strcmp(argv[1], "2") != 0))) { 607 ap[1] = '\01'; 608 warnx("-a option used without an argument; " 609 "reverting to historical behavior"); 610 } 611 break; 612 case 'j': 613 /* 614 * The original join allowed "-j[12] arg" and "-j arg". 615 * Convert the former to "-[12] arg". Don't convert 616 * the latter since getopt(3) can handle it. 617 */ 618 switch(ap[2]) { 619 case '1': 620 if (ap[3] != '\0') 621 goto jbad; 622 ap[1] = '1'; 623 ap[2] = '\0'; 624 break; 625 case '2': 626 if (ap[3] != '\0') 627 goto jbad; 628 ap[1] = '2'; 629 ap[2] = '\0'; 630 break; 631 case '\0': 632 break; 633 default: 634 jbad: errx(1, "illegal option -- %s", ap); 635 usage(); 636 } 637 break; 638 case 'o': 639 /* 640 * The original join allowed "-o arg arg". 641 * Convert to "-o arg -o arg". 642 */ 643 if (ap[2] != '\0') 644 break; 645 for (p = argv + 2; *p; ++p) { 646 if (p[0][0] == '0' || ((p[0][0] != '1' && 647 p[0][0] != '2') || p[0][1] != '.')) 648 break; 649 len = strlen(*p); 650 if (len - 2 != strspn(*p + 2, "0123456789")) 651 break; 652 if ((t = malloc(len + 3)) == NULL) 653 err(1, NULL); 654 t[0] = '-'; 655 t[1] = 'o'; 656 memmove(t + 2, *p, len + 1); 657 *p = t; 658 } 659 argv = p - 1; 660 break; 661 } 662 } 663 } 664 665 static void 666 usage(void) 667 { 668 (void)fprintf(stderr, "%s %s\n%s\n", 669 "usage: join [-a fileno | -v fileno ] [-e string] [-1 field]", 670 "[-2 field]", 671 " [-o list] [-t char] file1 file2"); 672 exit(1); 673 } 674