xref: /freebsd/usr.bin/m4/gnum4.c (revision ea906c4152774dff300bb26fbfc1e4188351c89a) 
1 /* $OpenBSD: gnum4.c,v 1.18 2002/04/26 16:15:16 espie 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 #include <sys/cdefs.h>
29 __FBSDID("$FreeBSD$");
30 
31 /*
32  * functions needed to support gnu-m4 extensions, including a fake freezing
33  */
34 
35 #include <sys/param.h>
36 #include <sys/types.h>
37 #include <sys/wait.h>
38 #include <ctype.h>
39 #include <paths.h>
40 #include <regex.h>
41 #include <stddef.h>
42 #include <stdlib.h>
43 #include <stdio.h>
44 #include <string.h>
45 #include <err.h>
46 #include <errno.h>
47 #include <unistd.h>
48 #include "mdef.h"
49 #include "stdd.h"
50 #include "extern.h"
51 
52 
53 int mimic_gnu = 0;
54 
55 /*
56  * Support for include path search
57  * First search in the the current directory.
58  * If not found, and the path is not absolute, include path kicks in.
59  * First, -I options, in the order found on the command line.
60  * Then M4PATH env variable
61  */
62 
63 struct path_entry {
64 	char *name;
65 	struct path_entry *next;
66 } *first, *last;
67 
68 static struct path_entry *new_path_entry(const char *);
69 static void ensure_m4path(void);
70 static struct input_file *dopath(struct input_file *, const char *);
71 
72 static struct path_entry *
73 new_path_entry(const char *dirname)
74 {
75 	struct path_entry *n;
76 
77 	n = malloc(sizeof(struct path_entry));
78 	if (!n)
79 		errx(1, "out of memory");
80 	n->name = strdup(dirname);
81 	if (!n->name)
82 		errx(1, "out of memory");
83 	n->next = 0;
84 	return n;
85 }
86 
87 void
88 addtoincludepath(const char *dirname)
89 {
90 	struct path_entry *n;
91 
92 	n = new_path_entry(dirname);
93 
94 	if (last) {
95 		last->next = n;
96 		last = n;
97 	}
98 	else
99 		last = first = n;
100 }
101 
102 static void
103 ensure_m4path(void)
104 {
105 	static int envpathdone = 0;
106 	char *envpath;
107 	char *sweep;
108 	char *path;
109 
110 	if (envpathdone)
111 		return;
112 	envpathdone = TRUE;
113 	envpath = getenv("M4PATH");
114 	if (!envpath)
115 		return;
116 	/* for portability: getenv result is read-only */
117 	envpath = strdup(envpath);
118 	if (!envpath)
119 		errx(1, "out of memory");
120 	for (sweep = envpath;
121 	    (path = strsep(&sweep, ":")) != NULL;)
122 	    addtoincludepath(path);
123 	free(envpath);
124 }
125 
126 static
127 struct input_file *
128 dopath(struct input_file *i, const char *filename)
129 {
130 	char path[MAXPATHLEN];
131 	struct path_entry *pe;
132 	FILE *f;
133 
134 	for (pe = first; pe; pe = pe->next) {
135 		snprintf(path, sizeof(path), "%s/%s", pe->name, filename);
136 		if ((f = fopen(path, "r")) != NULL) {
137 			set_input(i, f, path);
138 			return i;
139 		}
140 	}
141 	return NULL;
142 }
143 
144 struct input_file *
145 fopen_trypath(struct input_file *i, const char *filename)
146 {
147 	FILE *f;
148 
149 	f = fopen(filename, "r");
150 	if (f != NULL) {
151 		set_input(i, f, filename);
152 		return i;
153 	}
154 	if (filename[0] == '/')
155 		return NULL;
156 
157 	ensure_m4path();
158 
159 	return dopath(i, filename);
160 }
161 
162 void
163 doindir(const char *argv[], int argc)
164 {
165 	ndptr p;
166 
167 	p = lookup(argv[2]);
168 	if (p == NULL)
169 		errx(1, "undefined macro %s", argv[2]);
170 	argv[1] = p->defn;
171 	eval(argv+1, argc-1, p->type);
172 }
173 
174 void
175 dobuiltin(const char *argv[], int argc)
176 {
177 	int n;
178 	argv[1] = NULL;
179 	n = builtin_type(argv[2]);
180 	if (n != -1)
181 		eval(argv+1, argc-1, n);
182 	else
183 		errx(1, "unknown builtin %s", argv[2]);
184 }
185 
186 
187 /* We need some temporary buffer space, as pb pushes BACK and substitution
188  * proceeds forward... */
189 static char *buffer;
190 static size_t bufsize = 0;
191 static size_t current = 0;
192 
193 static void addchars(const char *, size_t);
194 static void addchar(int);
195 static char *twiddle(const char *);
196 static char *getstring(void);
197 static void exit_regerror(int, regex_t *);
198 static void do_subst(const char *, regex_t *, const char *, regmatch_t *);
199 static void do_regexpindex(const char *, regex_t *, regmatch_t *);
200 static void do_regexp(const char *, regex_t *, const char *, regmatch_t *);
201 static void add_sub(size_t, const char *, regex_t *, regmatch_t *);
202 static void add_replace(const char *, regex_t *, const char *, regmatch_t *);
203 #define addconstantstring(s) addchars((s), sizeof(s)-1)
204 
205 static void
206 addchars(const char *c, size_t n)
207 {
208 	if (n == 0)
209 		return;
210 	while (current + n > bufsize) {
211 		if (bufsize == 0)
212 			bufsize = 1024;
213 		else
214 			bufsize *= 2;
215 		buffer = realloc(buffer, bufsize);
216 		if (buffer == NULL)
217 			errx(1, "out of memory");
218 	}
219 	memcpy(buffer+current, c, n);
220 	current += n;
221 }
222 
223 static void
224 addchar(int c)
225 {
226 	if (current +1 > bufsize) {
227 		if (bufsize == 0)
228 			bufsize = 1024;
229 		else
230 			bufsize *= 2;
231 		buffer = realloc(buffer, bufsize);
232 		if (buffer == NULL)
233 			errx(1, "out of memory");
234 	}
235 	buffer[current++] = c;
236 }
237 
238 static char *
239 getstring(void)
240 {
241 	addchar('\0');
242 	current = 0;
243 	return buffer;
244 }
245 
246 
247 static void
248 exit_regerror(int er, regex_t *re)
249 {
250 	size_t 	errlen;
251 	char 	*errbuf;
252 
253 	errlen = regerror(er, re, NULL, 0);
254 	errbuf = xalloc(errlen);
255 	regerror(er, re, errbuf, errlen);
256 	errx(1, "regular expression error: %s", errbuf);
257 }
258 
259 static void
260 add_sub(size_t n, const char *string, regex_t *re, regmatch_t *pm)
261 {
262 	if (n > re->re_nsub)
263 		warnx("No subexpression %zu", n);
264 	/* Subexpressions that did not match are
265 	 * not an error.  */
266 	else if (pm[n].rm_so != -1 &&
267 	    pm[n].rm_eo != -1) {
268 		addchars(string + pm[n].rm_so,
269 			pm[n].rm_eo - pm[n].rm_so);
270 	}
271 }
272 
273 /* Add replacement string to the output buffer, recognizing special
274  * constructs and replacing them with substrings of the original string.
275  */
276 static void
277 add_replace(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
278 {
279 	const char *p;
280 
281 	for (p = replace; *p != '\0'; p++) {
282 		if (*p == '&' && !mimic_gnu) {
283 			add_sub(0, string, re, pm);
284 			continue;
285 		}
286 		if (*p == '\\') {
287 			if (p[1] == '\\') {
288 				addchar(p[1]);
289 				p++;
290 				continue;
291 			}
292 			if (p[1] == '&') {
293 				if (mimic_gnu)
294 					add_sub(0, string, re, pm);
295 				else
296 					addchar(p[1]);
297 				p++;
298 				continue;
299 			}
300 			if (isdigit(p[1])) {
301 				add_sub(*(++p) - '0', string, re, pm);
302 				continue;
303 			}
304 		}
305 	    	addchar(*p);
306 	}
307 }
308 
309 static void
310 do_subst(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
311 {
312 	int error;
313 	int flags = 0;
314 	const char *last_match = NULL;
315 
316 	while ((error = regexec(re, string, re->re_nsub+1, pm, flags)) == 0) {
317 		if (pm[0].rm_eo != 0) {
318 			if (string[pm[0].rm_eo-1] == '\n')
319 				flags = 0;
320 			else
321 				flags = REG_NOTBOL;
322 		}
323 
324 		/* NULL length matches are special... We use the `vi-mode'
325 		 * rule: don't allow a NULL-match at the last match
326 		 * position.
327 		 */
328 		if (pm[0].rm_so == pm[0].rm_eo &&
329 		    string + pm[0].rm_so == last_match) {
330 			if (*string == '\0')
331 				return;
332 			addchar(*string);
333 			if (*string++ == '\n')
334 				flags = 0;
335 			else
336 				flags = REG_NOTBOL;
337 			continue;
338 		}
339 		last_match = string + pm[0].rm_so;
340 		addchars(string, pm[0].rm_so);
341 		add_replace(string, re, replace, pm);
342 		string += pm[0].rm_eo;
343 	}
344 	if (error != REG_NOMATCH)
345 		exit_regerror(error, re);
346 	pbstr(string);
347 }
348 
349 static void
350 do_regexp(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
351 {
352 	int error;
353 
354 	switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
355 	case 0:
356 		add_replace(string, re, replace, pm);
357 		pbstr(getstring());
358 		break;
359 	case REG_NOMATCH:
360 		break;
361 	default:
362 		exit_regerror(error, re);
363 	}
364 }
365 
366 static void
367 do_regexpindex(const char *string, regex_t *re, regmatch_t *pm)
368 {
369 	int error;
370 
371 	switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
372 	case 0:
373 		pbunsigned(pm[0].rm_so);
374 		break;
375 	case REG_NOMATCH:
376 		pbnum(-1);
377 		break;
378 	default:
379 		exit_regerror(error, re);
380 	}
381 }
382 
383 /* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
384  * says. So we twiddle with the regexp before passing it to regcomp.
385  */
386 static char *
387 twiddle(const char *p)
388 {
389 	/* This could use strcspn for speed... */
390 	while (*p != '\0') {
391 		if (*p == '\\') {
392 			switch(p[1]) {
393 			case '(':
394 			case ')':
395 			case '|':
396 				addchar(p[1]);
397 				break;
398 			case 'w':
399 				addconstantstring("[_a-zA-Z0-9]");
400 				break;
401 			case 'W':
402 				addconstantstring("[^_a-zA-Z0-9]");
403 				break;
404 			case '<':
405 				addconstantstring("[[:<:]]");
406 				break;
407 			case '>':
408 				addconstantstring("[[:>:]]");
409 				break;
410 			default:
411 				addchars(p, 2);
412 				break;
413 			}
414 			p+=2;
415 			continue;
416 		}
417 		if (*p == '(' || *p == ')' || *p == '|')
418 			addchar('\\');
419 
420 		addchar(*p);
421 		p++;
422 	}
423 	return getstring();
424 }
425 
426 /* patsubst(string, regexp, opt replacement) */
427 /* argv[2]: string
428  * argv[3]: regexp
429  * argv[4]: opt rep
430  */
431 void
432 dopatsubst(const char *argv[], int argc)
433 {
434 	int error;
435 	regex_t re;
436 	regmatch_t *pmatch;
437 
438 	if (argc <= 3) {
439 		warnx("Too few arguments to patsubst");
440 		return;
441 	}
442 	error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
443 	    REG_NEWLINE | REG_EXTENDED);
444 	if (error != 0)
445 		exit_regerror(error, &re);
446 
447 	pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
448 	do_subst(argv[2], &re,
449 	    argc != 4 && argv[4] != NULL ? argv[4] : "", pmatch);
450 	pbstr(getstring());
451 	free(pmatch);
452 	regfree(&re);
453 }
454 
455 void
456 doregexp(const char *argv[], int argc)
457 {
458 	int error;
459 	regex_t re;
460 	regmatch_t *pmatch;
461 
462 	if (argc <= 3) {
463 		warnx("Too few arguments to regexp");
464 		return;
465 	}
466 	error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
467 	    REG_EXTENDED);
468 	if (error != 0)
469 		exit_regerror(error, &re);
470 
471 	pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
472 	if (argv[4] == NULL || argc == 4)
473 		do_regexpindex(argv[2], &re, pmatch);
474 	else
475 		do_regexp(argv[2], &re, argv[4], pmatch);
476 	free(pmatch);
477 	regfree(&re);
478 }
479 
480 void
481 doesyscmd(const char *cmd)
482 {
483 	int p[2];
484 	pid_t pid, cpid;
485 	int cc;
486 	int status;
487 
488 	/* Follow gnu m4 documentation: first flush buffers. */
489 	fflush(NULL);
490 
491 	/* Just set up standard output, share stderr and stdin with m4 */
492 	if (pipe(p) == -1)
493 		err(1, "bad pipe");
494 	switch(cpid = fork()) {
495 	case -1:
496 		err(1, "bad fork");
497 		/* NOTREACHED */
498 	case 0:
499 		(void) close(p[0]);
500 		(void) dup2(p[1], 1);
501 		(void) close(p[1]);
502 		execl(_PATH_BSHELL, "sh", "-c", cmd, (char *)NULL);
503 		exit(1);
504 	default:
505 		/* Read result in two stages, since m4's buffer is
506 		 * pushback-only. */
507 		(void) close(p[1]);
508 		do {
509 			char result[BUFSIZE];
510 			cc = read(p[0], result, sizeof result);
511 			if (cc > 0)
512 				addchars(result, cc);
513 		} while (cc > 0 || (cc == -1 && errno == EINTR));
514 
515 		(void) close(p[0]);
516 		while ((pid = wait(&status)) != cpid && pid >= 0)
517 			continue;
518 		pbstr(getstring());
519 	}
520 }
521