1 /* $OpenBSD: join.c,v 1.33 2020/07/23 20:13:01 martijn Exp $ */ 2 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 #include <err.h> 37 #include <errno.h> 38 #include <limits.h> 39 #include <locale.h> 40 #include <stdio.h> 41 #include <stdlib.h> 42 #include <string.h> 43 #include <unistd.h> 44 #include <wchar.h> 45 46 /* 47 * There's a structure per input file which encapsulates the state of the 48 * file. We repeatedly read lines from each file until we've read in all 49 * the consecutive lines from the file with a common join field. Then we 50 * compare the set of lines with an equivalent set from the other file. 51 */ 52 typedef struct { 53 char *line; /* line */ 54 size_t linealloc; /* bytes allocated for line */ 55 char **fields; /* line field(s) */ 56 u_long fieldcnt; /* line field(s) count */ 57 u_long fieldalloc; /* line field(s) allocated count */ 58 } LINE; 59 60 typedef struct { 61 FILE *fp; /* file descriptor */ 62 u_long joinf; /* join field (-1, -2, -j) */ 63 int unpair; /* output unpairable lines (-a) */ 64 u_long number; /* 1 for file 1, 2 for file 2 */ 65 LINE *set; /* set of lines with same field */ 66 int pushbool; /* if pushback is set */ 67 u_long pushback; /* line on the stack */ 68 u_long setcnt; /* set count */ 69 u_long setalloc; /* set allocated count */ 70 } INPUT; 71 INPUT input1 = { NULL, 0, 0, 1, NULL, 0, 0, 0, 0 }, 72 input2 = { NULL, 0, 0, 2, NULL, 0, 0, 0, 0 }; 73 74 typedef struct { 75 u_long filenum; /* file number */ 76 u_long fieldno; /* field number */ 77 } OLIST; 78 OLIST *olist; /* output field list */ 79 u_long olistcnt; /* output field list count */ 80 u_long olistalloc; /* output field allocated count */ 81 82 int joinout = 1; /* show lines with matched join fields (-v) */ 83 int needsep; /* need separator character */ 84 int spans = 1; /* span multiple delimiters (-t) */ 85 char *empty; /* empty field replacement string (-e) */ 86 wchar_t tabchar[] = L" \t"; /* delimiter characters (-t) */ 87 88 int cmp(LINE *, u_long, LINE *, u_long); 89 void fieldarg(char *); 90 void joinlines(INPUT *, INPUT *); 91 char *mbssep(char **, const wchar_t *); 92 void obsolete(char **); 93 void outfield(LINE *, u_long, int); 94 void outoneline(INPUT *, LINE *); 95 void outtwoline(INPUT *, LINE *, INPUT *, LINE *); 96 void slurp(INPUT *); 97 void usage(void); 98 99 int 100 main(int argc, char *argv[]) 101 { 102 INPUT *F1, *F2; 103 int aflag, ch, cval, vflag; 104 char *end; 105 106 setlocale(LC_CTYPE, ""); 107 108 if (pledge("stdio rpath", NULL) == -1) 109 err(1, "pledge"); 110 111 F1 = &input1; 112 F2 = &input2; 113 114 aflag = vflag = 0; 115 obsolete(argv); 116 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 117 switch (ch) { 118 case '\01': /* See comment in obsolete(). */ 119 aflag = 1; 120 F1->unpair = F2->unpair = 1; 121 break; 122 case '1': 123 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) 124 errx(1, "-1 option field number less than 1"); 125 if (*end) 126 errx(1, "illegal field number -- %s", optarg); 127 --F1->joinf; 128 break; 129 case '2': 130 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) 131 errx(1, "-2 option field number less than 1"); 132 if (*end) 133 errx(1, "illegal field number -- %s", optarg); 134 --F2->joinf; 135 break; 136 case 'a': 137 aflag = 1; 138 switch(strtol(optarg, &end, 10)) { 139 case 1: 140 F1->unpair = 1; 141 break; 142 case 2: 143 F2->unpair = 1; 144 break; 145 default: 146 errx(1, "-a option file number not 1 or 2"); 147 break; 148 } 149 if (*end) 150 errx(1, "illegal file number -- %s", optarg); 151 break; 152 case 'e': 153 empty = optarg; 154 break; 155 case 'j': 156 if ((F1->joinf = F2->joinf = strtol(optarg, &end, 10)) < 1) 157 errx(1, "-j option field number less than 1"); 158 if (*end) 159 errx(1, "illegal field number -- %s", optarg); 160 --F1->joinf; 161 --F2->joinf; 162 break; 163 case 'o': 164 fieldarg(optarg); 165 break; 166 case 't': 167 spans = 0; 168 if (mbtowc(tabchar, optarg, MB_CUR_MAX) != 169 strlen(optarg)) 170 errx(1, "illegal tab character specification"); 171 tabchar[1] = L'\0'; 172 break; 173 case 'v': 174 vflag = 1; 175 joinout = 0; 176 switch (strtol(optarg, &end, 10)) { 177 case 1: 178 F1->unpair = 1; 179 break; 180 case 2: 181 F2->unpair = 1; 182 break; 183 default: 184 errx(1, "-v option file number not 1 or 2"); 185 break; 186 } 187 if (*end) 188 errx(1, "illegal file number -- %s", optarg); 189 break; 190 case '?': 191 default: 192 usage(); 193 } 194 } 195 argc -= optind; 196 argv += optind; 197 198 if (aflag && vflag) 199 errx(1, "the -a and -v options are mutually exclusive"); 200 201 if (argc != 2) 202 usage(); 203 204 /* Open the files; "-" means stdin. */ 205 if (!strcmp(*argv, "-")) 206 F1->fp = stdin; 207 else if ((F1->fp = fopen(*argv, "r")) == NULL) 208 err(1, "%s", *argv); 209 ++argv; 210 if (!strcmp(*argv, "-")) 211 F2->fp = stdin; 212 else if ((F2->fp = fopen(*argv, "r")) == NULL) 213 err(1, "%s", *argv); 214 if (F1->fp == stdin && F2->fp == stdin) 215 errx(1, "only one input file may be stdin"); 216 217 if (pledge("stdio", NULL) == -1) 218 err(1, "pledge"); 219 220 slurp(F1); 221 slurp(F2); 222 223 /* 224 * We try to let the files have the same field value, advancing 225 * whoever falls behind and always advancing the file(s) we output 226 * from. 227 */ 228 while (F1->setcnt && F2->setcnt) { 229 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 230 if (cval == 0) { 231 /* Oh joy, oh rapture, oh beauty divine! */ 232 if (joinout) 233 joinlines(F1, F2); 234 slurp(F1); 235 slurp(F2); 236 } else if (cval < 0) { 237 /* File 1 takes the lead... */ 238 if (F1->unpair) 239 joinlines(F1, NULL); 240 slurp(F1); 241 } else { 242 /* File 2 takes the lead... */ 243 if (F2->unpair) 244 joinlines(F2, NULL); 245 slurp(F2); 246 } 247 } 248 249 /* 250 * Now that one of the files is used up, optionally output any 251 * remaining lines from the other file. 252 */ 253 if (F1->unpair) 254 while (F1->setcnt) { 255 joinlines(F1, NULL); 256 slurp(F1); 257 } 258 if (F2->unpair) 259 while (F2->setcnt) { 260 joinlines(F2, NULL); 261 slurp(F2); 262 } 263 264 return 0; 265 } 266 267 void 268 slurp(INPUT *F) 269 { 270 LINE *lp, *lastlp, tmp; 271 ssize_t len; 272 u_long cnt; 273 char *bp, *fieldp; 274 275 /* 276 * Read all of the lines from an input file that have the same 277 * join field. 278 */ 279 280 F->setcnt = 0; 281 for (lastlp = NULL; ; ++F->setcnt) { 282 /* 283 * If we're out of space to hold line structures, allocate 284 * more. Initialize the structure so that we know that this 285 * is new space. 286 */ 287 if (F->setcnt == F->setalloc) { 288 LINE *p; 289 u_long newsize = F->setalloc + 50; 290 cnt = F->setalloc; 291 if ((p = reallocarray(F->set, newsize, sizeof(LINE))) 292 == NULL) 293 err(1, NULL); 294 F->set = p; 295 F->setalloc = newsize; 296 memset(F->set + cnt, 0, 50 * sizeof(LINE)); 297 /* re-set lastlp in case it moved */ 298 if (lastlp != NULL) 299 lastlp = &F->set[F->setcnt - 1]; 300 } 301 /* 302 * Get any pushed back line, else get the next line. Allocate 303 * space as necessary. If taking the line from the stack swap 304 * the two structures so that we don't lose space allocated to 305 * either structure. This could be avoided by doing another 306 * level of indirection, but it's probably okay as is. 307 */ 308 lp = &F->set[F->setcnt]; 309 if (F->setcnt) 310 lastlp = &F->set[F->setcnt - 1]; 311 if (F->pushbool) { 312 tmp = F->set[F->setcnt]; 313 F->set[F->setcnt] = F->set[F->pushback]; 314 F->set[F->pushback] = tmp; 315 F->pushbool = 0; 316 continue; 317 } 318 if ((len = getline(&(lp->line), &(lp->linealloc), F->fp)) == -1) 319 break; 320 321 /* Remove the trailing newline, if any. */ 322 if (lp->line[len - 1] == '\n') 323 lp->line[--len] = '\0'; 324 325 /* Split the line into fields, allocate space as necessary. */ 326 lp->fieldcnt = 0; 327 bp = lp->line; 328 while ((fieldp = mbssep(&bp, tabchar)) != NULL) { 329 if (spans && *fieldp == '\0') 330 continue; 331 if (lp->fieldcnt == lp->fieldalloc) { 332 char **p; 333 u_long newsize = lp->fieldalloc + 50; 334 if ((p = reallocarray(lp->fields, newsize, 335 sizeof(char *))) == NULL) 336 err(1, NULL); 337 lp->fields = p; 338 lp->fieldalloc = newsize; 339 } 340 lp->fields[lp->fieldcnt++] = fieldp; 341 } 342 343 /* See if the join field value has changed. */ 344 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) { 345 F->pushbool = 1; 346 F->pushback = F->setcnt; 347 break; 348 } 349 } 350 } 351 352 char * 353 mbssep(char **stringp, const wchar_t *wcdelim) 354 { 355 char *s, *p; 356 size_t ndelim; 357 int i; 358 /* tabchar is never more than 2 */ 359 char mbdelim[2][MB_LEN_MAX + 1]; 360 size_t mblen[2]; 361 362 if ((s = *stringp) == NULL) 363 return NULL; 364 ndelim = wcslen(wcdelim); 365 for (i = 0; i < ndelim; i++) { 366 /* wcdelim generated via mbtowc */ 367 mblen[i] = wctomb(mbdelim[i], wcdelim[i]); 368 } 369 for (p = s; *p != '\0'; p++) { 370 for (i = 0; i < ndelim; i++) { 371 if (strncmp(p, mbdelim[i], mblen[i]) == 0) { 372 *p = '\0'; 373 *stringp = p + mblen[i]; 374 return s; 375 } 376 } 377 } 378 *stringp = NULL; 379 return s; 380 } 381 382 int 383 cmp(LINE *lp1, u_long fieldno1, LINE *lp2, u_long fieldno2) 384 { 385 if (lp1->fieldcnt <= fieldno1) 386 return lp2->fieldcnt <= fieldno2 ? 0 : -1; 387 if (lp2->fieldcnt <= fieldno2) 388 return 1; 389 return strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2]); 390 } 391 392 void 393 joinlines(INPUT *F1, INPUT *F2) 394 { 395 u_long cnt1, cnt2; 396 397 /* 398 * Output the results of a join comparison. The output may be from 399 * either file 1 or file 2 (in which case the first argument is the 400 * file from which to output) or from both. 401 */ 402 if (F2 == NULL) { 403 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 404 outoneline(F1, &F1->set[cnt1]); 405 return; 406 } 407 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 408 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 409 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 410 } 411 412 void 413 outoneline(INPUT *F, LINE *lp) 414 { 415 u_long cnt; 416 417 /* 418 * Output a single line from one of the files, according to the 419 * join rules. This happens when we are writing unmatched single 420 * lines. Output empty fields in the right places. 421 */ 422 if (olist) 423 for (cnt = 0; cnt < olistcnt; ++cnt) { 424 if (olist[cnt].filenum == F->number) 425 outfield(lp, olist[cnt].fieldno, 0); 426 else if (olist[cnt].filenum == 0) 427 outfield(lp, F->joinf, 0); 428 else 429 outfield(lp, 0, 1); 430 } 431 else { 432 /* 433 * Output the join field, then the remaining fields from F 434 */ 435 outfield(lp, F->joinf, 0); 436 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 437 if (F->joinf != cnt) 438 outfield(lp, cnt, 0); 439 } 440 441 putchar('\n'); 442 if (ferror(stdout)) 443 err(1, "stdout"); 444 needsep = 0; 445 } 446 447 void 448 outtwoline(INPUT *F1, LINE *lp1, INPUT *F2, LINE *lp2) 449 { 450 u_long cnt; 451 452 /* Output a pair of lines according to the join list (if any). */ 453 if (olist) { 454 for (cnt = 0; cnt < olistcnt; ++cnt) 455 if (olist[cnt].filenum == 0) { 456 if (lp1->fieldcnt >= F1->joinf) 457 outfield(lp1, F1->joinf, 0); 458 else 459 outfield(lp2, F2->joinf, 0); 460 } else if (olist[cnt].filenum == 1) 461 outfield(lp1, olist[cnt].fieldno, 0); 462 else /* if (olist[cnt].filenum == 2) */ 463 outfield(lp2, olist[cnt].fieldno, 0); 464 } else { 465 /* 466 * Output the join field, then the remaining fields from F1 467 * and F2. 468 */ 469 outfield(lp1, F1->joinf, 0); 470 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 471 if (F1->joinf != cnt) 472 outfield(lp1, cnt, 0); 473 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 474 if (F2->joinf != cnt) 475 outfield(lp2, cnt, 0); 476 } 477 putchar('\n'); 478 if (ferror(stdout)) 479 err(1, "stdout"); 480 needsep = 0; 481 } 482 483 void 484 outfield(LINE *lp, u_long fieldno, int out_empty) 485 { 486 if (needsep++) 487 putwchar(*tabchar); 488 if (!ferror(stdout)) { 489 if (lp->fieldcnt <= fieldno || out_empty) { 490 if (empty != NULL) 491 fputs(empty, stdout); 492 } else { 493 if (*lp->fields[fieldno] == '\0') 494 return; 495 fputs(lp->fields[fieldno], stdout); 496 } 497 } 498 if (ferror(stdout)) 499 err(1, "stdout"); 500 } 501 502 /* 503 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 504 * fields. 505 */ 506 void 507 fieldarg(char *option) 508 { 509 u_long fieldno, filenum; 510 char *end, *token; 511 512 while ((token = strsep(&option, ", \t")) != NULL) { 513 if (*token == '\0') 514 continue; 515 if (token[0] == '0') 516 filenum = fieldno = 0; 517 else if ((token[0] == '1' || token[0] == '2') && 518 token[1] == '.') { 519 filenum = token[0] - '0'; 520 fieldno = strtol(token + 2, &end, 10); 521 if (*end) 522 errx(1, "malformed -o option field"); 523 if (fieldno == 0) 524 errx(1, "field numbers are 1 based"); 525 --fieldno; 526 } else 527 errx(1, "malformed -o option field"); 528 if (olistcnt == olistalloc) { 529 OLIST *p; 530 u_long newsize = olistalloc + 50; 531 if ((p = reallocarray(olist, newsize, sizeof(OLIST))) 532 == NULL) 533 err(1, NULL); 534 olist = p; 535 olistalloc = newsize; 536 } 537 olist[olistcnt].filenum = filenum; 538 olist[olistcnt].fieldno = fieldno; 539 ++olistcnt; 540 } 541 } 542 543 void 544 obsolete(char **argv) 545 { 546 size_t len; 547 char **p, *ap, *t; 548 549 while ((ap = *++argv) != NULL) { 550 /* Return if "--". */ 551 if (ap[0] == '-' && ap[1] == '-') 552 return; 553 /* skip if not an option */ 554 if (ap[0] != '-') 555 continue; 556 switch (ap[1]) { 557 case 'a': 558 /* 559 * The original join allowed "-a", which meant the 560 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 561 * only specifies this as "-a 1" and "a -2", so we 562 * have to use another option flag, one that is 563 * unlikely to ever be used or accidentally entered 564 * on the command line. (Well, we could reallocate 565 * the argv array, but that hardly seems worthwhile.) 566 */ 567 if (ap[2] == '\0' && (argv[1] == NULL || 568 (strcmp(argv[1], "1") != 0 && 569 strcmp(argv[1], "2") != 0))) { 570 ap[1] = '\01'; 571 warnx("-a option used without an argument; " 572 "reverting to historical behavior"); 573 } 574 break; 575 case 'j': 576 /* 577 * The original join allowed "-j[12] arg" and "-j arg". 578 * Convert the former to "-[12] arg". Don't convert 579 * the latter since getopt(3) can handle it. 580 */ 581 switch(ap[2]) { 582 case '1': 583 case '2': 584 if (ap[3] != '\0') 585 goto jbad; 586 ap[1] = ap[2]; 587 ap[2] = '\0'; 588 break; 589 case '\0': 590 break; 591 default: 592 jbad: warnx("unknown option -- %s", ap + 1); 593 usage(); 594 } 595 break; 596 case 'o': 597 /* 598 * The original join allowed "-o arg arg". 599 * Convert to "-o arg -o arg". 600 */ 601 if (ap[2] != '\0' || argv[1] == NULL) 602 break; 603 for (p = argv + 2; *p != NULL; ++p) { 604 if (p[0][0] == '0' || ((p[0][0] != '1' && 605 p[0][0] != '2') || p[0][1] != '.')) 606 break; 607 len = strlen(*p); 608 if (len - 2 != strspn(*p + 2, "0123456789")) 609 break; 610 if ((t = malloc(len + 3)) == NULL) 611 err(1, NULL); 612 t[0] = '-'; 613 t[1] = 'o'; 614 memmove(t + 2, *p, len + 1); 615 *p = t; 616 } 617 argv = p - 1; 618 break; 619 } 620 } 621 } 622 623 void 624 usage(void) 625 { 626 int len; 627 extern char *__progname; 628 629 len = strlen(__progname) + sizeof("usage: "); 630 (void)fprintf(stderr, "usage: %s [-1 field] [-2 field] " 631 "[-a file_number | -v file_number] [-e string]\n" 632 "%*s[-o list] [-t char] file1 file2\n", 633 __progname, len, ""); 634 exit(1); 635 } 636