1 /*-
2 * SPDX-License-Identifier: BSD-3-Clause
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. Neither the name of the University nor the names of its contributors
20 * may be used to endorse or promote products derived from this software
21 * without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * SUCH DAMAGE.
34 */
35
36 #include <sys/cdefs.h>
37 __FBSDID("$FreeBSD$");
38
39 #ifndef lint
40 static const char sccsid[] = "@(#)compile.c 8.1 (Berkeley) 6/6/93";
41 #endif
42
43 #include <sys/types.h>
44 #include <sys/stat.h>
45
46 #include <ctype.h>
47 #include <err.h>
48 #include <errno.h>
49 #include <fcntl.h>
50 #include <limits.h>
51 #include <regex.h>
52 #include <stdio.h>
53 #include <stdlib.h>
54 #include <string.h>
55 #include <wchar.h>
56
57 #include "defs.h"
58 #include "extern.h"
59
60 #define LHSZ 128
61 #define LHMASK (LHSZ - 1)
62 static struct labhash {
63 struct labhash *lh_next;
64 u_int lh_hash;
65 struct s_command *lh_cmd;
66 int lh_ref;
67 } *labels[LHSZ];
68
69 static char *compile_addr(char *, struct s_addr *);
70 static char *compile_ccl(char **, char *);
71 static char *compile_delimited(char *, char *, int);
72 static char *compile_flags(char *, struct s_subst *);
73 static regex_t *compile_re(char *, int);
74 static char *compile_subst(char *, struct s_subst *);
75 static char *compile_text(void);
76 static char *compile_tr(char *, struct s_tr **);
77 static struct s_command
78 **compile_stream(struct s_command **);
79 static char *duptoeol(char *, const char *);
80 static void enterlabel(struct s_command *);
81 static struct s_command
82 *findlabel(char *);
83 static void fixuplabel(struct s_command *, struct s_command *);
84 static void uselabel(void);
85
86 /*
87 * Command specification. This is used to drive the command parser.
88 */
89 struct s_format {
90 char code; /* Command code */
91 int naddr; /* Number of address args */
92 enum e_args args; /* Argument type */
93 };
94
95 static struct s_format cmd_fmts[] = {
96 {'{', 2, GROUP},
97 {'}', 0, ENDGROUP},
98 {'a', 1, TEXT},
99 {'b', 2, BRANCH},
100 {'c', 2, TEXT},
101 {'d', 2, EMPTY},
102 {'D', 2, EMPTY},
103 {'g', 2, EMPTY},
104 {'G', 2, EMPTY},
105 {'h', 2, EMPTY},
106 {'H', 2, EMPTY},
107 {'i', 1, TEXT},
108 {'l', 2, EMPTY},
109 {'n', 2, EMPTY},
110 {'N', 2, EMPTY},
111 {'p', 2, EMPTY},
112 {'P', 2, EMPTY},
113 {'q', 1, EMPTY},
114 {'r', 1, RFILE},
115 {'s', 2, SUBST},
116 {'t', 2, BRANCH},
117 {'w', 2, WFILE},
118 {'x', 2, EMPTY},
119 {'y', 2, TR},
120 {'!', 2, NONSEL},
121 {':', 0, LABEL},
122 {'#', 0, COMMENT},
123 {'=', 1, EMPTY},
124 {'\0', 0, COMMENT},
125 };
126
127 /* The compiled program. */
128 struct s_command *prog;
129
130 /*
131 * Compile the program into prog.
132 * Initialise appends.
133 */
134 void
compile(void)135 compile(void)
136 {
137 *compile_stream(&prog) = NULL;
138 fixuplabel(prog, NULL);
139 uselabel();
140 if (appendnum == 0)
141 appends = NULL;
142 else if ((appends = malloc(sizeof(struct s_appends) * appendnum)) ==
143 NULL)
144 err(1, "malloc");
145 if ((match = malloc((maxnsub + 1) * sizeof(regmatch_t))) == NULL)
146 err(1, "malloc");
147 }
148
149 #define EATSPACE() do { \
150 if (p) \
151 while (*p && isspace((unsigned char)*p)) \
152 p++; \
153 } while (0)
154
155 static struct s_command **
compile_stream(struct s_command ** link)156 compile_stream(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 char re[_POSIX2_LINE_MAX + 1];
163 int naddr; /* Number of addresses */
164
165 stack = NULL;
166 for (;;) {
167 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) {
168 if (stack != NULL)
169 errx(1, "%lu: %s: unexpected EOF (pending }'s)",
170 linenum, fname);
171 return (link);
172 }
173
174 semicolon: EATSPACE();
175 if (p) {
176 if (*p == '#' || *p == '\0')
177 continue;
178 else if (*p == ';') {
179 p++;
180 goto semicolon;
181 }
182 }
183 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL)
184 err(1, "malloc");
185 link = &cmd->next;
186 cmd->startline = cmd->nonsel = 0;
187 /* First parse the addresses */
188 naddr = 0;
189
190 /* Valid characters to start an address */
191 #define addrchar(c) (strchr("0123456789/\\$", (c)))
192 if (addrchar(*p)) {
193 naddr++;
194 if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL)
195 err(1, "malloc");
196 p = compile_addr(p, cmd->a1);
197 EATSPACE(); /* EXTENSION */
198 if (*p == ',') {
199 p++;
200 EATSPACE(); /* EXTENSION */
201 naddr++;
202 if ((cmd->a2 = malloc(sizeof(struct s_addr)))
203 == NULL)
204 err(1, "malloc");
205 p = compile_addr(p, cmd->a2);
206 EATSPACE();
207 } else
208 cmd->a2 = NULL;
209 } else
210 cmd->a1 = cmd->a2 = NULL;
211
212 nonsel: /* Now parse the command */
213 if (!*p)
214 errx(1, "%lu: %s: command expected", linenum, fname);
215 cmd->code = *p;
216 for (fp = cmd_fmts; fp->code; fp++)
217 if (fp->code == *p)
218 break;
219 if (!fp->code)
220 errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p);
221 if (naddr > fp->naddr)
222 errx(1,
223 "%lu: %s: command %c expects up to %d address(es), found %d",
224 linenum, fname, *p, fp->naddr, naddr);
225 switch (fp->args) {
226 case NONSEL: /* ! */
227 p++;
228 EATSPACE();
229 cmd->nonsel = 1;
230 goto nonsel;
231 case GROUP: /* { */
232 p++;
233 EATSPACE();
234 cmd->next = stack;
235 stack = cmd;
236 link = &cmd->u.c;
237 if (*p)
238 goto semicolon;
239 break;
240 case ENDGROUP:
241 /*
242 * Short-circuit command processing, since end of
243 * group is really just a noop.
244 */
245 cmd->nonsel = 1;
246 if (stack == NULL)
247 errx(1, "%lu: %s: unexpected }", linenum, fname);
248 cmd2 = stack;
249 stack = cmd2->next;
250 cmd2->next = cmd;
251 /*FALLTHROUGH*/
252 case EMPTY: /* d D g G h H l n N p P q x = \0 */
253 p++;
254 EATSPACE();
255 if (*p == ';') {
256 p++;
257 link = &cmd->next;
258 goto semicolon;
259 }
260 if (*p)
261 errx(1, "%lu: %s: extra characters at the end of %c command",
262 linenum, fname, cmd->code);
263 break;
264 case TEXT: /* a c i */
265 p++;
266 EATSPACE();
267 if (*p != '\\')
268 errx(1,
269 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code);
270 p++;
271 EATSPACE();
272 if (*p)
273 errx(1,
274 "%lu: %s: extra characters after \\ at the end of %c command",
275 linenum, fname, cmd->code);
276 cmd->t = compile_text();
277 break;
278 case COMMENT: /* \0 # */
279 break;
280 case WFILE: /* w */
281 p++;
282 EATSPACE();
283 if (*p == '\0')
284 errx(1, "%lu: %s: filename expected", linenum, fname);
285 cmd->t = duptoeol(p, "w command");
286 if (aflag)
287 cmd->u.fd = -1;
288 else if ((cmd->u.fd = open(p,
289 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
290 DEFFILEMODE)) == -1)
291 err(1, "%s", p);
292 break;
293 case RFILE: /* r */
294 p++;
295 EATSPACE();
296 if (*p == '\0')
297 errx(1, "%lu: %s: filename expected", linenum, fname);
298 else
299 cmd->t = duptoeol(p, "read command");
300 break;
301 case BRANCH: /* b t */
302 p++;
303 EATSPACE();
304 if (*p == '\0')
305 cmd->t = NULL;
306 else
307 cmd->t = duptoeol(p, "branch");
308 break;
309 case LABEL: /* : */
310 p++;
311 EATSPACE();
312 cmd->t = duptoeol(p, "label");
313 if (strlen(p) == 0)
314 errx(1, "%lu: %s: empty label", linenum, fname);
315 enterlabel(cmd);
316 break;
317 case SUBST: /* s */
318 p++;
319 if (*p == '\0' || *p == '\\')
320 errx(1,
321 "%lu: %s: substitute pattern can not be delimited by newline or backslash",
322 linenum, fname);
323 if ((cmd->u.s = calloc(1, sizeof(struct s_subst))) == NULL)
324 err(1, "malloc");
325 p = compile_delimited(p, re, 0);
326 if (p == NULL)
327 errx(1,
328 "%lu: %s: unterminated substitute pattern", linenum, fname);
329
330 /* Compile RE with no case sensitivity temporarily */
331 if (*re == '\0')
332 cmd->u.s->re = NULL;
333 else
334 cmd->u.s->re = compile_re(re, 0);
335 --p;
336 p = compile_subst(p, cmd->u.s);
337 p = compile_flags(p, cmd->u.s);
338
339 /* Recompile RE with case sensitivity from "I" flag if any */
340 if (*re == '\0')
341 cmd->u.s->re = NULL;
342 else
343 cmd->u.s->re = compile_re(re, cmd->u.s->icase);
344 EATSPACE();
345 if (*p == ';') {
346 p++;
347 link = &cmd->next;
348 goto semicolon;
349 }
350 break;
351 case TR: /* y */
352 p++;
353 p = compile_tr(p, &cmd->u.y);
354 EATSPACE();
355 if (*p == ';') {
356 p++;
357 link = &cmd->next;
358 goto semicolon;
359 }
360 if (*p)
361 errx(1,
362 "%lu: %s: extra text at the end of a transform command", linenum, fname);
363 break;
364 }
365 }
366 }
367
368 /*
369 * Get a delimited string. P points to the delimiter of the string; d points
370 * to a buffer area. Newline and delimiter escapes are processed; other
371 * escapes are ignored.
372 *
373 * Returns a pointer to the first character after the final delimiter or NULL
374 * in the case of a non-terminated string. The character array d is filled
375 * with the processed string.
376 */
377 static char *
compile_delimited(char * p,char * d,int is_tr)378 compile_delimited(char *p, char *d, int is_tr)
379 {
380 char c;
381
382 c = *p++;
383 if (c == '\0')
384 return (NULL);
385 else if (c == '\\')
386 errx(1, "%lu: %s: \\ can not be used as a string delimiter",
387 linenum, fname);
388 else if (c == '\n')
389 errx(1, "%lu: %s: newline can not be used as a string delimiter",
390 linenum, fname);
391 while (*p) {
392 if (*p == '[' && *p != c) {
393 if ((d = compile_ccl(&p, d)) == NULL)
394 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname);
395 continue;
396 } else if (*p == '\\' && p[1] == '[') {
397 *d++ = *p++;
398 } else if (*p == '\\' && p[1] == c)
399 p++;
400 else if (*p == '\\' && p[1] == 'n') {
401 *d++ = '\n';
402 p += 2;
403 continue;
404 } else if (*p == '\\' && p[1] == '\\') {
405 if (is_tr)
406 p++;
407 else
408 *d++ = *p++;
409 } else if (*p == c) {
410 *d = '\0';
411 return (p + 1);
412 }
413 *d++ = *p++;
414 }
415 return (NULL);
416 }
417
418
419 /* compile_ccl: expand a POSIX character class */
420 static char *
compile_ccl(char ** sp,char * t)421 compile_ccl(char **sp, char *t)
422 {
423 int c, d;
424 char *s = *sp;
425
426 *t++ = *s++;
427 if (*s == '^')
428 *t++ = *s++;
429 if (*s == ']')
430 *t++ = *s++;
431 for (; *s && (*t = *s) != ']'; s++, t++)
432 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
433 *++t = *++s, t++, s++;
434 for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
435 if ((c = *s) == '\0')
436 return NULL;
437 }
438 return (*s == ']') ? *sp = ++s, ++t : NULL;
439 }
440
441 /*
442 * Compiles the regular expression in RE and returns a pointer to the compiled
443 * regular expression.
444 * Cflags are passed to regcomp.
445 */
446 static regex_t *
compile_re(char * re,int case_insensitive)447 compile_re(char *re, int case_insensitive)
448 {
449 regex_t *rep;
450 int eval, flags;
451
452
453 flags = rflags;
454 if (case_insensitive)
455 flags |= REG_ICASE;
456 if ((rep = malloc(sizeof(regex_t))) == NULL)
457 err(1, "malloc");
458 if ((eval = regcomp(rep, re, flags)) != 0)
459 errx(1, "%lu: %s: RE error: %s",
460 linenum, fname, strregerror(eval, rep));
461 if (maxnsub < rep->re_nsub)
462 maxnsub = rep->re_nsub;
463 return (rep);
464 }
465
466 /*
467 * Compile the substitution string of a regular expression and set res to
468 * point to a saved copy of it. Nsub is the number of parenthesized regular
469 * expressions.
470 */
471 static char *
compile_subst(char * p,struct s_subst * s)472 compile_subst(char *p, struct s_subst *s)
473 {
474 static char lbuf[_POSIX2_LINE_MAX + 1];
475 int asize, size;
476 u_char ref;
477 char c, *text, *op, *sp;
478 int more = 1, sawesc = 0;
479
480 c = *p++; /* Terminator character */
481 if (c == '\0')
482 return (NULL);
483
484 s->maxbref = 0;
485 s->linenum = linenum;
486 asize = 2 * _POSIX2_LINE_MAX + 1;
487 if ((text = malloc(asize)) == NULL)
488 err(1, "malloc");
489 size = 0;
490 do {
491 op = sp = text + size;
492 for (; *p; p++) {
493 if (*p == '\\' || sawesc) {
494 /*
495 * If this is a continuation from the last
496 * buffer, we won't have a character to
497 * skip over.
498 */
499 if (sawesc)
500 sawesc = 0;
501 else
502 p++;
503
504 if (*p == '\0') {
505 /*
506 * This escaped character is continued
507 * in the next part of the line. Note
508 * this fact, then cause the loop to
509 * exit w/ normal EOL case and reenter
510 * above with the new buffer.
511 */
512 sawesc = 1;
513 p--;
514 continue;
515 } else if (strchr("123456789", *p) != NULL) {
516 *sp++ = '\\';
517 ref = *p - '0';
518 if (s->re != NULL &&
519 ref > s->re->re_nsub)
520 errx(1, "%lu: %s: \\%c not defined in the RE",
521 linenum, fname, *p);
522 if (s->maxbref < ref)
523 s->maxbref = ref;
524 } else if (*p == '&' || *p == '\\')
525 *sp++ = '\\';
526 } else if (*p == c) {
527 if (*++p == '\0' && more) {
528 if (cu_fgets(lbuf, sizeof(lbuf), &more))
529 p = lbuf;
530 }
531 *sp++ = '\0';
532 size += sp - op;
533 if ((s->new = realloc(text, size)) == NULL)
534 err(1, "realloc");
535 return (p);
536 } else if (*p == '\n') {
537 errx(1,
538 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname);
539 /* NOTREACHED */
540 }
541 *sp++ = *p;
542 }
543 size += sp - op;
544 if (asize - size < _POSIX2_LINE_MAX + 1) {
545 asize *= 2;
546 if ((text = realloc(text, asize)) == NULL)
547 err(1, "realloc");
548 }
549 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more) != NULL);
550 errx(1, "%lu: %s: unterminated substitute in regular expression",
551 linenum, fname);
552 /* NOTREACHED */
553 }
554
555 /*
556 * Compile the flags of the s command
557 */
558 static char *
compile_flags(char * p,struct s_subst * s)559 compile_flags(char *p, struct s_subst *s)
560 {
561 int gn; /* True if we have seen g or n */
562 unsigned long nval;
563 char wfile[_POSIX2_LINE_MAX + 1], *q, *eq;
564
565 s->n = 1; /* Default */
566 s->p = 0;
567 s->wfile = NULL;
568 s->wfd = -1;
569 s->icase = 0;
570 for (gn = 0;;) {
571 EATSPACE(); /* EXTENSION */
572 switch (*p) {
573 case 'g':
574 if (gn)
575 errx(1,
576 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
577 gn = 1;
578 s->n = 0;
579 break;
580 case '\0':
581 case '\n':
582 case ';':
583 return (p);
584 case 'p':
585 s->p = 1;
586 break;
587 case 'i':
588 case 'I':
589 s->icase = 1;
590 break;
591 case '1': case '2': case '3':
592 case '4': case '5': case '6':
593 case '7': case '8': case '9':
594 if (gn)
595 errx(1,
596 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
597 gn = 1;
598 errno = 0;
599 nval = strtol(p, &p, 10);
600 if (errno == ERANGE || nval > INT_MAX)
601 errx(1,
602 "%lu: %s: overflow in the 'N' substitute flag", linenum, fname);
603 s->n = nval;
604 p--;
605 break;
606 case 'w':
607 p++;
608 #ifdef HISTORIC_PRACTICE
609 if (*p != ' ') {
610 warnx("%lu: %s: space missing before w wfile", linenum, fname);
611 return (p);
612 }
613 #endif
614 EATSPACE();
615 q = wfile;
616 eq = wfile + sizeof(wfile) - 1;
617 while (*p) {
618 if (*p == '\n')
619 break;
620 if (q >= eq)
621 err(1, "wfile too long");
622 *q++ = *p++;
623 }
624 *q = '\0';
625 if (q == wfile)
626 errx(1, "%lu: %s: no wfile specified", linenum, fname);
627 s->wfile = strdup(wfile);
628 if (!aflag && (s->wfd = open(wfile,
629 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
630 DEFFILEMODE)) == -1)
631 err(1, "%s", wfile);
632 return (p);
633 default:
634 errx(1, "%lu: %s: bad flag in substitute command: '%c'",
635 linenum, fname, *p);
636 break;
637 }
638 p++;
639 }
640 }
641
642 /*
643 * Compile a translation set of strings into a lookup table.
644 */
645 static char *
compile_tr(char * p,struct s_tr ** py)646 compile_tr(char *p, struct s_tr **py)
647 {
648 struct s_tr *y;
649 int i;
650 const char *op, *np;
651 char old[_POSIX2_LINE_MAX + 1];
652 char new[_POSIX2_LINE_MAX + 1];
653 size_t oclen, oldlen, nclen, newlen;
654 mbstate_t mbs1, mbs2;
655
656 if ((*py = y = malloc(sizeof(*y))) == NULL)
657 err(1, NULL);
658 y->multis = NULL;
659 y->nmultis = 0;
660
661 if (*p == '\0' || *p == '\\')
662 errx(1,
663 "%lu: %s: transform pattern can not be delimited by newline or backslash",
664 linenum, fname);
665 p = compile_delimited(p, old, 1);
666 if (p == NULL)
667 errx(1, "%lu: %s: unterminated transform source string",
668 linenum, fname);
669 p = compile_delimited(p - 1, new, 1);
670 if (p == NULL)
671 errx(1, "%lu: %s: unterminated transform target string",
672 linenum, fname);
673 EATSPACE();
674 op = old;
675 oldlen = mbsrtowcs(NULL, &op, 0, NULL);
676 if (oldlen == (size_t)-1)
677 err(1, NULL);
678 np = new;
679 newlen = mbsrtowcs(NULL, &np, 0, NULL);
680 if (newlen == (size_t)-1)
681 err(1, NULL);
682 if (newlen != oldlen)
683 errx(1, "%lu: %s: transform strings are not the same length",
684 linenum, fname);
685 if (MB_CUR_MAX == 1) {
686 /*
687 * The single-byte encoding case is easy: generate a
688 * lookup table.
689 */
690 for (i = 0; i <= UCHAR_MAX; i++)
691 y->bytetab[i] = (char)i;
692 for (; *op; op++, np++)
693 y->bytetab[(u_char)*op] = *np;
694 } else {
695 /*
696 * Multi-byte encoding case: generate a lookup table as
697 * above, but only for single-byte characters. The first
698 * bytes of multi-byte characters have their lookup table
699 * entries set to 0, which causes do_tr() to search through
700 * an auxiliary vector of multi-byte mappings.
701 */
702 memset(&mbs1, 0, sizeof(mbs1));
703 memset(&mbs2, 0, sizeof(mbs2));
704 for (i = 0; i <= UCHAR_MAX; i++)
705 y->bytetab[i] = (btowc(i) != WEOF) ? i : 0;
706 while (*op != '\0') {
707 oclen = mbrlen(op, MB_LEN_MAX, &mbs1);
708 if (oclen == (size_t)-1 || oclen == (size_t)-2)
709 errc(1, EILSEQ, NULL);
710 nclen = mbrlen(np, MB_LEN_MAX, &mbs2);
711 if (nclen == (size_t)-1 || nclen == (size_t)-2)
712 errc(1, EILSEQ, NULL);
713 if (oclen == 1 && nclen == 1)
714 y->bytetab[(u_char)*op] = *np;
715 else {
716 y->bytetab[(u_char)*op] = 0;
717 y->multis = realloc(y->multis,
718 (y->nmultis + 1) * sizeof(*y->multis));
719 if (y->multis == NULL)
720 err(1, NULL);
721 i = y->nmultis++;
722 y->multis[i].fromlen = oclen;
723 memcpy(y->multis[i].from, op, oclen);
724 y->multis[i].tolen = nclen;
725 memcpy(y->multis[i].to, np, nclen);
726 }
727 op += oclen;
728 np += nclen;
729 }
730 }
731 return (p);
732 }
733
734 /*
735 * Compile the text following an a, c, or i command.
736 */
737 static char *
compile_text(void)738 compile_text(void)
739 {
740 int asize, esc_nl, size;
741 char *text, *p, *op, *s;
742 char lbuf[_POSIX2_LINE_MAX + 1];
743
744 asize = 2 * _POSIX2_LINE_MAX + 1;
745 if ((text = malloc(asize)) == NULL)
746 err(1, "malloc");
747 size = 0;
748 while (cu_fgets(lbuf, sizeof(lbuf), NULL) != NULL) {
749 op = s = text + size;
750 p = lbuf;
751 #ifdef LEGACY_BSDSED_COMPAT
752 EATSPACE();
753 #endif
754 for (esc_nl = 0; *p != '\0'; p++) {
755 if (*p == '\\' && p[1] != '\0' && *++p == '\n')
756 esc_nl = 1;
757 *s++ = *p;
758 }
759 size += s - op;
760 if (!esc_nl) {
761 *s = '\0';
762 break;
763 }
764 if (asize - size < _POSIX2_LINE_MAX + 1) {
765 asize *= 2;
766 if ((text = realloc(text, asize)) == NULL)
767 err(1, "realloc");
768 }
769 }
770 text[size] = '\0';
771 if ((p = realloc(text, size + 1)) == NULL)
772 err(1, "realloc");
773 return (p);
774 }
775
776 /*
777 * Get an address and return a pointer to the first character after
778 * it. Fill the structure pointed to according to the address.
779 */
780 static char *
compile_addr(char * p,struct s_addr * a)781 compile_addr(char *p, struct s_addr *a)
782 {
783 char *end, re[_POSIX2_LINE_MAX + 1];
784 int icase;
785
786 icase = 0;
787
788 a->type = 0;
789 switch (*p) {
790 case '\\': /* Context address */
791 ++p;
792 /* FALLTHROUGH */
793 case '/': /* Context address */
794 p = compile_delimited(p, re, 0);
795 if (p == NULL)
796 errx(1, "%lu: %s: unterminated regular expression", linenum, fname);
797 /* Check for case insensitive regexp flag */
798 if (*p == 'I') {
799 icase = 1;
800 p++;
801 }
802 if (*re == '\0')
803 a->u.r = NULL;
804 else
805 a->u.r = compile_re(re, icase);
806 a->type = AT_RE;
807 return (p);
808
809 case '$': /* Last line */
810 a->type = AT_LAST;
811 return (p + 1);
812
813 case '+': /* Relative line number */
814 a->type = AT_RELLINE;
815 p++;
816 /* FALLTHROUGH */
817 /* Line number */
818 case '0': case '1': case '2': case '3': case '4':
819 case '5': case '6': case '7': case '8': case '9':
820 if (a->type == 0)
821 a->type = AT_LINE;
822 a->u.l = strtol(p, &end, 10);
823 return (end);
824 default:
825 errx(1, "%lu: %s: expected context address", linenum, fname);
826 return (NULL);
827 }
828 }
829
830 /*
831 * duptoeol --
832 * Return a copy of all the characters up to \n or \0.
833 */
834 static char *
duptoeol(char * s,const char * ctype)835 duptoeol(char *s, const char *ctype)
836 {
837 size_t len;
838 int ws;
839 char *p, *start;
840
841 ws = 0;
842 for (start = s; *s != '\0' && *s != '\n'; ++s)
843 ws = isspace((unsigned char)*s);
844 *s = '\0';
845 if (ws)
846 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype);
847 len = s - start + 1;
848 if ((p = malloc(len)) == NULL)
849 err(1, "malloc");
850 return (memmove(p, start, len));
851 }
852
853 /*
854 * Convert goto label names to addresses, and count a and r commands, in
855 * the given subset of the script. Free the memory used by labels in b
856 * and t commands (but not by :).
857 *
858 * TODO: Remove } nodes
859 */
860 static void
fixuplabel(struct s_command * cp,struct s_command * end)861 fixuplabel(struct s_command *cp, struct s_command *end)
862 {
863
864 for (; cp != end; cp = cp->next)
865 switch (cp->code) {
866 case 'a':
867 case 'r':
868 appendnum++;
869 break;
870 case 'b':
871 case 't':
872 /* Resolve branch target. */
873 if (cp->t == NULL) {
874 cp->u.c = NULL;
875 break;
876 }
877 if ((cp->u.c = findlabel(cp->t)) == NULL)
878 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t);
879 free(cp->t);
880 break;
881 case '{':
882 /* Do interior commands. */
883 fixuplabel(cp->u.c, cp->next);
884 break;
885 }
886 }
887
888 /*
889 * Associate the given command label for later lookup.
890 */
891 static void
enterlabel(struct s_command * cp)892 enterlabel(struct s_command *cp)
893 {
894 struct labhash **lhp, *lh;
895 u_char *p;
896 u_int h, c;
897
898 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
899 h = (h << 5) + h + c;
900 lhp = &labels[h & LHMASK];
901 for (lh = *lhp; lh != NULL; lh = lh->lh_next)
902 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
903 errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t);
904 if ((lh = malloc(sizeof *lh)) == NULL)
905 err(1, "malloc");
906 lh->lh_next = *lhp;
907 lh->lh_hash = h;
908 lh->lh_cmd = cp;
909 lh->lh_ref = 0;
910 *lhp = lh;
911 }
912
913 /*
914 * Find the label contained in the command l in the command linked
915 * list cp. L is excluded from the search. Return NULL if not found.
916 */
917 static struct s_command *
findlabel(char * name)918 findlabel(char *name)
919 {
920 struct labhash *lh;
921 u_char *p;
922 u_int h, c;
923
924 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
925 h = (h << 5) + h + c;
926 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
927 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
928 lh->lh_ref = 1;
929 return (lh->lh_cmd);
930 }
931 }
932 return (NULL);
933 }
934
935 /*
936 * Warn about any unused labels. As a side effect, release the label hash
937 * table space.
938 */
939 static void
uselabel(void)940 uselabel(void)
941 {
942 struct labhash *lh, *next;
943 int i;
944
945 for (i = 0; i < LHSZ; i++) {
946 for (lh = labels[i]; lh != NULL; lh = next) {
947 next = lh->lh_next;
948 if (!lh->lh_ref)
949 warnx("%lu: %s: unused label '%s'",
950 linenum, fname, lh->lh_cmd->t);
951 free(lh);
952 }
953 }
954 }
955