parser.c revision 60593 
1/*-
2 * Copyright (c) 1991, 1993
3 *	The Regents of the University of California.  All rights reserved.
4 *
5 * This code is derived from software contributed to Berkeley by
6 * Kenneth Almquist.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 *    notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 *    notice, this list of conditions and the following disclaimer in the
15 *    documentation and/or other materials provided with the distribution.
16 * 3. All advertising materials mentioning features or use of this software
17 *    must display the following acknowledgement:
18 *	This product includes software developed by the University of
19 *	California, Berkeley and its contributors.
20 * 4. Neither the name of the University nor the names of its contributors
21 *    may be used to endorse or promote products derived from this software
22 *    without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 */
36
37#ifndef lint
38#if 0
39static char sccsid[] = "@(#)parser.c	8.7 (Berkeley) 5/16/95";
40#endif
41static const char rcsid[] =
42  "$FreeBSD: head/bin/sh/parser.c 60593 2000-05-15 13:02:07Z cracauer $";
43#endif /* not lint */
44
45#include <stdlib.h>
46
47#include "shell.h"
48#include "parser.h"
49#include "nodes.h"
50#include "expand.h"	/* defines rmescapes() */
51#include "redir.h"	/* defines copyfd() */
52#include "syntax.h"
53#include "options.h"
54#include "input.h"
55#include "output.h"
56#include "var.h"
57#include "error.h"
58#include "memalloc.h"
59#include "mystring.h"
60#include "alias.h"
61#include "show.h"
62#include "eval.h"
63#ifndef NO_HISTORY
64#include "myhistedit.h"
65#endif
66
67/*
68 * Shell command parser.
69 */
70
71#define EOFMARKLEN 79
72
73/* values returned by readtoken */
74#include "token.h"
75
76
77
78struct heredoc {
79	struct heredoc *next;	/* next here document in list */
80	union node *here;		/* redirection node */
81	char *eofmark;		/* string indicating end of input */
82	int striptabs;		/* if set, strip leading tabs */
83};
84
85
86
87struct heredoc *heredoclist;	/* list of here documents to read */
88int parsebackquote;		/* nonzero if we are inside backquotes */
89int doprompt;			/* if set, prompt the user */
90int needprompt;			/* true if interactive and at start of line */
91int lasttoken;			/* last token read */
92MKINIT int tokpushback;		/* last token pushed back */
93char *wordtext;			/* text of last word returned by readtoken */
94MKINIT int checkkwd;            /* 1 == check for kwds, 2 == also eat newlines */
95struct nodelist *backquotelist;
96union node *redirnode;
97struct heredoc *heredoc;
98int quoteflag;			/* set if (part of) last token was quoted */
99int startlinno;			/* line # where last token started */
100
101/* XXX When 'noaliases' is set to one, no alias expansion takes place. */
102static int noaliases = 0;
103
104#define GDB_HACK 1 /* avoid local declarations which gdb can't handle */
105#ifdef GDB_HACK
106static const char argvars[5] = {CTLVAR, VSNORMAL|VSQUOTE, '@', '=', '\0'};
107static const char types[] = "}-+?=";
108#endif
109
110
111STATIC union node *list __P((int));
112STATIC union node *andor __P((void));
113STATIC union node *pipeline __P((void));
114STATIC union node *command __P((void));
115STATIC union node *simplecmd __P((union node **, union node *));
116STATIC union node *makename __P((void));
117STATIC void parsefname __P((void));
118STATIC void parseheredoc __P((void));
119STATIC int peektoken __P((void));
120STATIC int readtoken __P((void));
121STATIC int xxreadtoken __P((void));
122STATIC int readtoken1 __P((int, char const *, char *, int));
123STATIC int noexpand __P((char *));
124STATIC void synexpect __P((int));
125STATIC void synerror __P((char *));
126STATIC void setprompt __P((int));
127
128
129/*
130 * Read and parse a command.  Returns NEOF on end of file.  (NULL is a
131 * valid parse tree indicating a blank line.)
132 */
133
134union node *
135parsecmd(interact)
136	int interact;
137{
138	int t;
139
140	tokpushback = 0;
141	doprompt = interact;
142	if (doprompt)
143		setprompt(1);
144	else
145		setprompt(0);
146	needprompt = 0;
147	t = readtoken();
148	if (t == TEOF)
149		return NEOF;
150	if (t == TNL)
151		return NULL;
152	tokpushback++;
153	return list(1);
154}
155
156
157STATIC union node *
158list(nlflag)
159	int nlflag;
160{
161	union node *n1, *n2, *n3;
162	int tok;
163
164	checkkwd = 2;
165	if (nlflag == 0 && tokendlist[peektoken()])
166		return NULL;
167	n1 = NULL;
168	for (;;) {
169		n2 = andor();
170		tok = readtoken();
171		if (tok == TBACKGND) {
172			if (n2->type == NCMD || n2->type == NPIPE) {
173				n2->ncmd.backgnd = 1;
174			} else if (n2->type == NREDIR) {
175				n2->type = NBACKGND;
176			} else {
177				n3 = (union node *)stalloc(sizeof (struct nredir));
178				n3->type = NBACKGND;
179				n3->nredir.n = n2;
180				n3->nredir.redirect = NULL;
181				n2 = n3;
182			}
183		}
184		if (n1 == NULL) {
185			n1 = n2;
186		}
187		else {
188			n3 = (union node *)stalloc(sizeof (struct nbinary));
189			n3->type = NSEMI;
190			n3->nbinary.ch1 = n1;
191			n3->nbinary.ch2 = n2;
192			n1 = n3;
193		}
194		switch (tok) {
195		case TBACKGND:
196		case TSEMI:
197			tok = readtoken();
198			/* fall through */
199		case TNL:
200			if (tok == TNL) {
201				parseheredoc();
202				if (nlflag)
203					return n1;
204			} else {
205				tokpushback++;
206			}
207			checkkwd = 2;
208			if (tokendlist[peektoken()])
209				return n1;
210			break;
211		case TEOF:
212			if (heredoclist)
213				parseheredoc();
214			else
215				pungetc();		/* push back EOF on input */
216			return n1;
217		default:
218			if (nlflag)
219				synexpect(-1);
220			tokpushback++;
221			return n1;
222		}
223	}
224}
225
226
227
228STATIC union node *
229andor() {
230	union node *n1, *n2, *n3;
231	int t;
232
233	n1 = pipeline();
234	for (;;) {
235		if ((t = readtoken()) == TAND) {
236			t = NAND;
237		} else if (t == TOR) {
238			t = NOR;
239		} else {
240			tokpushback++;
241			return n1;
242		}
243		n2 = pipeline();
244		n3 = (union node *)stalloc(sizeof (struct nbinary));
245		n3->type = t;
246		n3->nbinary.ch1 = n1;
247		n3->nbinary.ch2 = n2;
248		n1 = n3;
249	}
250}
251
252
253
254STATIC union node *
255pipeline() {
256	union node *n1, *pipenode, *notnode;
257	struct nodelist *lp, *prev;
258	int negate = 0;
259
260	TRACE(("pipeline: entered\n"));
261	while (readtoken() == TNOT) {
262		TRACE(("pipeline: TNOT recognized\n"));
263		negate = !negate;
264	}
265	tokpushback++;
266	n1 = command();
267	if (readtoken() == TPIPE) {
268		pipenode = (union node *)stalloc(sizeof (struct npipe));
269		pipenode->type = NPIPE;
270		pipenode->npipe.backgnd = 0;
271		lp = (struct nodelist *)stalloc(sizeof (struct nodelist));
272		pipenode->npipe.cmdlist = lp;
273		lp->n = n1;
274		do {
275			prev = lp;
276			lp = (struct nodelist *)stalloc(sizeof (struct nodelist));
277			lp->n = command();
278			prev->next = lp;
279		} while (readtoken() == TPIPE);
280		lp->next = NULL;
281		n1 = pipenode;
282	}
283	tokpushback++;
284	if (negate) {
285		notnode = (union node *)stalloc(sizeof(struct nnot));
286		notnode->type = NNOT;
287		notnode->nnot.com = n1;
288		n1 = notnode;
289	}
290	return n1;
291}
292
293
294
295STATIC union node *
296command() {
297	union node *n1, *n2;
298	union node *ap, **app;
299	union node *cp, **cpp;
300	union node *redir, **rpp;
301	int t;
302
303	checkkwd = 2;
304	redir = NULL;
305	n1 = NULL;
306	rpp = &redir;
307
308	/* Check for redirection which may precede command */
309	while (readtoken() == TREDIR) {
310		*rpp = n2 = redirnode;
311		rpp = &n2->nfile.next;
312		parsefname();
313	}
314	tokpushback++;
315
316	switch (readtoken()) {
317	case TIF:
318		n1 = (union node *)stalloc(sizeof (struct nif));
319		n1->type = NIF;
320		n1->nif.test = list(0);
321		if (readtoken() != TTHEN)
322			synexpect(TTHEN);
323		n1->nif.ifpart = list(0);
324		n2 = n1;
325		while (readtoken() == TELIF) {
326			n2->nif.elsepart = (union node *)stalloc(sizeof (struct nif));
327			n2 = n2->nif.elsepart;
328			n2->type = NIF;
329			n2->nif.test = list(0);
330			if (readtoken() != TTHEN)
331				synexpect(TTHEN);
332			n2->nif.ifpart = list(0);
333		}
334		if (lasttoken == TELSE)
335			n2->nif.elsepart = list(0);
336		else {
337			n2->nif.elsepart = NULL;
338			tokpushback++;
339		}
340		if (readtoken() != TFI)
341			synexpect(TFI);
342		checkkwd = 1;
343		break;
344	case TWHILE:
345	case TUNTIL: {
346		int got;
347		n1 = (union node *)stalloc(sizeof (struct nbinary));
348		n1->type = (lasttoken == TWHILE)? NWHILE : NUNTIL;
349		n1->nbinary.ch1 = list(0);
350		if ((got=readtoken()) != TDO) {
351TRACE(("expecting DO got %s %s\n", tokname[got], got == TWORD ? wordtext : ""));
352			synexpect(TDO);
353		}
354		n1->nbinary.ch2 = list(0);
355		if (readtoken() != TDONE)
356			synexpect(TDONE);
357		checkkwd = 1;
358		break;
359	}
360	case TFOR:
361		if (readtoken() != TWORD || quoteflag || ! goodname(wordtext))
362			synerror("Bad for loop variable");
363		n1 = (union node *)stalloc(sizeof (struct nfor));
364		n1->type = NFOR;
365		n1->nfor.var = wordtext;
366		if (readtoken() == TWORD && ! quoteflag && equal(wordtext, "in")) {
367			app = &ap;
368			while (readtoken() == TWORD) {
369				n2 = (union node *)stalloc(sizeof (struct narg));
370				n2->type = NARG;
371				n2->narg.text = wordtext;
372				n2->narg.backquote = backquotelist;
373				*app = n2;
374				app = &n2->narg.next;
375			}
376			*app = NULL;
377			n1->nfor.args = ap;
378			if (lasttoken != TNL && lasttoken != TSEMI)
379				synexpect(-1);
380		} else {
381#ifndef GDB_HACK
382			static const char argvars[5] = {CTLVAR, VSNORMAL|VSQUOTE,
383								   '@', '=', '\0'};
384#endif
385			n2 = (union node *)stalloc(sizeof (struct narg));
386			n2->type = NARG;
387			n2->narg.text = (char *)argvars;
388			n2->narg.backquote = NULL;
389			n2->narg.next = NULL;
390			n1->nfor.args = n2;
391			/*
392			 * Newline or semicolon here is optional (but note
393			 * that the original Bourne shell only allowed NL).
394			 */
395			if (lasttoken != TNL && lasttoken != TSEMI)
396				tokpushback++;
397		}
398		checkkwd = 2;
399		if ((t = readtoken()) == TDO)
400			t = TDONE;
401		else if (t == TBEGIN)
402			t = TEND;
403		else
404			synexpect(-1);
405		n1->nfor.body = list(0);
406		if (readtoken() != t)
407			synexpect(t);
408		checkkwd = 1;
409		break;
410	case TCASE:
411		n1 = (union node *)stalloc(sizeof (struct ncase));
412		n1->type = NCASE;
413		if (readtoken() != TWORD)
414			synexpect(TWORD);
415		n1->ncase.expr = n2 = (union node *)stalloc(sizeof (struct narg));
416		n2->type = NARG;
417		n2->narg.text = wordtext;
418		n2->narg.backquote = backquotelist;
419		n2->narg.next = NULL;
420		while (readtoken() == TNL);
421		if (lasttoken != TWORD || ! equal(wordtext, "in"))
422			synerror("expecting \"in\"");
423		cpp = &n1->ncase.cases;
424		noaliases = 1;	/* turn off alias expansion */
425		checkkwd = 2, readtoken();
426		do {
427			*cpp = cp = (union node *)stalloc(sizeof (struct nclist));
428			cp->type = NCLIST;
429			app = &cp->nclist.pattern;
430			for (;;) {
431				*app = ap = (union node *)stalloc(sizeof (struct narg));
432				ap->type = NARG;
433				ap->narg.text = wordtext;
434				ap->narg.backquote = backquotelist;
435				if (checkkwd = 2, readtoken() != TPIPE)
436					break;
437				app = &ap->narg.next;
438				readtoken();
439			}
440			ap->narg.next = NULL;
441			if (lasttoken != TRP)
442				noaliases = 0, synexpect(TRP);
443			cp->nclist.body = list(0);
444
445			checkkwd = 2;
446			if ((t = readtoken()) != TESAC) {
447				if (t != TENDCASE)
448					noaliases = 0, synexpect(TENDCASE);
449				else
450					checkkwd = 2, readtoken();
451			}
452			cpp = &cp->nclist.next;
453		} while(lasttoken != TESAC);
454		noaliases = 0;	/* reset alias expansion */
455		*cpp = NULL;
456		checkkwd = 1;
457		break;
458	case TLP:
459		n1 = (union node *)stalloc(sizeof (struct nredir));
460		n1->type = NSUBSHELL;
461		n1->nredir.n = list(0);
462		n1->nredir.redirect = NULL;
463		if (readtoken() != TRP)
464			synexpect(TRP);
465		checkkwd = 1;
466		break;
467	case TBEGIN:
468		n1 = list(0);
469		if (readtoken() != TEND)
470			synexpect(TEND);
471		checkkwd = 1;
472		break;
473	/* Handle an empty command like other simple commands.  */
474	case TSEMI:
475		/*
476		 * An empty command before a ; doesn't make much sense, and
477		 * should certainly be disallowed in the case of `if ;'.
478		 */
479		if (!redir)
480			synexpect(-1);
481	case TAND:
482	case TOR:
483	case TNL:
484	case TEOF:
485	case TWORD:
486	case TRP:
487		tokpushback++;
488		return simplecmd(rpp, redir);
489	default:
490		synexpect(-1);
491	}
492
493	/* Now check for redirection which may follow command */
494	while (readtoken() == TREDIR) {
495		*rpp = n2 = redirnode;
496		rpp = &n2->nfile.next;
497		parsefname();
498	}
499	tokpushback++;
500	*rpp = NULL;
501	if (redir) {
502		if (n1->type != NSUBSHELL) {
503			n2 = (union node *)stalloc(sizeof (struct nredir));
504			n2->type = NREDIR;
505			n2->nredir.n = n1;
506			n1 = n2;
507		}
508		n1->nredir.redirect = redir;
509	}
510	return n1;
511}
512
513
514STATIC union node *
515simplecmd(rpp, redir)
516	union node **rpp, *redir;
517	{
518	union node *args, **app;
519	union node **orig_rpp = rpp;
520	union node *n = NULL;
521
522	/* If we don't have any redirections already, then we must reset */
523	/* rpp to be the address of the local redir variable.  */
524	if (redir == 0)
525		rpp = &redir;
526
527	args = NULL;
528	app = &args;
529	/*
530	 * We save the incoming value, because we need this for shell
531	 * functions.  There can not be a redirect or an argument between
532	 * the function name and the open parenthesis.
533	 */
534	orig_rpp = rpp;
535
536	for (;;) {
537		if (readtoken() == TWORD) {
538			n = (union node *)stalloc(sizeof (struct narg));
539			n->type = NARG;
540			n->narg.text = wordtext;
541			n->narg.backquote = backquotelist;
542			*app = n;
543			app = &n->narg.next;
544		} else if (lasttoken == TREDIR) {
545			*rpp = n = redirnode;
546			rpp = &n->nfile.next;
547			parsefname();	/* read name of redirection file */
548		} else if (lasttoken == TLP && app == &args->narg.next
549					    && rpp == orig_rpp) {
550			/* We have a function */
551			if (readtoken() != TRP)
552				synexpect(TRP);
553#ifdef notdef
554			if (! goodname(n->narg.text))
555				synerror("Bad function name");
556#endif
557			n->type = NDEFUN;
558			n->narg.next = command();
559			return n;
560		} else {
561			tokpushback++;
562			break;
563		}
564	}
565	*app = NULL;
566	*rpp = NULL;
567	n = (union node *)stalloc(sizeof (struct ncmd));
568	n->type = NCMD;
569	n->ncmd.backgnd = 0;
570	n->ncmd.args = args;
571	n->ncmd.redirect = redir;
572	return n;
573}
574
575STATIC union node *
576makename() {
577	union node *n;
578
579	n = (union node *)stalloc(sizeof (struct narg));
580	n->type = NARG;
581	n->narg.next = NULL;
582	n->narg.text = wordtext;
583	n->narg.backquote = backquotelist;
584	return n;
585}
586
587void fixredir(n, text, err)
588	union node *n;
589	const char *text;
590	int err;
591	{
592	TRACE(("Fix redir %s %d\n", text, err));
593	if (!err)
594		n->ndup.vname = NULL;
595
596	if (is_digit(text[0]) && text[1] == '\0')
597		n->ndup.dupfd = digit_val(text[0]);
598	else if (text[0] == '-' && text[1] == '\0')
599		n->ndup.dupfd = -1;
600	else {
601
602		if (err)
603			synerror("Bad fd number");
604		else
605			n->ndup.vname = makename();
606	}
607}
608
609
610STATIC void
611parsefname() {
612	union node *n = redirnode;
613
614	if (readtoken() != TWORD)
615		synexpect(-1);
616	if (n->type == NHERE) {
617		struct heredoc *here = heredoc;
618		struct heredoc *p;
619		int i;
620
621		if (quoteflag == 0)
622			n->type = NXHERE;
623		TRACE(("Here document %d\n", n->type));
624		if (here->striptabs) {
625			while (*wordtext == '\t')
626				wordtext++;
627		}
628		if (! noexpand(wordtext) || (i = strlen(wordtext)) == 0 || i > EOFMARKLEN)
629			synerror("Illegal eof marker for << redirection");
630		rmescapes(wordtext);
631		here->eofmark = wordtext;
632		here->next = NULL;
633		if (heredoclist == NULL)
634			heredoclist = here;
635		else {
636			for (p = heredoclist ; p->next ; p = p->next);
637			p->next = here;
638		}
639	} else if (n->type == NTOFD || n->type == NFROMFD) {
640		fixredir(n, wordtext, 0);
641	} else {
642		n->nfile.fname = makename();
643	}
644}
645
646
647/*
648 * Input any here documents.
649 */
650
651STATIC void
652parseheredoc() {
653	struct heredoc *here;
654	union node *n;
655
656	while (heredoclist) {
657		here = heredoclist;
658		heredoclist = here->next;
659		if (needprompt) {
660			setprompt(2);
661			needprompt = 0;
662		}
663		readtoken1(pgetc(), here->here->type == NHERE? SQSYNTAX : DQSYNTAX,
664				here->eofmark, here->striptabs);
665		n = (union node *)stalloc(sizeof (struct narg));
666		n->narg.type = NARG;
667		n->narg.next = NULL;
668		n->narg.text = wordtext;
669		n->narg.backquote = backquotelist;
670		here->here->nhere.doc = n;
671	}
672}
673
674STATIC int
675peektoken() {
676	int t;
677
678	t = readtoken();
679	tokpushback++;
680	return (t);
681}
682
683STATIC int
684readtoken() {
685	int t;
686	int savecheckkwd = checkkwd;
687	struct alias *ap;
688#ifdef DEBUG
689	int alreadyseen = tokpushback;
690#endif
691
692	top:
693	t = xxreadtoken();
694
695	if (checkkwd) {
696		/*
697		 * eat newlines
698		 */
699		if (checkkwd == 2) {
700			checkkwd = 0;
701			while (t == TNL) {
702				parseheredoc();
703				t = xxreadtoken();
704			}
705		} else
706			checkkwd = 0;
707		/*
708		 * check for keywords and aliases
709		 */
710		if (t == TWORD && !quoteflag)
711		{
712			char * const *pp;
713
714			for (pp = (char **)parsekwd; *pp; pp++) {
715				if (**pp == *wordtext && equal(*pp, wordtext))
716				{
717					lasttoken = t = pp - parsekwd + KWDOFFSET;
718					TRACE(("keyword %s recognized\n", tokname[t]));
719					goto out;
720				}
721			}
722			if (noaliases == 0 &&
723			    (ap = lookupalias(wordtext, 1)) != NULL) {
724				pushstring(ap->val, strlen(ap->val), ap);
725				checkkwd = savecheckkwd;
726				goto top;
727			}
728		}
729out:
730		checkkwd = 0;
731	}
732#ifdef DEBUG
733	if (!alreadyseen)
734	    TRACE(("token %s %s\n", tokname[t], t == TWORD ? wordtext : ""));
735	else
736	    TRACE(("reread token %s %s\n", tokname[t], t == TWORD ? wordtext : ""));
737#endif
738	return (t);
739}
740
741
742/*
743 * Read the next input token.
744 * If the token is a word, we set backquotelist to the list of cmds in
745 *	backquotes.  We set quoteflag to true if any part of the word was
746 *	quoted.
747 * If the token is TREDIR, then we set redirnode to a structure containing
748 *	the redirection.
749 * In all cases, the variable startlinno is set to the number of the line
750 *	on which the token starts.
751 *
752 * [Change comment:  here documents and internal procedures]
753 * [Readtoken shouldn't have any arguments.  Perhaps we should make the
754 *  word parsing code into a separate routine.  In this case, readtoken
755 *  doesn't need to have any internal procedures, but parseword does.
756 *  We could also make parseoperator in essence the main routine, and
757 *  have parseword (readtoken1?) handle both words and redirection.]
758 */
759
760#define RETURN(token)	return lasttoken = token
761
762STATIC int
763xxreadtoken() {
764	int c;
765
766	if (tokpushback) {
767		tokpushback = 0;
768		return lasttoken;
769	}
770	if (needprompt) {
771		setprompt(2);
772		needprompt = 0;
773	}
774	startlinno = plinno;
775	for (;;) {	/* until token or start of word found */
776		c = pgetc_macro();
777		if (c == ' ' || c == '\t')
778			continue;		/* quick check for white space first */
779		switch (c) {
780		case ' ': case '\t':
781			continue;
782		case '#':
783			while ((c = pgetc()) != '\n' && c != PEOF);
784			pungetc();
785			continue;
786		case '\\':
787			if (pgetc() == '\n') {
788				startlinno = ++plinno;
789				if (doprompt)
790					setprompt(2);
791				else
792					setprompt(0);
793				continue;
794			}
795			pungetc();
796			goto breakloop;
797		case '\n':
798			plinno++;
799			needprompt = doprompt;
800			RETURN(TNL);
801		case PEOF:
802			RETURN(TEOF);
803		case '&':
804			if (pgetc() == '&')
805				RETURN(TAND);
806			pungetc();
807			RETURN(TBACKGND);
808		case '|':
809			if (pgetc() == '|')
810				RETURN(TOR);
811			pungetc();
812			RETURN(TPIPE);
813		case ';':
814			if (pgetc() == ';')
815				RETURN(TENDCASE);
816			pungetc();
817			RETURN(TSEMI);
818		case '(':
819			RETURN(TLP);
820		case ')':
821			RETURN(TRP);
822		default:
823			goto breakloop;
824		}
825	}
826breakloop:
827	return readtoken1(c, BASESYNTAX, (char *)NULL, 0);
828#undef RETURN
829}
830
831
832
833/*
834 * If eofmark is NULL, read a word or a redirection symbol.  If eofmark
835 * is not NULL, read a here document.  In the latter case, eofmark is the
836 * word which marks the end of the document and striptabs is true if
837 * leading tabs should be stripped from the document.  The argument firstc
838 * is the first character of the input token or document.
839 *
840 * Because C does not have internal subroutines, I have simulated them
841 * using goto's to implement the subroutine linkage.  The following macros
842 * will run code that appears at the end of readtoken1.
843 */
844
845#define CHECKEND()	{goto checkend; checkend_return:;}
846#define PARSEREDIR()	{goto parseredir; parseredir_return:;}
847#define PARSESUB()	{goto parsesub; parsesub_return:;}
848#define PARSEBACKQOLD()	{oldstyle = 1; goto parsebackq; parsebackq_oldreturn:;}
849#define PARSEBACKQNEW()	{oldstyle = 0; goto parsebackq; parsebackq_newreturn:;}
850#define	PARSEARITH()	{goto parsearith; parsearith_return:;}
851
852STATIC int
853readtoken1(firstc, syntax, eofmark, striptabs)
854	int firstc;
855	char const *syntax;
856	char *eofmark;
857	int striptabs;
858	{
859	int c = firstc;
860	char *out;
861	int len;
862	char line[EOFMARKLEN + 1];
863	struct nodelist *bqlist;
864	int quotef;
865	int dblquote;
866	int varnest;	/* levels of variables expansion */
867	int arinest;	/* levels of arithmetic expansion */
868	int parenlevel;	/* levels of parens in arithmetic */
869	int oldstyle;
870	char const *prevsyntax;	/* syntax before arithmetic */
871	int synentry;
872#if __GNUC__
873	/* Avoid longjmp clobbering */
874	(void) &out;
875	(void) &quotef;
876	(void) &dblquote;
877	(void) &varnest;
878	(void) &arinest;
879	(void) &parenlevel;
880	(void) &oldstyle;
881	(void) &prevsyntax;
882	(void) &syntax;
883	(void) &synentry;
884#endif
885
886	startlinno = plinno;
887	dblquote = 0;
888	if (syntax == DQSYNTAX)
889		dblquote = 1;
890	quotef = 0;
891	bqlist = NULL;
892	varnest = 0;
893	arinest = 0;
894	parenlevel = 0;
895
896	STARTSTACKSTR(out);
897	loop: {	/* for each line, until end of word */
898#if ATTY
899		if (c == '\034' && doprompt
900		 && attyset() && ! equal(termval(), "emacs")) {
901			attyline();
902			if (syntax == BASESYNTAX)
903				return readtoken();
904			c = pgetc();
905			goto loop;
906		}
907#endif
908		CHECKEND();	/* set c to PEOF if at end of here document */
909		for (;;) {	/* until end of line or end of word */
910			CHECKSTRSPACE(3, out);	/* permit 3 calls to USTPUTC */
911
912			if (c < 0 && c != PEOF)
913				synentry = CWORD;
914			else
915				synentry = syntax[c];
916
917			switch(synentry) {
918			case CNL:	/* '\n' */
919				if (syntax == BASESYNTAX)
920					goto endword;	/* exit outer loop */
921				USTPUTC(c, out);
922				plinno++;
923				if (doprompt)
924					setprompt(2);
925				else
926					setprompt(0);
927				c = pgetc();
928				goto loop;		/* continue outer loop */
929			case CWORD:
930				USTPUTC(c, out);
931				break;
932			case CCTL:
933				if (eofmark == NULL || dblquote)
934					USTPUTC(CTLESC, out);
935				USTPUTC(c, out);
936				break;
937			case CBACK:	/* backslash */
938				c = pgetc();
939				if (c == PEOF) {
940					USTPUTC('\\', out);
941					pungetc();
942				} else if (c == '\n') {
943					if (doprompt)
944						setprompt(2);
945					else
946						setprompt(0);
947				} else {
948					if (dblquote && c != '\\' &&
949					    c != '`' && c != '$' &&
950					    (c != '"' || eofmark != NULL))
951						USTPUTC('\\', out);
952					if (c >= 0 && SQSYNTAX[c] == CCTL)
953						USTPUTC(CTLESC, out);
954					else if (eofmark == NULL)
955						USTPUTC(CTLQUOTEMARK, out);
956					USTPUTC(c, out);
957					quotef++;
958				}
959				break;
960			case CSQUOTE:
961				if (eofmark == NULL)
962					USTPUTC(CTLQUOTEMARK, out);
963				syntax = SQSYNTAX;
964				break;
965			case CDQUOTE:
966				if (eofmark == NULL)
967					USTPUTC(CTLQUOTEMARK, out);
968				syntax = DQSYNTAX;
969				dblquote = 1;
970				break;
971			case CENDQUOTE:
972				if (eofmark != NULL && arinest == 0 &&
973				    varnest == 0) {
974					USTPUTC(c, out);
975				} else {
976					if (arinest) {
977						syntax = ARISYNTAX;
978						dblquote = 0;
979					} else if (eofmark == NULL) {
980						syntax = BASESYNTAX;
981						dblquote = 0;
982					}
983					quotef++;
984				}
985				break;
986			case CVAR:	/* '$' */
987				PARSESUB();		/* parse substitution */
988				break;
989			case CENDVAR:	/* '}' */
990				if (varnest > 0) {
991					varnest--;
992					USTPUTC(CTLENDVAR, out);
993				} else {
994					USTPUTC(c, out);
995				}
996				break;
997			case CLP:	/* '(' in arithmetic */
998				parenlevel++;
999				USTPUTC(c, out);
1000				break;
1001			case CRP:	/* ')' in arithmetic */
1002				if (parenlevel > 0) {
1003					USTPUTC(c, out);
1004					--parenlevel;
1005				} else {
1006					if (pgetc() == ')') {
1007						if (--arinest == 0) {
1008							USTPUTC(CTLENDARI, out);
1009							syntax = prevsyntax;
1010							if (syntax == DQSYNTAX)
1011								dblquote = 1;
1012							else
1013								dblquote = 0;
1014						} else
1015							USTPUTC(')', out);
1016					} else {
1017						/*
1018						 * unbalanced parens
1019						 *  (don't 2nd guess - no error)
1020						 */
1021						pungetc();
1022						USTPUTC(')', out);
1023					}
1024				}
1025				break;
1026			case CBQUOTE:	/* '`' */
1027				PARSEBACKQOLD();
1028				break;
1029			case CEOF:
1030				goto endword;		/* exit outer loop */
1031			default:
1032				if (varnest == 0)
1033					goto endword;	/* exit outer loop */
1034				USTPUTC(c, out);
1035			}
1036			c = pgetc_macro();
1037		}
1038	}
1039endword:
1040	if (syntax == ARISYNTAX)
1041		synerror("Missing '))'");
1042	if (syntax != BASESYNTAX && ! parsebackquote && eofmark == NULL)
1043		synerror("Unterminated quoted string");
1044	if (varnest != 0) {
1045		startlinno = plinno;
1046		synerror("Missing '}'");
1047	}
1048	USTPUTC('\0', out);
1049	len = out - stackblock();
1050	out = stackblock();
1051	if (eofmark == NULL) {
1052		if ((c == '>' || c == '<')
1053		 && quotef == 0
1054		 && len <= 2
1055		 && (*out == '\0' || is_digit(*out))) {
1056			PARSEREDIR();
1057			return lasttoken = TREDIR;
1058		} else {
1059			pungetc();
1060		}
1061	}
1062	quoteflag = quotef;
1063	backquotelist = bqlist;
1064	grabstackblock(len);
1065	wordtext = out;
1066	return lasttoken = TWORD;
1067/* end of readtoken routine */
1068
1069
1070
1071/*
1072 * Check to see whether we are at the end of the here document.  When this
1073 * is called, c is set to the first character of the next input line.  If
1074 * we are at the end of the here document, this routine sets the c to PEOF.
1075 */
1076
1077checkend: {
1078	if (eofmark) {
1079		if (striptabs) {
1080			while (c == '\t')
1081				c = pgetc();
1082		}
1083		if (c == *eofmark) {
1084			if (pfgets(line, sizeof line) != NULL) {
1085				char *p, *q;
1086
1087				p = line;
1088				for (q = eofmark + 1 ; *q && *p == *q ; p++, q++);
1089				if (*p == '\n' && *q == '\0') {
1090					c = PEOF;
1091					plinno++;
1092					needprompt = doprompt;
1093				} else {
1094					pushstring(line, strlen(line), NULL);
1095				}
1096			}
1097		}
1098	}
1099	goto checkend_return;
1100}
1101
1102
1103/*
1104 * Parse a redirection operator.  The variable "out" points to a string
1105 * specifying the fd to be redirected.  The variable "c" contains the
1106 * first character of the redirection operator.
1107 */
1108
1109parseredir: {
1110	char fd = *out;
1111	union node *np;
1112
1113	np = (union node *)stalloc(sizeof (struct nfile));
1114	if (c == '>') {
1115		np->nfile.fd = 1;
1116		c = pgetc();
1117		if (c == '>')
1118			np->type = NAPPEND;
1119		else if (c == '&')
1120			np->type = NTOFD;
1121		else {
1122			np->type = NTO;
1123			pungetc();
1124		}
1125	} else {	/* c == '<' */
1126		np->nfile.fd = 0;
1127		c = pgetc();
1128		if (c == '<') {
1129			if (sizeof (struct nfile) != sizeof (struct nhere)) {
1130				np = (union node *)stalloc(sizeof (struct nhere));
1131				np->nfile.fd = 0;
1132			}
1133			np->type = NHERE;
1134			heredoc = (struct heredoc *)stalloc(sizeof (struct heredoc));
1135			heredoc->here = np;
1136			if ((c = pgetc()) == '-') {
1137				heredoc->striptabs = 1;
1138			} else {
1139				heredoc->striptabs = 0;
1140				pungetc();
1141			}
1142		} else if (c == '&')
1143			np->type = NFROMFD;
1144		else {
1145			np->type = NFROM;
1146			pungetc();
1147		}
1148	}
1149	if (fd != '\0')
1150		np->nfile.fd = digit_val(fd);
1151	redirnode = np;
1152	goto parseredir_return;
1153}
1154
1155
1156/*
1157 * Parse a substitution.  At this point, we have read the dollar sign
1158 * and nothing else.
1159 */
1160
1161parsesub: {
1162	int subtype;
1163	int typeloc;
1164	int flags;
1165	char *p;
1166#ifndef GDB_HACK
1167	static const char types[] = "}-+?=";
1168#endif
1169       int bracketed_name = 0; /* used to handle ${[0-9]*} variables */
1170
1171	c = pgetc();
1172	if (c != '(' && c != '{' && !is_name(c) && !is_special(c)) {
1173		USTPUTC('$', out);
1174		pungetc();
1175	} else if (c == '(') {	/* $(command) or $((arith)) */
1176		if (pgetc() == '(') {
1177			PARSEARITH();
1178		} else {
1179			pungetc();
1180			PARSEBACKQNEW();
1181		}
1182	} else {
1183		USTPUTC(CTLVAR, out);
1184		typeloc = out - stackblock();
1185		USTPUTC(VSNORMAL, out);
1186		subtype = VSNORMAL;
1187		if (c == '{') {
1188			bracketed_name = 1;
1189			c = pgetc();
1190			if (c == '#') {
1191				if ((c = pgetc()) == '}')
1192					c = '#';
1193				else
1194					subtype = VSLENGTH;
1195			}
1196			else
1197				subtype = 0;
1198		}
1199		if (is_name(c)) {
1200			do {
1201				STPUTC(c, out);
1202				c = pgetc();
1203			} while (is_in_name(c));
1204		} else if (is_digit(c)) {
1205			if (bracketed_name) {
1206				do {
1207					STPUTC(c, out);
1208					c = pgetc();
1209				} while (is_digit(c));
1210			} else {
1211				STPUTC(c, out);
1212				c = pgetc();
1213			}
1214		} else {
1215			if (! is_special(c))
1216badsub:				synerror("Bad substitution");
1217			USTPUTC(c, out);
1218			c = pgetc();
1219		}
1220		STPUTC('=', out);
1221		flags = 0;
1222		if (subtype == 0) {
1223			switch (c) {
1224			case ':':
1225				flags = VSNUL;
1226				c = pgetc();
1227				/*FALLTHROUGH*/
1228			default:
1229				p = strchr(types, c);
1230				if (p == NULL)
1231					goto badsub;
1232				subtype = p - types + VSNORMAL;
1233				break;
1234			case '%':
1235			case '#':
1236				{
1237					int cc = c;
1238					subtype = c == '#' ? VSTRIMLEFT :
1239							     VSTRIMRIGHT;
1240					c = pgetc();
1241					if (c == cc)
1242						subtype++;
1243					else
1244						pungetc();
1245					break;
1246				}
1247			}
1248		} else {
1249			pungetc();
1250		}
1251		if (subtype != VSLENGTH && (dblquote || arinest))
1252			flags |= VSQUOTE;
1253		*(stackblock() + typeloc) = subtype | flags;
1254		if (subtype != VSNORMAL)
1255			varnest++;
1256	}
1257	goto parsesub_return;
1258}
1259
1260
1261/*
1262 * Called to parse command substitutions.  Newstyle is set if the command
1263 * is enclosed inside $(...); nlpp is a pointer to the head of the linked
1264 * list of commands (passed by reference), and savelen is the number of
1265 * characters on the top of the stack which must be preserved.
1266 */
1267
1268parsebackq: {
1269	struct nodelist **nlpp;
1270	int savepbq;
1271	union node *n;
1272	char *volatile str;
1273	struct jmploc jmploc;
1274	struct jmploc *volatile savehandler;
1275	int savelen;
1276	int saveprompt;
1277#if __GNUC__
1278	/* Avoid longjmp clobbering */
1279	(void) &saveprompt;
1280#endif
1281
1282	savepbq = parsebackquote;
1283	if (setjmp(jmploc.loc)) {
1284		if (str)
1285			ckfree(str);
1286		parsebackquote = 0;
1287		handler = savehandler;
1288		longjmp(handler->loc, 1);
1289	}
1290	INTOFF;
1291	str = NULL;
1292	savelen = out - stackblock();
1293	if (savelen > 0) {
1294		str = ckmalloc(savelen);
1295		memcpy(str, stackblock(), savelen);
1296	}
1297	savehandler = handler;
1298	handler = &jmploc;
1299	INTON;
1300        if (oldstyle) {
1301                /* We must read until the closing backquote, giving special
1302                   treatment to some slashes, and then push the string and
1303                   reread it as input, interpreting it normally.  */
1304                char *out;
1305                int c;
1306                int savelen;
1307                char *str;
1308
1309
1310                STARTSTACKSTR(out);
1311		for (;;) {
1312			if (needprompt) {
1313				setprompt(2);
1314				needprompt = 0;
1315			}
1316			switch (c = pgetc()) {
1317			case '`':
1318				goto done;
1319
1320			case '\\':
1321                                if ((c = pgetc()) == '\n') {
1322					plinno++;
1323					if (doprompt)
1324						setprompt(2);
1325					else
1326						setprompt(0);
1327					/*
1328					 * If eating a newline, avoid putting
1329					 * the newline into the new character
1330					 * stream (via the STPUTC after the
1331					 * switch).
1332					 */
1333					continue;
1334				}
1335                                if (c != '\\' && c != '`' && c != '$'
1336                                    && (!dblquote || c != '"'))
1337                                        STPUTC('\\', out);
1338				break;
1339
1340			case '\n':
1341				plinno++;
1342				needprompt = doprompt;
1343				break;
1344
1345			case PEOF:
1346			        startlinno = plinno;
1347				synerror("EOF in backquote substitution");
1348 				break;
1349
1350			default:
1351				break;
1352			}
1353			STPUTC(c, out);
1354                }
1355done:
1356                STPUTC('\0', out);
1357                savelen = out - stackblock();
1358                if (savelen > 0) {
1359                        str = ckmalloc(savelen);
1360                        memcpy(str, stackblock(), savelen);
1361			setinputstring(str, 1);
1362                }
1363        }
1364	nlpp = &bqlist;
1365	while (*nlpp)
1366		nlpp = &(*nlpp)->next;
1367	*nlpp = (struct nodelist *)stalloc(sizeof (struct nodelist));
1368	(*nlpp)->next = NULL;
1369	parsebackquote = oldstyle;
1370
1371	if (oldstyle) {
1372		saveprompt = doprompt;
1373		doprompt = 0;
1374	}
1375
1376	n = list(0);
1377
1378	if (oldstyle)
1379		doprompt = saveprompt;
1380	else {
1381		if (readtoken() != TRP)
1382			synexpect(TRP);
1383	}
1384
1385	(*nlpp)->n = n;
1386        if (oldstyle) {
1387		/*
1388		 * Start reading from old file again, ignoring any pushed back
1389		 * tokens left from the backquote parsing
1390		 */
1391                popfile();
1392		tokpushback = 0;
1393	}
1394	while (stackblocksize() <= savelen)
1395		growstackblock();
1396	STARTSTACKSTR(out);
1397	if (str) {
1398		memcpy(out, str, savelen);
1399		STADJUST(savelen, out);
1400		INTOFF;
1401		ckfree(str);
1402		str = NULL;
1403		INTON;
1404	}
1405	parsebackquote = savepbq;
1406	handler = savehandler;
1407	if (arinest || dblquote)
1408		USTPUTC(CTLBACKQ | CTLQUOTE, out);
1409	else
1410		USTPUTC(CTLBACKQ, out);
1411	if (oldstyle)
1412		goto parsebackq_oldreturn;
1413	else
1414		goto parsebackq_newreturn;
1415}
1416
1417/*
1418 * Parse an arithmetic expansion (indicate start of one and set state)
1419 */
1420parsearith: {
1421
1422	if (++arinest == 1) {
1423		prevsyntax = syntax;
1424		syntax = ARISYNTAX;
1425		USTPUTC(CTLARI, out);
1426		if (dblquote)
1427			USTPUTC('"',out);
1428		else
1429			USTPUTC(' ',out);
1430	} else {
1431		/*
1432		 * we collapse embedded arithmetic expansion to
1433		 * parenthesis, which should be equivalent
1434		 */
1435		USTPUTC('(', out);
1436	}
1437	goto parsearith_return;
1438}
1439
1440} /* end of readtoken */
1441
1442
1443
1444#ifdef mkinit
1445RESET {
1446	tokpushback = 0;
1447	checkkwd = 0;
1448}
1449#endif
1450
1451/*
1452 * Returns true if the text contains nothing to expand (no dollar signs
1453 * or backquotes).
1454 */
1455
1456STATIC int
1457noexpand(text)
1458	char *text;
1459	{
1460	char *p;
1461	char c;
1462
1463	p = text;
1464	while ((c = *p++) != '\0') {
1465		if ( c == CTLQUOTEMARK)
1466			continue;
1467		if (c == CTLESC)
1468			p++;
1469		else if (c >= 0 && BASESYNTAX[(int)c] == CCTL)
1470			return 0;
1471	}
1472	return 1;
1473}
1474
1475
1476/*
1477 * Return true if the argument is a legal variable name (a letter or
1478 * underscore followed by zero or more letters, underscores, and digits).
1479 */
1480
1481int
1482goodname(name)
1483	char *name;
1484	{
1485	char *p;
1486
1487	p = name;
1488	if (! is_name(*p))
1489		return 0;
1490	while (*++p) {
1491		if (! is_in_name(*p))
1492			return 0;
1493	}
1494	return 1;
1495}
1496
1497
1498/*
1499 * Called when an unexpected token is read during the parse.  The argument
1500 * is the token that is expected, or -1 if more than one type of token can
1501 * occur at this point.
1502 */
1503
1504STATIC void
1505synexpect(token)
1506	int token;
1507{
1508	char msg[64];
1509
1510	if (token >= 0) {
1511		fmtstr(msg, 64, "%s unexpected (expecting %s)",
1512			tokname[lasttoken], tokname[token]);
1513	} else {
1514		fmtstr(msg, 64, "%s unexpected", tokname[lasttoken]);
1515	}
1516	synerror(msg);
1517}
1518
1519
1520STATIC void
1521synerror(msg)
1522	char *msg;
1523	{
1524	if (commandname)
1525		outfmt(&errout, "%s: %d: ", commandname, startlinno);
1526	outfmt(&errout, "Syntax error: %s\n", msg);
1527	error((char *)NULL);
1528}
1529
1530STATIC void
1531setprompt(which)
1532	int which;
1533	{
1534	whichprompt = which;
1535
1536#ifndef NO_HISTORY
1537	if (!el)
1538#endif
1539		out2str(getprompt(NULL));
1540}
1541
1542/*
1543 * called by editline -- any expansions to the prompt
1544 *    should be added here.
1545 */
1546char *
1547getprompt(unused)
1548	void *unused __unused;
1549{
1550	switch (whichprompt) {
1551	case 0:
1552		return "";
1553	case 1:
1554		return ps1val();
1555	case 2:
1556		return ps2val();
1557	default:
1558		return "<internal prompt error>";
1559	}
1560}
1561