1 /* $NetBSD: expand.c,v 1.88 2012/12/22 20:15:22 dsl Exp $ */ 2 3 /*- 4 * Copyright (c) 1991, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Kenneth Almquist. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. Neither the name of the University nor the names of its contributors 19 * may be used to endorse or promote products derived from this software 20 * without specific prior written permission. 21 * 22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 32 * SUCH DAMAGE. 33 */ 34 35 #include <sys/cdefs.h> 36 #ifndef lint 37 #if 0 38 static char sccsid[] = "@(#)expand.c 8.5 (Berkeley) 5/15/95"; 39 #else 40 __RCSID("$NetBSD: expand.c,v 1.88 2012/12/22 20:15:22 dsl Exp $"); 41 #endif 42 #endif /* not lint */ 43 44 #include <sys/types.h> 45 #include <sys/time.h> 46 #include <sys/stat.h> 47 #include <errno.h> 48 #include <dirent.h> 49 #include <unistd.h> 50 #include <pwd.h> 51 #include <limits.h> 52 #include <stdlib.h> 53 #include <stdio.h> 54 55 /* 56 * Routines to expand arguments to commands. We have to deal with 57 * backquotes, shell variables, and file metacharacters. 58 */ 59 60 #include "shell.h" 61 #include "main.h" 62 #include "nodes.h" 63 #include "eval.h" 64 #include "expand.h" 65 #include "syntax.h" 66 #include "parser.h" 67 #include "jobs.h" 68 #include "options.h" 69 #include "builtins.h" 70 #include "var.h" 71 #include "input.h" 72 #include "output.h" 73 #include "memalloc.h" 74 #include "error.h" 75 #include "mystring.h" 76 #include "show.h" 77 78 /* 79 * Structure specifying which parts of the string should be searched 80 * for IFS characters. 81 */ 82 83 struct ifsregion { 84 struct ifsregion *next; /* next region in list */ 85 int begoff; /* offset of start of region */ 86 int endoff; /* offset of end of region */ 87 int inquotes; /* search for nul bytes only */ 88 }; 89 90 91 char *expdest; /* output of current string */ 92 struct nodelist *argbackq; /* list of back quote expressions */ 93 struct ifsregion ifsfirst; /* first struct in list of ifs regions */ 94 struct ifsregion *ifslastp; /* last struct in list */ 95 struct arglist exparg; /* holds expanded arg list */ 96 97 STATIC void argstr(char *, int); 98 STATIC char *exptilde(char *, int); 99 STATIC void expbackq(union node *, int, int); 100 STATIC int subevalvar(char *, char *, int, int, int, int); 101 STATIC char *evalvar(char *, int); 102 STATIC int varisset(char *, int); 103 STATIC void varvalue(char *, int, int, int); 104 STATIC void recordregion(int, int, int); 105 STATIC void removerecordregions(int); 106 STATIC void ifsbreakup(char *, struct arglist *); 107 STATIC void ifsfree(void); 108 STATIC void expandmeta(struct strlist *, int); 109 STATIC void expmeta(char *, char *); 110 STATIC void addfname(char *); 111 STATIC struct strlist *expsort(struct strlist *); 112 STATIC struct strlist *msort(struct strlist *, int); 113 STATIC int pmatch(char *, char *, int); 114 STATIC char *cvtnum(int, char *); 115 116 /* 117 * Expand shell variables and backquotes inside a here document. 118 */ 119 120 void 121 expandhere(union node *arg, int fd) 122 { 123 herefd = fd; 124 expandarg(arg, NULL, 0); 125 xwrite(fd, stackblock(), expdest - stackblock()); 126 } 127 128 129 /* 130 * Perform variable substitution and command substitution on an argument, 131 * placing the resulting list of arguments in arglist. If EXP_FULL is true, 132 * perform splitting and file name expansion. When arglist is NULL, perform 133 * here document expansion. 134 */ 135 136 void 137 expandarg(union node *arg, struct arglist *arglist, int flag) 138 { 139 struct strlist *sp; 140 char *p; 141 142 argbackq = arg->narg.backquote; 143 STARTSTACKSTR(expdest); 144 ifsfirst.next = NULL; 145 ifslastp = NULL; 146 argstr(arg->narg.text, flag); 147 if (arglist == NULL) { 148 return; /* here document expanded */ 149 } 150 STPUTC('\0', expdest); 151 p = grabstackstr(expdest); 152 exparg.lastp = &exparg.list; 153 /* 154 * TODO - EXP_REDIR 155 */ 156 if (flag & EXP_FULL) { 157 ifsbreakup(p, &exparg); 158 *exparg.lastp = NULL; 159 exparg.lastp = &exparg.list; 160 expandmeta(exparg.list, flag); 161 } else { 162 if (flag & EXP_REDIR) /*XXX - for now, just remove escapes */ 163 rmescapes(p); 164 sp = (struct strlist *)stalloc(sizeof (struct strlist)); 165 sp->text = p; 166 *exparg.lastp = sp; 167 exparg.lastp = &sp->next; 168 } 169 ifsfree(); 170 *exparg.lastp = NULL; 171 if (exparg.list) { 172 *arglist->lastp = exparg.list; 173 arglist->lastp = exparg.lastp; 174 } 175 } 176 177 178 179 /* 180 * Perform variable and command substitution. 181 * If EXP_FULL is set, output CTLESC characters to allow for further processing. 182 * Otherwise treat $@ like $* since no splitting will be performed. 183 */ 184 185 STATIC void 186 argstr(char *p, int flag) 187 { 188 char c; 189 int quotes = flag & (EXP_FULL | EXP_CASE); /* do CTLESC */ 190 int firsteq = 1; 191 const char *ifs = NULL; 192 int ifs_split = EXP_IFS_SPLIT; 193 194 if (flag & EXP_IFS_SPLIT) 195 ifs = ifsset() ? ifsval() : " \t\n"; 196 197 if (*p == '~' && (flag & (EXP_TILDE | EXP_VARTILDE))) 198 p = exptilde(p, flag); 199 for (;;) { 200 switch (c = *p++) { 201 case '\0': 202 case CTLENDVAR: /* end of expanding yyy in ${xxx-yyy} */ 203 return; 204 case CTLQUOTEMARK: 205 /* "$@" syntax adherence hack */ 206 if (p[0] == CTLVAR && p[2] == '@' && p[3] == '=') 207 break; 208 if ((flag & EXP_FULL) != 0) 209 STPUTC(c, expdest); 210 ifs_split = 0; 211 break; 212 case CTLQUOTEEND: 213 ifs_split = EXP_IFS_SPLIT; 214 break; 215 case CTLESC: 216 if (quotes) 217 STPUTC(c, expdest); 218 c = *p++; 219 STPUTC(c, expdest); 220 break; 221 case CTLVAR: 222 p = evalvar(p, (flag & ~EXP_IFS_SPLIT) | (flag & ifs_split)); 223 break; 224 case CTLBACKQ: 225 case CTLBACKQ|CTLQUOTE: 226 expbackq(argbackq->n, c & CTLQUOTE, flag); 227 argbackq = argbackq->next; 228 break; 229 case CTLENDARI: 230 expari(flag); 231 break; 232 case ':': 233 case '=': 234 /* 235 * sort of a hack - expand tildes in variable 236 * assignments (after the first '=' and after ':'s). 237 */ 238 STPUTC(c, expdest); 239 if (flag & EXP_VARTILDE && *p == '~') { 240 if (c == '=') { 241 if (firsteq) 242 firsteq = 0; 243 else 244 break; 245 } 246 p = exptilde(p, flag); 247 } 248 break; 249 default: 250 STPUTC(c, expdest); 251 if (flag & ifs_split && strchr(ifs, c) != NULL) { 252 /* We need to get the output split here... */ 253 recordregion(expdest - stackblock() - 1, 254 expdest - stackblock(), 0); 255 } 256 break; 257 } 258 } 259 } 260 261 STATIC char * 262 exptilde(char *p, int flag) 263 { 264 char c, *startp = p; 265 struct passwd *pw; 266 const char *home; 267 int quotes = flag & (EXP_FULL | EXP_CASE); 268 269 while ((c = *p) != '\0') { 270 switch(c) { 271 case CTLESC: 272 return (startp); 273 case CTLQUOTEMARK: 274 return (startp); 275 case ':': 276 if (flag & EXP_VARTILDE) 277 goto done; 278 break; 279 case '/': 280 goto done; 281 } 282 p++; 283 } 284 done: 285 *p = '\0'; 286 if (*(startp+1) == '\0') { 287 if ((home = lookupvar("HOME")) == NULL) 288 goto lose; 289 } else { 290 if ((pw = getpwnam(startp+1)) == NULL) 291 goto lose; 292 home = pw->pw_dir; 293 } 294 if (*home == '\0') 295 goto lose; 296 *p = c; 297 while ((c = *home++) != '\0') { 298 if (quotes && SQSYNTAX[(int)c] == CCTL) 299 STPUTC(CTLESC, expdest); 300 STPUTC(c, expdest); 301 } 302 return (p); 303 lose: 304 *p = c; 305 return (startp); 306 } 307 308 309 STATIC void 310 removerecordregions(int endoff) 311 { 312 if (ifslastp == NULL) 313 return; 314 315 if (ifsfirst.endoff > endoff) { 316 while (ifsfirst.next != NULL) { 317 struct ifsregion *ifsp; 318 INTOFF; 319 ifsp = ifsfirst.next->next; 320 ckfree(ifsfirst.next); 321 ifsfirst.next = ifsp; 322 INTON; 323 } 324 if (ifsfirst.begoff > endoff) 325 ifslastp = NULL; 326 else { 327 ifslastp = &ifsfirst; 328 ifsfirst.endoff = endoff; 329 } 330 return; 331 } 332 333 ifslastp = &ifsfirst; 334 while (ifslastp->next && ifslastp->next->begoff < endoff) 335 ifslastp=ifslastp->next; 336 while (ifslastp->next != NULL) { 337 struct ifsregion *ifsp; 338 INTOFF; 339 ifsp = ifslastp->next->next; 340 ckfree(ifslastp->next); 341 ifslastp->next = ifsp; 342 INTON; 343 } 344 if (ifslastp->endoff > endoff) 345 ifslastp->endoff = endoff; 346 } 347 348 349 /* 350 * Expand arithmetic expression. Backup to start of expression, 351 * evaluate, place result in (backed up) result, adjust string position. 352 */ 353 void 354 expari(int flag) 355 { 356 char *p, *start; 357 intmax_t result; 358 int adjustment; 359 int begoff; 360 int quotes = flag & (EXP_FULL | EXP_CASE); 361 int quoted; 362 363 /* ifsfree(); */ 364 365 /* 366 * This routine is slightly over-complicated for 367 * efficiency. First we make sure there is 368 * enough space for the result, which may be bigger 369 * than the expression if we add exponentation. Next we 370 * scan backwards looking for the start of arithmetic. If the 371 * next previous character is a CTLESC character, then we 372 * have to rescan starting from the beginning since CTLESC 373 * characters have to be processed left to right. 374 */ 375 /* SPACE_NEEDED is enough for all digits, plus possible "-", plus 2 (why?) */ 376 #define SPACE_NEEDED ((sizeof(intmax_t) * CHAR_BIT + 2) / 3 + 1 + 2) 377 CHECKSTRSPACE((int)(SPACE_NEEDED - 2), expdest); 378 USTPUTC('\0', expdest); 379 start = stackblock(); 380 p = expdest - 1; 381 while (*p != CTLARI && p >= start) 382 --p; 383 if (*p != CTLARI) 384 error("missing CTLARI (shouldn't happen)"); 385 if (p > start && *(p-1) == CTLESC) 386 for (p = start; *p != CTLARI; p++) 387 if (*p == CTLESC) 388 p++; 389 390 if (p[1] == '"') 391 quoted=1; 392 else 393 quoted=0; 394 begoff = p - start; 395 removerecordregions(begoff); 396 if (quotes) 397 rmescapes(p+2); 398 result = arith(p+2); 399 fmtstr(p, SPACE_NEEDED, "%"PRIdMAX, result); 400 401 while (*p++) 402 ; 403 404 if (quoted == 0) 405 recordregion(begoff, p - 1 - start, 0); 406 adjustment = expdest - p + 1; 407 STADJUST(-adjustment, expdest); 408 } 409 410 411 /* 412 * Expand stuff in backwards quotes. 413 */ 414 415 STATIC void 416 expbackq(union node *cmd, int quoted, int flag) 417 { 418 struct backcmd in; 419 int i; 420 char buf[128]; 421 char *p; 422 char *dest = expdest; 423 struct ifsregion saveifs, *savelastp; 424 struct nodelist *saveargbackq; 425 char lastc; 426 int startloc = dest - stackblock(); 427 char const *syntax = quoted? DQSYNTAX : BASESYNTAX; 428 int saveherefd; 429 int quotes = flag & (EXP_FULL | EXP_CASE); 430 int nnl; 431 432 INTOFF; 433 saveifs = ifsfirst; 434 savelastp = ifslastp; 435 saveargbackq = argbackq; 436 saveherefd = herefd; 437 herefd = -1; 438 p = grabstackstr(dest); 439 evalbackcmd(cmd, &in); 440 ungrabstackstr(p, dest); 441 ifsfirst = saveifs; 442 ifslastp = savelastp; 443 argbackq = saveargbackq; 444 herefd = saveherefd; 445 446 p = in.buf; 447 lastc = '\0'; 448 nnl = 0; 449 for (;;) { 450 if (--in.nleft < 0) { 451 if (in.fd < 0) 452 break; 453 while ((i = read(in.fd, buf, sizeof buf)) < 0 && errno == EINTR); 454 TRACE(("expbackq: read returns %d\n", i)); 455 if (i <= 0) 456 break; 457 p = buf; 458 in.nleft = i - 1; 459 } 460 lastc = *p++; 461 if (lastc != '\0') { 462 if (lastc == '\n') 463 nnl++; 464 else { 465 CHECKSTRSPACE(nnl + 2, dest); 466 while (nnl > 0) { 467 nnl--; 468 USTPUTC('\n', dest); 469 } 470 if (quotes && syntax[(int)lastc] == CCTL) 471 USTPUTC(CTLESC, dest); 472 USTPUTC(lastc, dest); 473 } 474 } 475 } 476 477 if (in.fd >= 0) 478 close(in.fd); 479 if (in.buf) 480 ckfree(in.buf); 481 if (in.jp) 482 back_exitstatus = waitforjob(in.jp); 483 if (quoted == 0) 484 recordregion(startloc, dest - stackblock(), 0); 485 TRACE(("evalbackq: size=%d: \"%.*s\"\n", 486 (int)((dest - stackblock()) - startloc), 487 (int)((dest - stackblock()) - startloc), 488 stackblock() + startloc)); 489 expdest = dest; 490 INTON; 491 } 492 493 494 495 STATIC int 496 subevalvar(char *p, char *str, int strloc, int subtype, int startloc, int varflags) 497 { 498 char *startp; 499 char *loc = NULL; 500 char *q; 501 int c = 0; 502 int saveherefd = herefd; 503 struct nodelist *saveargbackq = argbackq; 504 int amount, how; 505 506 herefd = -1; 507 switch (subtype) { 508 case VSTRIMLEFT: 509 case VSTRIMLEFTMAX: 510 case VSTRIMRIGHT: 511 case VSTRIMRIGHTMAX: 512 how = (varflags & VSQUOTE) ? 0 : EXP_CASE; 513 break; 514 default: 515 how = 0; 516 break; 517 } 518 argstr(p, how); 519 STACKSTRNUL(expdest); 520 herefd = saveherefd; 521 argbackq = saveargbackq; 522 startp = stackblock() + startloc; 523 if (str == NULL) 524 str = stackblock() + strloc; 525 526 switch (subtype) { 527 case VSASSIGN: 528 setvar(str, startp, 0); 529 amount = startp - expdest; 530 STADJUST(amount, expdest); 531 varflags &= ~VSNUL; 532 return 1; 533 534 case VSQUESTION: 535 if (*p != CTLENDVAR) { 536 outfmt(&errout, "%s\n", startp); 537 error(NULL); 538 } 539 error("%.*s: parameter %snot set", 540 (int)(p - str - 1), 541 str, (varflags & VSNUL) ? "null or " 542 : nullstr); 543 /* NOTREACHED */ 544 545 case VSTRIMLEFT: 546 for (loc = startp; loc < str; loc++) { 547 c = *loc; 548 *loc = '\0'; 549 if (patmatch(str, startp, varflags & VSQUOTE)) 550 goto recordleft; 551 *loc = c; 552 if ((varflags & VSQUOTE) && *loc == CTLESC) 553 loc++; 554 } 555 return 0; 556 557 case VSTRIMLEFTMAX: 558 for (loc = str - 1; loc >= startp;) { 559 c = *loc; 560 *loc = '\0'; 561 if (patmatch(str, startp, varflags & VSQUOTE)) 562 goto recordleft; 563 *loc = c; 564 loc--; 565 if ((varflags & VSQUOTE) && loc > startp && 566 *(loc - 1) == CTLESC) { 567 for (q = startp; q < loc; q++) 568 if (*q == CTLESC) 569 q++; 570 if (q > loc) 571 loc--; 572 } 573 } 574 return 0; 575 576 case VSTRIMRIGHT: 577 for (loc = str - 1; loc >= startp;) { 578 if (patmatch(str, loc, varflags & VSQUOTE)) 579 goto recordright; 580 loc--; 581 if ((varflags & VSQUOTE) && loc > startp && 582 *(loc - 1) == CTLESC) { 583 for (q = startp; q < loc; q++) 584 if (*q == CTLESC) 585 q++; 586 if (q > loc) 587 loc--; 588 } 589 } 590 return 0; 591 592 case VSTRIMRIGHTMAX: 593 for (loc = startp; loc < str - 1; loc++) { 594 if (patmatch(str, loc, varflags & VSQUOTE)) 595 goto recordright; 596 if ((varflags & VSQUOTE) && *loc == CTLESC) 597 loc++; 598 } 599 return 0; 600 601 default: 602 abort(); 603 } 604 605 recordleft: 606 *loc = c; 607 amount = ((str - 1) - (loc - startp)) - expdest; 608 STADJUST(amount, expdest); 609 while (loc != str - 1) 610 *startp++ = *loc++; 611 return 1; 612 613 recordright: 614 amount = loc - expdest; 615 STADJUST(amount, expdest); 616 STPUTC('\0', expdest); 617 STADJUST(-1, expdest); 618 return 1; 619 } 620 621 622 /* 623 * Expand a variable, and return a pointer to the next character in the 624 * input string. 625 */ 626 627 STATIC char * 628 evalvar(char *p, int flag) 629 { 630 int subtype; 631 int varflags; 632 char *var; 633 char *val; 634 int patloc; 635 int c; 636 int set; 637 int special; 638 int startloc; 639 int varlen; 640 int apply_ifs; 641 int quotes = flag & (EXP_FULL | EXP_CASE); 642 643 varflags = (unsigned char)*p++; 644 subtype = varflags & VSTYPE; 645 var = p; 646 special = !is_name(*p); 647 p = strchr(p, '=') + 1; 648 649 again: /* jump here after setting a variable with ${var=text} */ 650 if (special) { 651 set = varisset(var, varflags & VSNUL); 652 val = NULL; 653 } else { 654 val = lookupvar(var); 655 if (val == NULL || ((varflags & VSNUL) && val[0] == '\0')) { 656 val = NULL; 657 set = 0; 658 } else 659 set = 1; 660 } 661 662 varlen = 0; 663 startloc = expdest - stackblock(); 664 665 if (!set && uflag) { 666 switch (subtype) { 667 case VSNORMAL: 668 case VSTRIMLEFT: 669 case VSTRIMLEFTMAX: 670 case VSTRIMRIGHT: 671 case VSTRIMRIGHTMAX: 672 case VSLENGTH: 673 error("%.*s: parameter not set", 674 (int)(p - var - 1), var); 675 /* NOTREACHED */ 676 } 677 } 678 679 if (set && subtype != VSPLUS) { 680 /* insert the value of the variable */ 681 if (special) { 682 varvalue(var, varflags & VSQUOTE, subtype, flag); 683 if (subtype == VSLENGTH) { 684 varlen = expdest - stackblock() - startloc; 685 STADJUST(-varlen, expdest); 686 } 687 } else { 688 char const *syntax = (varflags & VSQUOTE) ? DQSYNTAX 689 : BASESYNTAX; 690 691 if (subtype == VSLENGTH) { 692 for (;*val; val++) 693 varlen++; 694 } else { 695 while (*val) { 696 if (quotes && syntax[(int)*val] == CCTL) 697 STPUTC(CTLESC, expdest); 698 STPUTC(*val++, expdest); 699 } 700 701 } 702 } 703 } 704 705 706 if (flag & EXP_IN_QUOTES) 707 apply_ifs = 0; 708 else if (varflags & VSQUOTE) { 709 if (*var == '@' && shellparam.nparam != 1) 710 apply_ifs = 1; 711 else { 712 /* 713 * Mark so that we don't apply IFS if we recurse through 714 * here expanding $bar from "${foo-$bar}". 715 */ 716 flag |= EXP_IN_QUOTES; 717 apply_ifs = 0; 718 } 719 } else 720 apply_ifs = 1; 721 722 switch (subtype) { 723 case VSLENGTH: 724 expdest = cvtnum(varlen, expdest); 725 break; 726 727 case VSNORMAL: 728 break; 729 730 case VSPLUS: 731 set = !set; 732 /* FALLTHROUGH */ 733 case VSMINUS: 734 if (!set) { 735 argstr(p, flag | (apply_ifs ? EXP_IFS_SPLIT : 0)); 736 /* 737 * ${x-a b c} doesn't get split, but removing the 738 * 'apply_ifs = 0' apparently breaks ${1+"$@"}.. 739 * ${x-'a b' c} should generate 2 args. 740 */ 741 /* We should have marked stuff already */ 742 apply_ifs = 0; 743 } 744 break; 745 746 case VSTRIMLEFT: 747 case VSTRIMLEFTMAX: 748 case VSTRIMRIGHT: 749 case VSTRIMRIGHTMAX: 750 if (!set) 751 break; 752 /* 753 * Terminate the string and start recording the pattern 754 * right after it 755 */ 756 STPUTC('\0', expdest); 757 patloc = expdest - stackblock(); 758 if (subevalvar(p, NULL, patloc, subtype, 759 startloc, varflags) == 0) { 760 int amount = (expdest - stackblock() - patloc) + 1; 761 STADJUST(-amount, expdest); 762 } 763 /* Remove any recorded regions beyond start of variable */ 764 removerecordregions(startloc); 765 apply_ifs = 1; 766 break; 767 768 case VSASSIGN: 769 case VSQUESTION: 770 if (set) 771 break; 772 if (subevalvar(p, var, 0, subtype, startloc, varflags)) { 773 varflags &= ~VSNUL; 774 /* 775 * Remove any recorded regions beyond 776 * start of variable 777 */ 778 removerecordregions(startloc); 779 goto again; 780 } 781 apply_ifs = 0; 782 break; 783 784 default: 785 abort(); 786 } 787 788 if (apply_ifs) 789 recordregion(startloc, expdest - stackblock(), 790 varflags & VSQUOTE); 791 792 if (subtype != VSNORMAL) { /* skip to end of alternative */ 793 int nesting = 1; 794 for (;;) { 795 if ((c = *p++) == CTLESC) 796 p++; 797 else if (c == CTLBACKQ || c == (CTLBACKQ|CTLQUOTE)) { 798 if (set) 799 argbackq = argbackq->next; 800 } else if (c == CTLVAR) { 801 if ((*p++ & VSTYPE) != VSNORMAL) 802 nesting++; 803 } else if (c == CTLENDVAR) { 804 if (--nesting == 0) 805 break; 806 } 807 } 808 } 809 return p; 810 } 811 812 813 814 /* 815 * Test whether a specialized variable is set. 816 */ 817 818 STATIC int 819 varisset(char *name, int nulok) 820 { 821 if (*name == '!') 822 return backgndpid != -1; 823 else if (*name == '@' || *name == '*') { 824 if (*shellparam.p == NULL) 825 return 0; 826 827 if (nulok) { 828 char **av; 829 830 for (av = shellparam.p; *av; av++) 831 if (**av != '\0') 832 return 1; 833 return 0; 834 } 835 } else if (is_digit(*name)) { 836 char *ap; 837 int num = atoi(name); 838 839 if (num > shellparam.nparam) 840 return 0; 841 842 if (num == 0) 843 ap = arg0; 844 else 845 ap = shellparam.p[num - 1]; 846 847 if (nulok && (ap == NULL || *ap == '\0')) 848 return 0; 849 } 850 return 1; 851 } 852 853 854 855 /* 856 * Add the value of a specialized variable to the stack string. 857 */ 858 859 STATIC void 860 varvalue(char *name, int quoted, int subtype, int flag) 861 { 862 int num; 863 char *p; 864 int i; 865 char sep; 866 char **ap; 867 char const *syntax; 868 869 #define STRTODEST(p) \ 870 do {\ 871 if (flag & (EXP_FULL | EXP_CASE) && subtype != VSLENGTH) { \ 872 syntax = quoted? DQSYNTAX : BASESYNTAX; \ 873 while (*p) { \ 874 if (syntax[(int)*p] == CCTL) \ 875 STPUTC(CTLESC, expdest); \ 876 STPUTC(*p++, expdest); \ 877 } \ 878 } else \ 879 while (*p) \ 880 STPUTC(*p++, expdest); \ 881 } while (0) 882 883 884 switch (*name) { 885 case '$': 886 num = rootpid; 887 goto numvar; 888 case '?': 889 num = exitstatus; 890 goto numvar; 891 case '#': 892 num = shellparam.nparam; 893 goto numvar; 894 case '!': 895 num = backgndpid; 896 numvar: 897 expdest = cvtnum(num, expdest); 898 break; 899 case '-': 900 for (i = 0; optlist[i].name; i++) { 901 if (optlist[i].val && optlist[i].letter) 902 STPUTC(optlist[i].letter, expdest); 903 } 904 break; 905 case '@': 906 if (flag & EXP_FULL && quoted) { 907 for (ap = shellparam.p ; (p = *ap++) != NULL ; ) { 908 STRTODEST(p); 909 if (*ap) 910 /* A NUL separates args inside "" */ 911 STPUTC('\0', expdest); 912 } 913 break; 914 } 915 /* fall through */ 916 case '*': 917 if (ifsset() != 0) 918 sep = ifsval()[0]; 919 else 920 sep = ' '; 921 for (ap = shellparam.p ; (p = *ap++) != NULL ; ) { 922 STRTODEST(p); 923 if (*ap && sep) 924 STPUTC(sep, expdest); 925 } 926 break; 927 case '0': 928 p = arg0; 929 STRTODEST(p); 930 break; 931 default: 932 if (is_digit(*name)) { 933 num = atoi(name); 934 if (num > 0 && num <= shellparam.nparam) { 935 p = shellparam.p[num - 1]; 936 STRTODEST(p); 937 } 938 } 939 break; 940 } 941 } 942 943 944 945 /* 946 * Record the fact that we have to scan this region of the 947 * string for IFS characters. 948 */ 949 950 STATIC void 951 recordregion(int start, int end, int inquotes) 952 { 953 struct ifsregion *ifsp; 954 955 if (ifslastp == NULL) { 956 ifsp = &ifsfirst; 957 } else { 958 if (ifslastp->endoff == start 959 && ifslastp->inquotes == inquotes) { 960 /* extend previous area */ 961 ifslastp->endoff = end; 962 return; 963 } 964 ifsp = (struct ifsregion *)ckmalloc(sizeof (struct ifsregion)); 965 ifslastp->next = ifsp; 966 } 967 ifslastp = ifsp; 968 ifslastp->next = NULL; 969 ifslastp->begoff = start; 970 ifslastp->endoff = end; 971 ifslastp->inquotes = inquotes; 972 } 973 974 975 976 /* 977 * Break the argument string into pieces based upon IFS and add the 978 * strings to the argument list. The regions of the string to be 979 * searched for IFS characters have been stored by recordregion. 980 */ 981 STATIC void 982 ifsbreakup(char *string, struct arglist *arglist) 983 { 984 struct ifsregion *ifsp; 985 struct strlist *sp; 986 char *start; 987 char *p; 988 char *q; 989 const char *ifs; 990 const char *ifsspc; 991 int had_param_ch = 0; 992 993 start = string; 994 995 if (ifslastp == NULL) { 996 /* Return entire argument, IFS doesn't apply to any of it */ 997 sp = (struct strlist *)stalloc(sizeof *sp); 998 sp->text = start; 999 *arglist->lastp = sp; 1000 arglist->lastp = &sp->next; 1001 return; 1002 } 1003 1004 ifs = ifsset() ? ifsval() : " \t\n"; 1005 1006 for (ifsp = &ifsfirst; ifsp != NULL; ifsp = ifsp->next) { 1007 p = string + ifsp->begoff; 1008 while (p < string + ifsp->endoff) { 1009 had_param_ch = 1; 1010 q = p; 1011 if (*p == CTLESC) 1012 p++; 1013 if (ifsp->inquotes) { 1014 /* Only NULs (should be from "$@") end args */ 1015 if (*p != 0) { 1016 p++; 1017 continue; 1018 } 1019 ifsspc = NULL; 1020 } else { 1021 if (!strchr(ifs, *p)) { 1022 p++; 1023 continue; 1024 } 1025 had_param_ch = 0; 1026 ifsspc = strchr(" \t\n", *p); 1027 1028 /* Ignore IFS whitespace at start */ 1029 if (q == start && ifsspc != NULL) { 1030 p++; 1031 start = p; 1032 continue; 1033 } 1034 } 1035 1036 /* Save this argument... */ 1037 *q = '\0'; 1038 sp = (struct strlist *)stalloc(sizeof *sp); 1039 sp->text = start; 1040 *arglist->lastp = sp; 1041 arglist->lastp = &sp->next; 1042 p++; 1043 1044 if (ifsspc != NULL) { 1045 /* Ignore further trailing IFS whitespace */ 1046 for (; p < string + ifsp->endoff; p++) { 1047 q = p; 1048 if (*p == CTLESC) 1049 p++; 1050 if (strchr(ifs, *p) == NULL) { 1051 p = q; 1052 break; 1053 } 1054 if (strchr(" \t\n", *p) == NULL) { 1055 p++; 1056 break; 1057 } 1058 } 1059 } 1060 start = p; 1061 } 1062 } 1063 1064 /* 1065 * Save anything left as an argument. 1066 * Traditionally we have treated 'IFS=':'; set -- x$IFS' as 1067 * generating 2 arguments, the second of which is empty. 1068 * Some recent clarification of the Posix spec say that it 1069 * should only generate one.... 1070 */ 1071 if (had_param_ch || *start != 0) { 1072 sp = (struct strlist *)stalloc(sizeof *sp); 1073 sp->text = start; 1074 *arglist->lastp = sp; 1075 arglist->lastp = &sp->next; 1076 } 1077 } 1078 1079 STATIC void 1080 ifsfree(void) 1081 { 1082 while (ifsfirst.next != NULL) { 1083 struct ifsregion *ifsp; 1084 INTOFF; 1085 ifsp = ifsfirst.next->next; 1086 ckfree(ifsfirst.next); 1087 ifsfirst.next = ifsp; 1088 INTON; 1089 } 1090 ifslastp = NULL; 1091 ifsfirst.next = NULL; 1092 } 1093 1094 1095 1096 /* 1097 * Expand shell metacharacters. At this point, the only control characters 1098 * should be escapes. The results are stored in the list exparg. 1099 */ 1100 1101 char *expdir; 1102 1103 1104 STATIC void 1105 expandmeta(struct strlist *str, int flag) 1106 { 1107 char *p; 1108 struct strlist **savelastp; 1109 struct strlist *sp; 1110 char c; 1111 /* TODO - EXP_REDIR */ 1112 1113 while (str) { 1114 if (fflag) 1115 goto nometa; 1116 p = str->text; 1117 for (;;) { /* fast check for meta chars */ 1118 if ((c = *p++) == '\0') 1119 goto nometa; 1120 if (c == '*' || c == '?' || c == '[' || c == '!') 1121 break; 1122 } 1123 savelastp = exparg.lastp; 1124 INTOFF; 1125 if (expdir == NULL) { 1126 int i = strlen(str->text); 1127 expdir = ckmalloc(i < 2048 ? 2048 : i); /* XXX */ 1128 } 1129 1130 expmeta(expdir, str->text); 1131 ckfree(expdir); 1132 expdir = NULL; 1133 INTON; 1134 if (exparg.lastp == savelastp) { 1135 /* 1136 * no matches 1137 */ 1138 nometa: 1139 *exparg.lastp = str; 1140 rmescapes(str->text); 1141 exparg.lastp = &str->next; 1142 } else { 1143 *exparg.lastp = NULL; 1144 *savelastp = sp = expsort(*savelastp); 1145 while (sp->next != NULL) 1146 sp = sp->next; 1147 exparg.lastp = &sp->next; 1148 } 1149 str = str->next; 1150 } 1151 } 1152 1153 1154 /* 1155 * Do metacharacter (i.e. *, ?, [...]) expansion. 1156 */ 1157 1158 STATIC void 1159 expmeta(char *enddir, char *name) 1160 { 1161 char *p; 1162 const char *cp; 1163 char *q; 1164 char *start; 1165 char *endname; 1166 int metaflag; 1167 struct stat statb; 1168 DIR *dirp; 1169 struct dirent *dp; 1170 int atend; 1171 int matchdot; 1172 1173 metaflag = 0; 1174 start = name; 1175 for (p = name ; ; p++) { 1176 if (*p == '*' || *p == '?') 1177 metaflag = 1; 1178 else if (*p == '[') { 1179 q = p + 1; 1180 if (*q == '!') 1181 q++; 1182 for (;;) { 1183 while (*q == CTLQUOTEMARK) 1184 q++; 1185 if (*q == CTLESC) 1186 q++; 1187 if (*q == '/' || *q == '\0') 1188 break; 1189 if (*++q == ']') { 1190 metaflag = 1; 1191 break; 1192 } 1193 } 1194 } else if (*p == '!' && p[1] == '!' && (p == name || p[-1] == '/')) { 1195 metaflag = 1; 1196 } else if (*p == '\0') 1197 break; 1198 else if (*p == CTLQUOTEMARK) 1199 continue; 1200 else if (*p == CTLESC) 1201 p++; 1202 if (*p == '/') { 1203 if (metaflag) 1204 break; 1205 start = p + 1; 1206 } 1207 } 1208 if (metaflag == 0) { /* we've reached the end of the file name */ 1209 if (enddir != expdir) 1210 metaflag++; 1211 for (p = name ; ; p++) { 1212 if (*p == CTLQUOTEMARK) 1213 continue; 1214 if (*p == CTLESC) 1215 p++; 1216 *enddir++ = *p; 1217 if (*p == '\0') 1218 break; 1219 } 1220 if (metaflag == 0 || lstat(expdir, &statb) >= 0) 1221 addfname(expdir); 1222 return; 1223 } 1224 endname = p; 1225 if (start != name) { 1226 p = name; 1227 while (p < start) { 1228 while (*p == CTLQUOTEMARK) 1229 p++; 1230 if (*p == CTLESC) 1231 p++; 1232 *enddir++ = *p++; 1233 } 1234 } 1235 if (enddir == expdir) { 1236 cp = "."; 1237 } else if (enddir == expdir + 1 && *expdir == '/') { 1238 cp = "/"; 1239 } else { 1240 cp = expdir; 1241 enddir[-1] = '\0'; 1242 } 1243 if ((dirp = opendir(cp)) == NULL) 1244 return; 1245 if (enddir != expdir) 1246 enddir[-1] = '/'; 1247 if (*endname == 0) { 1248 atend = 1; 1249 } else { 1250 atend = 0; 1251 *endname++ = '\0'; 1252 } 1253 matchdot = 0; 1254 p = start; 1255 while (*p == CTLQUOTEMARK) 1256 p++; 1257 if (*p == CTLESC) 1258 p++; 1259 if (*p == '.') 1260 matchdot++; 1261 while (! int_pending() && (dp = readdir(dirp)) != NULL) { 1262 if (dp->d_name[0] == '.' && ! matchdot) 1263 continue; 1264 if (patmatch(start, dp->d_name, 0)) { 1265 if (atend) { 1266 scopy(dp->d_name, enddir); 1267 addfname(expdir); 1268 } else { 1269 for (p = enddir, cp = dp->d_name; 1270 (*p++ = *cp++) != '\0';) 1271 continue; 1272 p[-1] = '/'; 1273 expmeta(p, endname); 1274 } 1275 } 1276 } 1277 closedir(dirp); 1278 if (! atend) 1279 endname[-1] = '/'; 1280 } 1281 1282 1283 /* 1284 * Add a file name to the list. 1285 */ 1286 1287 STATIC void 1288 addfname(char *name) 1289 { 1290 char *p; 1291 struct strlist *sp; 1292 1293 p = stalloc(strlen(name) + 1); 1294 scopy(name, p); 1295 sp = (struct strlist *)stalloc(sizeof *sp); 1296 sp->text = p; 1297 *exparg.lastp = sp; 1298 exparg.lastp = &sp->next; 1299 } 1300 1301 1302 /* 1303 * Sort the results of file name expansion. It calculates the number of 1304 * strings to sort and then calls msort (short for merge sort) to do the 1305 * work. 1306 */ 1307 1308 STATIC struct strlist * 1309 expsort(struct strlist *str) 1310 { 1311 int len; 1312 struct strlist *sp; 1313 1314 len = 0; 1315 for (sp = str ; sp ; sp = sp->next) 1316 len++; 1317 return msort(str, len); 1318 } 1319 1320 1321 STATIC struct strlist * 1322 msort(struct strlist *list, int len) 1323 { 1324 struct strlist *p, *q = NULL; 1325 struct strlist **lpp; 1326 int half; 1327 int n; 1328 1329 if (len <= 1) 1330 return list; 1331 half = len >> 1; 1332 p = list; 1333 for (n = half ; --n >= 0 ; ) { 1334 q = p; 1335 p = p->next; 1336 } 1337 q->next = NULL; /* terminate first half of list */ 1338 q = msort(list, half); /* sort first half of list */ 1339 p = msort(p, len - half); /* sort second half */ 1340 lpp = &list; 1341 for (;;) { 1342 if (strcmp(p->text, q->text) < 0) { 1343 *lpp = p; 1344 lpp = &p->next; 1345 if ((p = *lpp) == NULL) { 1346 *lpp = q; 1347 break; 1348 } 1349 } else { 1350 *lpp = q; 1351 lpp = &q->next; 1352 if ((q = *lpp) == NULL) { 1353 *lpp = p; 1354 break; 1355 } 1356 } 1357 } 1358 return list; 1359 } 1360 1361 1362 1363 /* 1364 * Returns true if the pattern matches the string. 1365 */ 1366 1367 int 1368 patmatch(char *pattern, char *string, int squoted) 1369 { 1370 #ifdef notdef 1371 if (pattern[0] == '!' && pattern[1] == '!') 1372 return 1 - pmatch(pattern + 2, string); 1373 else 1374 #endif 1375 return pmatch(pattern, string, squoted); 1376 } 1377 1378 1379 STATIC int 1380 pmatch(char *pattern, char *string, int squoted) 1381 { 1382 char *p, *q; 1383 char c; 1384 1385 p = pattern; 1386 q = string; 1387 for (;;) { 1388 switch (c = *p++) { 1389 case '\0': 1390 goto breakloop; 1391 case CTLESC: 1392 if (squoted && *q == CTLESC) 1393 q++; 1394 if (*q++ != *p++) 1395 return 0; 1396 break; 1397 case CTLQUOTEMARK: 1398 continue; 1399 case '?': 1400 if (squoted && *q == CTLESC) 1401 q++; 1402 if (*q++ == '\0') 1403 return 0; 1404 break; 1405 case '*': 1406 c = *p; 1407 while (c == CTLQUOTEMARK || c == '*') 1408 c = *++p; 1409 if (c != CTLESC && c != CTLQUOTEMARK && 1410 c != '?' && c != '*' && c != '[') { 1411 while (*q != c) { 1412 if (squoted && *q == CTLESC && 1413 q[1] == c) 1414 break; 1415 if (*q == '\0') 1416 return 0; 1417 if (squoted && *q == CTLESC) 1418 q++; 1419 q++; 1420 } 1421 } 1422 do { 1423 if (pmatch(p, q, squoted)) 1424 return 1; 1425 if (squoted && *q == CTLESC) 1426 q++; 1427 } while (*q++ != '\0'); 1428 return 0; 1429 case '[': { 1430 char *endp; 1431 int invert, found; 1432 char chr; 1433 1434 endp = p; 1435 if (*endp == '!') 1436 endp++; 1437 for (;;) { 1438 while (*endp == CTLQUOTEMARK) 1439 endp++; 1440 if (*endp == '\0') 1441 goto dft; /* no matching ] */ 1442 if (*endp == CTLESC) 1443 endp++; 1444 if (*++endp == ']') 1445 break; 1446 } 1447 invert = 0; 1448 if (*p == '!') { 1449 invert++; 1450 p++; 1451 } 1452 found = 0; 1453 chr = *q++; 1454 if (squoted && chr == CTLESC) 1455 chr = *q++; 1456 if (chr == '\0') 1457 return 0; 1458 c = *p++; 1459 do { 1460 if (c == CTLQUOTEMARK) 1461 continue; 1462 if (c == CTLESC) 1463 c = *p++; 1464 if (*p == '-' && p[1] != ']') { 1465 p++; 1466 while (*p == CTLQUOTEMARK) 1467 p++; 1468 if (*p == CTLESC) 1469 p++; 1470 if (chr >= c && chr <= *p) 1471 found = 1; 1472 p++; 1473 } else { 1474 if (chr == c) 1475 found = 1; 1476 } 1477 } while ((c = *p++) != ']'); 1478 if (found == invert) 1479 return 0; 1480 break; 1481 } 1482 dft: default: 1483 if (squoted && *q == CTLESC) 1484 q++; 1485 if (*q++ != c) 1486 return 0; 1487 break; 1488 } 1489 } 1490 breakloop: 1491 if (*q != '\0') 1492 return 0; 1493 return 1; 1494 } 1495 1496 1497 1498 /* 1499 * Remove any CTLESC characters from a string. 1500 */ 1501 1502 void 1503 rmescapes(char *str) 1504 { 1505 char *p, *q; 1506 1507 p = str; 1508 while (*p != CTLESC && *p != CTLQUOTEMARK) { 1509 if (*p++ == '\0') 1510 return; 1511 } 1512 q = p; 1513 while (*p) { 1514 if (*p == CTLQUOTEMARK) { 1515 p++; 1516 continue; 1517 } 1518 if (*p == CTLESC) 1519 p++; 1520 *q++ = *p++; 1521 } 1522 *q = '\0'; 1523 } 1524 1525 1526 1527 /* 1528 * See if a pattern matches in a case statement. 1529 */ 1530 1531 int 1532 casematch(union node *pattern, char *val) 1533 { 1534 struct stackmark smark; 1535 int result; 1536 char *p; 1537 1538 setstackmark(&smark); 1539 argbackq = pattern->narg.backquote; 1540 STARTSTACKSTR(expdest); 1541 ifslastp = NULL; 1542 argstr(pattern->narg.text, EXP_TILDE | EXP_CASE); 1543 STPUTC('\0', expdest); 1544 p = grabstackstr(expdest); 1545 result = patmatch(p, val, 0); 1546 popstackmark(&smark); 1547 return result; 1548 } 1549 1550 /* 1551 * Our own itoa(). 1552 */ 1553 1554 STATIC char * 1555 cvtnum(int num, char *buf) 1556 { 1557 char temp[32]; 1558 int neg = num < 0; 1559 char *p = temp + 31; 1560 1561 temp[31] = '\0'; 1562 1563 do { 1564 *--p = num % 10 + '0'; 1565 } while ((num /= 10) != 0); 1566 1567 if (neg) 1568 *--p = '-'; 1569 1570 while (*p) 1571 STPUTC(*p++, buf); 1572 return buf; 1573 } 1574 1575 /* 1576 * Do most of the work for wordexp(3). 1577 */ 1578 1579 int 1580 wordexpcmd(int argc, char **argv) 1581 { 1582 size_t len; 1583 int i; 1584 1585 out1fmt("%d", argc - 1); 1586 out1c('\0'); 1587 for (i = 1, len = 0; i < argc; i++) 1588 len += strlen(argv[i]); 1589 out1fmt("%zu", len); 1590 out1c('\0'); 1591 for (i = 1; i < argc; i++) { 1592 out1str(argv[i]); 1593 out1c('\0'); 1594 } 1595 return (0); 1596 } 1597