Annotation of src/usr.bin/sed/compile.c, Revision 1.40
1.40 ! jca 1: /* $OpenBSD: compile.c,v 1.39 2015/10/26 14:08:47 mmcc 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: #include <sys/types.h>
37: #include <sys/stat.h>
38:
39: #include <ctype.h>
40: #include <errno.h>
41: #include <fcntl.h>
42: #include <limits.h>
43: #include <regex.h>
44: #include <stdio.h>
45: #include <stdlib.h>
46: #include <string.h>
47:
48: #include "defs.h"
49: #include "extern.h"
50:
51: #define LHSZ 128
52: #define LHMASK (LHSZ - 1)
53: static struct labhash {
54: struct labhash *lh_next;
55: u_int lh_hash;
56: struct s_command *lh_cmd;
57: int lh_ref;
58: } *labels[LHSZ];
59:
1.11 millert 60: static char *compile_addr(char *, struct s_addr *);
61: static char *compile_ccl(char **, char *);
1.33 naddy 62: static char *compile_delimited(char *, char *, int);
1.11 millert 63: static char *compile_flags(char *, struct s_subst *);
64: static char *compile_re(char *, regex_t **);
65: static char *compile_subst(char *, struct s_subst *);
66: static char *compile_text(void);
67: static char *compile_tr(char *, char **);
1.1 deraadt 68: static struct s_command
1.11 millert 69: **compile_stream(struct s_command **);
1.16 deraadt 70: static char *duptoeol(char *, char *, char **);
1.11 millert 71: static void enterlabel(struct s_command *);
1.1 deraadt 72: static struct s_command
1.11 millert 73: *findlabel(char *);
74: static void fixuplabel(struct s_command *, struct s_command *);
75: static void uselabel(void);
1.1 deraadt 76:
77: /*
78: * Command specification. This is used to drive the command parser.
79: */
80: struct s_format {
81: char code; /* Command code */
82: int naddr; /* Number of address args */
83: enum e_args args; /* Argument type */
84: };
85:
86: static struct s_format cmd_fmts[] = {
87: {'{', 2, GROUP},
88: {'}', 0, ENDGROUP},
89: {'a', 1, TEXT},
90: {'b', 2, BRANCH},
91: {'c', 2, TEXT},
92: {'d', 2, EMPTY},
93: {'D', 2, EMPTY},
94: {'g', 2, EMPTY},
95: {'G', 2, EMPTY},
96: {'h', 2, EMPTY},
97: {'H', 2, EMPTY},
98: {'i', 1, TEXT},
99: {'l', 2, EMPTY},
100: {'n', 2, EMPTY},
101: {'N', 2, EMPTY},
102: {'p', 2, EMPTY},
103: {'P', 2, EMPTY},
104: {'q', 1, EMPTY},
105: {'r', 1, RFILE},
106: {'s', 2, SUBST},
107: {'t', 2, BRANCH},
108: {'w', 2, WFILE},
109: {'x', 2, EMPTY},
110: {'y', 2, TR},
111: {'!', 2, NONSEL},
112: {':', 0, LABEL},
113: {'#', 0, COMMENT},
114: {'=', 1, EMPTY},
115: {'\0', 0, COMMENT},
116: };
117:
118: /* The compiled program. */
119: struct s_command *prog;
120:
121: /*
122: * Compile the program into prog.
123: * Initialise appends.
124: */
125: void
1.15 deraadt 126: compile(void)
1.1 deraadt 127: {
128: *compile_stream(&prog) = NULL;
129: fixuplabel(prog, NULL);
130: uselabel();
1.36 deraadt 131: appends = xreallocarray(NULL, appendnum, sizeof(struct s_appends));
132: match = xreallocarray(NULL, maxnsub + 1, sizeof(regmatch_t));
1.1 deraadt 133: }
134:
135: #define EATSPACE() do { \
136: if (p) \
1.35 deraadt 137: while (isascii((unsigned char)*p) && \
138: isspace((unsigned char)*p)) \
1.1 deraadt 139: p++; \
140: } while (0)
141:
142: static struct s_command **
1.15 deraadt 143: compile_stream(struct s_command **link)
1.1 deraadt 144: {
1.10 mpech 145: char *p;
1.25 millert 146: static char *lbuf; /* To avoid excessive malloc calls */
1.28 millert 147: static size_t bufsize;
1.1 deraadt 148: struct s_command *cmd, *cmd2, *stack;
149: struct s_format *fp;
150: int naddr; /* Number of addresses */
151:
152: stack = 0;
153: for (;;) {
1.28 millert 154: if ((p = cu_fgets(&lbuf, &bufsize)) == NULL) {
1.1 deraadt 155: if (stack != 0)
1.39 mmcc 156: error(COMPILE, "unexpected EOF (pending }'s)");
1.1 deraadt 157: return (link);
158: }
159:
160: semicolon: EATSPACE();
1.23 otto 161: if (*p == '#' || *p == '\0')
1.1 deraadt 162: continue;
1.23 otto 163: if (*p == ';') {
164: p++;
165: goto semicolon;
166: }
1.1 deraadt 167: *link = cmd = xmalloc(sizeof(struct s_command));
168: link = &cmd->next;
169: cmd->nonsel = cmd->inrange = 0;
170: /* First parse the addresses */
171: naddr = 0;
172:
173: /* Valid characters to start an address */
174: #define addrchar(c) (strchr("0123456789/\\$", (c)))
175: if (addrchar(*p)) {
176: naddr++;
177: cmd->a1 = xmalloc(sizeof(struct s_addr));
178: p = compile_addr(p, cmd->a1);
179: EATSPACE(); /* EXTENSION */
180: if (*p == ',') {
181: p++;
182: EATSPACE(); /* EXTENSION */
183: naddr++;
184: cmd->a2 = xmalloc(sizeof(struct s_addr));
185: p = compile_addr(p, cmd->a2);
186: EATSPACE();
1.22 tedu 187: } else {
1.1 deraadt 188: cmd->a2 = 0;
1.22 tedu 189: }
190: } else {
1.1 deraadt 191: cmd->a1 = cmd->a2 = 0;
1.22 tedu 192: }
1.1 deraadt 193:
194: nonsel: /* Now parse the command */
195: if (!*p)
1.39 mmcc 196: error(COMPILE, "command expected");
1.1 deraadt 197: cmd->code = *p;
198: for (fp = cmd_fmts; fp->code; fp++)
199: if (fp->code == *p)
200: break;
201: if (!fp->code)
1.39 mmcc 202: error(COMPILE, "invalid command code %c", *p);
1.1 deraadt 203: if (naddr > fp->naddr)
1.39 mmcc 204: error(COMPILE,
1.22 tedu 205: "command %c expects up to %d address(es), found %d",
206: *p, fp->naddr, naddr);
1.1 deraadt 207: switch (fp->args) {
208: case NONSEL: /* ! */
209: p++;
210: EATSPACE();
1.38 millert 211: cmd->nonsel = 1;
1.1 deraadt 212: goto nonsel;
213: case GROUP: /* { */
214: p++;
215: EATSPACE();
216: cmd->next = stack;
217: stack = cmd;
218: link = &cmd->u.c;
219: if (*p)
220: goto semicolon;
221: break;
222: case ENDGROUP:
223: /*
224: * Short-circuit command processing, since end of
225: * group is really just a noop.
226: */
227: cmd->nonsel = 1;
228: if (stack == 0)
1.39 mmcc 229: error(COMPILE, "unexpected }");
1.1 deraadt 230: cmd2 = stack;
231: stack = cmd2->next;
232: cmd2->next = cmd;
233: /*FALLTHROUGH*/
234: case EMPTY: /* d D g G h H l n N p P q x = \0 */
235: p++;
236: EATSPACE();
237: if (*p == ';') {
238: p++;
239: link = &cmd->next;
240: goto semicolon;
241: }
242: if (*p)
1.39 mmcc 243: error(COMPILE,
1.1 deraadt 244: "extra characters at the end of %c command", cmd->code);
245: break;
246: case TEXT: /* a c i */
247: p++;
248: EATSPACE();
249: if (*p != '\\')
1.39 mmcc 250: error(COMPILE, "command %c expects \\ followed by"
1.22 tedu 251: " text", cmd->code);
1.1 deraadt 252: p++;
253: EATSPACE();
254: if (*p)
1.39 mmcc 255: error(COMPILE, "extra characters after \\ at the"
1.22 tedu 256: " end of %c command", cmd->code);
1.1 deraadt 257: cmd->t = compile_text();
258: break;
259: case COMMENT: /* \0 # */
260: break;
261: case WFILE: /* w */
262: p++;
263: EATSPACE();
264: if (*p == '\0')
1.39 mmcc 265: error(COMPILE, "filename expected");
1.16 deraadt 266: cmd->t = duptoeol(p, "w command", NULL);
1.1 deraadt 267: if (aflag)
268: cmd->u.fd = -1;
1.21 deraadt 269: else if ((cmd->u.fd = open(p,
1.1 deraadt 270: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
271: DEFFILEMODE)) == -1)
1.39 mmcc 272: error(FATAL, "%s: %s", p, strerror(errno));
1.1 deraadt 273: break;
274: case RFILE: /* r */
275: p++;
276: EATSPACE();
1.16 deraadt 277: cmd->t = duptoeol(p, "read command", NULL);
1.1 deraadt 278: break;
279: case BRANCH: /* b t */
280: p++;
281: EATSPACE();
282: if (*p == '\0')
283: cmd->t = NULL;
284: else
1.16 deraadt 285: cmd->t = duptoeol(p, "branch", &p);
286: if (*p == ';') {
287: p++;
288: goto semicolon;
289: }
1.1 deraadt 290: break;
291: case LABEL: /* : */
292: p++;
293: EATSPACE();
1.16 deraadt 294: cmd->t = duptoeol(p, "label", &p);
295: if (strlen(cmd->t) == 0)
1.39 mmcc 296: error(COMPILE, "empty label");
1.1 deraadt 297: enterlabel(cmd);
1.16 deraadt 298: if (*p == ';') {
299: p++;
300: goto semicolon;
301: }
1.1 deraadt 302: break;
303: case SUBST: /* s */
304: p++;
305: if (*p == '\0' || *p == '\\')
1.39 mmcc 306: error(COMPILE, "substitute pattern can not be"
1.22 tedu 307: " delimited by newline or backslash");
1.1 deraadt 308: cmd->u.s = xmalloc(sizeof(struct s_subst));
309: p = compile_re(p, &cmd->u.s->re);
310: if (p == NULL)
1.39 mmcc 311: error(COMPILE, "unterminated substitute pattern");
1.1 deraadt 312: --p;
313: p = compile_subst(p, cmd->u.s);
314: p = compile_flags(p, cmd->u.s);
315: EATSPACE();
316: if (*p == ';') {
317: p++;
318: link = &cmd->next;
319: goto semicolon;
320: }
321: break;
322: case TR: /* y */
323: p++;
324: p = compile_tr(p, (char **)&cmd->u.y);
325: EATSPACE();
326: if (*p == ';') {
327: p++;
328: link = &cmd->next;
329: goto semicolon;
330: }
331: if (*p)
1.39 mmcc 332: error(COMPILE, "extra text at the end of a"
1.22 tedu 333: " transform command");
1.1 deraadt 334: break;
335: }
336: }
337: }
338:
339: /*
340: * Get a delimited string. P points to the delimeter of the string; d points
341: * to a buffer area. Newline and delimiter escapes are processed; other
342: * escapes are ignored.
343: *
344: * Returns a pointer to the first character after the final delimiter or NULL
345: * in the case of a non-terminated string. The character array d is filled
346: * with the processed string.
347: */
348: static char *
1.33 naddy 349: compile_delimited(char *p, char *d, int is_tr)
1.1 deraadt 350: {
351: char c;
352:
353: c = *p++;
354: if (c == '\0')
355: return (NULL);
356: else if (c == '\\')
1.39 mmcc 357: error(COMPILE, "\\ can not be used as a string delimiter");
1.1 deraadt 358: else if (c == '\n')
1.39 mmcc 359: error(COMPILE, "newline can not be used as a string delimiter");
1.1 deraadt 360: while (*p) {
1.32 naddy 361: if (*p == '[' && *p != c) {
1.1 deraadt 362: if ((d = compile_ccl(&p, d)) == NULL)
1.39 mmcc 363: error(COMPILE, "unbalanced brackets ([])");
1.1 deraadt 364: continue;
365: } else if (*p == '\\' && p[1] == '[') {
366: *d++ = *p++;
1.22 tedu 367: } else if (*p == '\\' && p[1] == c) {
1.1 deraadt 368: p++;
1.22 tedu 369: } else if (*p == '\\' && p[1] == 'n') {
1.1 deraadt 370: *d++ = '\n';
371: p += 2;
372: continue;
1.22 tedu 373: } else if (*p == '\\' && p[1] == '\\') {
1.33 naddy 374: if (is_tr)
375: p++;
376: else
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;
1.25 millert 431: char *re;
1.1 deraadt 432:
1.25 millert 433: re = xmalloc(strlen(p) + 1); /* strlen(re) <= strlen(p) */
1.33 naddy 434: p = compile_delimited(p, re, 0);
1.1 deraadt 435: if (p && strlen(re) == 0) {
436: *repp = NULL;
1.25 millert 437: free(re);
1.1 deraadt 438: return (p);
439: }
440: *repp = xmalloc(sizeof(regex_t));
1.29 djm 441: if (p && (eval = regcomp(*repp, re, Eflag ? REG_EXTENDED : 0)) != 0)
1.39 mmcc 442: error(COMPILE, "RE error: %s", strregerror(eval, *repp));
1.1 deraadt 443: if (maxnsub < (*repp)->re_nsub)
444: maxnsub = (*repp)->re_nsub;
1.25 millert 445: free(re);
1.1 deraadt 446: return (p);
447: }
448:
449: /*
450: * Compile the substitution string of a regular expression and set res to
451: * point to a saved copy of it. Nsub is the number of parenthesized regular
452: * expressions.
453: */
454: static char *
1.15 deraadt 455: compile_subst(char *p, struct s_subst *s)
1.1 deraadt 456: {
1.25 millert 457: static char *lbuf;
1.28 millert 458: static size_t bufsize;
1.1 deraadt 459: int asize, ref, size;
460: char c, *text, *op, *sp;
1.19 otto 461: int sawesc = 0;
1.1 deraadt 462:
463: c = *p++; /* Terminator character */
464: if (c == '\0')
465: return (NULL);
466:
467: s->maxbref = 0;
468: s->linenum = linenum;
1.28 millert 469: text = NULL;
470: asize = size = 0;
1.1 deraadt 471: do {
1.28 millert 472: size_t len = ROUNDLEN(strlen(p) + 1);
473: if (asize - size < len) {
1.25 millert 474: do {
1.28 millert 475: asize += len;
476: } while (asize - size < len);
1.25 millert 477: text = xrealloc(text, asize);
478: }
1.1 deraadt 479: op = sp = text + size;
480: for (; *p; p++) {
1.19 otto 481: if (*p == '\\' || sawesc) {
482: /*
483: * If this is a continuation from the last
484: * buffer, we won't have a character to
485: * skip over.
486: */
487: if (sawesc)
488: sawesc = 0;
489: else
490: p++;
491:
492: if (*p == '\0') {
493: /*
494: * This escaped character is continued
495: * in the next part of the line. Note
496: * this fact, then cause the loop to
497: * exit w/ normal EOL case and reenter
498: * above with the new buffer.
499: */
500: sawesc = 1;
501: p--;
502: continue;
503: } else if (strchr("123456789", *p) != NULL) {
1.1 deraadt 504: *sp++ = '\\';
505: ref = *p - '0';
506: if (s->re != NULL &&
507: ref > s->re->re_nsub)
1.39 mmcc 508: error(COMPILE,
1.1 deraadt 509: "\\%c not defined in the RE", *p);
510: if (s->maxbref < ref)
511: s->maxbref = ref;
512: } else if (*p == '&' || *p == '\\')
513: *sp++ = '\\';
514: } else if (*p == c) {
515: p++;
516: *sp++ = '\0';
517: size += sp - op;
518: s->new = xrealloc(text, size);
519: return (p);
520: } else if (*p == '\n') {
1.39 mmcc 521: error(COMPILE,
1.1 deraadt 522: "unescaped newline inside substitute pattern");
523: /* NOTREACHED */
524: }
525: *sp++ = *p;
526: }
527: size += sp - op;
1.28 millert 528: } while ((p = cu_fgets(&lbuf, &bufsize)));
1.39 mmcc 529: error(COMPILE, "unterminated substitute in regular expression");
1.1 deraadt 530: /* NOTREACHED */
531: }
532:
533: /*
534: * Compile the flags of the s command
535: */
536: static char *
1.15 deraadt 537: compile_flags(char *p, struct s_subst *s)
1.1 deraadt 538: {
539: int gn; /* True if we have seen g or n */
1.17 otto 540: long l;
1.37 jsg 541: char wfile[PATH_MAX], *q, *eq;
1.1 deraadt 542:
543: s->n = 1; /* Default */
544: s->p = 0;
545: s->wfile = NULL;
546: s->wfd = -1;
547: for (gn = 0;;) {
548: EATSPACE(); /* EXTENSION */
549: switch (*p) {
550: case 'g':
551: if (gn)
1.39 mmcc 552: error(COMPILE, "more than one number or 'g' in"
1.22 tedu 553: " substitute flags");
1.1 deraadt 554: gn = 1;
555: s->n = 0;
556: break;
557: case '\0':
558: case '\n':
559: case ';':
560: return (p);
561: case 'p':
562: s->p = 1;
563: break;
564: case '1': case '2': case '3':
565: case '4': case '5': case '6':
566: case '7': case '8': case '9':
567: if (gn)
1.39 mmcc 568: error(COMPILE, "more than one number or 'g' in"
1.22 tedu 569: " substitute flags");
1.1 deraadt 570: gn = 1;
1.17 otto 571: l = strtol(p, &p, 10);
572: if (l <= 0 || l >= INT_MAX)
1.39 mmcc 573: error(COMPILE,
1.17 otto 574: "number in substitute flags out of range");
575: s->n = (int)l;
576: continue;
1.1 deraadt 577: case 'w':
578: p++;
579: #ifdef HISTORIC_PRACTICE
580: if (*p != ' ') {
1.39 mmcc 581: error(WARNING, "space missing before w wfile");
1.1 deraadt 582: return (p);
583: }
584: #endif
585: EATSPACE();
586: q = wfile;
1.37 jsg 587: eq = wfile + sizeof(wfile) - 1;
1.1 deraadt 588: while (*p) {
589: if (*p == '\n')
590: break;
1.37 jsg 591: if (q >= eq)
1.39 mmcc 592: error(COMPILE, "wfile too long");
1.1 deraadt 593: *q++ = *p++;
594: }
595: *q = '\0';
596: if (q == wfile)
1.39 mmcc 597: error(COMPILE, "no wfile specified");
1.1 deraadt 598: s->wfile = strdup(wfile);
599: if (!aflag && (s->wfd = open(wfile,
600: O_WRONLY|O_APPEND|O_CREAT|O_TRUNC,
601: DEFFILEMODE)) == -1)
1.39 mmcc 602: error(FATAL, "%s: %s", wfile, strerror(errno));
1.1 deraadt 603: return (p);
604: default:
1.39 mmcc 605: error(COMPILE,
1.1 deraadt 606: "bad flag in substitute command: '%c'", *p);
607: break;
608: }
609: p++;
610: }
611: }
612:
613: /*
614: * Compile a translation set of strings into a lookup table.
615: */
616: static char *
1.15 deraadt 617: compile_tr(char *p, char **transtab)
1.1 deraadt 618: {
619: int i;
620: char *lt, *op, *np;
1.25 millert 621: char *old = NULL, *new = NULL;
1.1 deraadt 622:
623: if (*p == '\0' || *p == '\\')
1.39 mmcc 624: error(COMPILE,
1.1 deraadt 625: "transform pattern can not be delimited by newline or backslash");
1.25 millert 626: old = xmalloc(strlen(p) + 1);
1.33 naddy 627: p = compile_delimited(p, old, 1);
1.1 deraadt 628: if (p == NULL) {
1.39 mmcc 629: error(COMPILE, "unterminated transform source string");
1.25 millert 630: goto bad;
1.1 deraadt 631: }
1.25 millert 632: new = xmalloc(strlen(p) + 1);
1.33 naddy 633: p = compile_delimited(--p, new, 1);
1.1 deraadt 634: if (p == NULL) {
1.39 mmcc 635: error(COMPILE, "unterminated transform target string");
1.25 millert 636: goto bad;
1.1 deraadt 637: }
638: EATSPACE();
639: if (strlen(new) != strlen(old)) {
1.39 mmcc 640: error(COMPILE, "transform strings are not the same length");
1.25 millert 641: goto bad;
1.1 deraadt 642: }
643: /* We assume characters are 8 bits */
1.18 deraadt 644: lt = xmalloc(UCHAR_MAX + 1);
1.1 deraadt 645: for (i = 0; i <= UCHAR_MAX; i++)
646: lt[i] = (char)i;
647: for (op = old, np = new; *op; op++, np++)
648: lt[(u_char)*op] = *np;
649: *transtab = lt;
1.25 millert 650: free(old);
651: free(new);
1.1 deraadt 652: return (p);
1.25 millert 653: bad:
654: free(old);
655: free(new);
656: return (NULL);
1.1 deraadt 657: }
658:
659: /*
1.7 deraadt 660: * Compile the text following an a, c, or i command.
1.1 deraadt 661: */
662: static char *
1.15 deraadt 663: compile_text(void)
1.1 deraadt 664: {
1.4 deraadt 665: int asize, esc_nl, size;
1.25 millert 666: char *lbuf, *text, *p, *op, *s;
1.28 millert 667: size_t bufsize;
1.1 deraadt 668:
1.28 millert 669: lbuf = text = NULL;
670: asize = size = 0;
671: while ((p = cu_fgets(&lbuf, &bufsize))) {
672: size_t len = ROUNDLEN(strlen(p) + 1);
673: if (asize - size < len) {
1.25 millert 674: do {
1.28 millert 675: asize += len;
676: } while (asize - size < len);
1.25 millert 677: text = xrealloc(text, asize);
678: }
1.1 deraadt 679: op = s = text + size;
1.4 deraadt 680: for (esc_nl = 0; *p != '\0'; p++) {
681: if (*p == '\\' && p[1] != '\0' && *++p == '\n')
682: esc_nl = 1;
1.1 deraadt 683: *s++ = *p;
684: }
685: size += s - op;
1.4 deraadt 686: if (!esc_nl) {
1.1 deraadt 687: *s = '\0';
688: break;
689: }
690: }
1.25 millert 691: free(lbuf);
1.34 millert 692: text = xrealloc(text, size + 1);
1.8 brian 693: text[size] = '\0';
1.34 millert 694: return (text);
1.1 deraadt 695: }
696:
697: /*
698: * Get an address and return a pointer to the first character after
699: * it. Fill the structure pointed to according to the address.
700: */
701: static char *
1.15 deraadt 702: compile_addr(char *p, struct s_addr *a)
1.1 deraadt 703: {
704: char *end;
705:
706: switch (*p) {
707: case '\\': /* Context address */
708: ++p;
709: /* FALLTHROUGH */
710: case '/': /* Context address */
711: p = compile_re(p, &a->u.r);
712: if (p == NULL)
1.39 mmcc 713: error(COMPILE, "unterminated regular expression");
1.1 deraadt 714: a->type = AT_RE;
715: return (p);
716:
717: case '$': /* Last line */
718: a->type = AT_LAST;
719: return (p + 1);
720: /* Line number */
1.21 deraadt 721: case '0': case '1': case '2': case '3': case '4':
1.1 deraadt 722: case '5': case '6': case '7': case '8': case '9':
723: a->type = AT_LINE;
1.3 millert 724: a->u.l = strtoul(p, &end, 10);
1.1 deraadt 725: return (end);
726: default:
1.39 mmcc 727: error(COMPILE, "expected context address");
1.1 deraadt 728: return (NULL);
729: }
730: }
731:
732: /*
733: * duptoeol --
734: * Return a copy of all the characters up to \n or \0.
735: */
736: static char *
1.16 deraadt 737: duptoeol(char *s, char *ctype, char **semi)
1.1 deraadt 738: {
739: size_t len;
740: int ws;
741: char *start;
742:
743: ws = 0;
1.16 deraadt 744: if (semi) {
745: for (start = s; *s != '\0' && *s != '\n' && *s != ';'; ++s)
1.35 deraadt 746: ws = isspace((unsigned char)*s);
1.16 deraadt 747: } else {
748: for (start = s; *s != '\0' && *s != '\n'; ++s)
1.35 deraadt 749: ws = isspace((unsigned char)*s);
1.16 deraadt 750: *s = '\0';
751: }
1.1 deraadt 752: if (ws)
1.39 mmcc 753: error(WARNING, "whitespace after %s", ctype);
1.1 deraadt 754: len = s - start + 1;
1.16 deraadt 755: if (semi)
756: *semi = s;
757: s = xmalloc(len);
758: strlcpy(s, start, len);
759: return (s);
1.1 deraadt 760: }
761:
762: /*
763: * Convert goto label names to addresses, and count a and r commands, in
764: * the given subset of the script. Free the memory used by labels in b
765: * and t commands (but not by :).
766: *
767: * TODO: Remove } nodes
768: */
769: static void
1.15 deraadt 770: fixuplabel(struct s_command *cp, struct s_command *end)
1.1 deraadt 771: {
772:
773: for (; cp != end; cp = cp->next)
774: switch (cp->code) {
775: case 'a':
776: case 'r':
777: appendnum++;
778: break;
779: case 'b':
780: case 't':
781: /* Resolve branch target. */
782: if (cp->t == NULL) {
783: cp->u.c = NULL;
784: break;
785: }
786: if ((cp->u.c = findlabel(cp->t)) == NULL)
1.40 ! jca 787: error(COMPILE, "undefined label '%s'", cp->t);
1.1 deraadt 788: free(cp->t);
789: break;
790: case '{':
791: /* Do interior commands. */
792: fixuplabel(cp->u.c, cp->next);
793: break;
794: }
795: }
796:
797: /*
798: * Associate the given command label for later lookup.
799: */
800: static void
1.15 deraadt 801: enterlabel(struct s_command *cp)
1.1 deraadt 802: {
1.10 mpech 803: struct labhash **lhp, *lh;
804: u_char *p;
805: u_int h, c;
1.1 deraadt 806:
807: for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++)
808: h = (h << 5) + h + c;
809: lhp = &labels[h & LHMASK];
810: for (lh = *lhp; lh != NULL; lh = lh->lh_next)
811: if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0)
1.40 ! jca 812: error(COMPILE, "duplicate label '%s'", cp->t);
1.1 deraadt 813: lh = xmalloc(sizeof *lh);
814: lh->lh_next = *lhp;
815: lh->lh_hash = h;
816: lh->lh_cmd = cp;
817: lh->lh_ref = 0;
818: *lhp = lh;
819: }
820:
821: /*
822: * Find the label contained in the command l in the command linked
823: * list cp. L is excluded from the search. Return NULL if not found.
824: */
825: static struct s_command *
1.15 deraadt 826: findlabel(char *name)
1.1 deraadt 827: {
1.10 mpech 828: struct labhash *lh;
829: u_char *p;
830: u_int h, c;
1.1 deraadt 831:
832: for (h = 0, p = (u_char *)name; (c = *p) != 0; p++)
833: h = (h << 5) + h + c;
834: for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) {
835: if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) {
836: lh->lh_ref = 1;
837: return (lh->lh_cmd);
838: }
839: }
840: return (NULL);
841: }
842:
1.21 deraadt 843: /*
1.1 deraadt 844: * Warn about any unused labels. As a side effect, release the label hash
845: * table space.
846: */
847: static void
1.15 deraadt 848: uselabel(void)
1.1 deraadt 849: {
1.10 mpech 850: struct labhash *lh, *next;
851: int i;
1.1 deraadt 852:
853: for (i = 0; i < LHSZ; i++) {
854: for (lh = labels[i]; lh != NULL; lh = next) {
855: next = lh->lh_next;
856: if (!lh->lh_ref)
1.39 mmcc 857: error(WARNING, "unused label '%s'",
1.1 deraadt 858: lh->lh_cmd->t);
859: free(lh);
860: }
861: }
862: }