xref: /dragonfly/usr.bin/m4/gnum4.c (revision a32bc35d) 
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  * $FreeBSD: src/usr.bin/m4/gnum4.c,v 1.9 2004/05/18 15:53:58 stefanf Exp $
28  */
29 
30 /*
31  * functions needed to support gnu-m4 extensions, including a fake freezing
32  */
33 
34 #include <sys/param.h>
35 #include <sys/types.h>
36 #include <sys/wait.h>
37 #include <ctype.h>
38 #include <paths.h>
39 #include <regex.h>
40 #include <stddef.h>
41 #include <stdlib.h>
42 #include <stdio.h>
43 #include <string.h>
44 #include <err.h>
45 #include <errno.h>
46 #include <unistd.h>
47 #include "mdef.h"
48 #include "stdd.h"
49 #include "extern.h"
50 
51 
52 int mimic_gnu = 0;
53 
54 /*
55  * Support for include path search
56  * First search in the the current directory.
57  * If not found, and the path is not absolute, include path kicks in.
58  * First, -I options, in the order found on the command line.
59  * Then M4PATH env variable
60  */
61 
62 struct path_entry {
63 	char *name;
64 	struct path_entry *next;
65 } *first, *last;
66 
67 static struct path_entry *new_path_entry(const char *);
68 static void ensure_m4path(void);
69 static struct input_file *dopath(struct input_file *, const char *);
70 
71 static struct path_entry *
72 new_path_entry(const char *dirname)
73 {
74 	struct path_entry *n;
75 
76 	n = malloc(sizeof(struct path_entry));
77 	if (!n)
78 		errx(1, "out of memory");
79 	n->name = strdup(dirname);
80 	if (!n->name)
81 		errx(1, "out of memory");
82 	n->next = NULL;
83 	return n;
84 }
85 
86 void
87 addtoincludepath(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(void)
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(struct input_file *i, const char *filename)
128 {
129 	char path[MAXPATHLEN];
130 	struct path_entry *pe;
131 	FILE *f;
132 
133 	for (pe = first; pe; pe = pe->next) {
134 		snprintf(path, sizeof(path), "%s/%s", pe->name, filename);
135 		if ((f = fopen(path, "r")) != NULL) {
136 			set_input(i, f, path);
137 			return i;
138 		}
139 	}
140 	return NULL;
141 }
142 
143 struct input_file *
144 fopen_trypath(struct input_file *i, const char *filename)
145 {
146 	FILE *f;
147 
148 	f = fopen(filename, "r");
149 	if (f != NULL) {
150 		set_input(i, f, filename);
151 		return i;
152 	}
153 	if (filename[0] == '/')
154 		return NULL;
155 
156 	ensure_m4path();
157 
158 	return dopath(i, filename);
159 }
160 
161 void
162 doindir(const char *argv[], int argc)
163 {
164 	ndptr p;
165 
166 	p = lookup(argv[2]);
167 	if (p == NULL)
168 		errx(1, "undefined macro %s", argv[2]);
169 	argv[1] = p->defn;
170 	eval(argv+1, argc-1, p->type);
171 }
172 
173 void
174 dobuiltin(const char *argv[], int argc)
175 {
176 	int n;
177 	argv[1] = NULL;
178 	n = builtin_type(argv[2]);
179 	if (n != -1)
180 		eval(argv+1, argc-1, n);
181 	else
182 		errx(1, "unknown builtin %s", argv[2]);
183 }
184 
185 
186 /* We need some temporary buffer space, as pb pushes BACK and substitution
187  * proceeds forward... */
188 static char *buffer;
189 static size_t bufsize = 0;
190 static size_t current = 0;
191 
192 static void addchars(const char *, size_t);
193 static void addchar(int);
194 static char *twiddle(const char *);
195 static char *getstring(void);
196 static void exit_regerror(int, regex_t *);
197 static void do_subst(const char *, regex_t *, const char *, regmatch_t *);
198 static void do_regexpindex(const char *, regex_t *, regmatch_t *);
199 static void do_regexp(const char *, regex_t *, const char *, regmatch_t *);
200 static void add_sub(size_t, const char *, regex_t *, regmatch_t *);
201 static void add_replace(const char *, regex_t *, const char *, regmatch_t *);
202 #define addconstantstring(s) addchars((s), sizeof(s)-1)
203 
204 static void
205 addchars(const char *c, size_t n)
206 {
207 	if (n == 0)
208 		return;
209 	while (current + n > bufsize) {
210 		if (bufsize == 0)
211 			bufsize = 1024;
212 		else
213 			bufsize *= 2;
214 		buffer = realloc(buffer, bufsize);
215 		if (buffer == NULL)
216 			errx(1, "out of memory");
217 	}
218 	memcpy(buffer+current, c, n);
219 	current += n;
220 }
221 
222 static void
223 addchar(int c)
224 {
225 	if (current +1 > bufsize) {
226 		if (bufsize == 0)
227 			bufsize = 1024;
228 		else
229 			bufsize *= 2;
230 		buffer = realloc(buffer, bufsize);
231 		if (buffer == NULL)
232 			errx(1, "out of memory");
233 	}
234 	buffer[current++] = c;
235 }
236 
237 static char *
238 getstring(void)
239 {
240 	addchar('\0');
241 	current = 0;
242 	return buffer;
243 }
244 
245 
246 static void
247 exit_regerror(int er, regex_t *re)
248 {
249 	size_t 	errlen;
250 	char 	*errbuf;
251 
252 	errlen = regerror(er, re, NULL, 0);
253 	errbuf = xalloc(errlen);
254 	regerror(er, re, errbuf, errlen);
255 	errx(1, "regular expression error: %s", errbuf);
256 }
257 
258 static void
259 add_sub(size_t n, const char *string, regex_t *re, regmatch_t *pm)
260 {
261 	if (n > re->re_nsub)
262 		warnx("No subexpression %zu", n);
263 	/* Subexpressions that did not match are
264 	 * not an error.  */
265 	else if (pm[n].rm_so != -1 &&
266 	    pm[n].rm_eo != -1) {
267 		addchars(string + pm[n].rm_so,
268 			pm[n].rm_eo - pm[n].rm_so);
269 	}
270 }
271 
272 /* Add replacement string to the output buffer, recognizing special
273  * constructs and replacing them with substrings of the original string.
274  */
275 static void
276 add_replace(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
277 {
278 	const char *p;
279 
280 	for (p = replace; *p != '\0'; p++) {
281 		if (*p == '&' && !mimic_gnu) {
282 			add_sub(0, string, re, pm);
283 			continue;
284 		}
285 		if (*p == '\\') {
286 			if (p[1] == '\\') {
287 				addchar(p[1]);
288 				p++;
289 				continue;
290 			}
291 			if (p[1] == '&') {
292 				if (mimic_gnu)
293 					add_sub(0, string, re, pm);
294 				else
295 					addchar(p[1]);
296 				p++;
297 				continue;
298 			}
299 			if (isdigit(p[1])) {
300 				add_sub(*(++p) - '0', string, re, pm);
301 				continue;
302 			}
303 		}
304 	    	addchar(*p);
305 	}
306 }
307 
308 static void
309 do_subst(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
310 {
311 	int error;
312 	int flags = 0;
313 	const char *last_match = NULL;
314 
315 	while ((error = regexec(re, string, re->re_nsub+1, pm, flags)) == 0) {
316 		if (pm[0].rm_eo != 0) {
317 			if (string[pm[0].rm_eo-1] == '\n')
318 				flags = 0;
319 			else
320 				flags = REG_NOTBOL;
321 		}
322 
323 		/* NULL length matches are special... We use the `vi-mode'
324 		 * rule: don't allow a NULL-match at the last match
325 		 * position.
326 		 */
327 		if (pm[0].rm_so == pm[0].rm_eo &&
328 		    string + pm[0].rm_so == last_match) {
329 			if (*string == '\0')
330 				return;
331 			addchar(*string);
332 			if (*string++ == '\n')
333 				flags = 0;
334 			else
335 				flags = REG_NOTBOL;
336 			continue;
337 		}
338 		last_match = string + pm[0].rm_so;
339 		addchars(string, pm[0].rm_so);
340 		add_replace(string, re, replace, pm);
341 		string += pm[0].rm_eo;
342 	}
343 	if (error != REG_NOMATCH)
344 		exit_regerror(error, re);
345 	pbstr(string);
346 }
347 
348 static void
349 do_regexp(const char *string, regex_t *re, const char *replace, regmatch_t *pm)
350 {
351 	int error;
352 
353 	switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
354 	case 0:
355 		add_replace(string, re, replace, pm);
356 		pbstr(getstring());
357 		break;
358 	case REG_NOMATCH:
359 		break;
360 	default:
361 		exit_regerror(error, re);
362 	}
363 }
364 
365 static void
366 do_regexpindex(const char *string, regex_t *re, regmatch_t *pm)
367 {
368 	int error;
369 
370 	switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
371 	case 0:
372 		pbunsigned(pm[0].rm_so);
373 		break;
374 	case REG_NOMATCH:
375 		pbnum(-1);
376 		break;
377 	default:
378 		exit_regerror(error, re);
379 	}
380 }
381 
382 /* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
383  * says. So we twiddle with the regexp before passing it to regcomp.
384  */
385 static char *
386 twiddle(const char *p)
387 {
388 	/* This could use strcspn for speed... */
389 	while (*p != '\0') {
390 		if (*p == '\\') {
391 			switch(p[1]) {
392 			case '(':
393 			case ')':
394 			case '|':
395 				addchar(p[1]);
396 				break;
397 			case 'w':
398 				addconstantstring("[_a-zA-Z0-9]");
399 				break;
400 			case 'W':
401 				addconstantstring("[^_a-zA-Z0-9]");
402 				break;
403 			case '<':
404 				addconstantstring("[[:<:]]");
405 				break;
406 			case '>':
407 				addconstantstring("[[:>:]]");
408 				break;
409 			default:
410 				addchars(p, 2);
411 				break;
412 			}
413 			p+=2;
414 			continue;
415 		}
416 		if (*p == '(' || *p == ')' || *p == '|')
417 			addchar('\\');
418 
419 		addchar(*p);
420 		p++;
421 	}
422 	return getstring();
423 }
424 
425 /* patsubst(string, regexp, opt replacement) */
426 /* argv[2]: string
427  * argv[3]: regexp
428  * argv[4]: opt rep
429  */
430 void
431 dopatsubst(const char *argv[], int argc)
432 {
433 	int error;
434 	regex_t re;
435 	regmatch_t *pmatch;
436 
437 	if (argc <= 3) {
438 		warnx("Too few arguments to patsubst");
439 		return;
440 	}
441 	error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
442 	    REG_NEWLINE | REG_EXTENDED);
443 	if (error != 0)
444 		exit_regerror(error, &re);
445 
446 	pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
447 	do_subst(argv[2], &re,
448 	    argc != 4 && argv[4] != NULL ? argv[4] : "", pmatch);
449 	pbstr(getstring());
450 	free(pmatch);
451 	regfree(&re);
452 }
453 
454 void
455 doregexp(const char *argv[], int argc)
456 {
457 	int error;
458 	regex_t re;
459 	regmatch_t *pmatch;
460 
461 	if (argc <= 3) {
462 		warnx("Too few arguments to regexp");
463 		return;
464 	}
465 	error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
466 	    REG_EXTENDED);
467 	if (error != 0)
468 		exit_regerror(error, &re);
469 
470 	pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
471 	if (argv[4] == NULL || argc == 4)
472 		do_regexpindex(argv[2], &re, pmatch);
473 	else
474 		do_regexp(argv[2], &re, argv[4], pmatch);
475 	free(pmatch);
476 	regfree(&re);
477 }
478 
479 void
480 doesyscmd(const char *cmd)
481 {
482 	int p[2];
483 	pid_t pid, cpid;
484 	int cc;
485 	int status;
486 
487 	/* Follow gnu m4 documentation: first flush buffers. */
488 	fflush(NULL);
489 
490 	/* Just set up standard output, share stderr and stdin with m4 */
491 	if (pipe(p) == -1)
492 		err(1, "bad pipe");
493 	switch(cpid = fork()) {
494 	case -1:
495 		err(1, "bad fork");
496 		/* NOTREACHED */
497 	case 0:
498 		(void) close(p[0]);
499 		(void) dup2(p[1], 1);
500 		(void) close(p[1]);
501 		execl(_PATH_BSHELL, "sh", "-c", cmd, NULL);
502 		exit(1);
503 	default:
504 		/* Read result in two stages, since m4's buffer is
505 		 * pushback-only. */
506 		(void) close(p[1]);
507 		do {
508 			char result[BUFSIZE];
509 			cc = read(p[0], result, sizeof result);
510 			if (cc > 0)
511 				addchars(result, cc);
512 		} while (cc > 0 || (cc == -1 && errno == EINTR));
513 
514 		(void) close(p[0]);
515 		while ((pid = wait(&status)) != cpid && pid >= 0)
516 			continue;
517 		pbstr(getstring());
518 	}
519 }
520