xref: /freebsd/usr.bin/m4/gnum4.c (revision eacee0ff7ec955b32e09515246bd97b6edcd2b0f) 
1 /* $OpenBSD: gnum4.c,v 1.16 2002/02/16 21:27:48 millert Exp $ */
2 
3 /*
4  * Copyright (c) 1999 Marc Espie
5  *
6  * Redistribution and use in source and binary forms, with or without
7  * modification, are permitted provided that the following conditions
8  * are met:
9  * 1. Redistributions of source code must retain the above copyright
10  *    notice, this list of conditions and the following disclaimer.
11  * 2. Redistributions in binary form must reproduce the above copyright
12  *    notice, this list of conditions and the following disclaimer in the
13  *    documentation and/or other materials provided with the distribution.
14  *
15  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
16  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
19  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25  * SUCH DAMAGE.
26  */
27 
28 /*
29  * functions needed to support gnu-m4 extensions, including a fake freezing
30  */
31 
32 #include <sys/param.h>
33 #include <sys/types.h>
34 #include <sys/wait.h>
35 #include <ctype.h>
36 #include <paths.h>
37 #include <regex.h>
38 #include <stddef.h>
39 #include <stdlib.h>
40 #include <stdio.h>
41 #include <string.h>
42 #include <err.h>
43 #include <errno.h>
44 #include <unistd.h>
45 #include "mdef.h"
46 #include "stdd.h"
47 #include "extern.h"
48 
49 
50 int mimic_gnu = 0;
51 
52 /*
53  * Support for include path search
54  * First search in the the current directory.
55  * If not found, and the path is not absolute, include path kicks in.
56  * First, -I options, in the order found on the command line.
57  * Then M4PATH env variable
58  */
59 
60 struct path_entry {
61 	char *name;
62 	struct path_entry *next;
63 } *first, *last;
64 
65 static struct path_entry *new_path_entry(const char *);
66 static void ensure_m4path(void);
67 static struct input_file *dopath(struct input_file *, const char *);
68 
69 static struct path_entry *
70 new_path_entry(dirname)
71 	const char *dirname;
72 {
73 	struct path_entry *n;
74 
75 	n = malloc(sizeof(struct path_entry));
76 	if (!n)
77 		errx(1, "out of memory");
78 	n->name = strdup(dirname);
79 	if (!n->name)
80 		errx(1, "out of memory");
81 	n->next = 0;
82 	return n;
83 }
84 
85 void
86 addtoincludepath(dirname)
87 	const char *dirname;
88 {
89 	struct path_entry *n;
90 
91 	n = new_path_entry(dirname);
92 
93 	if (last) {
94 		last->next = n;
95 		last = n;
96 	}
97 	else
98 		last = first = n;
99 }
100 
101 static void
102 ensure_m4path()
103 {
104 	static int envpathdone = 0;
105 	char *envpath;
106 	char *sweep;
107 	char *path;
108 
109 	if (envpathdone)
110 		return;
111 	envpathdone = TRUE;
112 	envpath = getenv("M4PATH");
113 	if (!envpath)
114 		return;
115 	/* for portability: getenv result is read-only */
116 	envpath = strdup(envpath);
117 	if (!envpath)
118 		errx(1, "out of memory");
119 	for (sweep = envpath;
120 	    (path = strsep(&sweep, ":")) != NULL;)
121 	    addtoincludepath(path);
122 	free(envpath);
123 }
124 
125 static
126 struct input_file *
127 dopath(i, filename)
128 	struct input_file *i;
129 	const char *filename;
130 {
131 	char path[MAXPATHLEN];
132 	struct path_entry *pe;
133 	FILE *f;
134 
135 	for (pe = first; pe; pe = pe->next) {
136 		snprintf(path, sizeof(path), "%s/%s", pe->name, filename);
137 		if ((f = fopen(path, "r")) != 0) {
138 			set_input(i, f, path);
139 			return i;
140 		}
141 	}
142 	return NULL;
143 }
144 
145 struct input_file *
146 fopen_trypath(i, filename)
147 	struct input_file *i;
148 	const char *filename;
149 {
150 	FILE *f;
151 
152 	f = fopen(filename, "r");
153 	if (f != NULL) {
154 		set_input(i, f, filename);
155 		return i;
156 	}
157 	if (filename[0] == '/')
158 		return NULL;
159 
160 	ensure_m4path();
161 
162 	return dopath(i, filename);
163 }
164 
165 void
166 doindir(argv, argc)
167 	const char *argv[];
168 	int argc;
169 {
170 	ndptr p;
171 
172 	p = lookup(argv[2]);
173 	if (p == NULL)
174 		errx(1, "undefined macro %s", argv[2]);
175 	argv[1] = p->defn;
176 	eval(argv+1, argc-1, p->type);
177 }
178 
179 void
180 dobuiltin(argv, argc)
181 	const char *argv[];
182 	int argc;
183 {
184 	int n;
185 	argv[1] = NULL;
186 	n = builtin_type(argv[2]);
187 	if (n != -1)
188 		eval(argv+1, argc-1, n);
189 	else
190 		errx(1, "unknown builtin %s", argv[2]);
191 }
192 
193 
194 /* We need some temporary buffer space, as pb pushes BACK and substitution
195  * proceeds forward... */
196 static char *buffer;
197 static size_t bufsize = 0;
198 static size_t current = 0;
199 
200 static void addchars(const char *, size_t);
201 static void addchar(char);
202 static char *twiddle(const char *);
203 static char *getstring(void);
204 static void exit_regerror(int, regex_t *);
205 static void do_subst(const char *, regex_t *, const char *, regmatch_t *);
206 static void do_regexpindex(const char *, regex_t *, regmatch_t *);
207 static void do_regexp(const char *, regex_t *, const char *, regmatch_t *);
208 static void add_sub(int, const char *, regex_t *, regmatch_t *);
209 static void add_replace(const char *, regex_t *, const char *, regmatch_t *);
210 #define addconstantstring(s) addchars((s), sizeof(s)-1)
211 
212 static void
213 addchars(c, n)
214 	const char *c;
215 	size_t n;
216 {
217 	if (n == 0)
218 		return;
219 	while (current + n > bufsize) {
220 		if (bufsize == 0)
221 			bufsize = 1024;
222 		else
223 			bufsize *= 2;
224 		buffer = realloc(buffer, bufsize);
225 		if (buffer == NULL)
226 			errx(1, "out of memory");
227 	}
228 	memcpy(buffer+current, c, n);
229 	current += n;
230 }
231 
232 static void
233 addchar(c)
234 	char c;
235 {
236 	if (current +1 > bufsize) {
237 		if (bufsize == 0)
238 			bufsize = 1024;
239 		else
240 			bufsize *= 2;
241 		buffer = realloc(buffer, bufsize);
242 		if (buffer == NULL)
243 			errx(1, "out of memory");
244 	}
245 	buffer[current++] = c;
246 }
247 
248 static char *
249 getstring()
250 {
251 	addchar('\0');
252 	current = 0;
253 	return buffer;
254 }
255 
256 
257 static void
258 exit_regerror(er, re)
259 	int er;
260 	regex_t *re;
261 {
262 	size_t 	errlen;
263 	char 	*errbuf;
264 
265 	errlen = regerror(er, re, NULL, 0);
266 	errbuf = xalloc(errlen);
267 	regerror(er, re, errbuf, errlen);
268 	errx(1, "regular expression error: %s", errbuf);
269 }
270 
271 static void
272 add_sub(n, string, re, pm)
273 	int n;
274 	const char *string;
275 	regex_t *re;
276 	regmatch_t *pm;
277 {
278 	if (n > re->re_nsub)
279 		warnx("No subexpression %d", n);
280 	/* Subexpressions that did not match are
281 	 * not an error.  */
282 	else if (pm[n].rm_so != -1 &&
283 	    pm[n].rm_eo != -1) {
284 		addchars(string + pm[n].rm_so,
285 			pm[n].rm_eo - pm[n].rm_so);
286 	}
287 }
288 
289 /* Add replacement string to the output buffer, recognizing special
290  * constructs and replacing them with substrings of the original string.
291  */
292 static void
293 add_replace(string, re, replace, pm)
294 	const char *string;
295 	regex_t *re;
296 	const char *replace;
297 	regmatch_t *pm;
298 {
299 	const char *p;
300 
301 	for (p = replace; *p != '\0'; p++) {
302 		if (*p == '&' && !mimic_gnu) {
303 			add_sub(0, string, re, pm);
304 			continue;
305 		}
306 		if (*p == '\\') {
307 			if (p[1] == '\\') {
308 				addchar(p[1]);
309 				p++;
310 				continue;
311 			}
312 			if (p[1] == '&') {
313 				if (mimic_gnu)
314 					add_sub(0, string, re, pm);
315 				else
316 					addchar(p[1]);
317 				p++;
318 				continue;
319 			}
320 			if (isdigit(p[1])) {
321 				add_sub(*(++p) - '0', string, re, pm);
322 				continue;
323 			}
324 		}
325 	    	addchar(*p);
326 	}
327 }
328 
329 static void
330 do_subst(string, re, replace, pm)
331 	const char *string;
332 	regex_t *re;
333 	const char *replace;
334 	regmatch_t *pm;
335 {
336 	int error;
337 	int flags = 0;
338 	const char *last_match = NULL;
339 
340 	while ((error = regexec(re, string, re->re_nsub+1, pm, flags)) == 0) {
341 		if (pm[0].rm_eo != 0) {
342 			if (string[pm[0].rm_eo-1] == '\n')
343 				flags = 0;
344 			else
345 				flags = REG_NOTBOL;
346 		}
347 
348 		/* NULL length matches are special... We use the `vi-mode'
349 		 * rule: don't allow a NULL-match at the last match
350 		 * position.
351 		 */
352 		if (pm[0].rm_so == pm[0].rm_eo &&
353 		    string + pm[0].rm_so == last_match) {
354 			if (*string == '\0')
355 				return;
356 			addchar(*string);
357 			if (*string++ == '\n')
358 				flags = 0;
359 			else
360 				flags = REG_NOTBOL;
361 			continue;
362 		}
363 		last_match = string + pm[0].rm_so;
364 		addchars(string, pm[0].rm_so);
365 		add_replace(string, re, replace, pm);
366 		string += pm[0].rm_eo;
367 	}
368 	if (error != REG_NOMATCH)
369 		exit_regerror(error, re);
370 	pbstr(string);
371 }
372 
373 static void
374 do_regexp(string, re, replace, pm)
375 	const char *string;
376 	regex_t *re;
377 	const char *replace;
378 	regmatch_t *pm;
379 {
380 	int error;
381 
382 	switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
383 	case 0:
384 		add_replace(string, re, replace, pm);
385 		pbstr(getstring());
386 		break;
387 	case REG_NOMATCH:
388 		break;
389 	default:
390 		exit_regerror(error, re);
391 	}
392 }
393 
394 static void
395 do_regexpindex(string, re, pm)
396 	const char *string;
397 	regex_t *re;
398 	regmatch_t *pm;
399 {
400 	int error;
401 
402 	switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
403 	case 0:
404 		pbunsigned(pm[0].rm_so);
405 		break;
406 	case REG_NOMATCH:
407 		pbnum(-1);
408 		break;
409 	default:
410 		exit_regerror(error, re);
411 	}
412 }
413 
414 /* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
415  * says. So we twiddle with the regexp before passing it to regcomp.
416  */
417 static char *
418 twiddle(p)
419 	const char *p;
420 {
421 	/* This could use strcspn for speed... */
422 	while (*p != '\0') {
423 		if (*p == '\\') {
424 			switch(p[1]) {
425 			case '(':
426 			case ')':
427 			case '|':
428 				addchar(p[1]);
429 				break;
430 			case 'w':
431 				addconstantstring("[_a-zA-Z0-9]");
432 				break;
433 			case 'W':
434 				addconstantstring("[^_a-zA-Z0-9]");
435 				break;
436 			case '<':
437 				addconstantstring("[[:<:]]");
438 				break;
439 			case '>':
440 				addconstantstring("[[:>:]]");
441 				break;
442 			default:
443 				addchars(p, 2);
444 				break;
445 			}
446 			p+=2;
447 			continue;
448 		}
449 		if (*p == '(' || *p == ')' || *p == '|')
450 			addchar('\\');
451 
452 		addchar(*p);
453 		p++;
454 	}
455 	return getstring();
456 }
457 
458 /* patsubst(string, regexp, opt replacement) */
459 /* argv[2]: string
460  * argv[3]: regexp
461  * argv[4]: opt rep
462  */
463 void
464 dopatsubst(argv, argc)
465 	const char *argv[];
466 	int argc;
467 {
468 	int error;
469 	regex_t re;
470 	regmatch_t *pmatch;
471 
472 	if (argc <= 3) {
473 		warnx("Too few arguments to patsubst");
474 		return;
475 	}
476 	error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
477 	    REG_NEWLINE | REG_EXTENDED);
478 	if (error != 0)
479 		exit_regerror(error, &re);
480 
481 	pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
482 	do_subst(argv[2], &re,
483 	    argc != 4 && argv[4] != NULL ? argv[4] : "", pmatch);
484 	pbstr(getstring());
485 	free(pmatch);
486 	regfree(&re);
487 }
488 
489 void
490 doregexp(argv, argc)
491 	const char *argv[];
492 	int argc;
493 {
494 	int error;
495 	regex_t re;
496 	regmatch_t *pmatch;
497 
498 	if (argc <= 3) {
499 		warnx("Too few arguments to regexp");
500 		return;
501 	}
502 	error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
503 	    REG_EXTENDED);
504 	if (error != 0)
505 		exit_regerror(error, &re);
506 
507 	pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
508 	if (argv[4] == NULL || argc == 4)
509 		do_regexpindex(argv[2], &re, pmatch);
510 	else
511 		do_regexp(argv[2], &re, argv[4], pmatch);
512 	free(pmatch);
513 	regfree(&re);
514 }
515 
516 void
517 doesyscmd(cmd)
518 	const char *cmd;
519 {
520 	int p[2];
521 	pid_t pid, cpid;
522 	char *argv[4];
523 	int cc;
524 	int status;
525 
526 	/* Follow gnu m4 documentation: first flush buffers. */
527 	fflush(NULL);
528 
529 	argv[0] = "sh";
530 	argv[1] = "-c";
531 	argv[2] = (char *)cmd;
532 	argv[3] = NULL;
533 
534 	/* Just set up standard output, share stderr and stdin with m4 */
535 	if (pipe(p) == -1)
536 		err(1, "bad pipe");
537 	switch(cpid = fork()) {
538 	case -1:
539 		err(1, "bad fork");
540 		/* NOTREACHED */
541 	case 0:
542 		(void) close(p[0]);
543 		(void) dup2(p[1], 1);
544 		(void) close(p[1]);
545 		execv(_PATH_BSHELL, argv);
546 		exit(1);
547 	default:
548 		/* Read result in two stages, since m4's buffer is
549 		 * pushback-only. */
550 		(void) close(p[1]);
551 		do {
552 			char result[BUFSIZE];
553 			cc = read(p[0], result, sizeof result);
554 			if (cc > 0)
555 				addchars(result, cc);
556 		} while (cc > 0 || (cc == -1 && errno == EINTR));
557 
558 		(void) close(p[0]);
559 		while ((pid = wait(&status)) != cpid && pid >= 0)
560 			continue;
561 		pbstr(getstring());
562 	}
563 }
564