Annotation of src/usr.bin/sed/compile.c, Revision 1.23
1.23 ! otto 1: /* $OpenBSD: compile.c,v 1.22 2006/10/09 00:23:56 tedu 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"; */
1.23 ! otto 38: static const char rcsid[] = "$OpenBSD: compile.c,v 1.22 2006/10/09 00:23:56 tedu 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) \
142: while (*p && isascii(*p) && isspace(*p)) \
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;
1.1 deraadt 150: static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */
151: struct s_command *cmd, *cmd2, *stack;
152: struct s_format *fp;
153: int naddr; /* Number of addresses */
154:
155: stack = 0;
156: for (;;) {
157: if ((p = cu_fgets(lbuf, sizeof(lbuf))) == NULL) {
158: if (stack != 0)
159: err(COMPILE, "unexpected EOF (pending }'s)");
160: return (link);
161: }
162:
163: semicolon: EATSPACE();
1.23 ! otto 164: if (*p == '#' || *p == '\0')
1.1 deraadt 165: continue;
1.23 ! otto 166: if (*p == ';') {
! 167: p++;
! 168: goto semicolon;
! 169: }
1.1 deraadt 170: *link = cmd = xmalloc(sizeof(struct s_command));
171: link = &cmd->next;
172: cmd->nonsel = cmd->inrange = 0;
173: /* First parse the addresses */
174: naddr = 0;
175:
176: /* Valid characters to start an address */
177: #define addrchar(c) (strchr("0123456789/\\$", (c)))
178: if (addrchar(*p)) {
179: naddr++;
180: cmd->a1 = xmalloc(sizeof(struct s_addr));
181: p = compile_addr(p, cmd->a1);
182: EATSPACE(); /* EXTENSION */
183: if (*p == ',') {
184: p++;
185: EATSPACE(); /* EXTENSION */
186: naddr++;
187: cmd->a2 = xmalloc(sizeof(struct s_addr));
188: p = compile_addr(p, cmd->a2);
189: EATSPACE();
1.22 tedu 190: } else {
1.1 deraadt 191: cmd->a2 = 0;
1.22 tedu 192: }
193: } else {
1.1 deraadt 194: cmd->a1 = cmd->a2 = 0;
1.22 tedu 195: }
1.1 deraadt 196:
197: nonsel: /* Now parse the command */
198: if (!*p)
199: err(COMPILE, "command expected");
200: cmd->code = *p;
201: for (fp = cmd_fmts; fp->code; fp++)
202: if (fp->code == *p)
203: break;
204: if (!fp->code)
205: err(COMPILE, "invalid command code %c", *p);
206: if (naddr > fp->naddr)
207: err(COMPILE,
1.22 tedu 208: "command %c expects up to %d address(es), found %d",
209: *p, fp->naddr, naddr);
1.1 deraadt 210: switch (fp->args) {
211: case NONSEL: /* ! */
212: p++;
213: EATSPACE();
214: cmd->nonsel = ! cmd->nonsel;
215: goto nonsel;
216: case GROUP: /* { */
217: p++;
218: EATSPACE();
219: cmd->next = stack;
220: stack = cmd;
221: link = &cmd->u.c;
222: if (*p)
223: goto semicolon;
224: break;
225: case ENDGROUP:
226: /*
227: * Short-circuit command processing, since end of
228: * group is really just a noop.
229: */
230: cmd->nonsel = 1;
231: if (stack == 0)
232: err(COMPILE, "unexpected }");
233: cmd2 = stack;
234: stack = cmd2->next;
235: cmd2->next = cmd;
236: /*FALLTHROUGH*/
237: case EMPTY: /* d D g G h H l n N p P q x = \0 */
238: p++;
239: EATSPACE();
240: if (*p == ';') {
241: p++;
242: link = &cmd->next;
243: goto semicolon;
244: }
245: if (*p)
246: err(COMPILE,
247: "extra characters at the end of %c command", cmd->code);
248: break;
249: case TEXT: /* a c i */
250: p++;
251: EATSPACE();
252: if (*p != '\\')
1.22 tedu 253: err(COMPILE, "command %c expects \\ followed by"
254: " text", cmd->code);
1.1 deraadt 255: p++;
256: EATSPACE();
257: if (*p)
1.22 tedu 258: err(COMPILE, "extra characters after \\ at the"
259: " end of %c command", cmd->code);
1.1 deraadt 260: cmd->t = compile_text();
261: break;
262: case COMMENT: /* \0 # */
263: break;
264: case WFILE: /* w */
265: p++;
266: EATSPACE();
267: if (*p == '\0')
268: err(COMPILE, "filename expected");
1.16 deraadt 269: cmd->t = duptoeol(p, "w command", NULL);
1.1 deraadt 270: if (aflag)
271: cmd->u.fd = -1;
1.21 deraadt 272: else if ((cmd->u.fd = open(p,
1.1 deraadt 273: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
274: DEFFILEMODE)) == -1)
1.12 jsyn 275: err(FATAL, "%s: %s", p, strerror(errno));
1.1 deraadt 276: break;
277: case RFILE: /* r */
278: p++;
279: EATSPACE();
1.16 deraadt 280: cmd->t = duptoeol(p, "read command", NULL);
1.1 deraadt 281: break;
282: case BRANCH: /* b t */
283: p++;
284: EATSPACE();
285: if (*p == '\0')
286: cmd->t = NULL;
287: else
1.16 deraadt 288: cmd->t = duptoeol(p, "branch", &p);
289: if (*p == ';') {
290: p++;
291: goto semicolon;
292: }
1.1 deraadt 293: break;
294: case LABEL: /* : */
295: p++;
296: EATSPACE();
1.16 deraadt 297: cmd->t = duptoeol(p, "label", &p);
298: if (strlen(cmd->t) == 0)
1.1 deraadt 299: err(COMPILE, "empty label");
300: enterlabel(cmd);
1.16 deraadt 301: if (*p == ';') {
302: p++;
303: goto semicolon;
304: }
1.1 deraadt 305: break;
306: case SUBST: /* s */
307: p++;
308: if (*p == '\0' || *p == '\\')
1.22 tedu 309: err(COMPILE, "substitute pattern can not be"
310: " delimited by newline or backslash");
1.1 deraadt 311: cmd->u.s = xmalloc(sizeof(struct s_subst));
312: p = compile_re(p, &cmd->u.s->re);
313: if (p == NULL)
314: err(COMPILE, "unterminated substitute pattern");
315: --p;
316: p = compile_subst(p, cmd->u.s);
317: p = compile_flags(p, cmd->u.s);
318: EATSPACE();
319: if (*p == ';') {
320: p++;
321: link = &cmd->next;
322: goto semicolon;
323: }
324: break;
325: case TR: /* y */
326: p++;
327: p = compile_tr(p, (char **)&cmd->u.y);
328: EATSPACE();
329: if (*p == ';') {
330: p++;
331: link = &cmd->next;
332: goto semicolon;
333: }
334: if (*p)
1.22 tedu 335: err(COMPILE, "extra text at the end of a"
336: " transform command");
1.1 deraadt 337: break;
338: }
339: }
340: }
341:
342: /*
343: * Get a delimited string. P points to the delimeter of the string; d points
344: * to a buffer area. Newline and delimiter escapes are processed; other
345: * escapes are ignored.
346: *
347: * Returns a pointer to the first character after the final delimiter or NULL
348: * in the case of a non-terminated string. The character array d is filled
349: * with the processed string.
350: */
351: static char *
1.15 deraadt 352: compile_delimited(char *p, char *d)
1.1 deraadt 353: {
354: char c;
355:
356: c = *p++;
357: if (c == '\0')
358: return (NULL);
359: else if (c == '\\')
360: err(COMPILE, "\\ can not be used as a string delimiter");
361: else if (c == '\n')
362: err(COMPILE, "newline can not be used as a string delimiter");
363: while (*p) {
364: if (*p == '[') {
365: if ((d = compile_ccl(&p, d)) == NULL)
366: err(COMPILE, "unbalanced brackets ([])");
367: continue;
368: } else if (*p == '\\' && p[1] == '[') {
369: *d++ = *p++;
1.22 tedu 370: } else if (*p == '\\' && p[1] == c) {
1.1 deraadt 371: p++;
1.22 tedu 372: } else if (*p == '\\' && p[1] == 'n') {
1.1 deraadt 373: *d++ = '\n';
374: p += 2;
375: continue;
1.22 tedu 376: } else if (*p == '\\' && p[1] == '\\') {
1.1 deraadt 377: *d++ = *p++;
1.22 tedu 378: } else if (*p == c) {
1.1 deraadt 379: *d = '\0';
380: return (p + 1);
381: }
382: *d++ = *p++;
383: }
384: return (NULL);
385: }
386:
387:
388: /* compile_ccl: expand a POSIX character class */
389: static char *
1.15 deraadt 390: compile_ccl(char **sp, char *t)
1.1 deraadt 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;
1.22 tedu 406: } else if (*s == '\\' && s[1] == 'n') {
407: *t = '\n';
408: s++;
409: }
410: if (*s == ']') {
411: *sp = ++s;
412: return (++t);
413: } else {
414: return (NULL);
415: }
1.1 deraadt 416: }
417:
418: /*
419: * Get a regular expression. P points to the delimiter of the regular
420: * expression; repp points to the address of a regexp pointer. Newline
421: * and delimiter escapes are processed; other escapes are ignored.
422: * Returns a pointer to the first character after the final delimiter
423: * or NULL in the case of a non terminated regular expression. The regexp
424: * pointer is set to the compiled regular expression.
425: * Cflags are passed to regcomp.
426: */
427: static char *
1.15 deraadt 428: compile_re(char *p, regex_t **repp)
1.1 deraadt 429: {
430: int eval;
431: char re[_POSIX2_LINE_MAX + 1];
432:
433: p = compile_delimited(p, re);
434: if (p && strlen(re) == 0) {
435: *repp = NULL;
436: return (p);
437: }
438: *repp = xmalloc(sizeof(regex_t));
1.6 millert 439: if (p && (eval = regcomp(*repp, re, 0)) != 0)
1.1 deraadt 440: err(COMPILE, "RE error: %s", strregerror(eval, *repp));
441: if (maxnsub < (*repp)->re_nsub)
442: maxnsub = (*repp)->re_nsub;
443: return (p);
444: }
445:
446: /*
447: * Compile the substitution string of a regular expression and set res to
448: * point to a saved copy of it. Nsub is the number of parenthesized regular
449: * expressions.
450: */
451: static char *
1.15 deraadt 452: compile_subst(char *p, struct s_subst *s)
1.1 deraadt 453: {
454: static char lbuf[_POSIX2_LINE_MAX + 1];
455: int asize, ref, size;
456: char c, *text, *op, *sp;
1.19 otto 457: int sawesc = 0;
1.1 deraadt 458:
459: c = *p++; /* Terminator character */
460: if (c == '\0')
461: return (NULL);
462:
463: s->maxbref = 0;
464: s->linenum = linenum;
465: asize = 2 * _POSIX2_LINE_MAX + 1;
466: text = xmalloc(asize);
467: size = 0;
468: do {
469: op = sp = text + size;
470: for (; *p; p++) {
1.19 otto 471: if (*p == '\\' || sawesc) {
472: /*
473: * If this is a continuation from the last
474: * buffer, we won't have a character to
475: * skip over.
476: */
477: if (sawesc)
478: sawesc = 0;
479: else
480: p++;
481:
482: if (*p == '\0') {
483: /*
484: * This escaped character is continued
485: * in the next part of the line. Note
486: * this fact, then cause the loop to
487: * exit w/ normal EOL case and reenter
488: * above with the new buffer.
489: */
490: sawesc = 1;
491: p--;
492: continue;
493: } else if (strchr("123456789", *p) != NULL) {
1.1 deraadt 494: *sp++ = '\\';
495: ref = *p - '0';
496: if (s->re != NULL &&
497: ref > s->re->re_nsub)
498: err(COMPILE,
499: "\\%c not defined in the RE", *p);
500: if (s->maxbref < ref)
501: s->maxbref = ref;
502: } else if (*p == '&' || *p == '\\')
503: *sp++ = '\\';
504: } else if (*p == c) {
505: p++;
506: *sp++ = '\0';
507: size += sp - op;
508: s->new = xrealloc(text, size);
509: return (p);
510: } else if (*p == '\n') {
511: err(COMPILE,
512: "unescaped newline inside substitute pattern");
513: /* NOTREACHED */
514: }
515: *sp++ = *p;
516: }
517: size += sp - op;
518: if (asize - size < _POSIX2_LINE_MAX + 1) {
519: asize *= 2;
1.20 otto 520: text = xrealloc(text, asize);
1.1 deraadt 521: }
522: } while (cu_fgets(p = lbuf, sizeof(lbuf)));
523: err(COMPILE, "unterminated substitute in regular expression");
524: /* NOTREACHED */
525: }
526:
527: /*
528: * Compile the flags of the s command
529: */
530: static char *
1.15 deraadt 531: compile_flags(char *p, struct s_subst *s)
1.1 deraadt 532: {
533: int gn; /* True if we have seen g or n */
1.17 otto 534: long l;
1.1 deraadt 535: char wfile[_POSIX2_LINE_MAX + 1], *q;
536:
537: s->n = 1; /* Default */
538: s->p = 0;
539: s->wfile = NULL;
540: s->wfd = -1;
541: for (gn = 0;;) {
542: EATSPACE(); /* EXTENSION */
543: switch (*p) {
544: case 'g':
545: if (gn)
1.22 tedu 546: err(COMPILE, "more than one number or 'g' in"
547: " substitute flags");
1.1 deraadt 548: gn = 1;
549: s->n = 0;
550: break;
551: case '\0':
552: case '\n':
553: case ';':
554: return (p);
555: case 'p':
556: s->p = 1;
557: break;
558: case '1': case '2': case '3':
559: case '4': case '5': case '6':
560: case '7': case '8': case '9':
561: if (gn)
1.22 tedu 562: err(COMPILE, "more than one number or 'g' in"
563: " substitute flags");
1.1 deraadt 564: gn = 1;
1.17 otto 565: l = strtol(p, &p, 10);
566: if (l <= 0 || l >= INT_MAX)
567: err(COMPILE,
568: "number in substitute flags out of range");
569: s->n = (int)l;
570: continue;
1.1 deraadt 571: case 'w':
572: p++;
573: #ifdef HISTORIC_PRACTICE
574: if (*p != ' ') {
575: err(WARNING, "space missing before w wfile");
576: return (p);
577: }
578: #endif
579: EATSPACE();
580: q = wfile;
581: while (*p) {
582: if (*p == '\n')
583: break;
584: *q++ = *p++;
585: }
586: *q = '\0';
587: if (q == wfile)
588: err(COMPILE, "no wfile specified");
589: s->wfile = strdup(wfile);
590: if (!aflag && (s->wfd = open(wfile,
591: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
592: DEFFILEMODE)) == -1)
1.12 jsyn 593: err(FATAL, "%s: %s", wfile, strerror(errno));
1.1 deraadt 594: return (p);
595: default:
596: err(COMPILE,
597: "bad flag in substitute command: '%c'", *p);
598: break;
599: }
600: p++;
601: }
602: }
603:
604: /*
605: * Compile a translation set of strings into a lookup table.
606: */
607: static char *
1.15 deraadt 608: compile_tr(char *p, char **transtab)
1.1 deraadt 609: {
610: int i;
611: char *lt, *op, *np;
612: char old[_POSIX2_LINE_MAX + 1];
613: char new[_POSIX2_LINE_MAX + 1];
614:
615: if (*p == '\0' || *p == '\\')
616: err(COMPILE,
617: "transform pattern can not be delimited by newline or backslash");
618: p = compile_delimited(p, old);
619: if (p == NULL) {
620: err(COMPILE, "unterminated transform source string");
621: return (NULL);
622: }
623: p = compile_delimited(--p, new);
624: if (p == NULL) {
625: err(COMPILE, "unterminated transform target string");
626: return (NULL);
627: }
628: EATSPACE();
629: if (strlen(new) != strlen(old)) {
630: err(COMPILE, "transform strings are not the same length");
631: return (NULL);
632: }
633: /* We assume characters are 8 bits */
1.18 deraadt 634: lt = xmalloc(UCHAR_MAX + 1);
1.1 deraadt 635: for (i = 0; i <= UCHAR_MAX; i++)
636: lt[i] = (char)i;
637: for (op = old, np = new; *op; op++, np++)
638: lt[(u_char)*op] = *np;
639: *transtab = lt;
640: return (p);
641: }
642:
643: /*
1.7 deraadt 644: * Compile the text following an a, c, or i command.
1.1 deraadt 645: */
646: static char *
1.15 deraadt 647: compile_text(void)
1.1 deraadt 648: {
1.4 deraadt 649: int asize, esc_nl, size;
1.1 deraadt 650: char *text, *p, *op, *s;
651: char lbuf[_POSIX2_LINE_MAX + 1];
652:
653: asize = 2 * _POSIX2_LINE_MAX + 1;
654: text = xmalloc(asize);
655: size = 0;
656: while (cu_fgets(lbuf, sizeof(lbuf))) {
657: op = s = text + size;
658: p = lbuf;
659: EATSPACE();
1.4 deraadt 660: for (esc_nl = 0; *p != '\0'; p++) {
661: if (*p == '\\' && p[1] != '\0' && *++p == '\n')
662: esc_nl = 1;
1.1 deraadt 663: *s++ = *p;
664: }
665: size += s - op;
1.4 deraadt 666: if (!esc_nl) {
1.1 deraadt 667: *s = '\0';
668: break;
669: }
670: if (asize - size < _POSIX2_LINE_MAX + 1) {
671: asize *= 2;
672: text = xmalloc(asize);
673: }
674: }
1.8 brian 675: text[size] = '\0';
1.1 deraadt 676: return (xrealloc(text, size + 1));
677: }
678:
679: /*
680: * Get an address and return a pointer to the first character after
681: * it. Fill the structure pointed to according to the address.
682: */
683: static char *
1.15 deraadt 684: compile_addr(char *p, struct s_addr *a)
1.1 deraadt 685: {
686: char *end;
687:
688: switch (*p) {
689: case '\\': /* Context address */
690: ++p;
691: /* FALLTHROUGH */
692: case '/': /* Context address */
693: p = compile_re(p, &a->u.r);
694: if (p == NULL)
695: err(COMPILE, "unterminated regular expression");
696: a->type = AT_RE;
697: return (p);
698:
699: case '$': /* Last line */
700: a->type = AT_LAST;
701: return (p + 1);
702: /* Line number */
1.21 deraadt 703: case '0': case '1': case '2': case '3': case '4':
1.1 deraadt 704: case '5': case '6': case '7': case '8': case '9':
705: a->type = AT_LINE;
1.3 millert 706: a->u.l = strtoul(p, &end, 10);
1.1 deraadt 707: return (end);
708: default:
709: err(COMPILE, "expected context address");
710: return (NULL);
711: }
712: }
713:
714: /*
715: * duptoeol --
716: * Return a copy of all the characters up to \n or \0.
717: */
718: static char *
1.16 deraadt 719: duptoeol(char *s, char *ctype, char **semi)
1.1 deraadt 720: {
721: size_t len;
722: int ws;
723: char *start;
724:
725: ws = 0;
1.16 deraadt 726: if (semi) {
727: for (start = s; *s != '\0' && *s != '\n' && *s != ';'; ++s)
728: ws = isspace(*s);
729: } else {
730: for (start = s; *s != '\0' && *s != '\n'; ++s)
731: ws = isspace(*s);
732: *s = '\0';
733: }
1.1 deraadt 734: if (ws)
735: err(WARNING, "whitespace after %s", ctype);
736: len = s - start + 1;
1.16 deraadt 737: if (semi)
738: *semi = s;
739: s = xmalloc(len);
740: strlcpy(s, start, len);
741: return (s);
1.1 deraadt 742: }
743:
744: /*
745: * Convert goto label names to addresses, and count a and r commands, in
746: * the given subset of the script. Free the memory used by labels in b
747: * and t commands (but not by :).
748: *
749: * TODO: Remove } nodes
750: */
751: static void
1.15 deraadt 752: fixuplabel(struct s_command *cp, struct s_command *end)
1.1 deraadt 753: {
754:
755: for (; cp != end; cp = cp->next)
756: switch (cp->code) {
757: case 'a':
758: case 'r':
759: appendnum++;
760: break;
761: case 'b':
762: case 't':
763: /* Resolve branch target. */
764: if (cp->t == NULL) {
765: cp->u.c = NULL;
766: break;
767: }
768: if ((cp->u.c = findlabel(cp->t)) == NULL)
769: err(COMPILE2, "undefined label '%s'", cp->t);
770: free(cp->t);
771: break;
772: case '{':
773: /* Do interior commands. */
774: fixuplabel(cp->u.c, cp->next);
775: break;
776: }
777: }
778:
779: /*
780: * Associate the given command label for later lookup.
781: */
782: static void
1.15 deraadt 783: enterlabel(struct s_command *cp)
1.1 deraadt 784: {
1.10 mpech 785: struct labhash **lhp, *lh;
786: u_char *p;
787: u_int h, c;
1.1 deraadt 788:
789: for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
790: h = (h << 5) + h + c;
791: lhp = &labels[h & LHMASK];
792: for (lh = *lhp; lh != NULL; lh = lh->lh_next)
793: if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
794: err(COMPILE2, "duplicate label '%s'", cp->t);
795: lh = xmalloc(sizeof *lh);
796: lh->lh_next = *lhp;
797: lh->lh_hash = h;
798: lh->lh_cmd = cp;
799: lh->lh_ref = 0;
800: *lhp = lh;
801: }
802:
803: /*
804: * Find the label contained in the command l in the command linked
805: * list cp. L is excluded from the search. Return NULL if not found.
806: */
807: static struct s_command *
1.15 deraadt 808: findlabel(char *name)
1.1 deraadt 809: {
1.10 mpech 810: struct labhash *lh;
811: u_char *p;
812: u_int h, c;
1.1 deraadt 813:
814: for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
815: h = (h << 5) + h + c;
816: for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
817: if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
818: lh->lh_ref = 1;
819: return (lh->lh_cmd);
820: }
821: }
822: return (NULL);
823: }
824:
1.21 deraadt 825: /*
1.1 deraadt 826: * Warn about any unused labels. As a side effect, release the label hash
827: * table space.
828: */
829: static void
1.15 deraadt 830: uselabel(void)
1.1 deraadt 831: {
1.10 mpech 832: struct labhash *lh, *next;
833: int i;
1.1 deraadt 834:
835: for (i = 0; i < LHSZ; i++) {
836: for (lh = labels[i]; lh != NULL; lh = next) {
837: next = lh->lh_next;
838: if (!lh->lh_ref)
839: err(WARNING, "unused label '%s'",
840: lh->lh_cmd->t);
841: free(lh);
842: }
843: }
844: }