expand.c revision 215783 
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 215783 2010-11-23 22:17:39Z jilles $");
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, 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 ${var+word} etc.
209 * If EXP_FULL, EXP_CASE or EXP_REDIR are set, keep and/or generate CTLESC
210 * characters to allow for further processing.
211 * If EXP_FULL is set, also preserve CTLQUOTEMARK characters.
212 */
213static void
214argstr(char *p, int flag)
215{
216	char c;
217	int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR);	/* do CTLESC */
218	int firsteq = 1;
219	int split_lit;
220	int lit_quoted;
221
222	split_lit = flag & EXP_SPLIT_LIT;
223	lit_quoted = flag & EXP_LIT_QUOTED;
224	flag &= ~(EXP_SPLIT_LIT | EXP_LIT_QUOTED);
225	if (*p == '~' && (flag & (EXP_TILDE | EXP_VARTILDE)))
226		p = exptilde(p, flag);
227	for (;;) {
228		CHECKSTRSPACE(2, expdest);
229		switch (c = *p++) {
230		case '\0':
231		case CTLENDVAR:
232			goto breakloop;
233		case CTLQUOTEMARK:
234			lit_quoted = 1;
235			/* "$@" syntax adherence hack */
236			if (p[0] == CTLVAR && p[2] == '@' && p[3] == '=')
237				break;
238			if ((flag & EXP_FULL) != 0)
239				USTPUTC(c, expdest);
240			break;
241		case CTLQUOTEEND:
242			lit_quoted = 0;
243			break;
244		case CTLESC:
245			if (quotes)
246				USTPUTC(c, expdest);
247			c = *p++;
248			USTPUTC(c, expdest);
249			if (split_lit && !lit_quoted)
250				recordregion(expdest - stackblock() -
251				    (quotes ? 2 : 1),
252				    expdest - stackblock(), 0);
253			break;
254		case CTLVAR:
255			p = evalvar(p, flag);
256			break;
257		case CTLBACKQ:
258		case CTLBACKQ|CTLQUOTE:
259			expbackq(argbackq->n, c & CTLQUOTE, flag);
260			argbackq = argbackq->next;
261			break;
262		case CTLENDARI:
263			expari(flag);
264			break;
265		case ':':
266		case '=':
267			/*
268			 * sort of a hack - expand tildes in variable
269			 * assignments (after the first '=' and after ':'s).
270			 */
271			USTPUTC(c, expdest);
272			if (split_lit && !lit_quoted)
273				recordregion(expdest - stackblock() - 1,
274				    expdest - stackblock(), 0);
275			if (flag & EXP_VARTILDE && *p == '~' &&
276			    (c != '=' || firsteq)) {
277				if (c == '=')
278					firsteq = 0;
279				p = exptilde(p, flag);
280			}
281			break;
282		default:
283			USTPUTC(c, expdest);
284			if (split_lit && !lit_quoted)
285				recordregion(expdest - stackblock() - 1,
286				    expdest - stackblock(), 0);
287		}
288	}
289breakloop:;
290}
291
292/*
293 * Perform tilde expansion, placing the result in the stack string and
294 * returning the next position in the input string to process.
295 */
296static char *
297exptilde(char *p, int flag)
298{
299	char c, *startp = p;
300	struct passwd *pw;
301	char *home;
302	int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR);
303
304	while ((c = *p) != '\0') {
305		switch(c) {
306		case CTLESC: /* This means CTL* are always considered quoted. */
307		case CTLVAR:
308		case CTLBACKQ:
309		case CTLBACKQ | CTLQUOTE:
310		case CTLARI:
311		case CTLENDARI:
312		case CTLQUOTEMARK:
313			return (startp);
314		case ':':
315			if (flag & EXP_VARTILDE)
316				goto done;
317			break;
318		case '/':
319		case CTLENDVAR:
320			goto done;
321		}
322		p++;
323	}
324done:
325	*p = '\0';
326	if (*(startp+1) == '\0') {
327		if ((home = lookupvar("HOME")) == NULL)
328			goto lose;
329	} else {
330		if ((pw = getpwnam(startp+1)) == NULL)
331			goto lose;
332		home = pw->pw_dir;
333	}
334	if (*home == '\0')
335		goto lose;
336	*p = c;
337	while ((c = *home++) != '\0') {
338		if (quotes && SQSYNTAX[(int)c] == CCTL)
339			STPUTC(CTLESC, expdest);
340		STPUTC(c, expdest);
341	}
342	return (p);
343lose:
344	*p = c;
345	return (startp);
346}
347
348
349static void
350removerecordregions(int endoff)
351{
352	if (ifslastp == NULL)
353		return;
354
355	if (ifsfirst.endoff > endoff) {
356		while (ifsfirst.next != NULL) {
357			struct ifsregion *ifsp;
358			INTOFF;
359			ifsp = ifsfirst.next->next;
360			ckfree(ifsfirst.next);
361			ifsfirst.next = ifsp;
362			INTON;
363		}
364		if (ifsfirst.begoff > endoff)
365			ifslastp = NULL;
366		else {
367			ifslastp = &ifsfirst;
368			ifsfirst.endoff = endoff;
369		}
370		return;
371	}
372
373	ifslastp = &ifsfirst;
374	while (ifslastp->next && ifslastp->next->begoff < endoff)
375		ifslastp=ifslastp->next;
376	while (ifslastp->next != NULL) {
377		struct ifsregion *ifsp;
378		INTOFF;
379		ifsp = ifslastp->next->next;
380		ckfree(ifslastp->next);
381		ifslastp->next = ifsp;
382		INTON;
383	}
384	if (ifslastp->endoff > endoff)
385		ifslastp->endoff = endoff;
386}
387
388/*
389 * Expand arithmetic expression.  Backup to start of expression,
390 * evaluate, place result in (backed up) result, adjust string position.
391 */
392void
393expari(int flag)
394{
395	char *p, *q, *start;
396	arith_t result;
397	int begoff;
398	int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR);
399	int quoted;
400
401	/*
402	 * This routine is slightly over-complicated for
403	 * efficiency.  First we make sure there is
404	 * enough space for the result, which may be bigger
405	 * than the expression.  Next we
406	 * scan backwards looking for the start of arithmetic.  If the
407	 * next previous character is a CTLESC character, then we
408	 * have to rescan starting from the beginning since CTLESC
409	 * characters have to be processed left to right.
410	 */
411	CHECKSTRSPACE(DIGITS(result) - 2, expdest);
412	USTPUTC('\0', expdest);
413	start = stackblock();
414	p = expdest - 2;
415	while (p >= start && *p != CTLARI)
416		--p;
417	if (p < start || *p != CTLARI)
418		error("missing CTLARI (shouldn't happen)");
419	if (p > start && *(p - 1) == CTLESC)
420		for (p = start; *p != CTLARI; p++)
421			if (*p == CTLESC)
422				p++;
423
424	if (p[1] == '"')
425		quoted=1;
426	else
427		quoted=0;
428	begoff = p - start;
429	removerecordregions(begoff);
430	if (quotes)
431		rmescapes(p+2);
432	q = grabstackstr(expdest);
433	result = arith(p+2);
434	ungrabstackstr(q, expdest);
435	fmtstr(p, DIGITS(result), ARITH_FORMAT_STR, result);
436	while (*p++)
437		;
438	if (quoted == 0)
439		recordregion(begoff, p - 1 - start, 0);
440	result = expdest - p + 1;
441	STADJUST(-result, expdest);
442}
443
444
445/*
446 * Perform command substitution.
447 */
448static void
449expbackq(union node *cmd, int quoted, int flag)
450{
451	struct backcmd in;
452	int i;
453	char buf[128];
454	char *p;
455	char *dest = expdest;
456	struct ifsregion saveifs, *savelastp;
457	struct nodelist *saveargbackq;
458	char lastc;
459	int startloc = dest - stackblock();
460	char const *syntax = quoted? DQSYNTAX : BASESYNTAX;
461	int saveherefd;
462	int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR);
463	int nnl;
464
465	INTOFF;
466	saveifs = ifsfirst;
467	savelastp = ifslastp;
468	saveargbackq = argbackq;
469	saveherefd = herefd;
470	herefd = -1;
471	p = grabstackstr(dest);
472	evalbackcmd(cmd, &in);
473	ungrabstackstr(p, dest);
474	ifsfirst = saveifs;
475	ifslastp = savelastp;
476	argbackq = saveargbackq;
477	herefd = saveherefd;
478
479	p = in.buf;
480	lastc = '\0';
481	nnl = 0;
482	/* Don't copy trailing newlines */
483	for (;;) {
484		if (--in.nleft < 0) {
485			if (in.fd < 0)
486				break;
487			while ((i = read(in.fd, buf, sizeof buf)) < 0 && errno == EINTR);
488			TRACE(("expbackq: read returns %d\n", i));
489			if (i <= 0)
490				break;
491			p = buf;
492			in.nleft = i - 1;
493		}
494		lastc = *p++;
495		if (lastc != '\0') {
496			if (quotes && syntax[(int)lastc] == CCTL)
497				STPUTC(CTLESC, dest);
498			if (lastc == '\n') {
499				nnl++;
500			} else {
501				while (nnl > 0) {
502					nnl--;
503					STPUTC('\n', dest);
504				}
505				STPUTC(lastc, dest);
506			}
507		}
508	}
509
510	if (in.fd >= 0)
511		close(in.fd);
512	if (in.buf)
513		ckfree(in.buf);
514	if (in.jp)
515		exitstatus = waitforjob(in.jp, (int *)NULL);
516	if (quoted == 0)
517		recordregion(startloc, dest - stackblock(), 0);
518	TRACE(("expbackq: size=%td: \"%.*s\"\n",
519		((dest - stackblock()) - startloc),
520		(int)((dest - stackblock()) - startloc),
521		stackblock() + startloc));
522	expdest = dest;
523	INTON;
524}
525
526
527
528static int
529subevalvar(char *p, char *str, int strloc, int subtype, int startloc,
530  int varflags, int quotes)
531{
532	char *startp;
533	char *loc = NULL;
534	char *q;
535	int c = 0;
536	int saveherefd = herefd;
537	struct nodelist *saveargbackq = argbackq;
538	int amount;
539
540	herefd = -1;
541	argstr(p, (subtype == VSTRIMLEFT || subtype == VSTRIMLEFTMAX ||
542	    subtype == VSTRIMRIGHT || subtype == VSTRIMRIGHTMAX ?
543	    EXP_CASE : 0) | EXP_TILDE);
544	STACKSTRNUL(expdest);
545	herefd = saveherefd;
546	argbackq = saveargbackq;
547	startp = stackblock() + startloc;
548	if (str == NULL)
549	    str = stackblock() + strloc;
550
551	switch (subtype) {
552	case VSASSIGN:
553		setvar(str, startp, 0);
554		amount = startp - expdest;
555		STADJUST(amount, expdest);
556		varflags &= ~VSNUL;
557		if (c != 0)
558			*loc = c;
559		return 1;
560
561	case VSQUESTION:
562		if (*p != CTLENDVAR) {
563			outfmt(out2, "%s\n", startp);
564			error((char *)NULL);
565		}
566		error("%.*s: parameter %snot set", (int)(p - str - 1),
567		      str, (varflags & VSNUL) ? "null or "
568					      : nullstr);
569		return 0;
570
571	case VSTRIMLEFT:
572		for (loc = startp; loc < str; loc++) {
573			c = *loc;
574			*loc = '\0';
575			if (patmatch(str, startp, quotes)) {
576				*loc = c;
577				goto recordleft;
578			}
579			*loc = c;
580			if (quotes && *loc == CTLESC)
581				loc++;
582		}
583		return 0;
584
585	case VSTRIMLEFTMAX:
586		for (loc = str - 1; loc >= startp;) {
587			c = *loc;
588			*loc = '\0';
589			if (patmatch(str, startp, quotes)) {
590				*loc = c;
591				goto recordleft;
592			}
593			*loc = c;
594			loc--;
595			if (quotes && loc > startp && *(loc - 1) == CTLESC) {
596				for (q = startp; q < loc; q++)
597					if (*q == CTLESC)
598						q++;
599				if (q > loc)
600					loc--;
601			}
602		}
603		return 0;
604
605	case VSTRIMRIGHT:
606		for (loc = str - 1; loc >= startp;) {
607			if (patmatch(str, loc, quotes)) {
608				amount = loc - expdest;
609				STADJUST(amount, expdest);
610				return 1;
611			}
612			loc--;
613			if (quotes && loc > startp && *(loc - 1) == CTLESC) {
614				for (q = startp; q < loc; q++)
615					if (*q == CTLESC)
616						q++;
617				if (q > loc)
618					loc--;
619			}
620		}
621		return 0;
622
623	case VSTRIMRIGHTMAX:
624		for (loc = startp; loc < str - 1; loc++) {
625			if (patmatch(str, loc, quotes)) {
626				amount = loc - expdest;
627				STADJUST(amount, expdest);
628				return 1;
629			}
630			if (quotes && *loc == CTLESC)
631				loc++;
632		}
633		return 0;
634
635
636	default:
637		abort();
638	}
639
640recordleft:
641	amount = ((str - 1) - (loc - startp)) - expdest;
642	STADJUST(amount, expdest);
643	while (loc != str - 1)
644		*startp++ = *loc++;
645	return 1;
646}
647
648
649/*
650 * Expand a variable, and return a pointer to the next character in the
651 * input string.
652 */
653
654static char *
655evalvar(char *p, int flag)
656{
657	int subtype;
658	int varflags;
659	char *var;
660	char *val;
661	int patloc;
662	int c;
663	int set;
664	int special;
665	int startloc;
666	int varlen;
667	int easy;
668	int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR);
669
670	varflags = (unsigned char)*p++;
671	subtype = varflags & VSTYPE;
672	var = p;
673	special = 0;
674	if (! is_name(*p))
675		special = 1;
676	p = strchr(p, '=') + 1;
677again: /* jump here after setting a variable with ${var=text} */
678	if (varflags & VSLINENO) {
679		set = 1;
680		special = 0;
681		val = var;
682		p[-1] = '\0';	/* temporarily overwrite '=' to have \0
683				   terminated string */
684	} else if (special) {
685		set = varisset(var, varflags & VSNUL);
686		val = NULL;
687	} else {
688		val = bltinlookup(var, 1);
689		if (val == NULL || ((varflags & VSNUL) && val[0] == '\0')) {
690			val = NULL;
691			set = 0;
692		} else
693			set = 1;
694	}
695	varlen = 0;
696	startloc = expdest - stackblock();
697	if (!set && uflag && *var != '@' && *var != '*') {
698		switch (subtype) {
699		case VSNORMAL:
700		case VSTRIMLEFT:
701		case VSTRIMLEFTMAX:
702		case VSTRIMRIGHT:
703		case VSTRIMRIGHTMAX:
704		case VSLENGTH:
705			error("%.*s: parameter not set", (int)(p - var - 1),
706			    var);
707		}
708	}
709	if (set && subtype != VSPLUS) {
710		/* insert the value of the variable */
711		if (special) {
712			varvalue(var, varflags & VSQUOTE, subtype, flag);
713			if (subtype == VSLENGTH) {
714				varlen = expdest - stackblock() - startloc;
715				STADJUST(-varlen, expdest);
716			}
717		} else {
718			char const *syntax = (varflags & VSQUOTE) ? DQSYNTAX
719								  : BASESYNTAX;
720
721			if (subtype == VSLENGTH) {
722				for (;*val; val++)
723					varlen++;
724			}
725			else {
726				while (*val) {
727					if (quotes &&
728					    syntax[(int)*val] == CCTL)
729						STPUTC(CTLESC, expdest);
730					STPUTC(*val++, expdest);
731				}
732
733			}
734		}
735	}
736
737	if (subtype == VSPLUS)
738		set = ! set;
739
740	easy = ((varflags & VSQUOTE) == 0 ||
741		(*var == '@' && shellparam.nparam != 1));
742
743
744	switch (subtype) {
745	case VSLENGTH:
746		expdest = cvtnum(varlen, expdest);
747		goto record;
748
749	case VSNORMAL:
750		if (!easy)
751			break;
752record:
753		recordregion(startloc, expdest - stackblock(),
754			     varflags & VSQUOTE);
755		break;
756
757	case VSPLUS:
758	case VSMINUS:
759		if (!set) {
760			argstr(p, flag | (flag & EXP_FULL ? EXP_SPLIT_LIT : 0) |
761			    (varflags & VSQUOTE ? EXP_LIT_QUOTED : 0));
762			break;
763		}
764		if (easy)
765			goto record;
766		break;
767
768	case VSTRIMLEFT:
769	case VSTRIMLEFTMAX:
770	case VSTRIMRIGHT:
771	case VSTRIMRIGHTMAX:
772		if (!set)
773			break;
774		/*
775		 * Terminate the string and start recording the pattern
776		 * right after it
777		 */
778		STPUTC('\0', expdest);
779		patloc = expdest - stackblock();
780		if (subevalvar(p, NULL, patloc, subtype,
781		    startloc, varflags, quotes) == 0) {
782			int amount = (expdest - stackblock() - patloc) + 1;
783			STADJUST(-amount, expdest);
784		}
785		/* Remove any recorded regions beyond start of variable */
786		removerecordregions(startloc);
787		goto record;
788
789	case VSASSIGN:
790	case VSQUESTION:
791		if (!set) {
792			if (subevalvar(p, var, 0, subtype, startloc, varflags,
793			    quotes)) {
794				varflags &= ~VSNUL;
795				/*
796				 * Remove any recorded regions beyond
797				 * start of variable
798				 */
799				removerecordregions(startloc);
800				goto again;
801			}
802			break;
803		}
804		if (easy)
805			goto record;
806		break;
807
808	case VSERROR:
809		c = p - var - 1;
810		error("${%.*s%s}: Bad substitution", c, var,
811		    (c > 0 && *p != CTLENDVAR) ? "..." : "");
812
813	default:
814		abort();
815	}
816	p[-1] = '=';	/* recover overwritten '=' */
817
818	if (subtype != VSNORMAL) {	/* skip to end of alternative */
819		int nesting = 1;
820		for (;;) {
821			if ((c = *p++) == CTLESC)
822				p++;
823			else if (c == CTLBACKQ || c == (CTLBACKQ|CTLQUOTE)) {
824				if (set)
825					argbackq = argbackq->next;
826			} else if (c == CTLVAR) {
827				if ((*p++ & VSTYPE) != VSNORMAL)
828					nesting++;
829			} else if (c == CTLENDVAR) {
830				if (--nesting == 0)
831					break;
832			}
833		}
834	}
835	return p;
836}
837
838
839
840/*
841 * Test whether a specialized variable is set.
842 */
843
844static int
845varisset(char *name, int nulok)
846{
847
848	if (*name == '!')
849		return backgndpidset();
850	else if (*name == '@' || *name == '*') {
851		if (*shellparam.p == NULL)
852			return 0;
853
854		if (nulok) {
855			char **av;
856
857			for (av = shellparam.p; *av; av++)
858				if (**av != '\0')
859					return 1;
860			return 0;
861		}
862	} else if (is_digit(*name)) {
863		char *ap;
864		int num = atoi(name);
865
866		if (num > shellparam.nparam)
867			return 0;
868
869		if (num == 0)
870			ap = arg0;
871		else
872			ap = shellparam.p[num - 1];
873
874		if (nulok && (ap == NULL || *ap == '\0'))
875			return 0;
876	}
877	return 1;
878}
879
880
881
882/*
883 * Add the value of a specialized variable to the stack string.
884 */
885
886static void
887varvalue(char *name, int quoted, int subtype, int flag)
888{
889	int num;
890	char *p;
891	int i;
892	char sep;
893	char **ap;
894	char const *syntax;
895
896#define STRTODEST(p) \
897	do {\
898	if (flag & (EXP_FULL | EXP_CASE) && subtype != VSLENGTH) { \
899		syntax = quoted? DQSYNTAX : BASESYNTAX; \
900		while (*p) { \
901			if (syntax[(int)*p] == CCTL) \
902				STPUTC(CTLESC, expdest); \
903			STPUTC(*p++, expdest); \
904		} \
905	} else \
906		STPUTS(p, expdest); \
907	} while (0)
908
909
910	switch (*name) {
911	case '$':
912		num = rootpid;
913		goto numvar;
914	case '?':
915		num = oexitstatus;
916		goto numvar;
917	case '#':
918		num = shellparam.nparam;
919		goto numvar;
920	case '!':
921		num = backgndpidval();
922numvar:
923		expdest = cvtnum(num, expdest);
924		break;
925	case '-':
926		for (i = 0 ; i < NOPTS ; i++) {
927			if (optlist[i].val)
928				STPUTC(optlist[i].letter, expdest);
929		}
930		break;
931	case '@':
932		if (flag & EXP_FULL && quoted) {
933			for (ap = shellparam.p ; (p = *ap++) != NULL ; ) {
934				STRTODEST(p);
935				if (*ap)
936					STPUTC('\0', expdest);
937			}
938			break;
939		}
940		/* FALLTHROUGH */
941	case '*':
942		if (ifsset())
943			sep = ifsval()[0];
944		else
945			sep = ' ';
946		for (ap = shellparam.p ; (p = *ap++) != NULL ; ) {
947			STRTODEST(p);
948			if (*ap && sep)
949				STPUTC(sep, expdest);
950		}
951		break;
952	case '0':
953		p = arg0;
954		STRTODEST(p);
955		break;
956	default:
957		if (is_digit(*name)) {
958			num = atoi(name);
959			if (num > 0 && num <= shellparam.nparam) {
960				p = shellparam.p[num - 1];
961				STRTODEST(p);
962			}
963		}
964		break;
965	}
966}
967
968
969
970/*
971 * Record the the fact that we have to scan this region of the
972 * string for IFS characters.
973 */
974
975static void
976recordregion(int start, int end, int inquotes)
977{
978	struct ifsregion *ifsp;
979
980	if (ifslastp == NULL) {
981		ifsp = &ifsfirst;
982	} else {
983		if (ifslastp->endoff == start
984		    && ifslastp->inquotes == inquotes) {
985			/* extend previous area */
986			ifslastp->endoff = end;
987			return;
988		}
989		ifsp = (struct ifsregion *)ckmalloc(sizeof (struct ifsregion));
990		ifslastp->next = ifsp;
991	}
992	ifslastp = ifsp;
993	ifslastp->next = NULL;
994	ifslastp->begoff = start;
995	ifslastp->endoff = end;
996	ifslastp->inquotes = inquotes;
997}
998
999
1000
1001/*
1002 * Break the argument string into pieces based upon IFS and add the
1003 * strings to the argument list.  The regions of the string to be
1004 * searched for IFS characters have been stored by recordregion.
1005 * CTLESC characters are preserved but have little effect in this pass
1006 * other than escaping CTL* characters.  In particular, they do not escape
1007 * IFS characters: that should be done with the ifsregion mechanism.
1008 * CTLQUOTEMARK characters are used to preserve empty quoted strings.
1009 * This pass treats them as a regular character, making the string non-empty.
1010 * Later, they are removed along with the other CTL* characters.
1011 */
1012static void
1013ifsbreakup(char *string, struct arglist *arglist)
1014{
1015	struct ifsregion *ifsp;
1016	struct strlist *sp;
1017	char *start;
1018	char *p;
1019	char *q;
1020	const char *ifs;
1021	const char *ifsspc;
1022	int had_param_ch = 0;
1023
1024	start = string;
1025
1026	if (ifslastp == NULL) {
1027		/* Return entire argument, IFS doesn't apply to any of it */
1028		sp = (struct strlist *)stalloc(sizeof *sp);
1029		sp->text = start;
1030		*arglist->lastp = sp;
1031		arglist->lastp = &sp->next;
1032		return;
1033	}
1034
1035	ifs = ifsset() ? ifsval() : " \t\n";
1036
1037	for (ifsp = &ifsfirst; ifsp != NULL; ifsp = ifsp->next) {
1038		p = string + ifsp->begoff;
1039		while (p < string + ifsp->endoff) {
1040			q = p;
1041			if (*p == CTLESC)
1042				p++;
1043			if (ifsp->inquotes) {
1044				/* Only NULs (should be from "$@") end args */
1045				had_param_ch = 1;
1046				if (*p != 0) {
1047					p++;
1048					continue;
1049				}
1050				ifsspc = NULL;
1051			} else {
1052				if (!strchr(ifs, *p)) {
1053					had_param_ch = 1;
1054					p++;
1055					continue;
1056				}
1057				ifsspc = strchr(" \t\n", *p);
1058
1059				/* Ignore IFS whitespace at start */
1060				if (q == start && ifsspc != NULL) {
1061					p++;
1062					start = p;
1063					continue;
1064				}
1065				had_param_ch = 0;
1066			}
1067
1068			/* Save this argument... */
1069			*q = '\0';
1070			sp = (struct strlist *)stalloc(sizeof *sp);
1071			sp->text = start;
1072			*arglist->lastp = sp;
1073			arglist->lastp = &sp->next;
1074			p++;
1075
1076			if (ifsspc != NULL) {
1077				/* Ignore further trailing IFS whitespace */
1078				for (; p < string + ifsp->endoff; p++) {
1079					q = p;
1080					if (*p == CTLESC)
1081						p++;
1082					if (strchr(ifs, *p) == NULL) {
1083						p = q;
1084						break;
1085					}
1086					if (strchr(" \t\n", *p) == NULL) {
1087						p++;
1088						break;
1089					}
1090				}
1091			}
1092			start = p;
1093		}
1094	}
1095
1096	/*
1097	 * Save anything left as an argument.
1098	 * Traditionally we have treated 'IFS=':'; set -- x$IFS' as
1099	 * generating 2 arguments, the second of which is empty.
1100	 * Some recent clarification of the Posix spec say that it
1101	 * should only generate one....
1102	 */
1103	if (had_param_ch || *start != 0) {
1104		sp = (struct strlist *)stalloc(sizeof *sp);
1105		sp->text = start;
1106		*arglist->lastp = sp;
1107		arglist->lastp = &sp->next;
1108	}
1109}
1110
1111
1112static char expdir[PATH_MAX];
1113#define expdir_end (expdir + sizeof(expdir))
1114
1115/*
1116 * Perform pathname generation and remove control characters.
1117 * At this point, the only control characters should be CTLESC and CTLQUOTEMARK.
1118 * The results are stored in the list exparg.
1119 */
1120static void
1121expandmeta(struct strlist *str, int flag __unused)
1122{
1123	char *p;
1124	struct strlist **savelastp;
1125	struct strlist *sp;
1126	char c;
1127	/* TODO - EXP_REDIR */
1128
1129	while (str) {
1130		if (fflag)
1131			goto nometa;
1132		p = str->text;
1133		for (;;) {			/* fast check for meta chars */
1134			if ((c = *p++) == '\0')
1135				goto nometa;
1136			if (c == '*' || c == '?' || c == '[')
1137				break;
1138		}
1139		savelastp = exparg.lastp;
1140		INTOFF;
1141		expmeta(expdir, str->text);
1142		INTON;
1143		if (exparg.lastp == savelastp) {
1144			/*
1145			 * no matches
1146			 */
1147nometa:
1148			*exparg.lastp = str;
1149			rmescapes(str->text);
1150			exparg.lastp = &str->next;
1151		} else {
1152			*exparg.lastp = NULL;
1153			*savelastp = sp = expsort(*savelastp);
1154			while (sp->next != NULL)
1155				sp = sp->next;
1156			exparg.lastp = &sp->next;
1157		}
1158		str = str->next;
1159	}
1160}
1161
1162
1163/*
1164 * Do metacharacter (i.e. *, ?, [...]) expansion.
1165 */
1166
1167static void
1168expmeta(char *enddir, char *name)
1169{
1170	char *p;
1171	char *q;
1172	char *start;
1173	char *endname;
1174	int metaflag;
1175	struct stat statb;
1176	DIR *dirp;
1177	struct dirent *dp;
1178	int atend;
1179	int matchdot;
1180	int esc;
1181
1182	metaflag = 0;
1183	start = name;
1184	for (p = name; esc = 0, *p; p += esc + 1) {
1185		if (*p == '*' || *p == '?')
1186			metaflag = 1;
1187		else if (*p == '[') {
1188			q = p + 1;
1189			if (*q == '!' || *q == '^')
1190				q++;
1191			for (;;) {
1192				while (*q == CTLQUOTEMARK)
1193					q++;
1194				if (*q == CTLESC)
1195					q++;
1196				if (*q == '/' || *q == '\0')
1197					break;
1198				if (*++q == ']') {
1199					metaflag = 1;
1200					break;
1201				}
1202			}
1203		} else if (*p == '\0')
1204			break;
1205		else if (*p == CTLQUOTEMARK)
1206			continue;
1207		else {
1208			if (*p == CTLESC)
1209				esc++;
1210			if (p[esc] == '/') {
1211				if (metaflag)
1212					break;
1213				start = p + esc + 1;
1214			}
1215		}
1216	}
1217	if (metaflag == 0) {	/* we've reached the end of the file name */
1218		if (enddir != expdir)
1219			metaflag++;
1220		for (p = name ; ; p++) {
1221			if (*p == CTLQUOTEMARK)
1222				continue;
1223			if (*p == CTLESC)
1224				p++;
1225			*enddir++ = *p;
1226			if (*p == '\0')
1227				break;
1228			if (enddir == expdir_end)
1229				return;
1230		}
1231		if (metaflag == 0 || lstat(expdir, &statb) >= 0)
1232			addfname(expdir);
1233		return;
1234	}
1235	endname = p;
1236	if (start != name) {
1237		p = name;
1238		while (p < start) {
1239			while (*p == CTLQUOTEMARK)
1240				p++;
1241			if (*p == CTLESC)
1242				p++;
1243			*enddir++ = *p++;
1244			if (enddir == expdir_end)
1245				return;
1246		}
1247	}
1248	if (enddir == expdir) {
1249		p = ".";
1250	} else if (enddir == expdir + 1 && *expdir == '/') {
1251		p = "/";
1252	} else {
1253		p = expdir;
1254		enddir[-1] = '\0';
1255	}
1256	if ((dirp = opendir(p)) == NULL)
1257		return;
1258	if (enddir != expdir)
1259		enddir[-1] = '/';
1260	if (*endname == 0) {
1261		atend = 1;
1262	} else {
1263		atend = 0;
1264		*endname = '\0';
1265		endname += esc + 1;
1266	}
1267	matchdot = 0;
1268	p = start;
1269	while (*p == CTLQUOTEMARK)
1270		p++;
1271	if (*p == CTLESC)
1272		p++;
1273	if (*p == '.')
1274		matchdot++;
1275	while (! int_pending() && (dp = readdir(dirp)) != NULL) {
1276		if (dp->d_name[0] == '.' && ! matchdot)
1277			continue;
1278		if (patmatch(start, dp->d_name, 0)) {
1279			if (enddir + dp->d_namlen + 1 > expdir_end)
1280				continue;
1281			memcpy(enddir, dp->d_name, dp->d_namlen + 1);
1282			if (atend)
1283				addfname(expdir);
1284			else {
1285				if (enddir + dp->d_namlen + 2 > expdir_end)
1286					continue;
1287				enddir[dp->d_namlen] = '/';
1288				enddir[dp->d_namlen + 1] = '\0';
1289				expmeta(enddir + dp->d_namlen + 1, endname);
1290			}
1291		}
1292	}
1293	closedir(dirp);
1294	if (! atend)
1295		endname[-esc - 1] = esc ? CTLESC : '/';
1296}
1297
1298
1299/*
1300 * Add a file name to the list.
1301 */
1302
1303static void
1304addfname(char *name)
1305{
1306	char *p;
1307	struct strlist *sp;
1308
1309	p = stalloc(strlen(name) + 1);
1310	scopy(name, p);
1311	sp = (struct strlist *)stalloc(sizeof *sp);
1312	sp->text = p;
1313	*exparg.lastp = sp;
1314	exparg.lastp = &sp->next;
1315}
1316
1317
1318/*
1319 * Sort the results of file name expansion.  It calculates the number of
1320 * strings to sort and then calls msort (short for merge sort) to do the
1321 * work.
1322 */
1323
1324static struct strlist *
1325expsort(struct strlist *str)
1326{
1327	int len;
1328	struct strlist *sp;
1329
1330	len = 0;
1331	for (sp = str ; sp ; sp = sp->next)
1332		len++;
1333	return msort(str, len);
1334}
1335
1336
1337static struct strlist *
1338msort(struct strlist *list, int len)
1339{
1340	struct strlist *p, *q = NULL;
1341	struct strlist **lpp;
1342	int half;
1343	int n;
1344
1345	if (len <= 1)
1346		return list;
1347	half = len >> 1;
1348	p = list;
1349	for (n = half ; --n >= 0 ; ) {
1350		q = p;
1351		p = p->next;
1352	}
1353	q->next = NULL;			/* terminate first half of list */
1354	q = msort(list, half);		/* sort first half of list */
1355	p = msort(p, len - half);		/* sort second half */
1356	lpp = &list;
1357	for (;;) {
1358		if (strcmp(p->text, q->text) < 0) {
1359			*lpp = p;
1360			lpp = &p->next;
1361			if ((p = *lpp) == NULL) {
1362				*lpp = q;
1363				break;
1364			}
1365		} else {
1366			*lpp = q;
1367			lpp = &q->next;
1368			if ((q = *lpp) == NULL) {
1369				*lpp = p;
1370				break;
1371			}
1372		}
1373	}
1374	return list;
1375}
1376
1377
1378
1379/*
1380 * Returns true if the pattern matches the string.
1381 */
1382
1383int
1384patmatch(const char *pattern, const char *string, int squoted)
1385{
1386	const char *p, *q;
1387	char c;
1388
1389	p = pattern;
1390	q = string;
1391	for (;;) {
1392		switch (c = *p++) {
1393		case '\0':
1394			goto breakloop;
1395		case CTLESC:
1396			if (squoted && *q == CTLESC)
1397				q++;
1398			if (*q++ != *p++)
1399				return 0;
1400			break;
1401		case CTLQUOTEMARK:
1402			continue;
1403		case '?':
1404			if (squoted && *q == CTLESC)
1405				q++;
1406			if (*q++ == '\0')
1407				return 0;
1408			break;
1409		case '*':
1410			c = *p;
1411			while (c == CTLQUOTEMARK || c == '*')
1412				c = *++p;
1413			if (c != CTLESC &&  c != CTLQUOTEMARK &&
1414			    c != '?' && c != '*' && c != '[') {
1415				while (*q != c) {
1416					if (squoted && *q == CTLESC &&
1417					    q[1] == c)
1418						break;
1419					if (*q == '\0')
1420						return 0;
1421					if (squoted && *q == CTLESC)
1422						q++;
1423					q++;
1424				}
1425			}
1426			do {
1427				if (patmatch(p, q, squoted))
1428					return 1;
1429				if (squoted && *q == CTLESC)
1430					q++;
1431			} while (*q++ != '\0');
1432			return 0;
1433		case '[': {
1434			const char *endp;
1435			int invert, found;
1436			char chr;
1437
1438			endp = p;
1439			if (*endp == '!' || *endp == '^')
1440				endp++;
1441			for (;;) {
1442				while (*endp == CTLQUOTEMARK)
1443					endp++;
1444				if (*endp == '\0')
1445					goto dft;		/* no matching ] */
1446				if (*endp == CTLESC)
1447					endp++;
1448				if (*++endp == ']')
1449					break;
1450			}
1451			invert = 0;
1452			if (*p == '!' || *p == '^') {
1453				invert++;
1454				p++;
1455			}
1456			found = 0;
1457			chr = *q++;
1458			if (squoted && chr == CTLESC)
1459				chr = *q++;
1460			if (chr == '\0')
1461				return 0;
1462			c = *p++;
1463			do {
1464				if (c == CTLQUOTEMARK)
1465					continue;
1466				if (c == CTLESC)
1467					c = *p++;
1468				if (*p == '-' && p[1] != ']') {
1469					p++;
1470					while (*p == CTLQUOTEMARK)
1471						p++;
1472					if (*p == CTLESC)
1473						p++;
1474					if (   collate_range_cmp(chr, c) >= 0
1475					    && collate_range_cmp(chr, *p) <= 0
1476					   )
1477						found = 1;
1478					p++;
1479				} else {
1480					if (chr == c)
1481						found = 1;
1482				}
1483			} while ((c = *p++) != ']');
1484			if (found == invert)
1485				return 0;
1486			break;
1487		}
1488dft:	        default:
1489			if (squoted && *q == CTLESC)
1490				q++;
1491			if (*q++ != c)
1492				return 0;
1493			break;
1494		}
1495	}
1496breakloop:
1497	if (*q != '\0')
1498		return 0;
1499	return 1;
1500}
1501
1502
1503
1504/*
1505 * Remove any CTLESC and CTLQUOTEMARK characters from a string.
1506 */
1507
1508void
1509rmescapes(char *str)
1510{
1511	char *p, *q;
1512
1513	p = str;
1514	while (*p != CTLESC && *p != CTLQUOTEMARK && *p != CTLQUOTEEND) {
1515		if (*p++ == '\0')
1516			return;
1517	}
1518	q = p;
1519	while (*p) {
1520		if (*p == CTLQUOTEMARK || *p == CTLQUOTEEND) {
1521			p++;
1522			continue;
1523		}
1524		if (*p == CTLESC)
1525			p++;
1526		*q++ = *p++;
1527	}
1528	*q = '\0';
1529}
1530
1531
1532
1533/*
1534 * See if a pattern matches in a case statement.
1535 */
1536
1537int
1538casematch(union node *pattern, const char *val)
1539{
1540	struct stackmark smark;
1541	int result;
1542	char *p;
1543
1544	setstackmark(&smark);
1545	argbackq = pattern->narg.backquote;
1546	STARTSTACKSTR(expdest);
1547	ifslastp = NULL;
1548	argstr(pattern->narg.text, EXP_TILDE | EXP_CASE);
1549	STPUTC('\0', expdest);
1550	p = grabstackstr(expdest);
1551	result = patmatch(p, val, 0);
1552	popstackmark(&smark);
1553	return result;
1554}
1555
1556/*
1557 * Our own itoa().
1558 */
1559
1560static char *
1561cvtnum(int num, char *buf)
1562{
1563	char temp[32];
1564	int neg = num < 0;
1565	char *p = temp + 31;
1566
1567	temp[31] = '\0';
1568
1569	do {
1570		*--p = num % 10 + '0';
1571	} while ((num /= 10) != 0);
1572
1573	if (neg)
1574		*--p = '-';
1575
1576	STPUTS(p, buf);
1577	return buf;
1578}
1579
1580/*
1581 * Do most of the work for wordexp(3).
1582 */
1583
1584int
1585wordexpcmd(int argc, char **argv)
1586{
1587	size_t len;
1588	int i;
1589
1590	out1fmt("%08x", argc - 1);
1591	for (i = 1, len = 0; i < argc; i++)
1592		len += strlen(argv[i]);
1593	out1fmt("%08x", (int)len);
1594	for (i = 1; i < argc; i++)
1595		outbin(argv[i], strlen(argv[i]) + 1, out1);
1596        return (0);
1597}
1598