Annotation of src/usr.bin/sed/compile.c, Revision 1.27
1.27 ! millert 1: /* $OpenBSD: compile.c,v 1.26 2008/10/09 16:40:56 millert 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"; */
1.27 ! millert 38: static const char rcsid[] = "$OpenBSD: compile.c,v 1.26 2008/10/09 16:40:56 millert 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 */
151: static size_t len = _POSIX2_LINE_MAX;
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;
1.25 millert 157: if (!lbuf)
158: lbuf = xmalloc(len);
1.1 deraadt 159: for (;;) {
1.25 millert 160: if ((p = cu_fgets(&lbuf, &len)) == NULL) {
1.1 deraadt 161: if (stack != 0)
162: err(COMPILE, "unexpected EOF (pending }'s)");
163: return (link);
164: }
165:
166: semicolon: EATSPACE();
1.23 otto 167: if (*p == '#' || *p == '\0')
1.1 deraadt 168: continue;
1.23 otto 169: if (*p == ';') {
170: p++;
171: goto semicolon;
172: }
1.1 deraadt 173: *link = cmd = xmalloc(sizeof(struct s_command));
174: link = &cmd->next;
175: cmd->nonsel = cmd->inrange = 0;
176: /* First parse the addresses */
177: naddr = 0;
178:
179: /* Valid characters to start an address */
180: #define addrchar(c) (strchr("0123456789/\\$", (c)))
181: if (addrchar(*p)) {
182: naddr++;
183: cmd->a1 = xmalloc(sizeof(struct s_addr));
184: p = compile_addr(p, cmd->a1);
185: EATSPACE(); /* EXTENSION */
186: if (*p == ',') {
187: p++;
188: EATSPACE(); /* EXTENSION */
189: naddr++;
190: cmd->a2 = xmalloc(sizeof(struct s_addr));
191: p = compile_addr(p, cmd->a2);
192: EATSPACE();
1.22 tedu 193: } else {
1.1 deraadt 194: cmd->a2 = 0;
1.22 tedu 195: }
196: } else {
1.1 deraadt 197: cmd->a1 = cmd->a2 = 0;
1.22 tedu 198: }
1.1 deraadt 199:
200: nonsel: /* Now parse the command */
201: if (!*p)
202: err(COMPILE, "command expected");
203: cmd->code = *p;
204: for (fp = cmd_fmts; fp->code; fp++)
205: if (fp->code == *p)
206: break;
207: if (!fp->code)
208: err(COMPILE, "invalid command code %c", *p);
209: if (naddr > fp->naddr)
210: err(COMPILE,
1.22 tedu 211: "command %c expects up to %d address(es), found %d",
212: *p, fp->naddr, naddr);
1.1 deraadt 213: switch (fp->args) {
214: case NONSEL: /* ! */
215: p++;
216: EATSPACE();
217: cmd->nonsel = ! cmd->nonsel;
218: goto nonsel;
219: case GROUP: /* { */
220: p++;
221: EATSPACE();
222: cmd->next = stack;
223: stack = cmd;
224: link = &cmd->u.c;
225: if (*p)
226: goto semicolon;
227: break;
228: case ENDGROUP:
229: /*
230: * Short-circuit command processing, since end of
231: * group is really just a noop.
232: */
233: cmd->nonsel = 1;
234: if (stack == 0)
235: err(COMPILE, "unexpected }");
236: cmd2 = stack;
237: stack = cmd2->next;
238: cmd2->next = cmd;
239: /*FALLTHROUGH*/
240: case EMPTY: /* d D g G h H l n N p P q x = \0 */
241: p++;
242: EATSPACE();
243: if (*p == ';') {
244: p++;
245: link = &cmd->next;
246: goto semicolon;
247: }
248: if (*p)
249: err(COMPILE,
250: "extra characters at the end of %c command", cmd->code);
251: break;
252: case TEXT: /* a c i */
253: p++;
254: EATSPACE();
255: if (*p != '\\')
1.22 tedu 256: err(COMPILE, "command %c expects \\ followed by"
257: " text", cmd->code);
1.1 deraadt 258: p++;
259: EATSPACE();
260: if (*p)
1.22 tedu 261: err(COMPILE, "extra characters after \\ at the"
262: " end of %c command", cmd->code);
1.1 deraadt 263: cmd->t = compile_text();
264: break;
265: case COMMENT: /* \0 # */
266: break;
267: case WFILE: /* w */
268: p++;
269: EATSPACE();
270: if (*p == '\0')
271: err(COMPILE, "filename expected");
1.16 deraadt 272: cmd->t = duptoeol(p, "w command", NULL);
1.1 deraadt 273: if (aflag)
274: cmd->u.fd = -1;
1.21 deraadt 275: else if ((cmd->u.fd = open(p,
1.1 deraadt 276: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
277: DEFFILEMODE)) == -1)
1.12 jsyn 278: err(FATAL, "%s: %s", p, strerror(errno));
1.1 deraadt 279: break;
280: case RFILE: /* r */
281: p++;
282: EATSPACE();
1.16 deraadt 283: cmd->t = duptoeol(p, "read command", NULL);
1.1 deraadt 284: break;
285: case BRANCH: /* b t */
286: p++;
287: EATSPACE();
288: if (*p == '\0')
289: cmd->t = NULL;
290: else
1.16 deraadt 291: cmd->t = duptoeol(p, "branch", &p);
292: if (*p == ';') {
293: p++;
294: goto semicolon;
295: }
1.1 deraadt 296: break;
297: case LABEL: /* : */
298: p++;
299: EATSPACE();
1.16 deraadt 300: cmd->t = duptoeol(p, "label", &p);
301: if (strlen(cmd->t) == 0)
1.1 deraadt 302: err(COMPILE, "empty label");
303: enterlabel(cmd);
1.16 deraadt 304: if (*p == ';') {
305: p++;
306: goto semicolon;
307: }
1.1 deraadt 308: break;
309: case SUBST: /* s */
310: p++;
311: if (*p == '\0' || *p == '\\')
1.22 tedu 312: err(COMPILE, "substitute pattern can not be"
313: " delimited by newline or backslash");
1.1 deraadt 314: cmd->u.s = xmalloc(sizeof(struct s_subst));
315: p = compile_re(p, &cmd->u.s->re);
316: if (p == NULL)
317: err(COMPILE, "unterminated substitute pattern");
318: --p;
319: p = compile_subst(p, cmd->u.s);
320: p = compile_flags(p, cmd->u.s);
321: EATSPACE();
322: if (*p == ';') {
323: p++;
324: link = &cmd->next;
325: goto semicolon;
326: }
327: break;
328: case TR: /* y */
329: p++;
330: p = compile_tr(p, (char **)&cmd->u.y);
331: EATSPACE();
332: if (*p == ';') {
333: p++;
334: link = &cmd->next;
335: goto semicolon;
336: }
337: if (*p)
1.22 tedu 338: err(COMPILE, "extra text at the end of a"
339: " transform command");
1.1 deraadt 340: break;
341: }
342: }
343: }
344:
345: /*
346: * Get a delimited string. P points to the delimeter of the string; d points
347: * to a buffer area. Newline and delimiter escapes are processed; other
348: * escapes are ignored.
349: *
350: * Returns a pointer to the first character after the final delimiter or NULL
351: * in the case of a non-terminated string. The character array d is filled
352: * with the processed string.
353: */
354: static char *
1.15 deraadt 355: compile_delimited(char *p, char *d)
1.1 deraadt 356: {
357: char c;
358:
359: c = *p++;
360: if (c == '\0')
361: return (NULL);
362: else if (c == '\\')
363: err(COMPILE, "\\ can not be used as a string delimiter");
364: else if (c == '\n')
365: err(COMPILE, "newline can not be used as a string delimiter");
366: while (*p) {
367: if (*p == '[') {
368: if ((d = compile_ccl(&p, d)) == NULL)
369: err(COMPILE, "unbalanced brackets ([])");
370: continue;
371: } else if (*p == '\\' && p[1] == '[') {
372: *d++ = *p++;
1.22 tedu 373: } else if (*p == '\\' && p[1] == c) {
1.1 deraadt 374: p++;
1.22 tedu 375: } else if (*p == '\\' && p[1] == 'n') {
1.1 deraadt 376: *d++ = '\n';
377: p += 2;
378: continue;
1.22 tedu 379: } else if (*p == '\\' && p[1] == '\\') {
1.1 deraadt 380: *d++ = *p++;
1.22 tedu 381: } else if (*p == c) {
1.1 deraadt 382: *d = '\0';
383: return (p + 1);
384: }
385: *d++ = *p++;
386: }
387: return (NULL);
388: }
389:
390:
391: /* compile_ccl: expand a POSIX character class */
392: static char *
1.15 deraadt 393: compile_ccl(char **sp, char *t)
1.1 deraadt 394: {
395: int c, d;
396: char *s = *sp;
397:
398: *t++ = *s++;
399: if (*s == '^')
400: *t++ = *s++;
401: if (*s == ']')
402: *t++ = *s++;
403: for (; *s && (*t = *s) != ']'; s++, t++)
404: if (*s == '[' && ((d = *(s+1)) == '.' || d == ':' || d == '=')) {
405: *++t = *++s, t++, s++;
406: for (c = *s; (*t = *s) != ']' || c != d; s++, t++)
407: if ((c = *s) == '\0')
408: return NULL;
1.22 tedu 409: } else if (*s == '\\' && s[1] == 'n') {
410: *t = '\n';
411: s++;
412: }
413: if (*s == ']') {
414: *sp = ++s;
415: return (++t);
416: } else {
417: return (NULL);
418: }
1.1 deraadt 419: }
420:
421: /*
422: * Get a regular expression. P points to the delimiter of the regular
423: * expression; repp points to the address of a regexp pointer. Newline
424: * and delimiter escapes are processed; other escapes are ignored.
425: * Returns a pointer to the first character after the final delimiter
426: * or NULL in the case of a non terminated regular expression. The regexp
427: * pointer is set to the compiled regular expression.
428: * Cflags are passed to regcomp.
429: */
430: static char *
1.15 deraadt 431: compile_re(char *p, regex_t **repp)
1.1 deraadt 432: {
433: int eval;
1.25 millert 434: char *re;
1.1 deraadt 435:
1.25 millert 436: re = xmalloc(strlen(p) + 1); /* strlen(re) <= strlen(p) */
1.1 deraadt 437: p = compile_delimited(p, re);
438: if (p && strlen(re) == 0) {
439: *repp = NULL;
1.25 millert 440: free(re);
1.1 deraadt 441: return (p);
442: }
443: *repp = xmalloc(sizeof(regex_t));
1.6 millert 444: if (p && (eval = regcomp(*repp, re, 0)) != 0)
1.1 deraadt 445: err(COMPILE, "RE error: %s", strregerror(eval, *repp));
446: if (maxnsub < (*repp)->re_nsub)
447: maxnsub = (*repp)->re_nsub;
1.25 millert 448: free(re);
1.1 deraadt 449: return (p);
450: }
451:
452: /*
453: * Compile the substitution string of a regular expression and set res to
454: * point to a saved copy of it. Nsub is the number of parenthesized regular
455: * expressions.
456: */
457: static char *
1.15 deraadt 458: compile_subst(char *p, struct s_subst *s)
1.1 deraadt 459: {
1.25 millert 460: static char *lbuf;
1.26 millert 461: static size_t len;
1.1 deraadt 462: int asize, ref, size;
463: char c, *text, *op, *sp;
1.19 otto 464: int sawesc = 0;
1.1 deraadt 465:
466: c = *p++; /* Terminator character */
467: if (c == '\0')
468: return (NULL);
469:
1.26 millert 470: len = strlen(p);
471: if (len < _POSIX2_LINE_MAX)
472: len = _POSIX2_LINE_MAX;
1.27 ! millert 473: if (!lbuf)
! 474: lbuf = xmalloc(len);
! 475:
1.1 deraadt 476: s->maxbref = 0;
477: s->linenum = linenum;
1.25 millert 478: asize = 2 * len + 1;
1.1 deraadt 479: text = xmalloc(asize);
480: size = 0;
481: do {
1.25 millert 482: if (asize - size < len + 1) {
483: do {
484: asize *= 2;
485: } while (asize - size < len + 1);
486: text = xrealloc(text, asize);
487: }
1.1 deraadt 488: op = sp = text + size;
489: for (; *p; p++) {
1.19 otto 490: if (*p == '\\' || sawesc) {
491: /*
492: * If this is a continuation from the last
493: * buffer, we won't have a character to
494: * skip over.
495: */
496: if (sawesc)
497: sawesc = 0;
498: else
499: p++;
500:
501: if (*p == '\0') {
502: /*
503: * This escaped character is continued
504: * in the next part of the line. Note
505: * this fact, then cause the loop to
506: * exit w/ normal EOL case and reenter
507: * above with the new buffer.
508: */
509: sawesc = 1;
510: p--;
511: continue;
512: } else if (strchr("123456789", *p) != NULL) {
1.1 deraadt 513: *sp++ = '\\';
514: ref = *p - '0';
515: if (s->re != NULL &&
516: ref > s->re->re_nsub)
517: err(COMPILE,
518: "\\%c not defined in the RE", *p);
519: if (s->maxbref < ref)
520: s->maxbref = ref;
521: } else if (*p == '&' || *p == '\\')
522: *sp++ = '\\';
523: } else if (*p == c) {
524: p++;
525: *sp++ = '\0';
526: size += sp - op;
527: s->new = xrealloc(text, size);
528: return (p);
529: } else if (*p == '\n') {
530: err(COMPILE,
531: "unescaped newline inside substitute pattern");
532: /* NOTREACHED */
533: }
534: *sp++ = *p;
535: }
536: size += sp - op;
1.25 millert 537: } while ((p = cu_fgets(&lbuf, &len)));
1.1 deraadt 538: err(COMPILE, "unterminated substitute in regular expression");
539: /* NOTREACHED */
540: }
541:
542: /*
543: * Compile the flags of the s command
544: */
545: static char *
1.15 deraadt 546: compile_flags(char *p, struct s_subst *s)
1.1 deraadt 547: {
548: int gn; /* True if we have seen g or n */
1.17 otto 549: long l;
1.25 millert 550: char wfile[PATH_MAX], *q;
1.1 deraadt 551:
552: s->n = 1; /* Default */
553: s->p = 0;
554: s->wfile = NULL;
555: s->wfd = -1;
556: for (gn = 0;;) {
557: EATSPACE(); /* EXTENSION */
558: switch (*p) {
559: case 'g':
560: if (gn)
1.22 tedu 561: err(COMPILE, "more than one number or 'g' in"
562: " substitute flags");
1.1 deraadt 563: gn = 1;
564: s->n = 0;
565: break;
566: case '\0':
567: case '\n':
568: case ';':
569: return (p);
570: case 'p':
571: s->p = 1;
572: break;
573: case '1': case '2': case '3':
574: case '4': case '5': case '6':
575: case '7': case '8': case '9':
576: if (gn)
1.22 tedu 577: err(COMPILE, "more than one number or 'g' in"
578: " substitute flags");
1.1 deraadt 579: gn = 1;
1.17 otto 580: l = strtol(p, &p, 10);
581: if (l <= 0 || l >= INT_MAX)
582: err(COMPILE,
583: "number in substitute flags out of range");
584: s->n = (int)l;
585: continue;
1.1 deraadt 586: case 'w':
587: p++;
588: #ifdef HISTORIC_PRACTICE
589: if (*p != ' ') {
590: err(WARNING, "space missing before w wfile");
591: return (p);
592: }
593: #endif
594: EATSPACE();
595: q = wfile;
596: while (*p) {
597: if (*p == '\n')
598: break;
599: *q++ = *p++;
600: }
601: *q = '\0';
602: if (q == wfile)
603: err(COMPILE, "no wfile specified");
604: s->wfile = strdup(wfile);
605: if (!aflag && (s->wfd = open(wfile,
606: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
607: DEFFILEMODE)) == -1)
1.12 jsyn 608: err(FATAL, "%s: %s", wfile, strerror(errno));
1.1 deraadt 609: return (p);
610: default:
611: err(COMPILE,
612: "bad flag in substitute command: '%c'", *p);
613: break;
614: }
615: p++;
616: }
617: }
618:
619: /*
620: * Compile a translation set of strings into a lookup table.
621: */
622: static char *
1.15 deraadt 623: compile_tr(char *p, char **transtab)
1.1 deraadt 624: {
625: int i;
626: char *lt, *op, *np;
1.25 millert 627: char *old = NULL, *new = NULL;
1.1 deraadt 628:
629: if (*p == '\0' || *p == '\\')
630: err(COMPILE,
631: "transform pattern can not be delimited by newline or backslash");
1.25 millert 632: old = xmalloc(strlen(p) + 1);
1.1 deraadt 633: p = compile_delimited(p, old);
634: if (p == NULL) {
635: err(COMPILE, "unterminated transform source string");
1.25 millert 636: goto bad;
1.1 deraadt 637: }
1.25 millert 638: new = xmalloc(strlen(p) + 1);
1.1 deraadt 639: p = compile_delimited(--p, new);
640: if (p == NULL) {
641: err(COMPILE, "unterminated transform target string");
1.25 millert 642: goto bad;
1.1 deraadt 643: }
644: EATSPACE();
645: if (strlen(new) != strlen(old)) {
646: err(COMPILE, "transform strings are not the same length");
1.25 millert 647: goto bad;
1.1 deraadt 648: }
649: /* We assume characters are 8 bits */
1.18 deraadt 650: lt = xmalloc(UCHAR_MAX + 1);
1.1 deraadt 651: for (i = 0; i <= UCHAR_MAX; i++)
652: lt[i] = (char)i;
653: for (op = old, np = new; *op; op++, np++)
654: lt[(u_char)*op] = *np;
655: *transtab = lt;
1.25 millert 656: free(old);
657: free(new);
1.1 deraadt 658: return (p);
1.25 millert 659: bad:
660: free(old);
661: free(new);
662: return (NULL);
1.1 deraadt 663: }
664:
665: /*
1.7 deraadt 666: * Compile the text following an a, c, or i command.
1.1 deraadt 667: */
668: static char *
1.15 deraadt 669: compile_text(void)
1.1 deraadt 670: {
1.4 deraadt 671: int asize, esc_nl, size;
1.25 millert 672: char *lbuf, *text, *p, *op, *s;
673: size_t len = _POSIX2_LINE_MAX;
1.1 deraadt 674:
1.25 millert 675: lbuf = xmalloc(len);
676: asize = 2 * len + 1;
1.1 deraadt 677: text = xmalloc(asize);
678: size = 0;
1.25 millert 679: while (cu_fgets(&lbuf, &len)) {
680: if (asize - size < len + 1) {
681: do {
682: asize *= 2;
683: } while (asize - size < len + 1);
684: text = xrealloc(text, asize);
685: }
1.1 deraadt 686: op = s = text + size;
687: p = lbuf;
688: EATSPACE();
1.4 deraadt 689: for (esc_nl = 0; *p != '\0'; p++) {
690: if (*p == '\\' && p[1] != '\0' && *++p == '\n')
691: esc_nl = 1;
1.1 deraadt 692: *s++ = *p;
693: }
694: size += s - op;
1.4 deraadt 695: if (!esc_nl) {
1.1 deraadt 696: *s = '\0';
697: break;
698: }
699: }
1.25 millert 700: free(lbuf);
1.8 brian 701: text[size] = '\0';
1.1 deraadt 702: return (xrealloc(text, size + 1));
703: }
704:
705: /*
706: * Get an address and return a pointer to the first character after
707: * it. Fill the structure pointed to according to the address.
708: */
709: static char *
1.15 deraadt 710: compile_addr(char *p, struct s_addr *a)
1.1 deraadt 711: {
712: char *end;
713:
714: switch (*p) {
715: case '\\': /* Context address */
716: ++p;
717: /* FALLTHROUGH */
718: case '/': /* Context address */
719: p = compile_re(p, &a->u.r);
720: if (p == NULL)
721: err(COMPILE, "unterminated regular expression");
722: a->type = AT_RE;
723: return (p);
724:
725: case '$': /* Last line */
726: a->type = AT_LAST;
727: return (p + 1);
728: /* Line number */
1.21 deraadt 729: case '0': case '1': case '2': case '3': case '4':
1.1 deraadt 730: case '5': case '6': case '7': case '8': case '9':
731: a->type = AT_LINE;
1.3 millert 732: a->u.l = strtoul(p, &end, 10);
1.1 deraadt 733: return (end);
734: default:
735: err(COMPILE, "expected context address");
736: return (NULL);
737: }
738: }
739:
740: /*
741: * duptoeol --
742: * Return a copy of all the characters up to \n or \0.
743: */
744: static char *
1.16 deraadt 745: duptoeol(char *s, char *ctype, char **semi)
1.1 deraadt 746: {
747: size_t len;
748: int ws;
749: char *start;
750:
751: ws = 0;
1.16 deraadt 752: if (semi) {
753: for (start = s; *s != '\0' && *s != '\n' && *s != ';'; ++s)
754: ws = isspace(*s);
755: } else {
756: for (start = s; *s != '\0' && *s != '\n'; ++s)
757: ws = isspace(*s);
758: *s = '\0';
759: }
1.1 deraadt 760: if (ws)
761: err(WARNING, "whitespace after %s", ctype);
762: len = s - start + 1;
1.16 deraadt 763: if (semi)
764: *semi = s;
765: s = xmalloc(len);
766: strlcpy(s, start, len);
767: return (s);
1.1 deraadt 768: }
769:
770: /*
771: * Convert goto label names to addresses, and count a and r commands, in
772: * the given subset of the script. Free the memory used by labels in b
773: * and t commands (but not by :).
774: *
775: * TODO: Remove } nodes
776: */
777: static void
1.15 deraadt 778: fixuplabel(struct s_command *cp, struct s_command *end)
1.1 deraadt 779: {
780:
781: for (; cp != end; cp = cp->next)
782: switch (cp->code) {
783: case 'a':
784: case 'r':
785: appendnum++;
786: break;
787: case 'b':
788: case 't':
789: /* Resolve branch target. */
790: if (cp->t == NULL) {
791: cp->u.c = NULL;
792: break;
793: }
794: if ((cp->u.c = findlabel(cp->t)) == NULL)
795: err(COMPILE2, "undefined label '%s'", cp->t);
796: free(cp->t);
797: break;
798: case '{':
799: /* Do interior commands. */
800: fixuplabel(cp->u.c, cp->next);
801: break;
802: }
803: }
804:
805: /*
806: * Associate the given command label for later lookup.
807: */
808: static void
1.15 deraadt 809: enterlabel(struct s_command *cp)
1.1 deraadt 810: {
1.10 mpech 811: struct labhash **lhp, *lh;
812: u_char *p;
813: u_int h, c;
1.1 deraadt 814:
815: for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
816: h = (h << 5) + h + c;
817: lhp = &labels[h & LHMASK];
818: for (lh = *lhp; lh != NULL; lh = lh->lh_next)
819: if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
820: err(COMPILE2, "duplicate label '%s'", cp->t);
821: lh = xmalloc(sizeof *lh);
822: lh->lh_next = *lhp;
823: lh->lh_hash = h;
824: lh->lh_cmd = cp;
825: lh->lh_ref = 0;
826: *lhp = lh;
827: }
828:
829: /*
830: * Find the label contained in the command l in the command linked
831: * list cp. L is excluded from the search. Return NULL if not found.
832: */
833: static struct s_command *
1.15 deraadt 834: findlabel(char *name)
1.1 deraadt 835: {
1.10 mpech 836: struct labhash *lh;
837: u_char *p;
838: u_int h, c;
1.1 deraadt 839:
840: for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
841: h = (h << 5) + h + c;
842: for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
843: if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
844: lh->lh_ref = 1;
845: return (lh->lh_cmd);
846: }
847: }
848: return (NULL);
849: }
850:
1.21 deraadt 851: /*
1.1 deraadt 852: * Warn about any unused labels. As a side effect, release the label hash
853: * table space.
854: */
855: static void
1.15 deraadt 856: uselabel(void)
1.1 deraadt 857: {
1.10 mpech 858: struct labhash *lh, *next;
859: int i;
1.1 deraadt 860:
861: for (i = 0; i < LHSZ; i++) {
862: for (lh = labels[i]; lh != NULL; lh = next) {
863: next = lh->lh_next;
864: if (!lh->lh_ref)
865: err(WARNING, "unused label '%s'",
866: lh->lh_cmd->t);
867: free(lh);
868: }
869: }
870: }