xref: /netbsd-src/usr.bin/sed/compile.c (revision 481fca6e59249d8ffcf24fef7cfbe7b131bfb080)
1 /*	$NetBSD: compile.c,v 1.19 1998/12/19 22:12:11 christos Exp $	*/
2 
3 /*-
4  * Copyright (c) 1992 Diomidis Spinellis.
5  * Copyright (c) 1992, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This code is derived from software contributed to Berkeley by
9  * Diomidis Spinellis of Imperial College, University of London.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  * 3. All advertising materials mentioning features or use of this software
20  *    must display the following acknowledgement:
21  *	This product includes software developed by the University of
22  *	California, Berkeley and its contributors.
23  * 4. Neither the name of the University nor the names of its contributors
24  *    may be used to endorse or promote products derived from this software
25  *    without specific prior written permission.
26  *
27  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
37  * SUCH DAMAGE.
38  */
39 
40 #include <sys/cdefs.h>
41 #ifndef lint
42 #if 0
43 static char sccsid[] = "@(#)compile.c	8.2 (Berkeley) 4/28/95";
44 #else
45 __RCSID("$NetBSD: compile.c,v 1.19 1998/12/19 22:12:11 christos Exp $");
46 #endif
47 #endif /* not lint */
48 
49 #include <sys/types.h>
50 #include <sys/stat.h>
51 
52 #include <ctype.h>
53 #include <errno.h>
54 #include <fcntl.h>
55 #include <limits.h>
56 #include <regex.h>
57 #include <stdio.h>
58 #include <stdlib.h>
59 #include <string.h>
60 
61 #include "defs.h"
62 #include "extern.h"
63 
64 #define LHSZ	128
65 #define	LHMASK	(LHSZ - 1)
66 static struct labhash {
67 	struct	labhash *lh_next;
68 	u_int	lh_hash;
69 	struct	s_command *lh_cmd;
70 	int	lh_ref;
71 } *labels[LHSZ];
72 
73 static char	 *compile_addr __P((char *, struct s_addr *));
74 static char	 *compile_ccl __P((char **, char *));
75 static char	 *compile_delimited __P((char *, char *));
76 static char	 *compile_flags __P((char *, struct s_subst *));
77 static char	 *compile_re __P((char *, regex_t **));
78 static char	 *compile_subst __P((char *, struct s_subst *));
79 static char	 *compile_text __P((void));
80 static char	 *compile_tr __P((char *, char **));
81 static struct s_command
82 		**compile_stream __P((struct s_command **));
83 static char	 *duptoeol __P((char *, char *));
84 static void	  enterlabel __P((struct s_command *));
85 static struct s_command
86 		 *findlabel __P((char *));
87 static void	  fixuplabel __P((struct s_command *, struct s_command *));
88 static void	  uselabel __P((void));
89 
90 /*
91  * Command specification.  This is used to drive the command parser.
92  */
93 struct s_format {
94 	char code;				/* Command code */
95 	int naddr;				/* Number of address args */
96 	enum e_args args;			/* Argument type */
97 };
98 
99 static struct s_format cmd_fmts[] = {
100 	{'{', 2, GROUP},
101 	{'}', 0, ENDGROUP},
102 	{'a', 1, TEXT},
103 	{'b', 2, BRANCH},
104 	{'c', 2, TEXT},
105 	{'d', 2, EMPTY},
106 	{'D', 2, EMPTY},
107 	{'g', 2, EMPTY},
108 	{'G', 2, EMPTY},
109 	{'h', 2, EMPTY},
110 	{'H', 2, EMPTY},
111 	{'i', 1, TEXT},
112 	{'l', 2, EMPTY},
113 	{'n', 2, EMPTY},
114 	{'N', 2, EMPTY},
115 	{'p', 2, EMPTY},
116 	{'P', 2, EMPTY},
117 	{'q', 1, EMPTY},
118 	{'r', 1, RFILE},
119 	{'s', 2, SUBST},
120 	{'t', 2, BRANCH},
121 	{'w', 2, WFILE},
122 	{'x', 2, EMPTY},
123 	{'y', 2, TR},
124 	{'!', 2, NONSEL},
125 	{':', 0, LABEL},
126 	{'#', 0, COMMENT},
127 	{'=', 1, EMPTY},
128 	{'\0', 0, COMMENT},
129 };
130 
131 /* The compiled program. */
132 struct s_command *prog;
133 
134 /*
135  * Compile the program into prog.
136  * Initialise appends.
137  */
138 void
139 compile()
140 {
141 	*compile_stream(&prog) = NULL;
142 	fixuplabel(prog, NULL);
143 	uselabel();
144 	if (appendnum > 0)
145 		appends = xmalloc(sizeof(struct s_appends) * appendnum);
146 	match = xmalloc((maxnsub + 1) * sizeof(regmatch_t));
147 }
148 
149 #define EATSPACE() do {							\
150 	if (p)								\
151 		while (*p && isascii((unsigned char)*p) &&		\
152 		    isspace((unsigned char)*p))				\
153 			p++;						\
154 	} while (0)
155 
156 static struct s_command **
157 compile_stream(link)
158 	struct s_command **link;
159 {
160 	char *p;
161 	static char lbuf[_POSIX2_LINE_MAX + 1];	/* To save stack */
162 	struct s_command *cmd, *cmd2, *stack;
163 	struct s_format *fp;
164 	int naddr;				/* Number of addresses */
165 
166 	stack = 0;
167 	for (;;) {
168 		if ((p = cu_fgets(lbuf, sizeof(lbuf))) == NULL) {
169 			if (stack != 0)
170 				err(COMPILE, "unexpected EOF (pending }'s)");
171 			return (link);
172 		}
173 
174 semicolon:	EATSPACE();
175 		if (p && (*p == '#' || *p == '\0'))
176 			continue;
177 		*link = cmd = xmalloc(sizeof(struct s_command));
178 		link = &cmd->next;
179 		cmd->nonsel = cmd->inrange = 0;
180 		/* First parse the addresses */
181 		naddr = 0;
182 
183 /* Valid characters to start an address */
184 #define	addrchar(c)	(strchr("0123456789/\\$", (c)))
185 		if (addrchar(*p)) {
186 			naddr++;
187 			cmd->a1 = xmalloc(sizeof(struct s_addr));
188 			p = compile_addr(p, cmd->a1);
189 			EATSPACE();				/* EXTENSION */
190 			if (*p == ',') {
191 				p++;
192 				EATSPACE();			/* EXTENSION */
193 				naddr++;
194 				cmd->a2 = xmalloc(sizeof(struct s_addr));
195 				p = compile_addr(p, cmd->a2);
196 				EATSPACE();
197 			} else
198 				cmd->a2 = 0;
199 		} else
200 			cmd->a1 = cmd->a2 = 0;
201 
202 nonsel:		/* Now parse the command */
203 		if (!*p)
204 			err(COMPILE, "command expected");
205 		cmd->code = *p;
206 		for (fp = cmd_fmts; fp->code; fp++)
207 			if (fp->code == *p)
208 				break;
209 		if (!fp->code)
210 			err(COMPILE, "invalid command code %c", *p);
211 		if (naddr > fp->naddr)
212 			err(COMPILE,
213 "command %c expects up to %d address(es), found %d", *p, fp->naddr, naddr);
214 		switch (fp->args) {
215 		case NONSEL:			/* ! */
216 			p++;
217 			EATSPACE();
218 			cmd->nonsel = ! cmd->nonsel;
219 			goto nonsel;
220 		case GROUP:			/* { */
221 			p++;
222 			EATSPACE();
223 			cmd->next = stack;
224 			stack = cmd;
225 			link = &cmd->u.c;
226 			if (*p)
227 				goto semicolon;
228 			break;
229 		case ENDGROUP:
230 			/*
231 			 * Short-circuit command processing, since end of
232 			 * group is really just a noop.
233 			 */
234 			cmd->nonsel = 1;
235 			if (stack == 0)
236 				err(COMPILE, "unexpected }");
237 			cmd2 = stack;
238 			stack = cmd2->next;
239 			cmd2->next = cmd;
240 			/*FALLTHROUGH*/
241 		case EMPTY:		/* d D g G h H l n N p P q x = \0 */
242 			p++;
243 			EATSPACE();
244 			if (*p == ';') {
245 				p++;
246 				link = &cmd->next;
247 				goto semicolon;
248 			}
249 			if (*p)
250 				err(COMPILE,
251 "extra characters at the end of %c command", cmd->code);
252 			break;
253 		case TEXT:			/* a c i */
254 			p++;
255 			EATSPACE();
256 			if (*p != '\\')
257 				err(COMPILE,
258 "command %c expects \\ followed by text", cmd->code);
259 			p++;
260 			EATSPACE();
261 			if (*p)
262 				err(COMPILE,
263 "extra characters after \\ at the end of %c command", cmd->code);
264 			cmd->t = compile_text();
265 			break;
266 		case COMMENT:			/* \0 # */
267 			break;
268 		case WFILE:			/* w */
269 			p++;
270 			EATSPACE();
271 			if (*p == '\0')
272 				err(COMPILE, "filename expected");
273 			cmd->t = duptoeol(p, "w command");
274 			if (aflag)
275 				cmd->u.fd = -1;
276 			else if ((cmd->u.fd = open(p,
277 			    O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
278 			    DEFFILEMODE)) == -1)
279 				err(FATAL, "%s: %s\n", p, strerror(errno));
280 			break;
281 		case RFILE:			/* r */
282 			p++;
283 			EATSPACE();
284 			if (*p == '\0')
285 				err(COMPILE, "filename expected");
286 			else
287 				cmd->t = duptoeol(p, "read command");
288 			break;
289 		case BRANCH:			/* b t */
290 			p++;
291 			EATSPACE();
292 			if (*p == '\0')
293 				cmd->t = NULL;
294 			else
295 				cmd->t = duptoeol(p, "branch");
296 			break;
297 		case LABEL:			/* : */
298 			p++;
299 			EATSPACE();
300 			cmd->t = duptoeol(p, "label");
301 			if (strlen(p) == 0)
302 				err(COMPILE, "empty label");
303 			enterlabel(cmd);
304 			break;
305 		case SUBST:			/* s */
306 			p++;
307 			if (*p == '\0' || *p == '\\')
308 				err(COMPILE,
309 "substitute pattern can not be delimited by newline or backslash");
310 			cmd->u.s = xmalloc(sizeof(struct s_subst));
311 			p = compile_re(p, &cmd->u.s->re);
312 			if (p == NULL)
313 				err(COMPILE, "unterminated substitute pattern");
314 			--p;
315 			p = compile_subst(p, cmd->u.s);
316 			p = compile_flags(p, cmd->u.s);
317 			EATSPACE();
318 			if (*p == ';') {
319 				p++;
320 				link = &cmd->next;
321 				goto semicolon;
322 			}
323 			break;
324 		case TR:			/* y */
325 			p++;
326 			p = compile_tr(p, (char **)&cmd->u.y);
327 			EATSPACE();
328 			if (*p == ';') {
329 				p++;
330 				link = &cmd->next;
331 				goto semicolon;
332 			}
333 			if (*p)
334 				err(COMPILE,
335 "extra text at the end of a transform command");
336 			break;
337 		}
338 	}
339 }
340 
341 /*
342  * Get a delimited string.  P points to the delimeter of the string; d points
343  * to a buffer area.  Newline and delimiter escapes are processed; other
344  * escapes are ignored.
345  *
346  * Returns a pointer to the first character after the final delimiter or NULL
347  * in the case of a non-terminated string.  The character array d is filled
348  * with the processed string.
349  */
350 static char *
351 compile_delimited(p, d)
352 	char *p, *d;
353 {
354 	char c;
355 
356 	c = *p++;
357 	if (c == '\0')
358 		return (NULL);
359 	else if (c == '\\')
360 		err(COMPILE, "\\ can not be used as a string delimiter");
361 	else if (c == '\n')
362 		err(COMPILE, "newline can not be used as a string delimiter");
363 	while (*p) {
364 		if (*p == '[') {
365 			if ((d = compile_ccl(&p, d)) == NULL)
366 				err(COMPILE, "unbalanced brackets ([])");
367 			continue;
368 		} else if (*p == '\\' && p[1] == '[') {
369 			*d++ = *p++;
370 		} else if (*p == '\\' && p[1] == c)
371 			p++;
372 		else if (*p == '\\' && p[1] == 'n') {
373 			*d++ = '\n';
374 			p += 2;
375 			continue;
376 		} else if (*p == '\\' && p[1] == '\\')
377 			*d++ = *p++;
378 		else if (*p == c) {
379 			*d = '\0';
380 			return (p + 1);
381 		}
382 		*d++ = *p++;
383 	}
384 	return (NULL);
385 }
386 
387 
388 /* compile_ccl: expand a POSIX character class */
389 static char *
390 compile_ccl(sp, t)
391 	char **sp;
392 	char *t;
393 {
394 	int c, d;
395 	char *s = *sp;
396 
397 	*t++ = *s++;
398 	if (*s == '^')
399 		*t++ = *s++;
400 	if (*s == ']')
401 		*t++ = *s++;
402 	for (; *s && (*t = *s) != ']'; s++, t++)
403 		if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
404 			*++t = *++s, t++, s++;
405 			for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
406 				if ((c = *s) == '\0')
407 					return NULL;
408 		} else if (*s == '\\' && s[1] == 'n')
409 			    *t = '\n', s++;
410 	return (*s == ']') ? *sp = ++s, ++t : NULL;
411 }
412 
413 /*
414  * Get a regular expression.  P points to the delimiter of the regular
415  * expression; repp points to the address of a regexp pointer.  Newline
416  * and delimiter escapes are processed; other escapes are ignored.
417  * Returns a pointer to the first character after the final delimiter
418  * or NULL in the case of a non terminated regular expression.  The regexp
419  * pointer is set to the compiled regular expression.
420  * Cflags are passed to regcomp.
421  */
422 static char *
423 compile_re(p, repp)
424 	char *p;
425 	regex_t **repp;
426 {
427 	int eval;
428 	char re[_POSIX2_LINE_MAX + 1];
429 
430 	p = compile_delimited(p, re);
431 	if (p && strlen(re) == 0) {
432 		*repp = NULL;
433 		return (p);
434 	}
435 	*repp = xmalloc(sizeof(regex_t));
436 	if (p && (eval = regcomp(*repp, re, 0)) != 0)
437 		err(COMPILE, "RE error: %s", strregerror(eval, *repp));
438 	if (maxnsub < (*repp)->re_nsub)
439 		maxnsub = (*repp)->re_nsub;
440 	return (p);
441 }
442 
443 /*
444  * Compile the substitution string of a regular expression and set res to
445  * point to a saved copy of it.  Nsub is the number of parenthesized regular
446  * expressions.
447  */
448 static char *
449 compile_subst(p, s)
450 	char *p;
451 	struct s_subst *s;
452 {
453 	static char lbuf[_POSIX2_LINE_MAX + 1];
454 	int asize, ref, size;
455 	char c, *text, *op, *sp;
456 
457 	c = *p++;			/* Terminator character */
458 	if (c == '\0')
459 		return (NULL);
460 
461 	s->maxbref = 0;
462 	s->linenum = linenum;
463 	asize = 2 * _POSIX2_LINE_MAX + 1;
464 	text = xmalloc(asize);
465 	size = 0;
466 	do {
467 		op = sp = text + size;
468 		for (; *p; p++) {
469 			if (*p == '\\') {
470 				p++;
471 				if (strchr("123456789", *p) != NULL) {
472 					*sp++ = '\\';
473 					ref = *p - '0';
474 					if (s->re != NULL &&
475 					    ref > s->re->re_nsub)
476 						err(COMPILE,
477 "\\%c not defined in the RE", *p);
478 					if (s->maxbref < ref)
479 						s->maxbref = ref;
480 				} else if (*p == '&' || *p == '\\')
481 					*sp++ = '\\';
482 			} else if (*p == c) {
483 				p++;
484 				*sp++ = '\0';
485 				size += sp - op;
486 				s->new = xrealloc(text, size);
487 				return (p);
488 			} else if (*p == '\n') {
489 				err(COMPILE,
490 "unescaped newline inside substitute pattern");
491 				/* NOTREACHED */
492 			}
493 			*sp++ = *p;
494 		}
495 		size += sp - op;
496 		if (asize - size < _POSIX2_LINE_MAX + 1) {
497 			asize *= 2;
498 			text = xmalloc(asize);
499 		}
500 	} while (cu_fgets(p = lbuf, sizeof(lbuf)));
501 	err(COMPILE, "unterminated substitute in regular expression");
502 	/* NOTREACHED */
503 	return (NULL);
504 }
505 
506 /*
507  * Compile the flags of the s command
508  */
509 static char *
510 compile_flags(p, s)
511 	char *p;
512 	struct s_subst *s;
513 {
514 	int gn;			/* True if we have seen g or n */
515 	char wfile[_POSIX2_LINE_MAX + 1], *q;
516 
517 	s->n = 1;				/* Default */
518 	s->p = 0;
519 	s->wfile = NULL;
520 	s->wfd = -1;
521 	for (gn = 0;;) {
522 		EATSPACE();			/* EXTENSION */
523 		switch (*p) {
524 		case 'g':
525 			if (gn)
526 				err(COMPILE,
527 "more than one number or 'g' in substitute flags");
528 			gn = 1;
529 			s->n = 0;
530 			break;
531 		case '\0':
532 		case '\n':
533 		case ';':
534 			return (p);
535 		case 'p':
536 			s->p = 1;
537 			break;
538 		case '1': case '2': case '3':
539 		case '4': case '5': case '6':
540 		case '7': case '8': case '9':
541 			if (gn)
542 				err(COMPILE,
543 "more than one number or 'g' in substitute flags");
544 			gn = 1;
545 			/* XXX Check for overflow */
546 			s->n = (int)strtol(p, &p, 10);
547 			break;
548 		case 'w':
549 			p++;
550 #ifdef HISTORIC_PRACTICE
551 			if (*p != ' ') {
552 				err(WARNING, "space missing before w wfile");
553 				return (p);
554 			}
555 #endif
556 			EATSPACE();
557 			q = wfile;
558 			while (*p) {
559 				if (*p == '\n')
560 					break;
561 				*q++ = *p++;
562 			}
563 			*q = '\0';
564 			if (q == wfile)
565 				err(COMPILE, "no wfile specified");
566 			s->wfile = strdup(wfile);
567 			if (!aflag && (s->wfd = open(wfile,
568 			    O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
569 			    DEFFILEMODE)) == -1)
570 				err(FATAL, "%s: %s\n", wfile, strerror(errno));
571 			return (p);
572 		default:
573 			err(COMPILE,
574 			    "bad flag in substitute command: '%c'", *p);
575 			break;
576 		}
577 		p++;
578 	}
579 }
580 
581 /*
582  * Compile a translation set of strings into a lookup table.
583  */
584 static char *
585 compile_tr(p, transtab)
586 	char *p;
587 	char **transtab;
588 {
589 	int i;
590 	char *lt, *op, *np;
591 	char old[_POSIX2_LINE_MAX + 1];
592 	char new[_POSIX2_LINE_MAX + 1];
593 
594 	if (*p == '\0' || *p == '\\')
595 		err(COMPILE,
596 "transform pattern can not be delimited by newline or backslash");
597 	p = compile_delimited(p, old);
598 	if (p == NULL) {
599 		err(COMPILE, "unterminated transform source string");
600 		return (NULL);
601 	}
602 	p = compile_delimited(--p, new);
603 	if (p == NULL) {
604 		err(COMPILE, "unterminated transform target string");
605 		return (NULL);
606 	}
607 	EATSPACE();
608 	if (strlen(new) != strlen(old)) {
609 		err(COMPILE, "transform strings are not the same length");
610 		return (NULL);
611 	}
612 	/* We assume characters are 8 bits */
613 	lt = xmalloc(UCHAR_MAX);
614 	for (i = 0; i <= UCHAR_MAX; i++)
615 		lt[i] = (char)i;
616 	for (op = old, np = new; *op; op++, np++)
617 		lt[(u_char)*op] = *np;
618 	*transtab = lt;
619 	return (p);
620 }
621 
622 /*
623  * Compile the text following an a or i command.
624  */
625 static char *
626 compile_text()
627 {
628 	int asize, size;
629 	char *text, *p, *op, *s;
630 	char lbuf[_POSIX2_LINE_MAX + 1];
631 
632 	asize = 2 * _POSIX2_LINE_MAX + 1;
633 	text = xmalloc(asize);
634 	size = 0;
635 	while (cu_fgets(lbuf, sizeof(lbuf))) {
636 		op = s = text + size;
637 		p = lbuf;
638 		EATSPACE();
639 		for (; *p; p++) {
640 			if (*p == '\\')
641 				p++;
642 			*s++ = *p;
643 		}
644 		size += s - op;
645 		if (p[-2] != '\\') {
646 			*s = '\0';
647 			break;
648 		}
649 		if (asize - size < _POSIX2_LINE_MAX + 1) {
650 			asize *= 2;
651 			text = xmalloc(asize);
652 		}
653 	}
654 	return (xrealloc(text, size + 1));
655 }
656 
657 /*
658  * Get an address and return a pointer to the first character after
659  * it.  Fill the structure pointed to according to the address.
660  */
661 static char *
662 compile_addr(p, a)
663 	char *p;
664 	struct s_addr *a;
665 {
666 	char *end;
667 
668 	switch (*p) {
669 	case '\\':				/* Context address */
670 		++p;
671 		/* FALLTHROUGH */
672 	case '/':				/* Context address */
673 		p = compile_re(p, &a->u.r);
674 		if (p == NULL)
675 			err(COMPILE, "unterminated regular expression");
676 		a->type = AT_RE;
677 		return (p);
678 
679 	case '$':				/* Last line */
680 		a->type = AT_LAST;
681 		return (p + 1);
682 						/* Line number */
683 	case '0': case '1': case '2': case '3': case '4':
684 	case '5': case '6': case '7': case '8': case '9':
685 		a->type = AT_LINE;
686 		a->u.l = strtol(p, &end, 10);
687 		return (end);
688 	default:
689 		err(COMPILE, "expected context address");
690 		return (NULL);
691 	}
692 }
693 
694 /*
695  * duptoeol --
696  *	Return a copy of all the characters up to \n or \0.
697  */
698 static char *
699 duptoeol(s, ctype)
700 	char *s;
701 	char *ctype;
702 {
703 	size_t len;
704 	int ws;
705 	char *start;
706 
707 	ws = 0;
708 	for (start = s; *s != '\0' && *s != '\n'; ++s)
709 		ws = isspace((unsigned char)*s);
710 	*s = '\0';
711 	if (ws)
712 		err(WARNING, "whitespace after %s", ctype);
713 	len = s - start + 1;
714 	return (memmove(xmalloc(len), start, len));
715 }
716 
717 /*
718  * Convert goto label names to addresses, and count a and r commands, in
719  * the given subset of the script.  Free the memory used by labels in b
720  * and t commands (but not by :).
721  *
722  * TODO: Remove } nodes
723  */
724 static void
725 fixuplabel(cp, end)
726 	struct s_command *cp, *end;
727 {
728 
729 	for (; cp != end; cp = cp->next)
730 		switch (cp->code) {
731 		case 'a':
732 		case 'r':
733 			appendnum++;
734 			break;
735 		case 'b':
736 		case 't':
737 			/* Resolve branch target. */
738 			if (cp->t == NULL) {
739 				cp->u.c = NULL;
740 				break;
741 			}
742 			if ((cp->u.c = findlabel(cp->t)) == NULL)
743 				err(COMPILE2, "undefined label '%s'", cp->t);
744 			free(cp->t);
745 			break;
746 		case '{':
747 			/* Do interior commands. */
748 			fixuplabel(cp->u.c, cp->next);
749 			break;
750 		}
751 }
752 
753 /*
754  * Associate the given command label for later lookup.
755  */
756 static void
757 enterlabel(cp)
758 	struct s_command *cp;
759 {
760 	struct labhash **lhp, *lh;
761 	u_char *p;
762 	u_int h, c;
763 
764 	for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
765 		h = (h << 5) + h + c;
766 	lhp = &labels[h & LHMASK];
767 	for (lh = *lhp; lh != NULL; lh = lh->lh_next)
768 		if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
769 			err(COMPILE2, "duplicate label '%s'", cp->t);
770 	lh = xmalloc(sizeof *lh);
771 	lh->lh_next = *lhp;
772 	lh->lh_hash = h;
773 	lh->lh_cmd = cp;
774 	lh->lh_ref = 0;
775 	*lhp = lh;
776 }
777 
778 /*
779  * Find the label contained in the command l in the command linked
780  * list cp.  L is excluded from the search.  Return NULL if not found.
781  */
782 static struct s_command *
783 findlabel(name)
784 	char *name;
785 {
786 	struct labhash *lh;
787 	u_char *p;
788 	u_int h, c;
789 
790 	for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
791 		h = (h << 5) + h + c;
792 	for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
793 		if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
794 			lh->lh_ref = 1;
795 			return (lh->lh_cmd);
796 		}
797 	}
798 	return (NULL);
799 }
800 
801 /*
802  * Warn about any unused labels.  As a side effect, release the label hash
803  * table space.
804  */
805 static void
806 uselabel()
807 {
808 	struct labhash *lh, *next;
809 	int i;
810 
811 	for (i = 0; i < LHSZ; i++) {
812 		for (lh = labels[i]; lh != NULL; lh = next) {
813 			next = lh->lh_next;
814 			if (!lh->lh_ref)
815 				err(WARNING, "unused label '%s'",
816 				    lh->lh_cmd->t);
817 			free(lh);
818 		}
819 	}
820 }
821