Annotation of src/usr.bin/sed/compile.c, Revision 1.30
1.30 ! reyk 1: /* $OpenBSD: compile.c,v 1.29 2009/08/07 03:30:56 djm Exp $ */
1.2 deraadt 2:
1.1 deraadt 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.
1.14 millert 19: * 3. Neither the name of the University nor the names of its contributors
1.1 deraadt 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: #ifndef lint
1.9 millert 37: /* from: static char sccsid[] = "@(#)compile.c 8.2 (Berkeley) 4/28/95"; */
1.30 ! reyk 38: static const char rcsid[] = "$OpenBSD: compile.c,v 1.29 2009/08/07 03:30:56 djm Exp $";
1.1 deraadt 39: #endif /* not lint */
40:
41: #include <sys/types.h>
42: #include <sys/stat.h>
43:
44: #include <ctype.h>
45: #include <errno.h>
46: #include <fcntl.h>
47: #include <limits.h>
48: #include <regex.h>
49: #include <stdio.h>
50: #include <stdlib.h>
51: #include <string.h>
52:
53: #include "defs.h"
54: #include "extern.h"
55:
56: #define LHSZ 128
57: #define LHMASK (LHSZ - 1)
58: static struct labhash {
59: struct labhash *lh_next;
60: u_int lh_hash;
61: struct s_command *lh_cmd;
62: int lh_ref;
63: } *labels[LHSZ];
64:
1.11 millert 65: static char *compile_addr(char *, struct s_addr *);
66: static char *compile_ccl(char **, char *);
67: static char *compile_delimited(char *, char *);
68: static char *compile_flags(char *, struct s_subst *);
69: static char *compile_re(char *, regex_t **);
70: static char *compile_subst(char *, struct s_subst *);
71: static char *compile_text(void);
72: static char *compile_tr(char *, char **);
1.1 deraadt 73: static struct s_command
1.11 millert 74: **compile_stream(struct s_command **);
1.16 deraadt 75: static char *duptoeol(char *, char *, char **);
1.11 millert 76: static void enterlabel(struct s_command *);
1.1 deraadt 77: static struct s_command
1.11 millert 78: *findlabel(char *);
79: static void fixuplabel(struct s_command *, struct s_command *);
80: static void uselabel(void);
1.1 deraadt 81:
82: /*
83: * Command specification. This is used to drive the command parser.
84: */
85: struct s_format {
86: char code; /* Command code */
87: int naddr; /* Number of address args */
88: enum e_args args; /* Argument type */
89: };
90:
91: static struct s_format cmd_fmts[] = {
92: {'{', 2, GROUP},
93: {'}', 0, ENDGROUP},
94: {'a', 1, TEXT},
95: {'b', 2, BRANCH},
96: {'c', 2, TEXT},
97: {'d', 2, EMPTY},
98: {'D', 2, EMPTY},
99: {'g', 2, EMPTY},
100: {'G', 2, EMPTY},
101: {'h', 2, EMPTY},
102: {'H', 2, EMPTY},
103: {'i', 1, TEXT},
104: {'l', 2, EMPTY},
105: {'n', 2, EMPTY},
106: {'N', 2, EMPTY},
107: {'p', 2, EMPTY},
108: {'P', 2, EMPTY},
109: {'q', 1, EMPTY},
110: {'r', 1, RFILE},
111: {'s', 2, SUBST},
112: {'t', 2, BRANCH},
113: {'w', 2, WFILE},
114: {'x', 2, EMPTY},
115: {'y', 2, TR},
116: {'!', 2, NONSEL},
117: {':', 0, LABEL},
118: {'#', 0, COMMENT},
119: {'=', 1, EMPTY},
120: {'\0', 0, COMMENT},
121: };
122:
123: /* The compiled program. */
124: struct s_command *prog;
125:
126: /*
127: * Compile the program into prog.
128: * Initialise appends.
129: */
130: void
1.15 deraadt 131: compile(void)
1.1 deraadt 132: {
133: *compile_stream(&prog) = NULL;
134: fixuplabel(prog, NULL);
135: uselabel();
136: appends = xmalloc(sizeof(struct s_appends) * appendnum);
137: match = xmalloc((maxnsub + 1) * sizeof(regmatch_t));
138: }
139:
140: #define EATSPACE() do { \
141: if (p) \
1.24 tedu 142: while (isascii(*p) && isspace(*p)) \
1.1 deraadt 143: p++; \
144: } while (0)
145:
146: static struct s_command **
1.15 deraadt 147: compile_stream(struct s_command **link)
1.1 deraadt 148: {
1.10 mpech 149: char *p;
1.25 millert 150: static char *lbuf; /* To avoid excessive malloc calls */
1.28 millert 151: static size_t bufsize;
1.1 deraadt 152: struct s_command *cmd, *cmd2, *stack;
153: struct s_format *fp;
154: int naddr; /* Number of addresses */
155:
156: stack = 0;
157: for (;;) {
1.28 millert 158: if ((p = cu_fgets(&lbuf, &bufsize)) == NULL) {
1.1 deraadt 159: if (stack != 0)
160: err(COMPILE, "unexpected EOF (pending }'s)");
161: return (link);
162: }
163:
164: semicolon: EATSPACE();
1.23 otto 165: if (*p == '#' || *p == '\0')
1.1 deraadt 166: continue;
1.23 otto 167: if (*p == ';') {
168: p++;
169: goto semicolon;
170: }
1.1 deraadt 171: *link = cmd = xmalloc(sizeof(struct s_command));
172: link = &cmd->next;
173: cmd->nonsel = cmd->inrange = 0;
174: /* First parse the addresses */
175: naddr = 0;
176:
177: /* Valid characters to start an address */
178: #define addrchar(c) (strchr("0123456789/\\$", (c)))
179: if (addrchar(*p)) {
180: naddr++;
181: cmd->a1 = xmalloc(sizeof(struct s_addr));
182: p = compile_addr(p, cmd->a1);
183: EATSPACE(); /* EXTENSION */
184: if (*p == ',') {
185: p++;
186: EATSPACE(); /* EXTENSION */
187: naddr++;
188: cmd->a2 = xmalloc(sizeof(struct s_addr));
189: p = compile_addr(p, cmd->a2);
190: EATSPACE();
1.22 tedu 191: } else {
1.1 deraadt 192: cmd->a2 = 0;
1.22 tedu 193: }
194: } else {
1.1 deraadt 195: cmd->a1 = cmd->a2 = 0;
1.22 tedu 196: }
1.1 deraadt 197:
198: nonsel: /* Now parse the command */
199: if (!*p)
200: err(COMPILE, "command expected");
201: cmd->code = *p;
202: for (fp = cmd_fmts; fp->code; fp++)
203: if (fp->code == *p)
204: break;
205: if (!fp->code)
206: err(COMPILE, "invalid command code %c", *p);
207: if (naddr > fp->naddr)
208: err(COMPILE,
1.22 tedu 209: "command %c expects up to %d address(es), found %d",
210: *p, fp->naddr, naddr);
1.1 deraadt 211: switch (fp->args) {
212: case NONSEL: /* ! */
213: p++;
214: EATSPACE();
215: cmd->nonsel = ! cmd->nonsel;
216: goto nonsel;
217: case GROUP: /* { */
218: p++;
219: EATSPACE();
220: cmd->next = stack;
221: stack = cmd;
222: link = &cmd->u.c;
223: if (*p)
224: goto semicolon;
225: break;
226: case ENDGROUP:
227: /*
228: * Short-circuit command processing, since end of
229: * group is really just a noop.
230: */
231: cmd->nonsel = 1;
232: if (stack == 0)
233: err(COMPILE, "unexpected }");
234: cmd2 = stack;
235: stack = cmd2->next;
236: cmd2->next = cmd;
237: /*FALLTHROUGH*/
238: case EMPTY: /* d D g G h H l n N p P q x = \0 */
239: p++;
240: EATSPACE();
241: if (*p == ';') {
242: p++;
243: link = &cmd->next;
244: goto semicolon;
245: }
246: if (*p)
247: err(COMPILE,
248: "extra characters at the end of %c command", cmd->code);
249: break;
250: case TEXT: /* a c i */
251: p++;
252: EATSPACE();
253: if (*p != '\\')
1.22 tedu 254: err(COMPILE, "command %c expects \\ followed by"
255: " text", cmd->code);
1.1 deraadt 256: p++;
257: EATSPACE();
258: if (*p)
1.22 tedu 259: err(COMPILE, "extra characters after \\ at the"
260: " end of %c command", cmd->code);
1.1 deraadt 261: cmd->t = compile_text();
262: break;
263: case COMMENT: /* \0 # */
264: break;
265: case WFILE: /* w */
266: p++;
267: EATSPACE();
268: if (*p == '\0')
269: err(COMPILE, "filename expected");
1.16 deraadt 270: cmd->t = duptoeol(p, "w command", NULL);
1.1 deraadt 271: if (aflag)
272: cmd->u.fd = -1;
1.21 deraadt 273: else if ((cmd->u.fd = open(p,
1.1 deraadt 274: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
275: DEFFILEMODE)) == -1)
1.12 jsyn 276: err(FATAL, "%s: %s", p, strerror(errno));
1.1 deraadt 277: break;
278: case RFILE: /* r */
279: p++;
280: EATSPACE();
1.16 deraadt 281: cmd->t = duptoeol(p, "read command", NULL);
1.1 deraadt 282: break;
283: case BRANCH: /* b t */
284: p++;
285: EATSPACE();
286: if (*p == '\0')
287: cmd->t = NULL;
288: else
1.16 deraadt 289: cmd->t = duptoeol(p, "branch", &p);
290: if (*p == ';') {
291: p++;
292: goto semicolon;
293: }
1.1 deraadt 294: break;
295: case LABEL: /* : */
296: p++;
297: EATSPACE();
1.16 deraadt 298: cmd->t = duptoeol(p, "label", &p);
299: if (strlen(cmd->t) == 0)
1.1 deraadt 300: err(COMPILE, "empty label");
301: enterlabel(cmd);
1.16 deraadt 302: if (*p == ';') {
303: p++;
304: goto semicolon;
305: }
1.1 deraadt 306: break;
307: case SUBST: /* s */
308: p++;
309: if (*p == '\0' || *p == '\\')
1.22 tedu 310: err(COMPILE, "substitute pattern can not be"
311: " delimited by newline or backslash");
1.1 deraadt 312: cmd->u.s = xmalloc(sizeof(struct s_subst));
313: p = compile_re(p, &cmd->u.s->re);
314: if (p == NULL)
315: err(COMPILE, "unterminated substitute pattern");
316: --p;
317: p = compile_subst(p, cmd->u.s);
318: p = compile_flags(p, cmd->u.s);
319: EATSPACE();
320: if (*p == ';') {
321: p++;
322: link = &cmd->next;
323: goto semicolon;
324: }
325: break;
326: case TR: /* y */
327: p++;
328: p = compile_tr(p, (char **)&cmd->u.y);
329: EATSPACE();
330: if (*p == ';') {
331: p++;
332: link = &cmd->next;
333: goto semicolon;
334: }
335: if (*p)
1.22 tedu 336: err(COMPILE, "extra text at the end of a"
337: " transform command");
1.1 deraadt 338: break;
339: }
340: }
341: }
342:
343: /*
344: * Get a delimited string. P points to the delimeter of the string; d points
345: * to a buffer area. Newline and delimiter escapes are processed; other
346: * escapes are ignored.
347: *
348: * Returns a pointer to the first character after the final delimiter or NULL
349: * in the case of a non-terminated string. The character array d is filled
350: * with the processed string.
351: */
352: static char *
1.15 deraadt 353: compile_delimited(char *p, char *d)
1.1 deraadt 354: {
355: char c;
356:
357: c = *p++;
358: if (c == '\0')
359: return (NULL);
360: else if (c == '\\')
361: err(COMPILE, "\\ can not be used as a string delimiter");
362: else if (c == '\n')
363: err(COMPILE, "newline can not be used as a string delimiter");
364: while (*p) {
365: if (*p == '[') {
366: if ((d = compile_ccl(&p, d)) == NULL)
367: err(COMPILE, "unbalanced brackets ([])");
368: continue;
369: } else if (*p == '\\' && p[1] == '[') {
370: *d++ = *p++;
1.22 tedu 371: } else if (*p == '\\' && p[1] == c) {
1.1 deraadt 372: p++;
1.22 tedu 373: } else if (*p == '\\' && p[1] == 'n') {
1.1 deraadt 374: *d++ = '\n';
375: p += 2;
376: continue;
1.22 tedu 377: } else if (*p == '\\' && p[1] == '\\') {
1.1 deraadt 378: *d++ = *p++;
1.22 tedu 379: } else if (*p == c) {
1.1 deraadt 380: *d = '\0';
381: return (p + 1);
382: }
383: *d++ = *p++;
384: }
385: return (NULL);
386: }
387:
388:
389: /* compile_ccl: expand a POSIX character class */
390: static char *
1.15 deraadt 391: compile_ccl(char **sp, char *t)
1.1 deraadt 392: {
393: int c, d;
394: char *s = *sp;
395:
396: *t++ = *s++;
397: if (*s == '^')
398: *t++ = *s++;
399: if (*s == ']')
400: *t++ = *s++;
401: for (; *s && (*t = *s) != ']'; s++, t++)
402: if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
403: *++t = *++s, t++, s++;
404: for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
405: if ((c = *s) == '\0')
406: return NULL;
1.22 tedu 407: } else if (*s == '\\' && s[1] == 'n') {
408: *t = '\n';
409: s++;
410: }
411: if (*s == ']') {
412: *sp = ++s;
413: return (++t);
414: } else {
415: return (NULL);
416: }
1.1 deraadt 417: }
418:
419: /*
420: * Get a regular expression. P points to the delimiter of the regular
421: * expression; repp points to the address of a regexp pointer. Newline
422: * and delimiter escapes are processed; other escapes are ignored.
423: * Returns a pointer to the first character after the final delimiter
424: * or NULL in the case of a non terminated regular expression. The regexp
425: * pointer is set to the compiled regular expression.
426: * Cflags are passed to regcomp.
427: */
428: static char *
1.15 deraadt 429: compile_re(char *p, regex_t **repp)
1.1 deraadt 430: {
431: int eval;
1.25 millert 432: char *re;
1.1 deraadt 433:
1.25 millert 434: re = xmalloc(strlen(p) + 1); /* strlen(re) <= strlen(p) */
1.1 deraadt 435: p = compile_delimited(p, re);
436: if (p && strlen(re) == 0) {
437: *repp = NULL;
1.25 millert 438: free(re);
1.1 deraadt 439: return (p);
440: }
441: *repp = xmalloc(sizeof(regex_t));
1.29 djm 442: if (p && (eval = regcomp(*repp, re, Eflag ? REG_EXTENDED : 0)) != 0)
1.1 deraadt 443: err(COMPILE, "RE error: %s", strregerror(eval, *repp));
444: if (maxnsub < (*repp)->re_nsub)
445: maxnsub = (*repp)->re_nsub;
1.25 millert 446: free(re);
1.1 deraadt 447: return (p);
448: }
449:
450: /*
451: * Compile the substitution string of a regular expression and set res to
452: * point to a saved copy of it. Nsub is the number of parenthesized regular
453: * expressions.
454: */
455: static char *
1.15 deraadt 456: compile_subst(char *p, struct s_subst *s)
1.1 deraadt 457: {
1.25 millert 458: static char *lbuf;
1.28 millert 459: static size_t bufsize;
1.1 deraadt 460: int asize, ref, size;
461: char c, *text, *op, *sp;
1.19 otto 462: int sawesc = 0;
1.1 deraadt 463:
464: c = *p++; /* Terminator character */
465: if (c == '\0')
466: return (NULL);
467:
468: s->maxbref = 0;
469: s->linenum = linenum;
1.28 millert 470: text = NULL;
471: asize = size = 0;
1.1 deraadt 472: do {
1.28 millert 473: size_t len = ROUNDLEN(strlen(p) + 1);
474: if (asize - size < len) {
1.25 millert 475: do {
1.28 millert 476: asize += len;
477: } while (asize - size < len);
1.25 millert 478: text = xrealloc(text, asize);
479: }
1.1 deraadt 480: op = sp = text + size;
481: for (; *p; p++) {
1.19 otto 482: if (*p == '\\' || sawesc) {
483: /*
484: * If this is a continuation from the last
485: * buffer, we won't have a character to
486: * skip over.
487: */
488: if (sawesc)
489: sawesc = 0;
490: else
491: p++;
492:
493: if (*p == '\0') {
494: /*
495: * This escaped character is continued
496: * in the next part of the line. Note
497: * this fact, then cause the loop to
498: * exit w/ normal EOL case and reenter
499: * above with the new buffer.
500: */
501: sawesc = 1;
502: p--;
503: continue;
504: } else if (strchr("123456789", *p) != NULL) {
1.1 deraadt 505: *sp++ = '\\';
506: ref = *p - '0';
507: if (s->re != NULL &&
508: ref > s->re->re_nsub)
509: err(COMPILE,
510: "\\%c not defined in the RE", *p);
511: if (s->maxbref < ref)
512: s->maxbref = ref;
513: } else if (*p == '&' || *p == '\\')
514: *sp++ = '\\';
515: } else if (*p == c) {
516: p++;
517: *sp++ = '\0';
518: size += sp - op;
519: s->new = xrealloc(text, size);
520: return (p);
521: } else if (*p == '\n') {
522: err(COMPILE,
523: "unescaped newline inside substitute pattern");
524: /* NOTREACHED */
525: }
526: *sp++ = *p;
527: }
528: size += sp - op;
1.28 millert 529: } while ((p = cu_fgets(&lbuf, &bufsize)));
1.1 deraadt 530: err(COMPILE, "unterminated substitute in regular expression");
531: /* NOTREACHED */
532: }
533:
534: /*
535: * Compile the flags of the s command
536: */
537: static char *
1.15 deraadt 538: compile_flags(char *p, struct s_subst *s)
1.1 deraadt 539: {
540: int gn; /* True if we have seen g or n */
1.17 otto 541: long l;
1.25 millert 542: char wfile[PATH_MAX], *q;
1.1 deraadt 543:
544: s->n = 1; /* Default */
545: s->p = 0;
546: s->wfile = NULL;
547: s->wfd = -1;
548: for (gn = 0;;) {
549: EATSPACE(); /* EXTENSION */
550: switch (*p) {
551: case 'g':
552: if (gn)
1.22 tedu 553: err(COMPILE, "more than one number or 'g' in"
554: " substitute flags");
1.1 deraadt 555: gn = 1;
556: s->n = 0;
557: break;
558: case '\0':
559: case '\n':
560: case ';':
561: return (p);
562: case 'p':
563: s->p = 1;
564: break;
565: case '1': case '2': case '3':
566: case '4': case '5': case '6':
567: case '7': case '8': case '9':
568: if (gn)
1.22 tedu 569: err(COMPILE, "more than one number or 'g' in"
570: " substitute flags");
1.1 deraadt 571: gn = 1;
1.17 otto 572: l = strtol(p, &p, 10);
573: if (l <= 0 || l >= INT_MAX)
574: err(COMPILE,
575: "number in substitute flags out of range");
576: s->n = (int)l;
577: continue;
1.1 deraadt 578: case 'w':
579: p++;
580: #ifdef HISTORIC_PRACTICE
581: if (*p != ' ') {
582: err(WARNING, "space missing before w wfile");
583: return (p);
584: }
585: #endif
586: EATSPACE();
587: q = wfile;
588: while (*p) {
589: if (*p == '\n')
590: break;
591: *q++ = *p++;
592: }
593: *q = '\0';
594: if (q == wfile)
595: err(COMPILE, "no wfile specified");
596: s->wfile = strdup(wfile);
597: if (!aflag && (s->wfd = open(wfile,
598: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
599: DEFFILEMODE)) == -1)
1.12 jsyn 600: err(FATAL, "%s: %s", wfile, strerror(errno));
1.1 deraadt 601: return (p);
602: default:
603: err(COMPILE,
604: "bad flag in substitute command: '%c'", *p);
605: break;
606: }
607: p++;
608: }
609: }
610:
611: /*
612: * Compile a translation set of strings into a lookup table.
613: */
614: static char *
1.15 deraadt 615: compile_tr(char *p, char **transtab)
1.1 deraadt 616: {
617: int i;
618: char *lt, *op, *np;
1.25 millert 619: char *old = NULL, *new = NULL;
1.1 deraadt 620:
621: if (*p == '\0' || *p == '\\')
622: err(COMPILE,
623: "transform pattern can not be delimited by newline or backslash");
1.25 millert 624: old = xmalloc(strlen(p) + 1);
1.1 deraadt 625: p = compile_delimited(p, old);
626: if (p == NULL) {
627: err(COMPILE, "unterminated transform source string");
1.25 millert 628: goto bad;
1.1 deraadt 629: }
1.25 millert 630: new = xmalloc(strlen(p) + 1);
1.1 deraadt 631: p = compile_delimited(--p, new);
632: if (p == NULL) {
633: err(COMPILE, "unterminated transform target string");
1.25 millert 634: goto bad;
1.1 deraadt 635: }
636: EATSPACE();
637: if (strlen(new) != strlen(old)) {
638: err(COMPILE, "transform strings are not the same length");
1.25 millert 639: goto bad;
1.1 deraadt 640: }
641: /* We assume characters are 8 bits */
1.18 deraadt 642: lt = xmalloc(UCHAR_MAX + 1);
1.1 deraadt 643: for (i = 0; i <= UCHAR_MAX; i++)
644: lt[i] = (char)i;
645: for (op = old, np = new; *op; op++, np++)
646: lt[(u_char)*op] = *np;
647: *transtab = lt;
1.25 millert 648: free(old);
649: free(new);
1.1 deraadt 650: return (p);
1.25 millert 651: bad:
652: free(old);
653: free(new);
654: return (NULL);
1.1 deraadt 655: }
656:
657: /*
1.7 deraadt 658: * Compile the text following an a, c, or i command.
1.1 deraadt 659: */
660: static char *
1.15 deraadt 661: compile_text(void)
1.1 deraadt 662: {
1.4 deraadt 663: int asize, esc_nl, size;
1.25 millert 664: char *lbuf, *text, *p, *op, *s;
1.28 millert 665: size_t bufsize;
1.1 deraadt 666:
1.28 millert 667: lbuf = text = NULL;
668: asize = size = 0;
669: while ((p = cu_fgets(&lbuf, &bufsize))) {
670: size_t len = ROUNDLEN(strlen(p) + 1);
671: if (asize - size < len) {
1.25 millert 672: do {
1.28 millert 673: asize += len;
674: } while (asize - size < len);
1.25 millert 675: text = xrealloc(text, asize);
676: }
1.1 deraadt 677: op = s = text + size;
1.4 deraadt 678: for (esc_nl = 0; *p != '\0'; p++) {
679: if (*p == '\\' && p[1] != '\0' && *++p == '\n')
680: esc_nl = 1;
1.1 deraadt 681: *s++ = *p;
682: }
683: size += s - op;
1.4 deraadt 684: if (!esc_nl) {
1.1 deraadt 685: *s = '\0';
686: break;
687: }
688: }
1.25 millert 689: free(lbuf);
1.8 brian 690: text[size] = '\0';
1.1 deraadt 691: return (xrealloc(text, size + 1));
692: }
693:
694: /*
695: * Get an address and return a pointer to the first character after
696: * it. Fill the structure pointed to according to the address.
697: */
698: static char *
1.15 deraadt 699: compile_addr(char *p, struct s_addr *a)
1.1 deraadt 700: {
701: char *end;
702:
703: switch (*p) {
704: case '\\': /* Context address */
705: ++p;
706: /* FALLTHROUGH */
707: case '/': /* Context address */
708: p = compile_re(p, &a->u.r);
709: if (p == NULL)
710: err(COMPILE, "unterminated regular expression");
711: a->type = AT_RE;
712: return (p);
713:
714: case '$': /* Last line */
715: a->type = AT_LAST;
716: return (p + 1);
717: /* Line number */
1.21 deraadt 718: case '0': case '1': case '2': case '3': case '4':
1.1 deraadt 719: case '5': case '6': case '7': case '8': case '9':
720: a->type = AT_LINE;
1.3 millert 721: a->u.l = strtoul(p, &end, 10);
1.1 deraadt 722: return (end);
723: default:
724: err(COMPILE, "expected context address");
725: return (NULL);
726: }
727: }
728:
729: /*
730: * duptoeol --
731: * Return a copy of all the characters up to \n or \0.
732: */
733: static char *
1.16 deraadt 734: duptoeol(char *s, char *ctype, char **semi)
1.1 deraadt 735: {
736: size_t len;
737: int ws;
738: char *start;
739:
740: ws = 0;
1.16 deraadt 741: if (semi) {
742: for (start = s; *s != '\0' && *s != '\n' && *s != ';'; ++s)
743: ws = isspace(*s);
744: } else {
745: for (start = s; *s != '\0' && *s != '\n'; ++s)
746: ws = isspace(*s);
747: *s = '\0';
748: }
1.1 deraadt 749: if (ws)
750: err(WARNING, "whitespace after %s", ctype);
751: len = s - start + 1;
1.16 deraadt 752: if (semi)
753: *semi = s;
754: s = xmalloc(len);
755: strlcpy(s, start, len);
756: return (s);
1.1 deraadt 757: }
758:
759: /*
760: * Convert goto label names to addresses, and count a and r commands, in
761: * the given subset of the script. Free the memory used by labels in b
762: * and t commands (but not by :).
763: *
764: * TODO: Remove } nodes
765: */
766: static void
1.15 deraadt 767: fixuplabel(struct s_command *cp, struct s_command *end)
1.1 deraadt 768: {
769:
770: for (; cp != end; cp = cp->next)
771: switch (cp->code) {
772: case 'a':
773: case 'r':
774: appendnum++;
775: break;
776: case 'b':
777: case 't':
778: /* Resolve branch target. */
779: if (cp->t == NULL) {
780: cp->u.c = NULL;
781: break;
782: }
783: if ((cp->u.c = findlabel(cp->t)) == NULL)
784: err(COMPILE2, "undefined label '%s'", cp->t);
785: free(cp->t);
786: break;
787: case '{':
788: /* Do interior commands. */
789: fixuplabel(cp->u.c, cp->next);
790: break;
791: }
792: }
793:
794: /*
795: * Associate the given command label for later lookup.
796: */
797: static void
1.15 deraadt 798: enterlabel(struct s_command *cp)
1.1 deraadt 799: {
1.10 mpech 800: struct labhash **lhp, *lh;
801: u_char *p;
802: u_int h, c;
1.1 deraadt 803:
804: for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
805: h = (h << 5) + h + c;
806: lhp = &labels[h & LHMASK];
807: for (lh = *lhp; lh != NULL; lh = lh->lh_next)
808: if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
809: err(COMPILE2, "duplicate label '%s'", cp->t);
810: lh = xmalloc(sizeof *lh);
811: lh->lh_next = *lhp;
812: lh->lh_hash = h;
813: lh->lh_cmd = cp;
814: lh->lh_ref = 0;
815: *lhp = lh;
816: }
817:
818: /*
819: * Find the label contained in the command l in the command linked
820: * list cp. L is excluded from the search. Return NULL if not found.
821: */
822: static struct s_command *
1.15 deraadt 823: findlabel(char *name)
1.1 deraadt 824: {
1.10 mpech 825: struct labhash *lh;
826: u_char *p;
827: u_int h, c;
1.1 deraadt 828:
829: for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
830: h = (h << 5) + h + c;
831: for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
832: if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
833: lh->lh_ref = 1;
834: return (lh->lh_cmd);
835: }
836: }
837: return (NULL);
838: }
839:
1.21 deraadt 840: /*
1.1 deraadt 841: * Warn about any unused labels. As a side effect, release the label hash
842: * table space.
843: */
844: static void
1.15 deraadt 845: uselabel(void)
1.1 deraadt 846: {
1.10 mpech 847: struct labhash *lh, *next;
848: int i;
1.1 deraadt 849:
850: for (i = 0; i < LHSZ; i++) {
851: for (lh = labels[i]; lh != NULL; lh = next) {
852: next = lh->lh_next;
853: if (!lh->lh_ref)
854: err(WARNING, "unused label '%s'",
855: lh->lh_cmd->t);
856: free(lh);
857: }
858: }
859: }