xref: /netbsd-src/usr.bin/sed/main.c (revision da9817918ec7e88db2912a2882967c7570a83f47)
1 /*	$NetBSD: main.c,v 1.20 2009/04/13 07:29:55 lukem Exp $	*/
2 
3 /*-
4  * Copyright (c) 1992, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Diomidis Spinellis of Imperial College, University of London.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 /*-
36  * Copyright (c) 1992 Diomidis Spinellis.
37  *
38  * This code is derived from software contributed to Berkeley by
39  * Diomidis Spinellis of Imperial College, University of London.
40  *
41  * Redistribution and use in source and binary forms, with or without
42  * modification, are permitted provided that the following conditions
43  * are met:
44  * 1. Redistributions of source code must retain the above copyright
45  *    notice, this list of conditions and the following disclaimer.
46  * 2. Redistributions in binary form must reproduce the above copyright
47  *    notice, this list of conditions and the following disclaimer in the
48  *    documentation and/or other materials provided with the distribution.
49  * 3. All advertising materials mentioning features or use of this software
50  *    must display the following acknowledgement:
51  *	This product includes software developed by the University of
52  *	California, Berkeley and its contributors.
53  * 4. Neither the name of the University nor the names of its contributors
54  *    may be used to endorse or promote products derived from this software
55  *    without specific prior written permission.
56  *
57  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
58  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
59  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
60  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
61  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
62  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
63  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
64  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
65  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
66  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
67  * SUCH DAMAGE.
68  */
69 
70 #if HAVE_NBTOOL_CONFIG_H
71 #include "nbtool_config.h"
72 #endif
73 
74 #include <sys/cdefs.h>
75 #ifndef lint
76 __COPYRIGHT("@(#) Copyright (c) 1992, 1993\
77  The Regents of the University of California.  All rights reserved.");
78 #endif /* not lint */
79 
80 #ifndef lint
81 #if 0
82 static char sccsid[] = "@(#)main.c	8.2 (Berkeley) 1/3/94";
83 #else
84 __RCSID("$NetBSD: main.c,v 1.20 2009/04/13 07:29:55 lukem Exp $");
85 #endif
86 #endif /* not lint */
87 
88 #include <sys/types.h>
89 
90 #include <ctype.h>
91 #include <errno.h>
92 #include <fcntl.h>
93 #include <regex.h>
94 #include <stddef.h>
95 #include <stdio.h>
96 #include <stdlib.h>
97 #include <string.h>
98 #include <unistd.h>
99 
100 #include "defs.h"
101 #include "extern.h"
102 
103 /*
104  * Linked list of units (strings and files) to be compiled
105  */
106 struct s_compunit {
107 	struct s_compunit *next;
108 	enum e_cut {CU_FILE, CU_STRING} type;
109 	char *s;			/* Pointer to string or fname */
110 };
111 
112 /*
113  * Linked list pointer to compilation units and pointer to current
114  * next pointer.
115  */
116 static struct s_compunit *script, **cu_nextp = &script;
117 
118 /*
119  * Linked list of files to be processed
120  */
121 struct s_flist {
122 	char *fname;
123 	struct s_flist *next;
124 };
125 
126 /*
127  * Linked list pointer to files and pointer to current
128  * next pointer.
129  */
130 static struct s_flist *files, **fl_nextp = &files;
131 
132 int aflag, eflag, nflag, ere;
133 
134 /*
135  * Current file and line number; line numbers restart across compilation
136  * units, but span across input files.
137  */
138 const char *fname;			/* File name. */
139 u_long linenum;
140 int lastline;			/* TRUE on the last line of the last file */
141 
142 static void add_compunit(enum e_cut, char *);
143 static void add_file(char *);
144 int	main(int, char **);
145 
146 int
147 main(int argc, char *argv[])
148 {
149 	int c, fflag;
150 
151 	setprogname(*argv);
152 	fflag = 0;
153 	while ((c = getopt(argc, argv, "ae:f:nrE")) != -1)
154 		switch (c) {
155 		case 'a':
156 			aflag = 1;
157 			break;
158 		case 'e':
159 			eflag = 1;
160 			add_compunit(CU_STRING, optarg);
161 			break;
162 		case 'f':
163 			fflag = 1;
164 			add_compunit(CU_FILE, optarg);
165 			break;
166 		case 'n':
167 			nflag = 1;
168 			break;
169 		case 'r':
170 		case 'E':
171 			ere = REG_EXTENDED;
172 			break;
173 		default:
174 		case '?':
175 			(void)fprintf(stderr,
176 "usage:\t%s [-aEnr] script [file ...]\n\t%s [-aEnr] [-e script] ... [-f script_file] ... [file ...]\n",
177 			    getprogname(), getprogname());
178 			exit(1);
179 		}
180 	argc -= optind;
181 	argv += optind;
182 
183 	/* First usage case; script is the first arg */
184 	if (!eflag && !fflag && *argv) {
185 		add_compunit(CU_STRING, *argv);
186 		argv++;
187 	}
188 
189 	compile();
190 
191 	/* Continue with first and start second usage */
192 	if (*argv)
193 		for (; *argv; argv++)
194 			add_file(*argv);
195 	else
196 		add_file(NULL);
197 	process();
198 	cfclose(prog, NULL);
199 	if (fclose(stdout))
200 		err(FATAL, "stdout: %s", strerror(errno));
201 	exit (0);
202 }
203 
204 /*
205  * Like fgets, but go through the chain of compilation units chaining them
206  * together.  Empty strings and files are ignored.
207  */
208 char *
209 cu_fgets(char *buf, int n)
210 {
211 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
212 	static FILE *f;		/* Current open file */
213 	static char *s;		/* Current pointer inside string */
214 	static char string_ident[30];
215 	char *p;
216 
217 again:
218 	switch (state) {
219 	case ST_EOF:
220 		if (script == NULL)
221 			return (NULL);
222 		linenum = 0;
223 		switch (script->type) {
224 		case CU_FILE:
225 			if ((f = fopen(script->s, "r")) == NULL)
226 				err(FATAL,
227 				    "%s: %s", script->s, strerror(errno));
228 			fname = script->s;
229 			state = ST_FILE;
230 			goto again;
231 		case CU_STRING:
232 			if ((snprintf(string_ident,
233 			    sizeof(string_ident), "\"%s\"", script->s)) >=
234 			    (int)(sizeof(string_ident) - 1))
235 				(void)strcpy(string_ident +
236 				    sizeof(string_ident) - 6, " ...\"");
237 			fname = string_ident;
238 			s = script->s;
239 			state = ST_STRING;
240 			goto again;
241 		}
242 	case ST_FILE:
243 		if ((p = fgets(buf, n, f)) != NULL) {
244 			linenum++;
245 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
246 				nflag = 1;
247 			return (p);
248 		}
249 		script = script->next;
250 		(void)fclose(f);
251 		state = ST_EOF;
252 		goto again;
253 	case ST_STRING:
254 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
255 			nflag = 1;
256 		p = buf;
257 		for (;;) {
258 			if (n-- <= 1) {
259 				*p = '\0';
260 				linenum++;
261 				return (buf);
262 			}
263 			switch (*s) {
264 			case '\0':
265 				state = ST_EOF;
266 				if (s == script->s) {
267 					script = script->next;
268 					goto again;
269 				} else {
270 					script = script->next;
271 					*p = '\0';
272 					linenum++;
273 					return (buf);
274 				}
275 			case '\n':
276 				*p++ = '\n';
277 				*p = '\0';
278 				s++;
279 				linenum++;
280 				return (buf);
281 			default:
282 				*p++ = *s++;
283 			}
284 		}
285 	}
286 	/* NOTREACHED */
287 	return (NULL);
288 }
289 
290 /*
291  * Like fgets, but go through the list of files chaining them together.
292  * Set len to the length of the line.
293  */
294 int
295 mf_fgets(SPACE *sp, enum e_spflag spflag)
296 {
297 	static FILE *f;		/* Current open file */
298 	size_t len;
299 	char *p;
300 	int c;
301 
302 	if (f == NULL)
303 		/* Advance to first non-empty file */
304 		for (;;) {
305 			if (files == NULL) {
306 				lastline = 1;
307 				return (0);
308 			}
309 			if (files->fname == NULL) {
310 				f = stdin;
311 				fname = "stdin";
312 			} else {
313 				fname = files->fname;
314 				if ((f = fopen(fname, "r")) == NULL)
315 					err(FATAL, "%s: %s",
316 					    fname, strerror(errno));
317 			}
318 			if ((c = getc(f)) != EOF) {
319 				(void)ungetc(c, f);
320 				break;
321 			}
322 			(void)fclose(f);
323 			files = files->next;
324 		}
325 
326 	if (lastline) {
327 		sp->len = 0;
328 		return (0);
329 	}
330 
331 	/*
332 	 * Use fgetln so that we can handle essentially infinite input data.
333 	 * Can't use the pointer into the stdio buffer as the process space
334 	 * because the ungetc() can cause it to move.
335 	 */
336 	p = fgetln(f, &len);
337 	if (ferror(f))
338 		err(FATAL, "%s: %s", fname, strerror(errno ? errno : EIO));
339 	cspace(sp, p, len, spflag);
340 
341 	linenum++;
342 	/* Advance to next non-empty file */
343 	while ((c = getc(f)) == EOF) {
344 		(void)fclose(f);
345 		files = files->next;
346 		if (files == NULL) {
347 			lastline = 1;
348 			return (1);
349 		}
350 		if (files->fname == NULL) {
351 			f = stdin;
352 			fname = "stdin";
353 		} else {
354 			fname = files->fname;
355 			if ((f = fopen(fname, "r")) == NULL)
356 				err(FATAL, "%s: %s", fname, strerror(errno));
357 		}
358 	}
359 	(void)ungetc(c, f);
360 	return (1);
361 }
362 
363 /*
364  * Add a compilation unit to the linked list
365  */
366 static void
367 add_compunit(enum e_cut type, char *s)
368 {
369 	struct s_compunit *cu;
370 
371 	cu = xmalloc(sizeof(struct s_compunit));
372 	cu->type = type;
373 	cu->s = s;
374 	cu->next = NULL;
375 	*cu_nextp = cu;
376 	cu_nextp = &cu->next;
377 }
378 
379 /*
380  * Add a file to the linked list
381  */
382 static void
383 add_file(char *s)
384 {
385 	struct s_flist *fp;
386 
387 	fp = xmalloc(sizeof(struct s_flist));
388 	fp->next = NULL;
389 	*fl_nextp = fp;
390 	fp->fname = s;
391 	fl_nextp = &fp->next;
392 }
393