Commit | Line | Data |
---|---|---|
f2edd8cd KB |
1 | /*- |
2 | * Copyright (c) 1992 Diomidis Spinellis. | |
ad787160 C |
3 | * Copyright (c) 1992, 1993 |
4 | * The Regents of the University of California. All rights reserved. | |
f2edd8cd KB |
5 | * |
6 | * This code is derived from software contributed to Berkeley by | |
7 | * Diomidis Spinellis of Imperial College, University of London. | |
8 | * | |
ad787160 C |
9 | * Redistribution and use in source and binary forms, with or without |
10 | * modification, are permitted provided that the following conditions | |
11 | * are met: | |
12 | * 1. Redistributions of source code must retain the above copyright | |
13 | * notice, this list of conditions and the following disclaimer. | |
14 | * 2. Redistributions in binary form must reproduce the above copyright | |
15 | * notice, this list of conditions and the following disclaimer in the | |
16 | * documentation and/or other materials provided with the distribution. | |
17 | * 3. All advertising materials mentioning features or use of this software | |
18 | * must display the following acknowledgement: | |
19 | * This product includes software developed by the University of | |
20 | * California, Berkeley and its contributors. | |
21 | * 4. Neither the name of the University nor the names of its contributors | |
22 | * may be used to endorse or promote products derived from this software | |
23 | * without specific prior written permission. | |
24 | * | |
25 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND | |
26 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | |
27 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | |
28 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE | |
29 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | |
30 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS | |
31 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) | |
32 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT | |
33 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY | |
34 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF | |
35 | * SUCH DAMAGE. | |
f2edd8cd KB |
36 | */ |
37 | ||
38 | #ifndef lint | |
fd88f5c5 | 39 | static char sccsid[] = "@(#)compile.c 8.2 (Berkeley) 4/28/95"; |
f2edd8cd KB |
40 | #endif /* not lint */ |
41 | ||
42 | #include <sys/types.h> | |
f3e85330 | 43 | #include <sys/stat.h> |
f2edd8cd KB |
44 | |
45 | #include <ctype.h> | |
46 | #include <errno.h> | |
47 | #include <fcntl.h> | |
48 | #include <limits.h> | |
49 | #include <regex.h> | |
50 | #include <stdio.h> | |
51 | #include <stdlib.h> | |
52 | #include <string.h> | |
53 | ||
54 | #include "defs.h" | |
55 | #include "extern.h" | |
56 | ||
32cb677f CT |
57 | #define LHSZ 128 |
58 | #define LHMASK (LHSZ - 1) | |
59 | static struct labhash { | |
60 | struct labhash *lh_next; | |
61 | u_int lh_hash; | |
62 | struct s_command *lh_cmd; | |
63 | int lh_ref; | |
64 | } *labels[LHSZ]; | |
65 | ||
f2edd8cd KB |
66 | static char *compile_addr __P((char *, struct s_addr *)); |
67 | static char *compile_delimited __P((char *, char *)); | |
68 | static char *compile_flags __P((char *, struct s_subst *)); | |
25d34f03 | 69 | static char *compile_re __P((char *, regex_t **)); |
f3e85330 | 70 | static char *compile_subst __P((char *, struct s_subst *)); |
f2edd8cd KB |
71 | static char *compile_text __P((void)); |
72 | static char *compile_tr __P((char *, char **)); | |
73 | static struct s_command | |
74 | **compile_stream __P((char *, struct s_command **, char *)); | |
910ea097 | 75 | static char *duptoeol __P((char *, char *)); |
32cb677f | 76 | static void enterlabel __P((struct s_command *)); |
f2edd8cd | 77 | static struct s_command |
32cb677f CT |
78 | *findlabel __P((char *)); |
79 | static void fixuplabel __P((struct s_command *, struct s_command *)); | |
80 | static void uselabel __P((void)); | |
f2edd8cd KB |
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 | {'a', 1, TEXT}, | |
94 | {'b', 2, BRANCH}, | |
95 | {'c', 2, TEXT}, | |
96 | {'d', 2, EMPTY}, | |
97 | {'D', 2, EMPTY}, | |
98 | {'g', 2, EMPTY}, | |
99 | {'G', 2, EMPTY}, | |
100 | {'h', 2, EMPTY}, | |
101 | {'H', 2, EMPTY}, | |
102 | {'i', 1, TEXT}, | |
103 | {'l', 2, EMPTY}, | |
104 | {'n', 2, EMPTY}, | |
105 | {'N', 2, EMPTY}, | |
106 | {'p', 2, EMPTY}, | |
107 | {'P', 2, EMPTY}, | |
108 | {'q', 1, EMPTY}, | |
109 | {'r', 1, RFILE}, | |
110 | {'s', 2, SUBST}, | |
111 | {'t', 2, BRANCH}, | |
112 | {'w', 2, WFILE}, | |
113 | {'x', 2, EMPTY}, | |
114 | {'y', 2, TR}, | |
115 | {'!', 2, NONSEL}, | |
116 | {':', 0, LABEL}, | |
117 | {'#', 0, COMMENT}, | |
118 | {'=', 1, EMPTY}, | |
119 | {'\0', 0, COMMENT}, | |
120 | }; | |
121 | ||
f3e85330 | 122 | /* The compiled program. */ |
f2edd8cd KB |
123 | struct s_command *prog; |
124 | ||
125 | /* | |
126 | * Compile the program into prog. | |
f3e85330 | 127 | * Initialise appends. |
f2edd8cd KB |
128 | */ |
129 | void | |
130 | compile() | |
131 | { | |
132 | *compile_stream(NULL, &prog, NULL) = NULL; | |
32cb677f CT |
133 | fixuplabel(prog, NULL); |
134 | uselabel(); | |
f2edd8cd | 135 | appends = xmalloc(sizeof(struct s_appends) * appendnum); |
25d34f03 | 136 | match = xmalloc((maxnsub + 1) * sizeof(regmatch_t)); |
f2edd8cd KB |
137 | } |
138 | ||
139 | #define EATSPACE() do { \ | |
140 | if (p) \ | |
141 | while (*p && isascii(*p) && isspace(*p)) \ | |
142 | p++; \ | |
143 | } while (0) | |
144 | ||
145 | static struct s_command ** | |
146 | compile_stream(terminator, link, p) | |
147 | char *terminator; | |
148 | struct s_command **link; | |
149 | register char *p; | |
150 | { | |
151 | static char lbuf[_POSIX2_LINE_MAX + 1]; /* To save stack */ | |
152 | struct s_command *cmd, *cmd2; | |
153 | struct s_format *fp; | |
154 | int naddr; /* Number of addresses */ | |
155 | ||
156 | if (p != NULL) | |
157 | goto semicolon; | |
158 | for (;;) { | |
159 | if ((p = cu_fgets(lbuf, sizeof(lbuf))) == NULL) { | |
160 | if (terminator != NULL) | |
161 | err(COMPILE, "unexpected EOF (pending }'s)"); | |
162 | return (link); | |
163 | } | |
164 | ||
165 | semicolon: EATSPACE(); | |
166 | if (p && (*p == '#' || *p == '\0')) | |
167 | continue; | |
168 | if (*p == '}') { | |
169 | if (terminator == NULL) | |
170 | err(COMPILE, "unexpected }"); | |
171 | return (link); | |
172 | } | |
173 | *link = cmd = xmalloc(sizeof(struct s_command)); | |
174 | link = &cmd->next; | |
32cb677f | 175 | cmd->nonsel = cmd->inrange = 0; |
f2edd8cd KB |
176 | /* First parse the addresses */ |
177 | naddr = 0; | |
f2edd8cd KB |
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 == ',') { | |
f2edd8cd KB |
187 | p++; |
188 | EATSPACE(); /* EXTENSION */ | |
82daf77a | 189 | naddr++; |
f2edd8cd KB |
190 | cmd->a2 = xmalloc(sizeof(struct s_addr)); |
191 | p = compile_addr(p, cmd->a2); | |
82daf77a KB |
192 | EATSPACE(); |
193 | } else | |
194 | cmd->a2 = 0; | |
195 | } else | |
196 | cmd->a1 = cmd->a2 = 0; | |
f2edd8cd KB |
197 | |
198 | nonsel: /* Now parse the command */ | |
f2edd8cd KB |
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, | |
209 | "command %c expects up to %d address(es), found %d", *p, fp->naddr, naddr); | |
210 | switch (fp->args) { | |
211 | case NONSEL: /* ! */ | |
f2edd8cd | 212 | p++; |
82daf77a KB |
213 | EATSPACE(); |
214 | cmd->nonsel = ! cmd->nonsel; | |
f2edd8cd KB |
215 | goto nonsel; |
216 | case GROUP: /* { */ | |
217 | p++; | |
218 | EATSPACE(); | |
219 | if (!*p) | |
220 | p = NULL; | |
221 | cmd2 = xmalloc(sizeof(struct s_command)); | |
222 | cmd2->code = '}'; | |
223 | *compile_stream("}", &cmd->u.c, p) = cmd2; | |
224 | cmd->next = cmd2; | |
225 | link = &cmd2->next; | |
82daf77a KB |
226 | /* |
227 | * Short-circuit command processing, since end of | |
228 | * group is really just a noop. | |
229 | */ | |
230 | cmd2->nonsel = 1; | |
231 | cmd2->a1 = cmd2->a2 = 0; | |
f2edd8cd KB |
232 | break; |
233 | case EMPTY: /* d D g G h H l n N p P q x = \0 */ | |
234 | p++; | |
235 | EATSPACE(); | |
236 | if (*p == ';') { | |
237 | p++; | |
238 | link = &cmd->next; | |
239 | goto semicolon; | |
240 | } | |
241 | if (*p) | |
242 | err(COMPILE, | |
243 | "extra characters at the end of %c command", cmd->code); | |
244 | break; | |
245 | case TEXT: /* a c i */ | |
246 | p++; | |
247 | EATSPACE(); | |
248 | if (*p != '\\') | |
249 | err(COMPILE, | |
250 | "command %c expects \\ followed by text", cmd->code); | |
251 | p++; | |
252 | EATSPACE(); | |
253 | if (*p) | |
254 | err(COMPILE, | |
255 | "extra characters after \\ at the end of %c command", cmd->code); | |
256 | cmd->t = compile_text(); | |
257 | break; | |
258 | case COMMENT: /* \0 # */ | |
259 | break; | |
260 | case WFILE: /* w */ | |
261 | p++; | |
262 | EATSPACE(); | |
263 | if (*p == '\0') | |
264 | err(COMPILE, "filename expected"); | |
910ea097 | 265 | cmd->t = duptoeol(p, "w command"); |
f2edd8cd KB |
266 | if (aflag) |
267 | cmd->u.fd = -1; | |
268 | else if ((cmd->u.fd = open(p, | |
269 | O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, | |
270 | DEFFILEMODE)) == -1) | |
271 | err(FATAL, "%s: %s\n", p, strerror(errno)); | |
272 | break; | |
273 | case RFILE: /* r */ | |
274 | p++; | |
275 | EATSPACE(); | |
276 | if (*p == '\0') | |
277 | err(COMPILE, "filename expected"); | |
278 | else | |
910ea097 | 279 | cmd->t = duptoeol(p, "read command"); |
f2edd8cd KB |
280 | break; |
281 | case BRANCH: /* b t */ | |
282 | p++; | |
283 | EATSPACE(); | |
284 | if (*p == '\0') | |
285 | cmd->t = NULL; | |
286 | else | |
910ea097 | 287 | cmd->t = duptoeol(p, "branch"); |
f2edd8cd KB |
288 | break; |
289 | case LABEL: /* : */ | |
290 | p++; | |
291 | EATSPACE(); | |
910ea097 | 292 | cmd->t = duptoeol(p, "label"); |
f2edd8cd KB |
293 | if (strlen(p) == 0) |
294 | err(COMPILE, "empty label"); | |
32cb677f | 295 | enterlabel(cmd); |
f2edd8cd KB |
296 | break; |
297 | case SUBST: /* s */ | |
298 | p++; | |
299 | if (*p == '\0' || *p == '\\') | |
300 | err(COMPILE, | |
301 | "substitute pattern can not be delimited by newline or backslash"); | |
302 | cmd->u.s = xmalloc(sizeof(struct s_subst)); | |
25d34f03 | 303 | p = compile_re(p, &cmd->u.s->re); |
f2edd8cd | 304 | if (p == NULL) |
6c6995a7 | 305 | err(COMPILE, "unterminated substitute pattern"); |
f3e85330 KB |
306 | --p; |
307 | p = compile_subst(p, cmd->u.s); | |
f2edd8cd KB |
308 | p = compile_flags(p, cmd->u.s); |
309 | EATSPACE(); | |
310 | if (*p == ';') { | |
311 | p++; | |
312 | link = &cmd->next; | |
313 | goto semicolon; | |
314 | } | |
315 | break; | |
316 | case TR: /* y */ | |
317 | p++; | |
318 | p = compile_tr(p, (char **)&cmd->u.y); | |
319 | EATSPACE(); | |
320 | if (*p == ';') { | |
321 | p++; | |
322 | link = &cmd->next; | |
323 | goto semicolon; | |
324 | } | |
325 | if (*p) | |
326 | err(COMPILE, | |
327 | "extra text at the end of a transform command"); | |
328 | break; | |
329 | } | |
330 | } | |
331 | } | |
332 | ||
333 | /* | |
334 | * Get a delimited string. P points to the delimeter of the string; d points | |
335 | * to a buffer area. Newline and delimiter escapes are processed; other | |
336 | * escapes are ignored. | |
337 | * | |
338 | * Returns a pointer to the first character after the final delimiter or NULL | |
339 | * in the case of a non-terminated string. The character array d is filled | |
340 | * with the processed string. | |
341 | */ | |
342 | static char * | |
343 | compile_delimited(p, d) | |
344 | char *p, *d; | |
345 | { | |
346 | char c; | |
347 | ||
348 | c = *p++; | |
349 | if (c == '\0') | |
350 | return (NULL); | |
351 | else if (c == '\\') | |
352 | err(COMPILE, "\\ can not be used as a string delimiter"); | |
353 | else if (c == '\n') | |
354 | err(COMPILE, "newline can not be used as a string delimiter"); | |
355 | while (*p) { | |
356 | if (*p == '\\' && p[1] == c) | |
f3e85330 | 357 | p++; |
f2edd8cd | 358 | else if (*p == '\\' && p[1] == 'n') { |
f3e85330 KB |
359 | *d++ = '\n'; |
360 | p += 2; | |
361 | continue; | |
64aaecab KB |
362 | } else if (*p == '\\' && p[1] == '\\') |
363 | *d++ = *p++; | |
364 | else if (*p == c) { | |
f2edd8cd KB |
365 | *d = '\0'; |
366 | return (p + 1); | |
367 | } | |
368 | *d++ = *p++; | |
369 | } | |
370 | return (NULL); | |
371 | } | |
372 | ||
373 | /* | |
f3e85330 KB |
374 | * Get a regular expression. P points to the delimiter of the regular |
375 | * expression; repp points to the address of a regexp pointer. Newline | |
376 | * and delimiter escapes are processed; other escapes are ignored. | |
f2edd8cd | 377 | * Returns a pointer to the first character after the final delimiter |
f3e85330 KB |
378 | * or NULL in the case of a non terminated regular expression. The regexp |
379 | * pointer is set to the compiled regular expression. | |
f2edd8cd KB |
380 | * Cflags are passed to regcomp. |
381 | */ | |
382 | static char * | |
25d34f03 | 383 | compile_re(p, repp) |
f2edd8cd | 384 | char *p; |
f3e85330 | 385 | regex_t **repp; |
f2edd8cd KB |
386 | { |
387 | int eval; | |
388 | char re[_POSIX2_LINE_MAX + 1]; | |
389 | ||
390 | p = compile_delimited(p, re); | |
f3e85330 KB |
391 | if (p && strlen(re) == 0) { |
392 | *repp = NULL; | |
393 | return (p); | |
394 | } | |
395 | *repp = xmalloc(sizeof(regex_t)); | |
25d34f03 | 396 | if (p && (eval = regcomp(*repp, re, 0)) != 0) |
f3e85330 | 397 | err(COMPILE, "RE error: %s", strregerror(eval, *repp)); |
25d34f03 KB |
398 | if (maxnsub < (*repp)->re_nsub) |
399 | maxnsub = (*repp)->re_nsub; | |
f2edd8cd KB |
400 | return (p); |
401 | } | |
402 | ||
403 | /* | |
404 | * Compile the substitution string of a regular expression and set res to | |
405 | * point to a saved copy of it. Nsub is the number of parenthesized regular | |
406 | * expressions. | |
407 | */ | |
408 | static char * | |
f3e85330 KB |
409 | compile_subst(p, s) |
410 | char *p; | |
411 | struct s_subst *s; | |
f2edd8cd KB |
412 | { |
413 | static char lbuf[_POSIX2_LINE_MAX + 1]; | |
f3e85330 KB |
414 | int asize, ref, size; |
415 | char c, *text, *op, *sp; | |
f2edd8cd KB |
416 | |
417 | c = *p++; /* Terminator character */ | |
418 | if (c == '\0') | |
419 | return (NULL); | |
420 | ||
f3e85330 KB |
421 | s->maxbref = 0; |
422 | s->linenum = linenum; | |
f2edd8cd KB |
423 | asize = 2 * _POSIX2_LINE_MAX + 1; |
424 | text = xmalloc(asize); | |
425 | size = 0; | |
426 | do { | |
f3e85330 | 427 | op = sp = text + size; |
f2edd8cd KB |
428 | for (; *p; p++) { |
429 | if (*p == '\\') { | |
430 | p++; | |
431 | if (strchr("123456789", *p) != NULL) { | |
f3e85330 KB |
432 | *sp++ = '\\'; |
433 | ref = *p - '0'; | |
f3e85330 KB |
434 | if (s->re != NULL && |
435 | ref > s->re->re_nsub) | |
f2edd8cd | 436 | err(COMPILE, |
f3e85330 | 437 | "\\%c not defined in the RE", *p); |
25d34f03 KB |
438 | if (s->maxbref < ref) |
439 | s->maxbref = ref; | |
64aaecab | 440 | } else if (*p == '&' || *p == '\\') |
f3e85330 | 441 | *sp++ = '\\'; |
f2edd8cd KB |
442 | } else if (*p == c) { |
443 | p++; | |
f3e85330 KB |
444 | *sp++ = '\0'; |
445 | size += sp - op; | |
446 | s->new = xrealloc(text, size); | |
f2edd8cd KB |
447 | return (p); |
448 | } else if (*p == '\n') { | |
449 | err(COMPILE, | |
450 | "unescaped newline inside substitute pattern"); | |
f3e85330 | 451 | /* NOTREACHED */ |
f2edd8cd | 452 | } |
f3e85330 | 453 | *sp++ = *p; |
f2edd8cd | 454 | } |
f3e85330 | 455 | size += sp - op; |
f2edd8cd KB |
456 | if (asize - size < _POSIX2_LINE_MAX + 1) { |
457 | asize *= 2; | |
458 | text = xmalloc(asize); | |
459 | } | |
460 | } while (cu_fgets(p = lbuf, sizeof(lbuf))); | |
f3e85330 KB |
461 | err(COMPILE, "unterminated substitute in regular expression"); |
462 | /* NOTREACHED */ | |
f2edd8cd KB |
463 | } |
464 | ||
465 | /* | |
466 | * Compile the flags of the s command | |
467 | */ | |
468 | static char * | |
469 | compile_flags(p, s) | |
470 | char *p; | |
471 | struct s_subst *s; | |
472 | { | |
473 | int gn; /* True if we have seen g or n */ | |
474 | char wfile[_POSIX2_LINE_MAX + 1], *q; | |
475 | ||
476 | s->n = 1; /* Default */ | |
477 | s->p = 0; | |
478 | s->wfile = NULL; | |
479 | s->wfd = -1; | |
480 | for (gn = 0;;) { | |
481 | EATSPACE(); /* EXTENSION */ | |
482 | switch (*p) { | |
483 | case 'g': | |
484 | if (gn) | |
6c6995a7 KB |
485 | err(COMPILE, |
486 | "more than one number or 'g' in substitute flags"); | |
f2edd8cd KB |
487 | gn = 1; |
488 | s->n = 0; | |
489 | break; | |
490 | case '\0': | |
491 | case '\n': | |
492 | case ';': | |
493 | return (p); | |
494 | case 'p': | |
495 | s->p = 1; | |
496 | break; | |
497 | case '1': case '2': case '3': | |
498 | case '4': case '5': case '6': | |
499 | case '7': case '8': case '9': | |
500 | if (gn) | |
6c6995a7 KB |
501 | err(COMPILE, |
502 | "more than one number or 'g' in substitute flags"); | |
f2edd8cd KB |
503 | gn = 1; |
504 | /* XXX Check for overflow */ | |
505 | s->n = (int)strtol(p, &p, 10); | |
506 | break; | |
507 | case 'w': | |
508 | p++; | |
509 | #ifdef HISTORIC_PRACTICE | |
510 | if (*p != ' ') { | |
511 | err(WARNING, "space missing before w wfile"); | |
512 | return (p); | |
513 | } | |
514 | #endif | |
515 | EATSPACE(); | |
516 | q = wfile; | |
517 | while (*p) { | |
518 | if (*p == '\n') | |
519 | break; | |
520 | *q++ = *p++; | |
521 | } | |
522 | *q = '\0'; | |
523 | if (q == wfile) | |
6c6995a7 | 524 | err(COMPILE, "no wfile specified"); |
f2edd8cd KB |
525 | s->wfile = strdup(wfile); |
526 | if (!aflag && (s->wfd = open(wfile, | |
527 | O_WRONLY|O_APPEND|O_CREAT|O_TRUNC, | |
528 | DEFFILEMODE)) == -1) | |
529 | err(FATAL, "%s: %s\n", wfile, strerror(errno)); | |
530 | return (p); | |
531 | default: | |
532 | err(COMPILE, | |
6c6995a7 | 533 | "bad flag in substitute command: '%c'", *p); |
f2edd8cd KB |
534 | break; |
535 | } | |
536 | p++; | |
537 | } | |
538 | } | |
539 | ||
540 | /* | |
541 | * Compile a translation set of strings into a lookup table. | |
542 | */ | |
543 | static char * | |
544 | compile_tr(p, transtab) | |
545 | char *p; | |
546 | char **transtab; | |
547 | { | |
548 | int i; | |
549 | char *lt, *op, *np; | |
550 | char old[_POSIX2_LINE_MAX + 1]; | |
551 | char new[_POSIX2_LINE_MAX + 1]; | |
552 | ||
553 | if (*p == '\0' || *p == '\\') | |
554 | err(COMPILE, | |
555 | "transform pattern can not be delimited by newline or backslash"); | |
556 | p = compile_delimited(p, old); | |
557 | if (p == NULL) { | |
558 | err(COMPILE, "unterminated transform source string"); | |
559 | return (NULL); | |
560 | } | |
561 | p = compile_delimited(--p, new); | |
562 | if (p == NULL) { | |
563 | err(COMPILE, "unterminated transform target string"); | |
564 | return (NULL); | |
565 | } | |
566 | EATSPACE(); | |
567 | if (strlen(new) != strlen(old)) { | |
568 | err(COMPILE, "transform strings are not the same length"); | |
569 | return (NULL); | |
570 | } | |
571 | /* We assume characters are 8 bits */ | |
572 | lt = xmalloc(UCHAR_MAX); | |
573 | for (i = 0; i <= UCHAR_MAX; i++) | |
574 | lt[i] = (char)i; | |
575 | for (op = old, np = new; *op; op++, np++) | |
576 | lt[(u_char)*op] = *np; | |
577 | *transtab = lt; | |
578 | return (p); | |
579 | } | |
580 | ||
581 | /* | |
582 | * Compile the text following an a or i command. | |
583 | */ | |
584 | static char * | |
585 | compile_text() | |
586 | { | |
587 | int asize, size; | |
588 | char *text, *p, *op, *s; | |
589 | char lbuf[_POSIX2_LINE_MAX + 1]; | |
590 | ||
591 | asize = 2 * _POSIX2_LINE_MAX + 1; | |
592 | text = xmalloc(asize); | |
593 | size = 0; | |
594 | while (cu_fgets(lbuf, sizeof(lbuf))) { | |
595 | op = s = text + size; | |
596 | p = lbuf; | |
597 | EATSPACE(); | |
598 | for (; *p; p++) { | |
599 | if (*p == '\\') | |
600 | p++; | |
601 | *s++ = *p; | |
602 | } | |
603 | size += s - op; | |
604 | if (p[-2] != '\\') { | |
605 | *s = '\0'; | |
606 | break; | |
607 | } | |
608 | if (asize - size < _POSIX2_LINE_MAX + 1) { | |
609 | asize *= 2; | |
610 | text = xmalloc(asize); | |
611 | } | |
612 | } | |
613 | return (xrealloc(text, size + 1)); | |
614 | } | |
615 | ||
616 | /* | |
617 | * Get an address and return a pointer to the first character after | |
618 | * it. Fill the structure pointed to according to the address. | |
619 | */ | |
620 | static char * | |
621 | compile_addr(p, a) | |
622 | char *p; | |
623 | struct s_addr *a; | |
624 | { | |
f2edd8cd KB |
625 | char *end; |
626 | ||
627 | switch (*p) { | |
628 | case '\\': /* Context address */ | |
f3e85330 KB |
629 | ++p; |
630 | /* FALLTHROUGH */ | |
f2edd8cd | 631 | case '/': /* Context address */ |
25d34f03 | 632 | p = compile_re(p, &a->u.r); |
f2edd8cd KB |
633 | if (p == NULL) |
634 | err(COMPILE, "unterminated regular expression"); | |
635 | a->type = AT_RE; | |
636 | return (p); | |
f3e85330 | 637 | |
f2edd8cd KB |
638 | case '$': /* Last line */ |
639 | a->type = AT_LAST; | |
640 | return (p + 1); | |
641 | /* Line number */ | |
642 | case '0': case '1': case '2': case '3': case '4': | |
643 | case '5': case '6': case '7': case '8': case '9': | |
644 | a->type = AT_LINE; | |
645 | a->u.l = strtol(p, &end, 10); | |
646 | return (end); | |
647 | default: | |
648 | err(COMPILE, "expected context address"); | |
649 | return (NULL); | |
650 | } | |
651 | } | |
652 | ||
653 | /* | |
910ea097 KB |
654 | * duptoeol -- |
655 | * Return a copy of all the characters up to \n or \0. | |
f2edd8cd KB |
656 | */ |
657 | static char * | |
910ea097 | 658 | duptoeol(s, ctype) |
f2edd8cd | 659 | register char *s; |
910ea097 | 660 | char *ctype; |
f2edd8cd KB |
661 | { |
662 | size_t len; | |
910ea097 | 663 | int ws; |
f2edd8cd KB |
664 | char *start; |
665 | ||
910ea097 KB |
666 | ws = 0; |
667 | for (start = s; *s != '\0' && *s != '\n'; ++s) | |
668 | ws = isspace(*s); | |
f2edd8cd | 669 | *s = '\0'; |
910ea097 KB |
670 | if (ws) |
671 | err(WARNING, "whitespace after %s", ctype); | |
f2edd8cd KB |
672 | len = s - start + 1; |
673 | return (memmove(xmalloc(len), start, len)); | |
674 | } | |
675 | ||
676 | /* | |
32cb677f CT |
677 | * Convert goto label names to addresses, and count a and r commands, in |
678 | * the given subset of the script. Free the memory used by labels in b | |
679 | * and t commands (but not by :). | |
910ea097 | 680 | * |
f2edd8cd KB |
681 | * TODO: Remove } nodes |
682 | */ | |
683 | static void | |
32cb677f CT |
684 | fixuplabel(cp, end) |
685 | struct s_command *cp, *end; | |
f2edd8cd | 686 | { |
f2edd8cd | 687 | |
680fc01f | 688 | for (; cp != end; cp = cp->next) |
f2edd8cd KB |
689 | switch (cp->code) { |
690 | case 'a': | |
691 | case 'r': | |
692 | appendnum++; | |
693 | break; | |
694 | case 'b': | |
695 | case 't': | |
32cb677f | 696 | /* Resolve branch target. */ |
f2edd8cd KB |
697 | if (cp->t == NULL) { |
698 | cp->u.c = NULL; | |
699 | break; | |
700 | } | |
32cb677f | 701 | if ((cp->u.c = findlabel(cp->t)) == NULL) |
6c6995a7 | 702 | err(COMPILE2, "undefined label '%s'", cp->t); |
f2edd8cd | 703 | free(cp->t); |
f2edd8cd KB |
704 | break; |
705 | case '{': | |
32cb677f CT |
706 | /* Do interior commands. */ |
707 | fixuplabel(cp->u.c, cp->next); | |
f2edd8cd | 708 | break; |
f2edd8cd | 709 | } |
32cb677f CT |
710 | } |
711 | ||
712 | /* | |
713 | * Associate the given command label for later lookup. | |
714 | */ | |
715 | static void | |
716 | enterlabel(cp) | |
717 | struct s_command *cp; | |
718 | { | |
719 | register struct labhash **lhp, *lh; | |
720 | register u_char *p; | |
721 | register u_int h, c; | |
722 | ||
723 | for (h = 0, p = (u_char *)cp->t; (c = *p) != 0; p++) | |
724 | h = (h << 5) + h + c; | |
725 | lhp = &labels[h & LHMASK]; | |
726 | for (lh = *lhp; lh != NULL; lh = lh->lh_next) | |
727 | if (lh->lh_hash == h && strcmp(cp->t, lh->lh_cmd->t) == 0) | |
728 | err(COMPILE2, "duplicate label '%s'", cp->t); | |
729 | lh = xmalloc(sizeof *lh); | |
730 | lh->lh_next = *lhp; | |
731 | lh->lh_hash = h; | |
732 | lh->lh_cmd = cp; | |
733 | lh->lh_ref = 0; | |
734 | *lhp = lh; | |
910ea097 KB |
735 | } |
736 | ||
737 | /* | |
738 | * Find the label contained in the command l in the command linked | |
739 | * list cp. L is excluded from the search. Return NULL if not found. | |
740 | */ | |
741 | static struct s_command * | |
32cb677f CT |
742 | findlabel(name) |
743 | char *name; | |
910ea097 | 744 | { |
32cb677f CT |
745 | register struct labhash *lh; |
746 | register u_char *p; | |
747 | register u_int h, c; | |
748 | ||
749 | for (h = 0, p = (u_char *)name; (c = *p) != 0; p++) | |
750 | h = (h << 5) + h + c; | |
751 | for (lh = labels[h & LHMASK]; lh != NULL; lh = lh->lh_next) { | |
752 | if (lh->lh_hash == h && strcmp(name, lh->lh_cmd->t) == 0) { | |
753 | lh->lh_ref = 1; | |
754 | return (lh->lh_cmd); | |
755 | } | |
910ea097 KB |
756 | } |
757 | return (NULL); | |
758 | } | |
759 | ||
760 | /* | |
32cb677f CT |
761 | * Warn about any unused labels. As a side effect, release the label hash |
762 | * table space. | |
910ea097 KB |
763 | */ |
764 | static void | |
32cb677f | 765 | uselabel() |
910ea097 | 766 | { |
32cb677f CT |
767 | register struct labhash *lh, *next; |
768 | register int i; | |
769 | ||
770 | for (i = 0; i < LHSZ; i++) { | |
771 | for (lh = labels[i]; lh != NULL; lh = next) { | |
772 | next = lh->lh_next; | |
773 | if (!lh->lh_ref) | |
774 | err(WARNING, "unused label '%s'", | |
775 | lh->lh_cmd->t); | |
776 | free(lh); | |
777 | } | |
910ea097 | 778 | } |
f2edd8cd | 779 | } |