expand.c revision 213811
1/*- 2 * Copyright (c) 1991, 1993 3 * The Regents of the University of California. All rights reserved. 4 * Copyright (c) 1997-2005 5 * Herbert Xu <herbert@gondor.apana.org.au>. 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 * 4. 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#ifndef lint 36#if 0 37static char sccsid[] = "@(#)expand.c 8.5 (Berkeley) 5/15/95"; 38#endif 39#endif /* not lint */ 40#include <sys/cdefs.h> 41__FBSDID("$FreeBSD: head/bin/sh/expand.c 213811 2010-10-13 22:18:03Z obrien $"); 42 43#include <sys/types.h> 44#include <sys/time.h> 45#include <sys/stat.h> 46#include <dirent.h> 47#include <errno.h> 48#include <inttypes.h> 49#include <limits.h> 50#include <pwd.h> 51#include <stdio.h> 52#include <stdlib.h> 53#include <string.h> 54#include <unistd.h> 55 56/* 57 * Routines to expand arguments to commands. We have to deal with 58 * backquotes, shell variables, and file metacharacters. 59 */ 60 61#include "shell.h" 62#include "main.h" 63#include "nodes.h" 64#include "eval.h" 65#include "expand.h" 66#include "syntax.h" 67#include "parser.h" 68#include "jobs.h" 69#include "options.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 "arith.h" 77#include "show.h" 78 79/* 80 * Structure specifying which parts of the string should be searched 81 * for IFS characters. 82 */ 83 84struct ifsregion { 85 struct ifsregion *next; /* next region in list */ 86 int begoff; /* offset of start of region */ 87 int endoff; /* offset of end of region */ 88 int inquotes; /* search for nul bytes only */ 89}; 90 91 92static char *expdest; /* output of current string */ 93static struct nodelist *argbackq; /* list of back quote expressions */ 94static struct ifsregion ifsfirst; /* first struct in list of ifs regions */ 95static struct ifsregion *ifslastp; /* last struct in list */ 96static struct arglist exparg; /* holds expanded arg list */ 97 98static void argstr(char *, int); 99static char *exptilde(char *, int); 100static void expbackq(union node *, int, int); 101static int subevalvar(char *, char *, int, int, int, int); 102static char *evalvar(char *, int); 103static int varisset(char *, int); 104static void varvalue(char *, int, int, int); 105static void recordregion(int, int, int); 106static void removerecordregions(int); 107static void ifsbreakup(char *, struct arglist *); 108static void expandmeta(struct strlist *, int); 109static void expmeta(char *, char *); 110static void addfname(char *); 111static struct strlist *expsort(struct strlist *); 112static struct strlist *msort(struct strlist *, int); 113static char *cvtnum(int, char *); 114static int collate_range_cmp(int, int); 115 116static int 117collate_range_cmp(int c1, int c2) 118{ 119 static char s1[2], s2[2]; 120 121 s1[0] = c1; 122 s2[0] = c2; 123 return (strcoll(s1, s2)); 124} 125 126/* 127 * Expand shell variables and backquotes inside a here document. 128 * union node *arg the document 129 * int fd; where to write the expanded version 130 */ 131 132void 133expandhere(union node *arg, int fd) 134{ 135 herefd = fd; 136 expandarg(arg, (struct arglist *)NULL, 0); 137 xwrite(fd, stackblock(), expdest - stackblock()); 138} 139 140 141/* 142 * Perform expansions on an argument, placing the resulting list of arguments 143 * in arglist. Parameter expansion, command substitution and arithmetic 144 * expansion are always performed; additional expansions can be requested 145 * via flag (EXP_*). 146 * The result is left in the stack string. 147 * When arglist is NULL, perform here document expansion. A partial result 148 * may be written to herefd, which is then not included in the stack string. 149 * 150 * Caution: this function uses global state and is not reentrant. 151 * However, a new invocation after an interrupted invocation is safe 152 * and will reset the global state for the new call. 153 */ 154void 155expandarg(union node *arg, struct arglist *arglist, int flag) 156{ 157 struct strlist *sp; 158 char *p; 159 160 argbackq = arg->narg.backquote; 161 STARTSTACKSTR(expdest); 162 ifsfirst.next = NULL; 163 ifslastp = NULL; 164 argstr(arg->narg.text, flag); 165 if (arglist == NULL) { 166 return; /* here document expanded */ 167 } 168 STPUTC('\0', expdest); 169 p = grabstackstr(expdest); 170 exparg.lastp = &exparg.list; 171 /* 172 * TODO - EXP_REDIR 173 */ 174 if (flag & EXP_FULL) { 175 ifsbreakup(p, &exparg); 176 *exparg.lastp = NULL; 177 exparg.lastp = &exparg.list; 178 expandmeta(exparg.list, flag); 179 } else { 180 if (flag & EXP_REDIR) /*XXX - for now, just remove escapes */ 181 rmescapes(p); 182 sp = (struct strlist *)stalloc(sizeof (struct strlist)); 183 sp->text = p; 184 *exparg.lastp = sp; 185 exparg.lastp = &sp->next; 186 } 187 while (ifsfirst.next != NULL) { 188 struct ifsregion *ifsp; 189 INTOFF; 190 ifsp = ifsfirst.next->next; 191 ckfree(ifsfirst.next); 192 ifsfirst.next = ifsp; 193 INTON; 194 } 195 *exparg.lastp = NULL; 196 if (exparg.list) { 197 *arglist->lastp = exparg.list; 198 arglist->lastp = exparg.lastp; 199 } 200} 201 202 203 204/* 205 * Perform parameter expansion, command substitution and arithmetic 206 * expansion, and tilde expansion if requested via EXP_TILDE/EXP_VARTILDE. 207 * Processing ends at a CTLENDVAR character as well as '\0'. 208 * This is used to expand word in ${var+word} etc. 209 * If EXP_FULL, EXP_CASE or EXP_REDIR are set, keep and/or generate CTLESC 210 * characters to allow for further processing. 211 * If EXP_FULL is set, also preserve CTLQUOTEMARK characters. 212 */ 213static void 214argstr(char *p, int flag) 215{ 216 char c; 217 int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR); /* do CTLESC */ 218 int firsteq = 1; 219 220 if (*p == '~' && (flag & (EXP_TILDE | EXP_VARTILDE))) 221 p = exptilde(p, flag); 222 for (;;) { 223 switch (c = *p++) { 224 case '\0': 225 case CTLENDVAR: 226 goto breakloop; 227 case CTLQUOTEMARK: 228 /* "$@" syntax adherence hack */ 229 if (p[0] == CTLVAR && p[2] == '@' && p[3] == '=') 230 break; 231 if ((flag & EXP_FULL) != 0) 232 STPUTC(c, expdest); 233 break; 234 case CTLESC: 235 if (quotes) 236 STPUTC(c, expdest); 237 c = *p++; 238 STPUTC(c, expdest); 239 break; 240 case CTLVAR: 241 p = evalvar(p, flag); 242 break; 243 case CTLBACKQ: 244 case CTLBACKQ|CTLQUOTE: 245 expbackq(argbackq->n, c & CTLQUOTE, flag); 246 argbackq = argbackq->next; 247 break; 248 case CTLENDARI: 249 expari(flag); 250 break; 251 case ':': 252 case '=': 253 /* 254 * sort of a hack - expand tildes in variable 255 * assignments (after the first '=' and after ':'s). 256 */ 257 STPUTC(c, expdest); 258 if (flag & EXP_VARTILDE && *p == '~') { 259 if (c == '=') { 260 if (firsteq) 261 firsteq = 0; 262 else 263 break; 264 } 265 p = exptilde(p, flag); 266 } 267 break; 268 default: 269 STPUTC(c, expdest); 270 } 271 } 272breakloop:; 273} 274 275/* 276 * Perform tilde expansion, placing the result in the stack string and 277 * returning the next position in the input string to process. 278 */ 279static char * 280exptilde(char *p, int flag) 281{ 282 char c, *startp = p; 283 struct passwd *pw; 284 char *home; 285 int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR); 286 287 while ((c = *p) != '\0') { 288 switch(c) { 289 case CTLESC: /* This means CTL* are always considered quoted. */ 290 case CTLVAR: 291 case CTLBACKQ: 292 case CTLBACKQ | CTLQUOTE: 293 case CTLARI: 294 case CTLENDARI: 295 case CTLQUOTEMARK: 296 return (startp); 297 case ':': 298 if (flag & EXP_VARTILDE) 299 goto done; 300 break; 301 case '/': 302 case CTLENDVAR: 303 goto done; 304 } 305 p++; 306 } 307done: 308 *p = '\0'; 309 if (*(startp+1) == '\0') { 310 if ((home = lookupvar("HOME")) == NULL) 311 goto lose; 312 } else { 313 if ((pw = getpwnam(startp+1)) == NULL) 314 goto lose; 315 home = pw->pw_dir; 316 } 317 if (*home == '\0') 318 goto lose; 319 *p = c; 320 while ((c = *home++) != '\0') { 321 if (quotes && SQSYNTAX[(int)c] == CCTL) 322 STPUTC(CTLESC, expdest); 323 STPUTC(c, expdest); 324 } 325 return (p); 326lose: 327 *p = c; 328 return (startp); 329} 330 331 332static void 333removerecordregions(int endoff) 334{ 335 if (ifslastp == NULL) 336 return; 337 338 if (ifsfirst.endoff > endoff) { 339 while (ifsfirst.next != NULL) { 340 struct ifsregion *ifsp; 341 INTOFF; 342 ifsp = ifsfirst.next->next; 343 ckfree(ifsfirst.next); 344 ifsfirst.next = ifsp; 345 INTON; 346 } 347 if (ifsfirst.begoff > endoff) 348 ifslastp = NULL; 349 else { 350 ifslastp = &ifsfirst; 351 ifsfirst.endoff = endoff; 352 } 353 return; 354 } 355 356 ifslastp = &ifsfirst; 357 while (ifslastp->next && ifslastp->next->begoff < endoff) 358 ifslastp=ifslastp->next; 359 while (ifslastp->next != NULL) { 360 struct ifsregion *ifsp; 361 INTOFF; 362 ifsp = ifslastp->next->next; 363 ckfree(ifslastp->next); 364 ifslastp->next = ifsp; 365 INTON; 366 } 367 if (ifslastp->endoff > endoff) 368 ifslastp->endoff = endoff; 369} 370 371/* 372 * Expand arithmetic expression. Backup to start of expression, 373 * evaluate, place result in (backed up) result, adjust string position. 374 */ 375void 376expari(int flag) 377{ 378 char *p, *q, *start; 379 arith_t result; 380 int begoff; 381 int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR); 382 int quoted; 383 384 /* 385 * This routine is slightly over-complicated for 386 * efficiency. First we make sure there is 387 * enough space for the result, which may be bigger 388 * than the expression. Next we 389 * scan backwards looking for the start of arithmetic. If the 390 * next previous character is a CTLESC character, then we 391 * have to rescan starting from the beginning since CTLESC 392 * characters have to be processed left to right. 393 */ 394 CHECKSTRSPACE(DIGITS(result) - 2, expdest); 395 USTPUTC('\0', expdest); 396 start = stackblock(); 397 p = expdest - 2; 398 while (p >= start && *p != CTLARI) 399 --p; 400 if (p < start || *p != CTLARI) 401 error("missing CTLARI (shouldn't happen)"); 402 if (p > start && *(p - 1) == CTLESC) 403 for (p = start; *p != CTLARI; p++) 404 if (*p == CTLESC) 405 p++; 406 407 if (p[1] == '"') 408 quoted=1; 409 else 410 quoted=0; 411 begoff = p - start; 412 removerecordregions(begoff); 413 if (quotes) 414 rmescapes(p+2); 415 q = grabstackstr(expdest); 416 result = arith(p+2); 417 ungrabstackstr(q, expdest); 418 fmtstr(p, DIGITS(result), ARITH_FORMAT_STR, result); 419 while (*p++) 420 ; 421 if (quoted == 0) 422 recordregion(begoff, p - 1 - start, 0); 423 result = expdest - p + 1; 424 STADJUST(-result, expdest); 425} 426 427 428/* 429 * Perform command substitution. 430 */ 431static void 432expbackq(union node *cmd, int quoted, int flag) 433{ 434 struct backcmd in; 435 int i; 436 char buf[128]; 437 char *p; 438 char *dest = expdest; 439 struct ifsregion saveifs, *savelastp; 440 struct nodelist *saveargbackq; 441 char lastc; 442 int startloc = dest - stackblock(); 443 char const *syntax = quoted? DQSYNTAX : BASESYNTAX; 444 int saveherefd; 445 int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR); 446 int nnl; 447 448 INTOFF; 449 saveifs = ifsfirst; 450 savelastp = ifslastp; 451 saveargbackq = argbackq; 452 saveherefd = herefd; 453 herefd = -1; 454 p = grabstackstr(dest); 455 evalbackcmd(cmd, &in); 456 ungrabstackstr(p, dest); 457 ifsfirst = saveifs; 458 ifslastp = savelastp; 459 argbackq = saveargbackq; 460 herefd = saveherefd; 461 462 p = in.buf; 463 lastc = '\0'; 464 nnl = 0; 465 /* Don't copy trailing newlines */ 466 for (;;) { 467 if (--in.nleft < 0) { 468 if (in.fd < 0) 469 break; 470 while ((i = read(in.fd, buf, sizeof buf)) < 0 && errno == EINTR); 471 TRACE(("expbackq: read returns %d\n", i)); 472 if (i <= 0) 473 break; 474 p = buf; 475 in.nleft = i - 1; 476 } 477 lastc = *p++; 478 if (lastc != '\0') { 479 if (quotes && syntax[(int)lastc] == CCTL) 480 STPUTC(CTLESC, dest); 481 if (lastc == '\n') { 482 nnl++; 483 } else { 484 while (nnl > 0) { 485 nnl--; 486 STPUTC('\n', dest); 487 } 488 STPUTC(lastc, dest); 489 } 490 } 491 } 492 493 if (in.fd >= 0) 494 close(in.fd); 495 if (in.buf) 496 ckfree(in.buf); 497 if (in.jp) 498 exitstatus = waitforjob(in.jp, (int *)NULL); 499 if (quoted == 0) 500 recordregion(startloc, dest - stackblock(), 0); 501 TRACE(("expbackq: size=%td: \"%.*s\"\n", 502 ((dest - stackblock()) - startloc), 503 (int)((dest - stackblock()) - startloc), 504 stackblock() + startloc)); 505 expdest = dest; 506 INTON; 507} 508 509 510 511static int 512subevalvar(char *p, char *str, int strloc, int subtype, int startloc, 513 int varflags) 514{ 515 char *startp; 516 char *loc = NULL; 517 char *q; 518 int c = 0; 519 int saveherefd = herefd; 520 struct nodelist *saveargbackq = argbackq; 521 int amount; 522 523 herefd = -1; 524 argstr(p, (subtype == VSTRIMLEFT || subtype == VSTRIMLEFTMAX || 525 subtype == VSTRIMRIGHT || subtype == VSTRIMRIGHTMAX ? 526 EXP_CASE : 0) | EXP_TILDE); 527 STACKSTRNUL(expdest); 528 herefd = saveherefd; 529 argbackq = saveargbackq; 530 startp = stackblock() + startloc; 531 if (str == NULL) 532 str = stackblock() + strloc; 533 534 switch (subtype) { 535 case VSASSIGN: 536 setvar(str, startp, 0); 537 amount = startp - expdest; 538 STADJUST(amount, expdest); 539 varflags &= ~VSNUL; 540 if (c != 0) 541 *loc = c; 542 return 1; 543 544 case VSQUESTION: 545 if (*p != CTLENDVAR) { 546 outfmt(out2, "%s\n", startp); 547 error((char *)NULL); 548 } 549 error("%.*s: parameter %snot set", (int)(p - str - 1), 550 str, (varflags & VSNUL) ? "null or " 551 : nullstr); 552 return 0; 553 554 case VSTRIMLEFT: 555 for (loc = startp; loc < str; loc++) { 556 c = *loc; 557 *loc = '\0'; 558 if (patmatch(str, startp, varflags & VSQUOTE)) { 559 *loc = c; 560 goto recordleft; 561 } 562 *loc = c; 563 if ((varflags & VSQUOTE) && *loc == CTLESC) 564 loc++; 565 } 566 return 0; 567 568 case VSTRIMLEFTMAX: 569 for (loc = str - 1; loc >= startp;) { 570 c = *loc; 571 *loc = '\0'; 572 if (patmatch(str, startp, varflags & VSQUOTE)) { 573 *loc = c; 574 goto recordleft; 575 } 576 *loc = c; 577 loc--; 578 if ((varflags & VSQUOTE) && loc > startp && 579 *(loc - 1) == CTLESC) { 580 for (q = startp; q < loc; q++) 581 if (*q == CTLESC) 582 q++; 583 if (q > loc) 584 loc--; 585 } 586 } 587 return 0; 588 589 case VSTRIMRIGHT: 590 for (loc = str - 1; loc >= startp;) { 591 if (patmatch(str, loc, varflags & VSQUOTE)) { 592 amount = loc - expdest; 593 STADJUST(amount, expdest); 594 return 1; 595 } 596 loc--; 597 if ((varflags & VSQUOTE) && loc > startp && 598 *(loc - 1) == CTLESC) { 599 for (q = startp; q < loc; q++) 600 if (*q == CTLESC) 601 q++; 602 if (q > loc) 603 loc--; 604 } 605 } 606 return 0; 607 608 case VSTRIMRIGHTMAX: 609 for (loc = startp; loc < str - 1; loc++) { 610 if (patmatch(str, loc, varflags & VSQUOTE)) { 611 amount = loc - expdest; 612 STADJUST(amount, expdest); 613 return 1; 614 } 615 if ((varflags & VSQUOTE) && *loc == CTLESC) 616 loc++; 617 } 618 return 0; 619 620 621 default: 622 abort(); 623 } 624 625recordleft: 626 amount = ((str - 1) - (loc - startp)) - expdest; 627 STADJUST(amount, expdest); 628 while (loc != str - 1) 629 *startp++ = *loc++; 630 return 1; 631} 632 633 634/* 635 * Expand a variable, and return a pointer to the next character in the 636 * input string. 637 */ 638 639static char * 640evalvar(char *p, int flag) 641{ 642 int subtype; 643 int varflags; 644 char *var; 645 char *val; 646 int patloc; 647 int c; 648 int set; 649 int special; 650 int startloc; 651 int varlen; 652 int easy; 653 int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR); 654 655 varflags = (unsigned char)*p++; 656 subtype = varflags & VSTYPE; 657 var = p; 658 special = 0; 659 if (! is_name(*p)) 660 special = 1; 661 p = strchr(p, '=') + 1; 662again: /* jump here after setting a variable with ${var=text} */ 663 if (varflags & VSLINENO) { 664 set = 1; 665 special = 0; 666 val = var; 667 p[-1] = '\0'; /* temporarily overwrite '=' to have \0 668 terminated string */ 669 } else if (special) { 670 set = varisset(var, varflags & VSNUL); 671 val = NULL; 672 } else { 673 val = bltinlookup(var, 1); 674 if (val == NULL || ((varflags & VSNUL) && val[0] == '\0')) { 675 val = NULL; 676 set = 0; 677 } else 678 set = 1; 679 } 680 varlen = 0; 681 startloc = expdest - stackblock(); 682 if (!set && uflag && *var != '@' && *var != '*') { 683 switch (subtype) { 684 case VSNORMAL: 685 case VSTRIMLEFT: 686 case VSTRIMLEFTMAX: 687 case VSTRIMRIGHT: 688 case VSTRIMRIGHTMAX: 689 case VSLENGTH: 690 error("%.*s: parameter not set", (int)(p - var - 1), 691 var); 692 } 693 } 694 if (set && subtype != VSPLUS) { 695 /* insert the value of the variable */ 696 if (special) { 697 varvalue(var, varflags & VSQUOTE, subtype, flag); 698 if (subtype == VSLENGTH) { 699 varlen = expdest - stackblock() - startloc; 700 STADJUST(-varlen, expdest); 701 } 702 } else { 703 char const *syntax = (varflags & VSQUOTE) ? DQSYNTAX 704 : BASESYNTAX; 705 706 if (subtype == VSLENGTH) { 707 for (;*val; val++) 708 varlen++; 709 } 710 else { 711 while (*val) { 712 if (quotes && 713 syntax[(int)*val] == CCTL) 714 STPUTC(CTLESC, expdest); 715 STPUTC(*val++, expdest); 716 } 717 718 } 719 } 720 } 721 722 if (subtype == VSPLUS) 723 set = ! set; 724 725 easy = ((varflags & VSQUOTE) == 0 || 726 (*var == '@' && shellparam.nparam != 1)); 727 728 729 switch (subtype) { 730 case VSLENGTH: 731 expdest = cvtnum(varlen, expdest); 732 goto record; 733 734 case VSNORMAL: 735 if (!easy) 736 break; 737record: 738 recordregion(startloc, expdest - stackblock(), 739 varflags & VSQUOTE); 740 break; 741 742 case VSPLUS: 743 case VSMINUS: 744 if (!set) { 745 argstr(p, flag); 746 break; 747 } 748 if (easy) 749 goto record; 750 break; 751 752 case VSTRIMLEFT: 753 case VSTRIMLEFTMAX: 754 case VSTRIMRIGHT: 755 case VSTRIMRIGHTMAX: 756 if (!set) 757 break; 758 /* 759 * Terminate the string and start recording the pattern 760 * right after it 761 */ 762 STPUTC('\0', expdest); 763 patloc = expdest - stackblock(); 764 if (subevalvar(p, NULL, patloc, subtype, 765 startloc, varflags) == 0) { 766 int amount = (expdest - stackblock() - patloc) + 1; 767 STADJUST(-amount, expdest); 768 } 769 /* Remove any recorded regions beyond start of variable */ 770 removerecordregions(startloc); 771 goto record; 772 773 case VSASSIGN: 774 case VSQUESTION: 775 if (!set) { 776 if (subevalvar(p, var, 0, subtype, startloc, varflags)) { 777 varflags &= ~VSNUL; 778 /* 779 * Remove any recorded regions beyond 780 * start of variable 781 */ 782 removerecordregions(startloc); 783 goto again; 784 } 785 break; 786 } 787 if (easy) 788 goto record; 789 break; 790 791 case VSERROR: 792 c = p - var - 1; 793 error("${%.*s%s}: Bad substitution", c, var, 794 (c > 0 && *p != CTLENDVAR) ? "..." : ""); 795 796 default: 797 abort(); 798 } 799 p[-1] = '='; /* recover overwritten '=' */ 800 801 if (subtype != VSNORMAL) { /* skip to end of alternative */ 802 int nesting = 1; 803 for (;;) { 804 if ((c = *p++) == CTLESC) 805 p++; 806 else if (c == CTLBACKQ || c == (CTLBACKQ|CTLQUOTE)) { 807 if (set) 808 argbackq = argbackq->next; 809 } else if (c == CTLVAR) { 810 if ((*p++ & VSTYPE) != VSNORMAL) 811 nesting++; 812 } else if (c == CTLENDVAR) { 813 if (--nesting == 0) 814 break; 815 } 816 } 817 } 818 return p; 819} 820 821 822 823/* 824 * Test whether a specialized variable is set. 825 */ 826 827static int 828varisset(char *name, int nulok) 829{ 830 831 if (*name == '!') 832 return backgndpidset(); 833 else if (*name == '@' || *name == '*') { 834 if (*shellparam.p == NULL) 835 return 0; 836 837 if (nulok) { 838 char **av; 839 840 for (av = shellparam.p; *av; av++) 841 if (**av != '\0') 842 return 1; 843 return 0; 844 } 845 } else if (is_digit(*name)) { 846 char *ap; 847 int num = atoi(name); 848 849 if (num > shellparam.nparam) 850 return 0; 851 852 if (num == 0) 853 ap = arg0; 854 else 855 ap = shellparam.p[num - 1]; 856 857 if (nulok && (ap == NULL || *ap == '\0')) 858 return 0; 859 } 860 return 1; 861} 862 863 864 865/* 866 * Add the value of a specialized variable to the stack string. 867 */ 868 869static void 870varvalue(char *name, int quoted, int subtype, int flag) 871{ 872 int num; 873 char *p; 874 int i; 875 char sep; 876 char **ap; 877 char const *syntax; 878 879#define STRTODEST(p) \ 880 do {\ 881 if (flag & (EXP_FULL | EXP_CASE) && subtype != VSLENGTH) { \ 882 syntax = quoted? DQSYNTAX : BASESYNTAX; \ 883 while (*p) { \ 884 if (syntax[(int)*p] == CCTL) \ 885 STPUTC(CTLESC, expdest); \ 886 STPUTC(*p++, expdest); \ 887 } \ 888 } else \ 889 while (*p) \ 890 STPUTC(*p++, expdest); \ 891 } while (0) 892 893 894 switch (*name) { 895 case '$': 896 num = rootpid; 897 goto numvar; 898 case '?': 899 num = oexitstatus; 900 goto numvar; 901 case '#': 902 num = shellparam.nparam; 903 goto numvar; 904 case '!': 905 num = backgndpidval(); 906numvar: 907 expdest = cvtnum(num, expdest); 908 break; 909 case '-': 910 for (i = 0 ; i < NOPTS ; i++) { 911 if (optlist[i].val) 912 STPUTC(optlist[i].letter, expdest); 913 } 914 break; 915 case '@': 916 if (flag & EXP_FULL && quoted) { 917 for (ap = shellparam.p ; (p = *ap++) != NULL ; ) { 918 STRTODEST(p); 919 if (*ap) 920 STPUTC('\0', expdest); 921 } 922 break; 923 } 924 /* FALLTHROUGH */ 925 case '*': 926 if (ifsset()) 927 sep = ifsval()[0]; 928 else 929 sep = ' '; 930 for (ap = shellparam.p ; (p = *ap++) != NULL ; ) { 931 STRTODEST(p); 932 if (*ap && sep) 933 STPUTC(sep, expdest); 934 } 935 break; 936 case '0': 937 p = arg0; 938 STRTODEST(p); 939 break; 940 default: 941 if (is_digit(*name)) { 942 num = atoi(name); 943 if (num > 0 && num <= shellparam.nparam) { 944 p = shellparam.p[num - 1]; 945 STRTODEST(p); 946 } 947 } 948 break; 949 } 950} 951 952 953 954/* 955 * Record the the fact that we have to scan this region of the 956 * string for IFS characters. 957 */ 958 959static void 960recordregion(int start, int end, int inquotes) 961{ 962 struct ifsregion *ifsp; 963 964 if (ifslastp == NULL) { 965 ifsp = &ifsfirst; 966 } else { 967 if (ifslastp->endoff == start 968 && ifslastp->inquotes == inquotes) { 969 /* extend previous area */ 970 ifslastp->endoff = end; 971 return; 972 } 973 ifsp = (struct ifsregion *)ckmalloc(sizeof (struct ifsregion)); 974 ifslastp->next = ifsp; 975 } 976 ifslastp = ifsp; 977 ifslastp->next = NULL; 978 ifslastp->begoff = start; 979 ifslastp->endoff = end; 980 ifslastp->inquotes = inquotes; 981} 982 983 984 985/* 986 * Break the argument string into pieces based upon IFS and add the 987 * strings to the argument list. The regions of the string to be 988 * searched for IFS characters have been stored by recordregion. 989 * CTLESC characters are preserved but have little effect in this pass 990 * other than escaping CTL* characters. In particular, they do not escape 991 * IFS characters: that should be done with the ifsregion mechanism. 992 * CTLQUOTEMARK characters are used to preserve empty quoted strings. 993 * This pass treats them as a regular character, making the string non-empty. 994 * Later, they are removed along with the other CTL* characters. 995 */ 996static void 997ifsbreakup(char *string, struct arglist *arglist) 998{ 999 struct ifsregion *ifsp; 1000 struct strlist *sp; 1001 char *start; 1002 char *p; 1003 char *q; 1004 const char *ifs; 1005 const char *ifsspc; 1006 int had_param_ch = 0; 1007 1008 start = string; 1009 1010 if (ifslastp == NULL) { 1011 /* Return entire argument, IFS doesn't apply to any of it */ 1012 sp = (struct strlist *)stalloc(sizeof *sp); 1013 sp->text = start; 1014 *arglist->lastp = sp; 1015 arglist->lastp = &sp->next; 1016 return; 1017 } 1018 1019 ifs = ifsset() ? ifsval() : " \t\n"; 1020 1021 for (ifsp = &ifsfirst; ifsp != NULL; ifsp = ifsp->next) { 1022 p = string + ifsp->begoff; 1023 while (p < string + ifsp->endoff) { 1024 q = p; 1025 if (*p == CTLESC) 1026 p++; 1027 if (ifsp->inquotes) { 1028 /* Only NULs (should be from "$@") end args */ 1029 had_param_ch = 1; 1030 if (*p != 0) { 1031 p++; 1032 continue; 1033 } 1034 ifsspc = NULL; 1035 } else { 1036 if (!strchr(ifs, *p)) { 1037 had_param_ch = 1; 1038 p++; 1039 continue; 1040 } 1041 ifsspc = strchr(" \t\n", *p); 1042 1043 /* Ignore IFS whitespace at start */ 1044 if (q == start && ifsspc != NULL) { 1045 p++; 1046 start = p; 1047 continue; 1048 } 1049 had_param_ch = 0; 1050 } 1051 1052 /* Save this argument... */ 1053 *q = '\0'; 1054 sp = (struct strlist *)stalloc(sizeof *sp); 1055 sp->text = start; 1056 *arglist->lastp = sp; 1057 arglist->lastp = &sp->next; 1058 p++; 1059 1060 if (ifsspc != NULL) { 1061 /* Ignore further trailing IFS whitespace */ 1062 for (; p < string + ifsp->endoff; p++) { 1063 q = p; 1064 if (*p == CTLESC) 1065 p++; 1066 if (strchr(ifs, *p) == NULL) { 1067 p = q; 1068 break; 1069 } 1070 if (strchr(" \t\n", *p) == NULL) { 1071 p++; 1072 break; 1073 } 1074 } 1075 } 1076 start = p; 1077 } 1078 } 1079 1080 /* 1081 * Save anything left as an argument. 1082 * Traditionally we have treated 'IFS=':'; set -- x$IFS' as 1083 * generating 2 arguments, the second of which is empty. 1084 * Some recent clarification of the Posix spec say that it 1085 * should only generate one.... 1086 */ 1087 if (had_param_ch || *start != 0) { 1088 sp = (struct strlist *)stalloc(sizeof *sp); 1089 sp->text = start; 1090 *arglist->lastp = sp; 1091 arglist->lastp = &sp->next; 1092 } 1093} 1094 1095 1096static char expdir[PATH_MAX]; 1097#define expdir_end (expdir + sizeof(expdir)) 1098 1099/* 1100 * Perform pathname generation and remove control characters. 1101 * At this point, the only control characters should be CTLESC and CTLQUOTEMARK. 1102 * The results are stored in the list exparg. 1103 */ 1104static void 1105expandmeta(struct strlist *str, int flag __unused) 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 == '[') 1121 break; 1122 } 1123 savelastp = exparg.lastp; 1124 INTOFF; 1125 expmeta(expdir, str->text); 1126 INTON; 1127 if (exparg.lastp == savelastp) { 1128 /* 1129 * no matches 1130 */ 1131nometa: 1132 *exparg.lastp = str; 1133 rmescapes(str->text); 1134 exparg.lastp = &str->next; 1135 } else { 1136 *exparg.lastp = NULL; 1137 *savelastp = sp = expsort(*savelastp); 1138 while (sp->next != NULL) 1139 sp = sp->next; 1140 exparg.lastp = &sp->next; 1141 } 1142 str = str->next; 1143 } 1144} 1145 1146 1147/* 1148 * Do metacharacter (i.e. *, ?, [...]) expansion. 1149 */ 1150 1151static void 1152expmeta(char *enddir, char *name) 1153{ 1154 char *p; 1155 char *q; 1156 char *start; 1157 char *endname; 1158 int metaflag; 1159 struct stat statb; 1160 DIR *dirp; 1161 struct dirent *dp; 1162 int atend; 1163 int matchdot; 1164 int esc; 1165 1166 metaflag = 0; 1167 start = name; 1168 for (p = name; esc = 0, *p; p += esc + 1) { 1169 if (*p == '*' || *p == '?') 1170 metaflag = 1; 1171 else if (*p == '[') { 1172 q = p + 1; 1173 if (*q == '!' || *q == '^') 1174 q++; 1175 for (;;) { 1176 while (*q == CTLQUOTEMARK) 1177 q++; 1178 if (*q == CTLESC) 1179 q++; 1180 if (*q == '/' || *q == '\0') 1181 break; 1182 if (*++q == ']') { 1183 metaflag = 1; 1184 break; 1185 } 1186 } 1187 } else if (*p == '\0') 1188 break; 1189 else if (*p == CTLQUOTEMARK) 1190 continue; 1191 else { 1192 if (*p == CTLESC) 1193 esc++; 1194 if (p[esc] == '/') { 1195 if (metaflag) 1196 break; 1197 start = p + esc + 1; 1198 } 1199 } 1200 } 1201 if (metaflag == 0) { /* we've reached the end of the file name */ 1202 if (enddir != expdir) 1203 metaflag++; 1204 for (p = name ; ; p++) { 1205 if (*p == CTLQUOTEMARK) 1206 continue; 1207 if (*p == CTLESC) 1208 p++; 1209 *enddir++ = *p; 1210 if (*p == '\0') 1211 break; 1212 if (enddir == expdir_end) 1213 return; 1214 } 1215 if (metaflag == 0 || lstat(expdir, &statb) >= 0) 1216 addfname(expdir); 1217 return; 1218 } 1219 endname = p; 1220 if (start != name) { 1221 p = name; 1222 while (p < start) { 1223 while (*p == CTLQUOTEMARK) 1224 p++; 1225 if (*p == CTLESC) 1226 p++; 1227 *enddir++ = *p++; 1228 if (enddir == expdir_end) 1229 return; 1230 } 1231 } 1232 if (enddir == expdir) { 1233 p = "."; 1234 } else if (enddir == expdir + 1 && *expdir == '/') { 1235 p = "/"; 1236 } else { 1237 p = expdir; 1238 enddir[-1] = '\0'; 1239 } 1240 if ((dirp = opendir(p)) == NULL) 1241 return; 1242 if (enddir != expdir) 1243 enddir[-1] = '/'; 1244 if (*endname == 0) { 1245 atend = 1; 1246 } else { 1247 atend = 0; 1248 *endname = '\0'; 1249 endname += esc + 1; 1250 } 1251 matchdot = 0; 1252 p = start; 1253 while (*p == CTLQUOTEMARK) 1254 p++; 1255 if (*p == CTLESC) 1256 p++; 1257 if (*p == '.') 1258 matchdot++; 1259 while (! int_pending() && (dp = readdir(dirp)) != NULL) { 1260 if (dp->d_name[0] == '.' && ! matchdot) 1261 continue; 1262 if (patmatch(start, dp->d_name, 0)) { 1263 if (enddir + dp->d_namlen + 1 > expdir_end) 1264 continue; 1265 memcpy(enddir, dp->d_name, dp->d_namlen + 1); 1266 if (atend) 1267 addfname(expdir); 1268 else { 1269 if (enddir + dp->d_namlen + 2 > expdir_end) 1270 continue; 1271 enddir[dp->d_namlen] = '/'; 1272 enddir[dp->d_namlen + 1] = '\0'; 1273 expmeta(enddir + dp->d_namlen + 1, endname); 1274 } 1275 } 1276 } 1277 closedir(dirp); 1278 if (! atend) 1279 endname[-esc - 1] = esc ? CTLESC : '/'; 1280} 1281 1282 1283/* 1284 * Add a file name to the list. 1285 */ 1286 1287static void 1288addfname(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 1308static struct strlist * 1309expsort(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 1321static struct strlist * 1322msort(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 1367int 1368patmatch(const char *pattern, const char *string, int squoted) 1369{ 1370 const char *p, *q; 1371 char c; 1372 1373 p = pattern; 1374 q = string; 1375 for (;;) { 1376 switch (c = *p++) { 1377 case '\0': 1378 goto breakloop; 1379 case CTLESC: 1380 if (squoted && *q == CTLESC) 1381 q++; 1382 if (*q++ != *p++) 1383 return 0; 1384 break; 1385 case CTLQUOTEMARK: 1386 continue; 1387 case '?': 1388 if (squoted && *q == CTLESC) 1389 q++; 1390 if (*q++ == '\0') 1391 return 0; 1392 break; 1393 case '*': 1394 c = *p; 1395 while (c == CTLQUOTEMARK || c == '*') 1396 c = *++p; 1397 if (c != CTLESC && c != CTLQUOTEMARK && 1398 c != '?' && c != '*' && c != '[') { 1399 while (*q != c) { 1400 if (squoted && *q == CTLESC && 1401 q[1] == c) 1402 break; 1403 if (*q == '\0') 1404 return 0; 1405 if (squoted && *q == CTLESC) 1406 q++; 1407 q++; 1408 } 1409 } 1410 do { 1411 if (patmatch(p, q, squoted)) 1412 return 1; 1413 if (squoted && *q == CTLESC) 1414 q++; 1415 } while (*q++ != '\0'); 1416 return 0; 1417 case '[': { 1418 const char *endp; 1419 int invert, found; 1420 char chr; 1421 1422 endp = p; 1423 if (*endp == '!' || *endp == '^') 1424 endp++; 1425 for (;;) { 1426 while (*endp == CTLQUOTEMARK) 1427 endp++; 1428 if (*endp == '\0') 1429 goto dft; /* no matching ] */ 1430 if (*endp == CTLESC) 1431 endp++; 1432 if (*++endp == ']') 1433 break; 1434 } 1435 invert = 0; 1436 if (*p == '!' || *p == '^') { 1437 invert++; 1438 p++; 1439 } 1440 found = 0; 1441 chr = *q++; 1442 if (squoted && chr == CTLESC) 1443 chr = *q++; 1444 if (chr == '\0') 1445 return 0; 1446 c = *p++; 1447 do { 1448 if (c == CTLQUOTEMARK) 1449 continue; 1450 if (c == CTLESC) 1451 c = *p++; 1452 if (*p == '-' && p[1] != ']') { 1453 p++; 1454 while (*p == CTLQUOTEMARK) 1455 p++; 1456 if (*p == CTLESC) 1457 p++; 1458 if ( collate_range_cmp(chr, c) >= 0 1459 && collate_range_cmp(chr, *p) <= 0 1460 ) 1461 found = 1; 1462 p++; 1463 } else { 1464 if (chr == c) 1465 found = 1; 1466 } 1467 } while ((c = *p++) != ']'); 1468 if (found == invert) 1469 return 0; 1470 break; 1471 } 1472dft: default: 1473 if (squoted && *q == CTLESC) 1474 q++; 1475 if (*q++ != c) 1476 return 0; 1477 break; 1478 } 1479 } 1480breakloop: 1481 if (*q != '\0') 1482 return 0; 1483 return 1; 1484} 1485 1486 1487 1488/* 1489 * Remove any CTLESC and CTLQUOTEMARK characters from a string. 1490 */ 1491 1492void 1493rmescapes(char *str) 1494{ 1495 char *p, *q; 1496 1497 p = str; 1498 while (*p != CTLESC && *p != CTLQUOTEMARK) { 1499 if (*p++ == '\0') 1500 return; 1501 } 1502 q = p; 1503 while (*p) { 1504 if (*p == CTLQUOTEMARK) { 1505 p++; 1506 continue; 1507 } 1508 if (*p == CTLESC) 1509 p++; 1510 *q++ = *p++; 1511 } 1512 *q = '\0'; 1513} 1514 1515 1516 1517/* 1518 * See if a pattern matches in a case statement. 1519 */ 1520 1521int 1522casematch(union node *pattern, const char *val) 1523{ 1524 struct stackmark smark; 1525 int result; 1526 char *p; 1527 1528 setstackmark(&smark); 1529 argbackq = pattern->narg.backquote; 1530 STARTSTACKSTR(expdest); 1531 ifslastp = NULL; 1532 argstr(pattern->narg.text, EXP_TILDE | EXP_CASE); 1533 STPUTC('\0', expdest); 1534 p = grabstackstr(expdest); 1535 result = patmatch(p, val, 0); 1536 popstackmark(&smark); 1537 return result; 1538} 1539 1540/* 1541 * Our own itoa(). 1542 */ 1543 1544static char * 1545cvtnum(int num, char *buf) 1546{ 1547 char temp[32]; 1548 int neg = num < 0; 1549 char *p = temp + 31; 1550 1551 temp[31] = '\0'; 1552 1553 do { 1554 *--p = num % 10 + '0'; 1555 } while ((num /= 10) != 0); 1556 1557 if (neg) 1558 *--p = '-'; 1559 1560 while (*p) 1561 STPUTC(*p++, buf); 1562 return buf; 1563} 1564 1565/* 1566 * Do most of the work for wordexp(3). 1567 */ 1568 1569int 1570wordexpcmd(int argc, char **argv) 1571{ 1572 size_t len; 1573 int i; 1574 1575 out1fmt("%08x", argc - 1); 1576 for (i = 1, len = 0; i < argc; i++) 1577 len += strlen(argv[i]); 1578 out1fmt("%08x", (int)len); 1579 for (i = 1; i < argc; i++) { 1580 out1str(argv[i]); 1581 out1c('\0'); 1582 } 1583 return (0); 1584} 1585