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 == '\\' &&
401 (p[1] == 'n' || p[1] == 'r' || p[1] == 't')) {
415 } else if (*p == '\\' && p[1] == '\\') {
420 } else if (*p == c) {
430 /* compile_ccl: expand a POSIX character class */
432 compile_ccl(char **sp, char *t)
442 for (; *s && (*t = *s) != ']'; s++, t++) {
443 if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
444 *++t = *++s, t++, s++;
445 for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
446 if ((c = *s) == '\0')
448 } else if (*s == '\\') {
465 return (*s == ']') ? *sp = ++s, ++t : NULL;
469 * Compiles the regular expression in RE and returns a pointer to the compiled
470 * regular expression.
471 * Cflags are passed to regcomp.
474 compile_re(char *re, int case_insensitive)
481 if (case_insensitive)
483 if ((rep = malloc(sizeof(regex_t))) == NULL)
485 if ((eval = regcomp(rep, re, flags)) != 0)
486 errx(1, "%lu: %s: RE error: %s",
487 linenum, fname, strregerror(eval, rep));
488 if (maxnsub < rep->re_nsub)
489 maxnsub = rep->re_nsub;
494 * Compile the substitution string of a regular expression and set res to
495 * point to a saved copy of it. Nsub is the number of parenthesized regular
499 compile_subst(char *p, struct s_subst *s)
501 static char lbuf[_POSIX2_LINE_MAX + 1];
504 char c, *text, *op, *sp;
505 int more = 1, sawesc = 0;
507 c = *p++; /* Terminator character */
512 s->linenum = linenum;
513 asize = 2 * _POSIX2_LINE_MAX + 1;
514 if ((text = malloc(asize)) == NULL)
518 op = sp = text + size;
520 if (*p == '\\' || sawesc) {
522 * If this is a continuation from the last
523 * buffer, we won't have a character to
533 * This escaped character is continued
534 * in the next part of the line. Note
535 * this fact, then cause the loop to
536 * exit w/ normal EOL case and reenter
537 * above with the new buffer.
542 } else if (strchr("123456789", *p) != NULL) {
546 ref > s->re->re_nsub)
547 errx(1, "%lu: %s: \\%c not defined in the RE",
549 if (s->maxbref < ref)
568 } else if (*p == c) {
569 if (*++p == '\0' && more) {
570 if (cu_fgets(lbuf, sizeof(lbuf), &more))
575 if ((s->new = realloc(text, size)) == NULL)
578 } else if (*p == '\n') {
580 "%lu: %s: unescaped newline inside substitute pattern", linenum, fname);
586 if (asize - size < _POSIX2_LINE_MAX + 1) {
588 if ((text = realloc(text, asize)) == NULL)
591 } while (cu_fgets(p = lbuf, sizeof(lbuf), &more) != NULL);
592 errx(1, "%lu: %s: unterminated substitute in regular expression",
598 * Compile the flags of the s command
601 compile_flags(char *p, struct s_subst *s)
603 int gn; /* True if we have seen g or n */
605 char wfile[_POSIX2_LINE_MAX + 1], *q, *eq;
607 s->n = 1; /* Default */
613 EATSPACE(); /* EXTENSION */
618 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
633 case '1': case '2': case '3':
634 case '4': case '5': case '6':
635 case '7': case '8': case '9':
638 "%lu: %s: more than one number or 'g' in substitute flags", linenum, fname);
641 nval = strtol(p, &p, 10);
642 if (errno == ERANGE || nval > INT_MAX)
644 "%lu: %s: overflow in the 'N' substitute flag", linenum, fname);
650 #ifdef HISTORIC_PRACTICE
652 warnx("%lu: %s: space missing before w wfile", linenum, fname);
658 eq = wfile + sizeof(wfile) - 1;
663 err(1, "wfile too long");
668 errx(1, "%lu: %s: no wfile specified", linenum, fname);
669 s->wfile = strdup(wfile);
670 if (!aflag && (s->wfd = open(wfile,
671 O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
676 errx(1, "%lu: %s: bad flag in substitute command: '%c'",
685 * Compile a translation set of strings into a lookup table.
688 compile_tr(char *p, struct s_tr **py)
693 char old[_POSIX2_LINE_MAX + 1];
694 char new[_POSIX2_LINE_MAX + 1];
695 size_t oclen, oldlen, nclen, newlen;
696 mbstate_t mbs1, mbs2;
698 if ((*py = y = malloc(sizeof(*y))) == NULL)
703 if (*p == '\0' || *p == '\\')
705 "%lu: %s: transform pattern can not be delimited by newline or backslash",
707 p = compile_delimited(p, old, 1);
709 errx(1, "%lu: %s: unterminated transform source string",
711 p = compile_delimited(p - 1, new, 1);
713 errx(1, "%lu: %s: unterminated transform target string",
717 oldlen = mbsrtowcs(NULL, &op, 0, NULL);
718 if (oldlen == (size_t)-1)
721 newlen = mbsrtowcs(NULL, &np, 0, NULL);
722 if (newlen == (size_t)-1)
724 if (newlen != oldlen)
725 errx(1, "%lu: %s: transform strings are not the same length",
727 if (MB_CUR_MAX == 1) {
729 * The single-byte encoding case is easy: generate a
732 for (i = 0; i <= UCHAR_MAX; i++)
733 y->bytetab[i] = (char)i;
734 for (; *op; op++, np++)
735 y->bytetab[(u_char)*op] = *np;
738 * Multi-byte encoding case: generate a lookup table as
739 * above, but only for single-byte characters. The first
740 * bytes of multi-byte characters have their lookup table
741 * entries set to 0, which causes do_tr() to search through
742 * an auxiliary vector of multi-byte mappings.
744 memset(&mbs1, 0, sizeof(mbs1));
745 memset(&mbs2, 0, sizeof(mbs2));
746 for (i = 0; i <= UCHAR_MAX; i++)
747 y->bytetab[i] = (btowc(i) != WEOF) ? i : 0;
748 while (*op != '\0') {
749 oclen = mbrlen(op, MB_LEN_MAX, &mbs1);
750 if (oclen == (size_t)-1 || oclen == (size_t)-2)
751 errc(1, EILSEQ, NULL);
752 nclen = mbrlen(np, MB_LEN_MAX, &mbs2);
753 if (nclen == (size_t)-1 || nclen == (size_t)-2)
754 errc(1, EILSEQ, NULL);
755 if (oclen == 1 && nclen == 1)
756 y->bytetab[(u_char)*op] = *np;
758 y->bytetab[(u_char)*op] = 0;
759 y->multis = realloc(y->multis,
760 (y->nmultis + 1) * sizeof(*y->multis));
761 if (y->multis == NULL)
764 y->multis[i].fromlen = oclen;
765 memcpy(y->multis[i].from, op, oclen);
766 y->multis[i].tolen = nclen;
767 memcpy(y->multis[i].to, np, nclen);
777 * Compile the text following an a, c, or i command.
782 int asize, esc_nl, size;
783 char *text, *p, *op, *s;
784 char lbuf[_POSIX2_LINE_MAX + 1];
786 asize = 2 * _POSIX2_LINE_MAX + 1;
787 if ((text = malloc(asize)) == NULL)
790 while (cu_fgets(lbuf, sizeof(lbuf), NULL) != NULL) {
791 op = s = text + size;
793 #ifdef LEGACY_BSDSED_COMPAT
796 for (esc_nl = 0; *p != '\0'; p++) {
797 if (*p == '\\' && p[1] != '\0' && *++p == '\n')
806 if (asize - size < _POSIX2_LINE_MAX + 1) {
808 if ((text = realloc(text, asize)) == NULL)
813 if ((p = realloc(text, size + 1)) == NULL)
819 * Get an address and return a pointer to the first character after
820 * it. Fill the structure pointed to according to the address.
823 compile_addr(char *p, struct s_addr *a)
825 char *end, re[_POSIX2_LINE_MAX + 1];
832 case '\\': /* Context address */
835 case '/': /* Context address */
836 p = compile_delimited(p, re, 0);
838 errx(1, "%lu: %s: unterminated regular expression", linenum, fname);
839 /* Check for case insensitive regexp flag */
847 a->u.r = compile_re(re, icase);
851 case '$': /* Last line */
855 case '+': /* Relative line number */
856 a->type = AT_RELLINE;
860 case '0': case '1': case '2': case '3': case '4':
861 case '5': case '6': case '7': case '8': case '9':
864 a->u.l = strtol(p, &end, 10);
867 errx(1, "%lu: %s: expected context address", linenum, fname);
874 * Return a copy of all the characters up to \n or \0.
877 duptoeol(char *s, const char *ctype)
884 for (start = s; *s != '\0' && *s != '\n'; ++s)
885 ws = isspace((unsigned char)*s);
888 warnx("%lu: %s: whitespace after %s", linenum, fname, ctype);
890 if ((p = malloc(len)) == NULL)
892 return (memmove(p, start, len));
896 * Convert goto label names to addresses, and count a and r commands, in
897 * the given subset of the script. Free the memory used by labels in b
898 * and t commands (but not by :).
900 * TODO: Remove } nodes
903 fixuplabel(struct s_command *cp, struct s_command *end)
906 for (; cp != end; cp = cp->next)
914 /* Resolve branch target. */
919 if ((cp->u.c = findlabel(cp->t)) == NULL)
920 errx(1, "%lu: %s: undefined label '%s'", linenum, fname, cp->t);
924 /* Do interior commands. */
925 fixuplabel(cp->u.c, cp->next);
931 * Associate the given command label for later lookup.
934 enterlabel(struct s_command *cp)
936 struct labhash **lhp, *lh;
940 for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
941 h = (h << 5) + h + c;
942 lhp = &labels[h & LHMASK];
943 for (lh = *lhp; lh != NULL; lh = lh->lh_next)
944 if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
945 errx(1, "%lu: %s: duplicate label '%s'", linenum, fname, cp->t);
946 if ((lh = malloc(sizeof *lh)) == NULL)
956 * Find the label contained in the command l in the command linked
957 * list cp. L is excluded from the search. Return NULL if not found.
959 static struct s_command *
960 findlabel(char *name)
966 for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
967 h = (h << 5) + h + c;
968 for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
969 if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
978 * Warn about any unused labels. As a side effect, release the label hash
984 struct labhash *lh, *next;
987 for (i = 0; i < LHSZ; i++) {
988 for (lh = labels[i]; lh != NULL; lh = next) {
991 warnx("%lu: %s: unused label '%s'",
992 linenum, fname, lh->lh_cmd->t);