xref: /netbsd-src/usr.bin/sed/compile.c (revision f5d3fbbc6ff4a77159fb268d247bd94cb7d7e332)
1 /*	$NetBSD: compile.c,v 1.17 1997/10/19 23:05:11 lukem 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 1.17 1997/10/19 23:05:11 lukem 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 	appends = xmalloc(sizeof(struct s_appends) * appendnum);
145 	match = xmalloc((maxnsub + 1) * sizeof(regmatch_t));
146 }
147 
148 #define EATSPACE() do {							\
149 	if (p)								\
150 		while (*p && isascii(*p) && isspace(*p))		\
151 			p++;						\
152 	} while (0)
153 
154 static struct s_command **
155 compile_stream(link)
156 	struct s_command **link;
157 {
158 	char *p;
159 	static char lbuf[_POSIX2_LINE_MAX + 1];	/* To save stack */
160 	struct s_command *cmd, *cmd2, *stack;
161 	struct s_format *fp;
162 	int naddr;				/* Number of addresses */
163 
164 	stack = 0;
165 	for (;;) {
166 		if ((p = cu_fgets(lbuf, sizeof(lbuf))) == NULL) {
167 			if (stack != 0)
168 				err(COMPILE, "unexpected EOF (pending }'s)");
169 			return (link);
170 		}
171 
172 semicolon:	EATSPACE();
173 		if (p && (*p == '#' || *p == '\0'))
174 			continue;
175 		*link = cmd = xmalloc(sizeof(struct s_command));
176 		link = &cmd->next;
177 		cmd->nonsel = cmd->inrange = 0;
178 		/* First parse the addresses */
179 		naddr = 0;
180 
181 /* Valid characters to start an address */
182 #define	addrchar(c)	(strchr("0123456789/\\$", (c)))
183 		if (addrchar(*p)) {
184 			naddr++;
185 			cmd->a1 = xmalloc(sizeof(struct s_addr));
186 			p = compile_addr(p, cmd->a1);
187 			EATSPACE();				/* EXTENSION */
188 			if (*p == ',') {
189 				p++;
190 				EATSPACE();			/* EXTENSION */
191 				naddr++;
192 				cmd->a2 = xmalloc(sizeof(struct s_addr));
193 				p = compile_addr(p, cmd->a2);
194 				EATSPACE();
195 			} else
196 				cmd->a2 = 0;
197 		} else
198 			cmd->a1 = cmd->a2 = 0;
199 
200 nonsel:		/* Now parse the command */
201 		if (!*p)
202 			err(COMPILE, "command expected");
203 		cmd->code = *p;
204 		for (fp = cmd_fmts; fp->code; fp++)
205 			if (fp->code == *p)
206 				break;
207 		if (!fp->code)
208 			err(COMPILE, "invalid command code %c", *p);
209 		if (naddr > fp->naddr)
210 			err(COMPILE,
211 "command %c expects up to %d address(es), found %d", *p, fp->naddr, naddr);
212 		switch (fp->args) {
213 		case NONSEL:			/* ! */
214 			p++;
215 			EATSPACE();
216 			cmd->nonsel = ! cmd->nonsel;
217 			goto nonsel;
218 		case GROUP:			/* { */
219 			p++;
220 			EATSPACE();
221 			cmd->next = stack;
222 			stack = cmd;
223 			link = &cmd->u.c;
224 			if (*p)
225 				goto semicolon;
226 			break;
227 		case ENDGROUP:
228 			/*
229 			 * Short-circuit command processing, since end of
230 			 * group is really just a noop.
231 			 */
232 			cmd->nonsel = 1;
233 			if (stack == 0)
234 				err(COMPILE, "unexpected }");
235 			cmd2 = stack;
236 			stack = cmd2->next;
237 			cmd2->next = cmd;
238 			/*FALLTHROUGH*/
239 		case EMPTY:		/* d D g G h H l n N p P q x = \0 */
240 			p++;
241 			EATSPACE();
242 			if (*p == ';') {
243 				p++;
244 				link = &cmd->next;
245 				goto semicolon;
246 			}
247 			if (*p)
248 				err(COMPILE,
249 "extra characters at the end of %c command", cmd->code);
250 			break;
251 		case TEXT:			/* a c i */
252 			p++;
253 			EATSPACE();
254 			if (*p != '\\')
255 				err(COMPILE,
256 "command %c expects \\ followed by text", cmd->code);
257 			p++;
258 			EATSPACE();
259 			if (*p)
260 				err(COMPILE,
261 "extra characters after \\ at the end of %c command", cmd->code);
262 			cmd->t = compile_text();
263 			break;
264 		case COMMENT:			/* \0 # */
265 			break;
266 		case WFILE:			/* w */
267 			p++;
268 			EATSPACE();
269 			if (*p == '\0')
270 				err(COMPILE, "filename expected");
271 			cmd->t = duptoeol(p, "w command");
272 			if (aflag)
273 				cmd->u.fd = -1;
274 			else if ((cmd->u.fd = open(p,
275 			    O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
276 			    DEFFILEMODE)) == -1)
277 				err(FATAL, "%s: %s\n", p, strerror(errno));
278 			break;
279 		case RFILE:			/* r */
280 			p++;
281 			EATSPACE();
282 			if (*p == '\0')
283 				err(COMPILE, "filename expected");
284 			else
285 				cmd->t = duptoeol(p, "read command");
286 			break;
287 		case BRANCH:			/* b t */
288 			p++;
289 			EATSPACE();
290 			if (*p == '\0')
291 				cmd->t = NULL;
292 			else
293 				cmd->t = duptoeol(p, "branch");
294 			break;
295 		case LABEL:			/* : */
296 			p++;
297 			EATSPACE();
298 			cmd->t = duptoeol(p, "label");
299 			if (strlen(p) == 0)
300 				err(COMPILE, "empty label");
301 			enterlabel(cmd);
302 			break;
303 		case SUBST:			/* s */
304 			p++;
305 			if (*p == '\0' || *p == '\\')
306 				err(COMPILE,
307 "substitute pattern can not be delimited by newline or backslash");
308 			cmd->u.s = xmalloc(sizeof(struct s_subst));
309 			p = compile_re(p, &cmd->u.s->re);
310 			if (p == NULL)
311 				err(COMPILE, "unterminated substitute pattern");
312 			--p;
313 			p = compile_subst(p, cmd->u.s);
314 			p = compile_flags(p, cmd->u.s);
315 			EATSPACE();
316 			if (*p == ';') {
317 				p++;
318 				link = &cmd->next;
319 				goto semicolon;
320 			}
321 			break;
322 		case TR:			/* y */
323 			p++;
324 			p = compile_tr(p, (char **)&cmd->u.y);
325 			EATSPACE();
326 			if (*p == ';') {
327 				p++;
328 				link = &cmd->next;
329 				goto semicolon;
330 			}
331 			if (*p)
332 				err(COMPILE,
333 "extra text at the end of a transform command");
334 			break;
335 		}
336 	}
337 }
338 
339 /*
340  * Get a delimited string.  P points to the delimeter of the string; d points
341  * to a buffer area.  Newline and delimiter escapes are processed; other
342  * escapes are ignored.
343  *
344  * Returns a pointer to the first character after the final delimiter or NULL
345  * in the case of a non-terminated string.  The character array d is filled
346  * with the processed string.
347  */
348 static char *
349 compile_delimited(p, d)
350 	char *p, *d;
351 {
352 	char c;
353 
354 	c = *p++;
355 	if (c == '\0')
356 		return (NULL);
357 	else if (c == '\\')
358 		err(COMPILE, "\\ can not be used as a string delimiter");
359 	else if (c == '\n')
360 		err(COMPILE, "newline can not be used as a string delimiter");
361 	while (*p) {
362 		if (*p == '[') {
363 			if ((d = compile_ccl(&p, d)) == NULL)
364 				err(COMPILE, "unbalanced brackets ([])");
365 			continue;
366 		} else if (*p == '\\' && p[1] == '[') {
367 			*d++ = *p++;
368 		} else if (*p == '\\' && p[1] == c)
369 			p++;
370 		else if (*p == '\\' && p[1] == 'n') {
371 			*d++ = '\n';
372 			p += 2;
373 			continue;
374 		} else if (*p == '\\' && p[1] == '\\')
375 			*d++ = *p++;
376 		else if (*p == c) {
377 			*d = '\0';
378 			return (p + 1);
379 		}
380 		*d++ = *p++;
381 	}
382 	return (NULL);
383 }
384 
385 
386 /* compile_ccl: expand a POSIX character class */
387 static char *
388 compile_ccl(sp, t)
389 	char **sp;
390 	char *t;
391 {
392 	int c, d;
393 	char *s = *sp;
394 
395 	*t++ = *s++;
396 	if (*s == '^')
397 		*t++ = *s++;
398 	if (*s == ']')
399 		*t++ = *s++;
400 	for (; *s && (*t = *s) != ']'; s++, t++)
401 		if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
402 			*++t = *++s, t++, s++;
403 			for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
404 				if ((c = *s) == '\0')
405 					return NULL;
406 		} else if (*s == '\\' && s[1] == 'n')
407 			    *t = '\n', s++;
408 	return (*s == ']') ? *sp = ++s, ++t : NULL;
409 }
410 
411 /*
412  * Get a regular expression.  P points to the delimiter of the regular
413  * expression; repp points to the address of a regexp pointer.  Newline
414  * and delimiter escapes are processed; other escapes are ignored.
415  * Returns a pointer to the first character after the final delimiter
416  * or NULL in the case of a non terminated regular expression.  The regexp
417  * pointer is set to the compiled regular expression.
418  * Cflags are passed to regcomp.
419  */
420 static char *
421 compile_re(p, repp)
422 	char *p;
423 	regex_t **repp;
424 {
425 	int eval;
426 	char re[_POSIX2_LINE_MAX + 1];
427 
428 	p = compile_delimited(p, re);
429 	if (p && strlen(re) == 0) {
430 		*repp = NULL;
431 		return (p);
432 	}
433 	*repp = xmalloc(sizeof(regex_t));
434 	if (p && (eval = regcomp(*repp, re, 0)) != 0)
435 		err(COMPILE, "RE error: %s", strregerror(eval, *repp));
436 	if (maxnsub < (*repp)->re_nsub)
437 		maxnsub = (*repp)->re_nsub;
438 	return (p);
439 }
440 
441 /*
442  * Compile the substitution string of a regular expression and set res to
443  * point to a saved copy of it.  Nsub is the number of parenthesized regular
444  * expressions.
445  */
446 static char *
447 compile_subst(p, s)
448 	char *p;
449 	struct s_subst *s;
450 {
451 	static char lbuf[_POSIX2_LINE_MAX + 1];
452 	int asize, ref, size;
453 	char c, *text, *op, *sp;
454 
455 	c = *p++;			/* Terminator character */
456 	if (c == '\0')
457 		return (NULL);
458 
459 	s->maxbref = 0;
460 	s->linenum = linenum;
461 	asize = 2 * _POSIX2_LINE_MAX + 1;
462 	text = xmalloc(asize);
463 	size = 0;
464 	do {
465 		op = sp = text + size;
466 		for (; *p; p++) {
467 			if (*p == '\\') {
468 				p++;
469 				if (strchr("123456789", *p) != NULL) {
470 					*sp++ = '\\';
471 					ref = *p - '0';
472 					if (s->re != NULL &&
473 					    ref > s->re->re_nsub)
474 						err(COMPILE,
475 "\\%c not defined in the RE", *p);
476 					if (s->maxbref < ref)
477 						s->maxbref = ref;
478 				} else if (*p == '&' || *p == '\\')
479 					*sp++ = '\\';
480 			} else if (*p == c) {
481 				p++;
482 				*sp++ = '\0';
483 				size += sp - op;
484 				s->new = xrealloc(text, size);
485 				return (p);
486 			} else if (*p == '\n') {
487 				err(COMPILE,
488 "unescaped newline inside substitute pattern");
489 				/* NOTREACHED */
490 			}
491 			*sp++ = *p;
492 		}
493 		size += sp - op;
494 		if (asize - size < _POSIX2_LINE_MAX + 1) {
495 			asize *= 2;
496 			text = xmalloc(asize);
497 		}
498 	} while (cu_fgets(p = lbuf, sizeof(lbuf)));
499 	err(COMPILE, "unterminated substitute in regular expression");
500 	/* NOTREACHED */
501 	return (NULL);
502 }
503 
504 /*
505  * Compile the flags of the s command
506  */
507 static char *
508 compile_flags(p, s)
509 	char *p;
510 	struct s_subst *s;
511 {
512 	int gn;			/* True if we have seen g or n */
513 	char wfile[_POSIX2_LINE_MAX + 1], *q;
514 
515 	s->n = 1;				/* Default */
516 	s->p = 0;
517 	s->wfile = NULL;
518 	s->wfd = -1;
519 	for (gn = 0;;) {
520 		EATSPACE();			/* EXTENSION */
521 		switch (*p) {
522 		case 'g':
523 			if (gn)
524 				err(COMPILE,
525 "more than one number or 'g' in substitute flags");
526 			gn = 1;
527 			s->n = 0;
528 			break;
529 		case '\0':
530 		case '\n':
531 		case ';':
532 			return (p);
533 		case 'p':
534 			s->p = 1;
535 			break;
536 		case '1': case '2': case '3':
537 		case '4': case '5': case '6':
538 		case '7': case '8': case '9':
539 			if (gn)
540 				err(COMPILE,
541 "more than one number or 'g' in substitute flags");
542 			gn = 1;
543 			/* XXX Check for overflow */
544 			s->n = (int)strtol(p, &p, 10);
545 			break;
546 		case 'w':
547 			p++;
548 #ifdef HISTORIC_PRACTICE
549 			if (*p != ' ') {
550 				err(WARNING, "space missing before w wfile");
551 				return (p);
552 			}
553 #endif
554 			EATSPACE();
555 			q = wfile;
556 			while (*p) {
557 				if (*p == '\n')
558 					break;
559 				*q++ = *p++;
560 			}
561 			*q = '\0';
562 			if (q == wfile)
563 				err(COMPILE, "no wfile specified");
564 			s->wfile = strdup(wfile);
565 			if (!aflag && (s->wfd = open(wfile,
566 			    O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
567 			    DEFFILEMODE)) == -1)
568 				err(FATAL, "%s: %s\n", wfile, strerror(errno));
569 			return (p);
570 		default:
571 			err(COMPILE,
572 			    "bad flag in substitute command: '%c'", *p);
573 			break;
574 		}
575 		p++;
576 	}
577 }
578 
579 /*
580  * Compile a translation set of strings into a lookup table.
581  */
582 static char *
583 compile_tr(p, transtab)
584 	char *p;
585 	char **transtab;
586 {
587 	int i;
588 	char *lt, *op, *np;
589 	char old[_POSIX2_LINE_MAX + 1];
590 	char new[_POSIX2_LINE_MAX + 1];
591 
592 	if (*p == '\0' || *p == '\\')
593 		err(COMPILE,
594 "transform pattern can not be delimited by newline or backslash");
595 	p = compile_delimited(p, old);
596 	if (p == NULL) {
597 		err(COMPILE, "unterminated transform source string");
598 		return (NULL);
599 	}
600 	p = compile_delimited(--p, new);
601 	if (p == NULL) {
602 		err(COMPILE, "unterminated transform target string");
603 		return (NULL);
604 	}
605 	EATSPACE();
606 	if (strlen(new) != strlen(old)) {
607 		err(COMPILE, "transform strings are not the same length");
608 		return (NULL);
609 	}
610 	/* We assume characters are 8 bits */
611 	lt = xmalloc(UCHAR_MAX);
612 	for (i = 0; i <= UCHAR_MAX; i++)
613 		lt[i] = (char)i;
614 	for (op = old, np = new; *op; op++, np++)
615 		lt[(u_char)*op] = *np;
616 	*transtab = lt;
617 	return (p);
618 }
619 
620 /*
621  * Compile the text following an a or i command.
622  */
623 static char *
624 compile_text()
625 {
626 	int asize, size;
627 	char *text, *p, *op, *s;
628 	char lbuf[_POSIX2_LINE_MAX + 1];
629 
630 	asize = 2 * _POSIX2_LINE_MAX + 1;
631 	text = xmalloc(asize);
632 	size = 0;
633 	while (cu_fgets(lbuf, sizeof(lbuf))) {
634 		op = s = text + size;
635 		p = lbuf;
636 		EATSPACE();
637 		for (; *p; p++) {
638 			if (*p == '\\')
639 				p++;
640 			*s++ = *p;
641 		}
642 		size += s - op;
643 		if (p[-2] != '\\') {
644 			*s = '\0';
645 			break;
646 		}
647 		if (asize - size < _POSIX2_LINE_MAX + 1) {
648 			asize *= 2;
649 			text = xmalloc(asize);
650 		}
651 	}
652 	return (xrealloc(text, size + 1));
653 }
654 
655 /*
656  * Get an address and return a pointer to the first character after
657  * it.  Fill the structure pointed to according to the address.
658  */
659 static char *
660 compile_addr(p, a)
661 	char *p;
662 	struct s_addr *a;
663 {
664 	char *end;
665 
666 	switch (*p) {
667 	case '\\':				/* Context address */
668 		++p;
669 		/* FALLTHROUGH */
670 	case '/':				/* Context address */
671 		p = compile_re(p, &a->u.r);
672 		if (p == NULL)
673 			err(COMPILE, "unterminated regular expression");
674 		a->type = AT_RE;
675 		return (p);
676 
677 	case '$':				/* Last line */
678 		a->type = AT_LAST;
679 		return (p + 1);
680 						/* Line number */
681 	case '0': case '1': case '2': case '3': case '4':
682 	case '5': case '6': case '7': case '8': case '9':
683 		a->type = AT_LINE;
684 		a->u.l = strtol(p, &end, 10);
685 		return (end);
686 	default:
687 		err(COMPILE, "expected context address");
688 		return (NULL);
689 	}
690 }
691 
692 /*
693  * duptoeol --
694  *	Return a copy of all the characters up to \n or \0.
695  */
696 static char *
697 duptoeol(s, ctype)
698 	char *s;
699 	char *ctype;
700 {
701 	size_t len;
702 	int ws;
703 	char *start;
704 
705 	ws = 0;
706 	for (start = s; *s != '\0' && *s != '\n'; ++s)
707 		ws = isspace(*s);
708 	*s = '\0';
709 	if (ws)
710 		err(WARNING, "whitespace after %s", ctype);
711 	len = s - start + 1;
712 	return (memmove(xmalloc(len), start, len));
713 }
714 
715 /*
716  * Convert goto label names to addresses, and count a and r commands, in
717  * the given subset of the script.  Free the memory used by labels in b
718  * and t commands (but not by :).
719  *
720  * TODO: Remove } nodes
721  */
722 static void
723 fixuplabel(cp, end)
724 	struct s_command *cp, *end;
725 {
726 
727 	for (; cp != end; cp = cp->next)
728 		switch (cp->code) {
729 		case 'a':
730 		case 'r':
731 			appendnum++;
732 			break;
733 		case 'b':
734 		case 't':
735 			/* Resolve branch target. */
736 			if (cp->t == NULL) {
737 				cp->u.c = NULL;
738 				break;
739 			}
740 			if ((cp->u.c = findlabel(cp->t)) == NULL)
741 				err(COMPILE2, "undefined label '%s'", cp->t);
742 			free(cp->t);
743 			break;
744 		case '{':
745 			/* Do interior commands. */
746 			fixuplabel(cp->u.c, cp->next);
747 			break;
748 		}
749 }
750 
751 /*
752  * Associate the given command label for later lookup.
753  */
754 static void
755 enterlabel(cp)
756 	struct s_command *cp;
757 {
758 	struct labhash **lhp, *lh;
759 	u_char *p;
760 	u_int h, c;
761 
762 	for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
763 		h = (h << 5) + h + c;
764 	lhp = &labels[h & LHMASK];
765 	for (lh = *lhp; lh != NULL; lh = lh->lh_next)
766 		if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
767 			err(COMPILE2, "duplicate label '%s'", cp->t);
768 	lh = xmalloc(sizeof *lh);
769 	lh->lh_next = *lhp;
770 	lh->lh_hash = h;
771 	lh->lh_cmd = cp;
772 	lh->lh_ref = 0;
773 	*lhp = lh;
774 }
775 
776 /*
777  * Find the label contained in the command l in the command linked
778  * list cp.  L is excluded from the search.  Return NULL if not found.
779  */
780 static struct s_command *
781 findlabel(name)
782 	char *name;
783 {
784 	struct labhash *lh;
785 	u_char *p;
786 	u_int h, c;
787 
788 	for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
789 		h = (h << 5) + h + c;
790 	for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
791 		if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
792 			lh->lh_ref = 1;
793 			return (lh->lh_cmd);
794 		}
795 	}
796 	return (NULL);
797 }
798 
799 /*
800  * Warn about any unused labels.  As a side effect, release the label hash
801  * table space.
802  */
803 static void
804 uselabel()
805 {
806 	struct labhash *lh, *next;
807 	int i;
808 
809 	for (i = 0; i < LHSZ; i++) {
810 		for (lh = labels[i]; lh != NULL; lh = next) {
811 			next = lh->lh_next;
812 			if (!lh->lh_ref)
813 				err(WARNING, "unused label '%s'",
814 				    lh->lh_cmd->t);
815 			free(lh);
816 		}
817 	}
818 }
819