2 * SPDX-License-Identifier: BSD-3-Clause
4 * Copyright (c) 1992 Diomidis Spinellis.
5 * Copyright (c) 1992, 1993
6 * The Regents of the University of California. All rights reserved.
8 * This code is derived from software contributed to Berkeley by
9 * Diomidis Spinellis of Imperial College, University of London.
11 * Redistribution and use in source and binary forms, with or without
12 * modification, are permitted provided that the following conditions
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.
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
36 #include <sys/cdefs.h>
37 __FBSDID("$FreeBSD$");
40 static const char sccsid[] = "@(#)compile.c 8.1 (Berkeley) 6/6/93";
43 #include <sys/types.h>
61 #define LHMASK (LHSZ - 1)
62 static struct labhash {
63 struct labhash *lh_next;
65 struct s_command *lh_cmd;
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
83 static void fixuplabel(struct s_command *, struct s_command *);
84 static void uselabel(void);
87 * Command specification. This is used to drive the command parser.
90 char code; /* Command code */
91 int naddr; /* Number of address args */
92 enum e_args args; /* Argument type */
95 static struct s_format cmd_fmts[] = {
127 /* The compiled program. */
128 struct s_command *prog;
131 * Compile the program into prog.
132 * Initialise appends.
137 *compile_stream(&prog) = NULL;
138 fixuplabel(prog, NULL);
142 else if ((appends = malloc(sizeof(struct s_appends) * appendnum)) ==
145 if ((match = malloc((maxnsub + 1) * sizeof(regmatch_t))) == NULL)
149 #define EATSPACE() do { \
151 while (*p && isspace((unsigned char)*p)) \
155 static struct s_command **
156 compile_stream(struct s_command **link)
159 static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */
160 struct s_command *cmd, *cmd2, *stack;
162 char re[_POSIX2_LINE_MAX + 1];
163 int naddr; /* Number of addresses */
167 if ((p = cu_fgets(lbuf, sizeof(lbuf), NULL)) == NULL) {
169 errx(1, "%lu: %s: unexpected EOF (pending }'s)",
174 semicolon: EATSPACE();
176 if (*p == '#' || *p == '\0')
178 else if (*p == ';') {
183 if ((*link = cmd = malloc(sizeof(struct s_command))) == NULL)
186 cmd->startline = cmd->nonsel = 0;
187 /* First parse the addresses */
190 /* Valid characters to start an address */
191 #define addrchar(c) (strchr("0123456789/\\$", (c)))
194 if ((cmd->a1 = malloc(sizeof(struct s_addr))) == NULL)
196 p = compile_addr(p, cmd->a1);
197 EATSPACE(); /* EXTENSION */
200 EATSPACE(); /* EXTENSION */
202 if ((cmd->a2 = malloc(sizeof(struct s_addr)))
205 p = compile_addr(p, cmd->a2);
210 cmd->a1 = cmd->a2 = NULL;
212 nonsel: /* Now parse the command */
214 errx(1, "%lu: %s: command expected", linenum, fname);
216 for (fp = cmd_fmts; fp->code; fp++)
220 errx(1, "%lu: %s: invalid command code %c", linenum, fname, *p);
221 if (naddr > fp->naddr)
223 "%lu: %s: command %c expects up to %d address(es), found %d",
224 linenum, fname, *p, fp->naddr, naddr);
242 * Short-circuit command processing, since end of
243 * group is really just a noop.
247 errx(1, "%lu: %s: unexpected }", linenum, fname);
252 case EMPTY: /* d D g G h H l n N p P q x = \0 */
261 errx(1, "%lu: %s: extra characters at the end of %c command",
262 linenum, fname, cmd->code);
264 case TEXT: /* a c i */
269 "%lu: %s: command %c expects \\ followed by text", linenum, fname, cmd->code);
274 "%lu: %s: extra characters after \\ at the end of %c command",
275 linenum, fname, cmd->code);
276 cmd->t = compile_text();
278 case COMMENT: /* \0 # */
284 errx(1, "%lu: %s: filename expected", linenum, fname);
285 cmd->t = duptoeol(p, "w command");
288 else if ((cmd->u.fd = open(p,
289 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
297 errx(1, "%lu: %s: filename expected", linenum, fname);
299 cmd->t = duptoeol(p, "read command");
301 case BRANCH: /* b t */
307 cmd->t = duptoeol(p, "branch");
312 cmd->t = duptoeol(p, "label");
314 errx(1, "%lu: %s: empty label", linenum, fname);
319 if (*p == '\0' || *p == '\\')
321 "%lu: %s: substitute pattern can not be delimited by newline or backslash",
323 if ((cmd->u.s = calloc(1, sizeof(struct s_subst))) == NULL)
325 p = compile_delimited(p, re, 0);
328 "%lu: %s: unterminated substitute pattern", linenum, fname);
330 /* Compile RE with no case sensitivity temporarily */
334 cmd->u.s->re = compile_re(re, 0);
336 p = compile_subst(p, cmd->u.s);
337 p = compile_flags(p, cmd->u.s);
339 /* Recompile RE with case sensitivity from "I" flag if any */
343 cmd->u.s->re = compile_re(re, cmd->u.s->icase);
353 p = compile_tr(p, &cmd->u.y);
362 "%lu: %s: extra text at the end of a transform command", linenum, fname);
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.
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.
378 compile_delimited(char *p, char *d, int is_tr)
386 errx(1, "%lu: %s: \\ can not be used as a string delimiter",
389 errx(1, "%lu: %s: newline can not be used as a string delimiter",
392 if (*p == '[' && *p != c) {
393 if ((d = compile_ccl(&p, d)) == NULL)
394 errx(1, "%lu: %s: unbalanced brackets ([])", linenum, fname);
396 } else if (*p == '\\' && p[1] == '[') {
398 } else if (*p == '\\' && p[1] == c)
400 else if (*p == '\\' && p[1] == 'n') {
404 } else if (*p == '\\' && p[1] == '\\') {
409 } else if (*p == c) {
419 /* compile_ccl: expand a POSIX character class */
421 compile_ccl(char **sp, char *t)
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')
438 return (*s == ']') ? *sp = ++s, ++t : NULL;
442 * Compiles the regular expression in RE and returns a pointer to the compiled
443 * regular expression.
444 * Cflags are passed to regcomp.
447 compile_re(char *re, int case_insensitive)
454 if (case_insensitive)
456 if ((rep = malloc(sizeof(regex_t))) == NULL)
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;
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
472 compile_subst(char *p, struct s_subst *s)
474 static char lbuf[_POSIX2_LINE_MAX + 1];
477 char c, *text, *op, *sp;
478 int more = 1, sawesc = 0;
480 c = *p++; /* Terminator character */
485 s->linenum = linenum;
486 asize = 2 * _POSIX2_LINE_MAX + 1;
487 if ((text = malloc(asize)) == NULL)
491 op = sp = text + size;
493 if (*p == '\\' || sawesc) {
495 * If this is a continuation from the last
496 * buffer, we won't have a character to
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.
515 } else if (strchr("123456789", *p) != NULL) {
519 ref > s->re->re_nsub)
520 errx(1, "%lu: %s: \\%c not defined in the RE",
522 if (s->maxbref < ref)
524 } else if (*p == '&' || *p == '\\')
526 } else if (*p == c) {
527 if (*++p == '\0' && more) {
528 if (cu_fgets(lbuf, sizeof(lbuf), &more))
533 if ((s->new = realloc(text, size)) == NULL)
536 } else if (*p == '\n') {
538 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname);
544 if (asize - size < _POSIX2_LINE_MAX + 1) {
546 if ((text = realloc(text, asize)) == NULL)
549 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more) != NULL);
550 errx(1, "%lu: %s: unterminated substitute in regular expression",
556 * Compile the flags of the s command
559 compile_flags(char *p, struct s_subst *s)
561 int gn; /* True if we have seen g or n */
563 char wfile[_POSIX2_LINE_MAX + 1], *q, *eq;
565 s->n = 1; /* Default */
571 EATSPACE(); /* EXTENSION */
576 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
591 case '1': case '2': case '3':
592 case '4': case '5': case '6':
593 case '7': case '8': case '9':
596 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
599 nval = strtol(p, &p, 10);
600 if (errno == ERANGE || nval > INT_MAX)
602 "%lu: %s: overflow in the 'N' substitute flag", linenum, fname);
608 #ifdef HISTORIC_PRACTICE
610 warnx("%lu: %s: space missing before w wfile", linenum, fname);
616 eq = wfile + sizeof(wfile) - 1;
621 err(1, "wfile too long");
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,
634 errx(1, "%lu: %s: bad flag in substitute command: '%c'",
643 * Compile a translation set of strings into a lookup table.
646 compile_tr(char *p, struct s_tr **py)
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;
656 if ((*py = y = malloc(sizeof(*y))) == NULL)
661 if (*p == '\0' || *p == '\\')
663 "%lu: %s: transform pattern can not be delimited by newline or backslash",
665 p = compile_delimited(p, old, 1);
667 errx(1, "%lu: %s: unterminated transform source string",
669 p = compile_delimited(p - 1, new, 1);
671 errx(1, "%lu: %s: unterminated transform target string",
675 oldlen = mbsrtowcs(NULL, &op, 0, NULL);
676 if (oldlen == (size_t)-1)
679 newlen = mbsrtowcs(NULL, &np, 0, NULL);
680 if (newlen == (size_t)-1)
682 if (newlen != oldlen)
683 errx(1, "%lu: %s: transform strings are not the same length",
685 if (MB_CUR_MAX == 1) {
687 * The single-byte encoding case is easy: generate a
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;
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.
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;
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)
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);
735 * Compile the text following an a, c, or i command.
740 int asize, esc_nl, size;
741 char *text, *p, *op, *s;
742 char lbuf[_POSIX2_LINE_MAX + 1];
744 asize = 2 * _POSIX2_LINE_MAX + 1;
745 if ((text = malloc(asize)) == NULL)
748 while (cu_fgets(lbuf, sizeof(lbuf), NULL) != NULL) {
749 op = s = text + size;
751 #ifdef LEGACY_BSDSED_COMPAT
754 for (esc_nl = 0; *p != '\0'; p++) {
755 if (*p == '\\' && p[1] != '\0' && *++p == '\n')
764 if (asize - size < _POSIX2_LINE_MAX + 1) {
766 if ((text = realloc(text, asize)) == NULL)
771 if ((p = realloc(text, size + 1)) == NULL)
777 * Get an address and return a pointer to the first character after
778 * it. Fill the structure pointed to according to the address.
781 compile_addr(char *p, struct s_addr *a)
783 char *end, re[_POSIX2_LINE_MAX + 1];
790 case '\\': /* Context address */
793 case '/': /* Context address */
794 p = compile_delimited(p, re, 0);
796 errx(1, "%lu: %s: unterminated regular expression", linenum, fname);
797 /* Check for case insensitive regexp flag */
805 a->u.r = compile_re(re, icase);
809 case '$': /* Last line */
813 case '+': /* Relative line number */
814 a->type = AT_RELLINE;
818 case '0': case '1': case '2': case '3': case '4':
819 case '5': case '6': case '7': case '8': case '9':
822 a->u.l = strtol(p, &end, 10);
825 errx(1, "%lu: %s: expected context address", linenum, fname);
832 * Return a copy of all the characters up to \n or \0.
835 duptoeol(char *s, const char *ctype)
842 for (start = s; *s != '\0' && *s != '\n'; ++s)
843 ws = isspace((unsigned char)*s);
846 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype);
848 if ((p = malloc(len)) == NULL)
850 return (memmove(p, start, len));
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 :).
858 * TODO: Remove } nodes
861 fixuplabel(struct s_command *cp, struct s_command *end)
864 for (; cp != end; cp = cp->next)
872 /* Resolve branch target. */
877 if ((cp->u.c = findlabel(cp->t)) == NULL)
878 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t);
882 /* Do interior commands. */
883 fixuplabel(cp->u.c, cp->next);
889 * Associate the given command label for later lookup.
892 enterlabel(struct s_command *cp)
894 struct labhash **lhp, *lh;
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)
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.
917 static struct s_command *
918 findlabel(char *name)
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) {
936 * Warn about any unused labels. As a side effect, release the label hash
942 struct labhash *lh, *next;
945 for (i = 0; i < LHSZ; i++) {
946 for (lh = labels[i]; lh != NULL; lh = next) {
949 warnx("%lu: %s: unused label '%s'",
950 linenum, fname, lh->lh_cmd->t);