xref: /netbsd-src/bin/pax/ftree.c (revision 3b01aba77a7a698587faaae455bbfe740923c1f5)
1 /*	$NetBSD: ftree.c,v 1.10 2000/02/17 03:12:24 itohy Exp $	*/
2 
3 /*-
4  * Copyright (c) 1992 Keith Muller.
5  * Copyright (c) 1992, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This code is derived from software contributed to Berkeley by
9  * Keith Muller of the University of California, San Diego.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  * 3. All advertising materials mentioning features or use of this software
20  *    must display the following acknowledgement:
21  *	This product includes software developed by the University of
22  *	California, Berkeley and its contributors.
23  * 4. Neither the name of the University nor the names of its contributors
24  *    may be used to endorse or promote products derived from this software
25  *    without specific prior written permission.
26  *
27  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
28  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
31  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
32  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
33  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
34  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
35  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
36  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
37  * SUCH DAMAGE.
38  */
39 
40 #include <sys/cdefs.h>
41 #ifndef lint
42 #if 0
43 static char sccsid[] = "@(#)ftree.c	8.2 (Berkeley) 4/18/94";
44 #else
45 __RCSID("$NetBSD: ftree.c,v 1.10 2000/02/17 03:12:24 itohy Exp $");
46 #endif
47 #endif /* not lint */
48 
49 #include <sys/types.h>
50 #include <sys/time.h>
51 #include <sys/stat.h>
52 #include <sys/param.h>
53 #include <unistd.h>
54 #include <string.h>
55 #include <stdio.h>
56 #include <ctype.h>
57 #include <errno.h>
58 #include <stdlib.h>
59 #include <fts.h>
60 #include "pax.h"
61 #include "ftree.h"
62 #include "extern.h"
63 
64 /*
65  * routines to interface with the fts library function.
66  *
67  * file args supplied to pax are stored on a single linked list (of type FTREE)
68  * and given to fts to be processed one at a time. pax "selects" files from
69  * the expansion of each arg into the corresponding file tree (if the arg is a
70  * directory, otherwise the node itself is just passed to pax). The selection
71  * is modified by the -n and -u flags. The user is informed when a specific
72  * file arg does not generate any selected files. -n keeps expanding the file
73  * tree arg until one of its files is selected, then skips to the next file
74  * arg. when the user does not supply the file trees as command line args to
75  * pax, they are read from stdin
76  */
77 
78 static FTS *ftsp = NULL;		/* current FTS handle */
79 static int ftsopts;			/* options to be used on fts_open */
80 static char *farray[2];			/* array for passing each arg to fts */
81 static FTREE *fthead = NULL;		/* head of linked list of file args */
82 static FTREE *fttail = NULL;		/* tail of linked list of file args */
83 static FTREE *ftcur = NULL;		/* current file arg being processed */
84 static FTSENT *ftent = NULL;		/* current file tree entry */
85 static int ftree_skip;			/* when set skip to next file arg */
86 
87 static int ftree_arg __P((void));
88 
89 /*
90  * ftree_start()
91  *	initialize the options passed to fts_open() during this run of pax
92  *	options are based on the selection of pax options by the user
93  *	fts_start() also calls fts_arg() to open the first valid file arg. We
94  *	also attempt to reset directory access times when -t (tflag) is set.
95  * Return:
96  *	0 if there is at least one valid file arg to process, -1 otherwise
97  */
98 
99 #if __STDC__
100 int
101 ftree_start(void)
102 #else
103 int
104 ftree_start()
105 #endif
106 {
107 	/*
108 	 * set up the operation mode of fts, open the first file arg. We must
109 	 * use FTS_NOCHDIR, as the user may have to open multiple archives and
110 	 * if fts did a chdir off into the boondocks, we may create an archive
111 	 * volume in an place where the user did not expect to.
112 	 */
113 	ftsopts = FTS_NOCHDIR;
114 
115 	/*
116 	 * optional user flags that effect file traversal
117 	 * -H command line symlink follow only (half follow)
118 	 * -L follow sylinks (logical)
119 	 * -P do not follow sylinks (physical). This is the default.
120 	 * -X do not cross over mount points
121 	 * -t preserve access times on files read.
122 	 * -n select only the first member of a file tree when a match is found
123 	 * -d do not extract subtrees rooted at a directory arg.
124 	 */
125 	if (Lflag)
126 		ftsopts |= FTS_LOGICAL;
127 	else
128 		ftsopts |= FTS_PHYSICAL;
129 	if (Hflag)
130 #	ifdef NET2_FTS
131 		tty_warn(0, "The -H flag is not supported on this version");
132 #	else
133 		ftsopts |= FTS_COMFOLLOW;
134 #	endif
135 	if (Xflag)
136 		ftsopts |= FTS_XDEV;
137 
138 	if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
139 		tty_warn(1, "Unable to allocate memory for file name buffer");
140 		return(-1);
141 	}
142 
143 	if (ftree_arg() < 0)
144 		return(-1);
145 	if (tflag && (atdir_start() < 0))
146 		return(-1);
147 	return(0);
148 }
149 
150 /*
151  * ftree_add()
152  *	add the arg to the linked list of files to process. Each will be
153  *	processed by fts one at a time
154  * Return:
155  *	0 if added to the linked list, -1 if failed
156  */
157 
158 #if __STDC__
159 int
160 ftree_add(char *str, int isdir)
161 #else
162 int
163 ftree_add(str, isdir)
164 	char *str;
165 	int isdir;
166 #endif
167 {
168 	FTREE *ft;
169 	int len;
170 
171 	/*
172 	 * simple check for bad args
173 	 */
174 	if ((str == NULL) || (*str == '\0')) {
175 		tty_warn(0, "Invalid file name arguement");
176 		return(-1);
177 	}
178 
179 	/*
180 	 * allocate FTREE node and add to the end of the linked list (args are
181 	 * processed in the same order they were passed to pax). Get rid of any
182 	 * trailing / the user may pass us. (watch out for / by itself).
183 	 */
184 	if ((ft = (FTREE *)malloc(sizeof(FTREE))) == NULL) {
185 		tty_warn(0, "Unable to allocate memory for filename");
186 		return(-1);
187 	}
188 
189 	if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
190 		str[len] = '\0';
191 	ft->fname = str;
192 	ft->refcnt = -isdir;
193 	ft->fow = NULL;
194 	if (fthead == NULL) {
195 		fttail = fthead = ft;
196 		return(0);
197 	}
198 	fttail->fow = ft;
199 	fttail = ft;
200 	return(0);
201 }
202 
203 /*
204  * ftree_sel()
205  *	this entry has been selected by pax. bump up reference count and handle
206  *	-n and -d processing.
207  */
208 
209 #if __STDC__
210 void
211 ftree_sel(ARCHD *arcn)
212 #else
213 void
214 ftree_sel(arcn)
215 	ARCHD *arcn;
216 #endif
217 {
218 	/*
219 	 * set reference bit for this pattern. This linked list is only used
220 	 * when file trees are supplied pax as args. The list is not used when
221 	 * the trees are read from stdin.
222 	 */
223 	if (ftcur != NULL)
224 		ftcur->refcnt = 1;
225 
226 	/*
227 	 * if -n we are done with this arg, force a skip to the next arg when
228 	 * pax asks for the next file in next_file().
229 	 * if -d we tell fts only to match the directory (if the arg is a dir)
230 	 * and not the entire file tree rooted at that point.
231 	 */
232 	if (nflag)
233 		ftree_skip = 1;
234 
235 	if (!dflag || (arcn->type != PAX_DIR))
236 		return;
237 
238 	if (ftent != NULL)
239 		(void)fts_set(ftsp, ftent, FTS_SKIP);
240 }
241 
242 /*
243  * ftree_chk()
244  *	called at end on pax execution. Prints all those file args that did not
245  *	have a selected member (reference count still 0)
246  */
247 
248 #if __STDC__
249 void
250 ftree_chk(void)
251 #else
252 void
253 ftree_chk()
254 #endif
255 {
256 	FTREE *ft;
257 	int wban = 0;
258 
259 	/*
260 	 * make sure all dir access times were reset.
261 	 */
262 	if (tflag)
263 		atdir_end();
264 
265 	/*
266 	 * walk down list and check reference count. Print out those members
267 	 * that never had a match
268 	 */
269 	for (ft = fthead; ft != NULL; ft = ft->fow) {
270 		if (ft->refcnt != 0)
271 			continue;
272 		if (wban == 0) {
273 			tty_warn(1,
274 			    "WARNING! These file names were not selected:");
275 			++wban;
276 		}
277 		(void)fprintf(stderr, "%s\n", ft->fname);
278 	}
279 }
280 
281 /*
282  * ftree_arg()
283  *	Get the next file arg for fts to process. Can be from either the linked
284  *	list or read from stdin when the user did not them as args to pax. Each
285  *	arg is processed until the first successful fts_open().
286  * Return:
287  *	0 when the next arg is ready to go, -1 if out of file args (or EOF on
288  *	stdin).
289  */
290 
291 #if __STDC__
292 static int
293 ftree_arg(void)
294 #else
295 static int
296 ftree_arg()
297 #endif
298 {
299 	char *pt;
300 
301 	/*
302 	 * close off the current file tree
303 	 */
304 	if (ftsp != NULL) {
305 		(void)fts_close(ftsp);
306 		ftsp = NULL;
307 	}
308 
309 	/*
310 	 * keep looping until we get a valid file tree to process. Stop when we
311 	 * reach the end of the list (or get an eof on stdin)
312 	 */
313 	for(;;) {
314 		if (fthead == NULL) {
315 			/*
316 			 * the user didn't supply any args, get the file trees
317 			 * to process from stdin;
318 			 */
319 			if (fgets(farray[0], PAXPATHLEN+1, stdin) == NULL)
320 				return(-1);
321 			if ((pt = strchr(farray[0], '\n')) != NULL)
322 				*pt = '\0';
323 		} else {
324 			/*
325 			 * the user supplied the file args as arguements to pax
326 			 */
327 			if (ftcur == NULL)
328 				ftcur = fthead;
329 			else if ((ftcur = ftcur->fow) == NULL)
330 				return(-1);
331 
332 			if (ftcur->refcnt < 0) {
333 				/*
334 				 * chdir entry.
335 				 * Change directory and retry loop.
336 				 */
337 				if (ar_dochdir(ftcur->fname))
338 					return (-1);
339 				continue;
340 			}
341 			farray[0] = ftcur->fname;
342 		}
343 
344 		/*
345 		 * watch it, fts wants the file arg stored in a array of char
346 		 * ptrs, with the last one a null. we use a two element array
347 		 * and set farray[0] to point at the buffer with the file name
348 		 * in it. We cannot pass all the file args to fts at one shot
349 		 * as we need to keep a handle on which file arg generates what
350 		 * files (the -n and -d flags need this). If the open is
351 		 * successful, return a 0.
352 		 */
353 		if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
354 			break;
355 	}
356 	return(0);
357 }
358 
359 /*
360  * next_file()
361  *	supplies the next file to process in the supplied archd structure.
362  * Return:
363  *	0 when contents of arcn have been set with the next file, -1 when done.
364  */
365 
366 #if __STDC__
367 int
368 next_file(ARCHD *arcn)
369 #else
370 int
371 next_file(arcn)
372 	ARCHD *arcn;
373 #endif
374 {
375 	int cnt;
376 	time_t atime;
377 	time_t mtime;
378 
379 	/*
380 	 * ftree_sel() might have set the ftree_skip flag if the user has the
381 	 * -n option and a file was selected from this file arg tree. (-n says
382 	 * only one member is matched for each pattern) ftree_skip being 1
383 	 * forces us to go to the next arg now.
384 	 */
385 	if (ftree_skip) {
386 		/*
387 		 * clear and go to next arg
388 		 */
389 		ftree_skip = 0;
390 		if (ftree_arg() < 0)
391 			return(-1);
392 	}
393 
394 	/*
395 	 * loop until we get a valid file to process
396 	 */
397 	for(;;) {
398 		if ((ftent = fts_read(ftsp)) == NULL) {
399 			/*
400 			 * out of files in this tree, go to next arg, if none
401 			 * we are done
402 			 */
403 			if (ftree_arg() < 0)
404 				return(-1);
405 			continue;
406 		}
407 
408 		/*
409 		 * handle each type of fts_read() flag
410 		 */
411 		switch(ftent->fts_info) {
412 		case FTS_D:
413 			/*
414 			 * cpio does *not* decend directories listed in the
415 			 * arguments, unlike pax/tar, so needs special handling
416 			 * here.  failure to do so results in massive amounts
417 			 * of duplicated files in the output.
418 			 */
419 			if (cpio_mode)
420 				continue;
421 			/* FALLTHROUGH */
422 		case FTS_DEFAULT:
423 		case FTS_F:
424 		case FTS_SL:
425 		case FTS_SLNONE:
426 			/*
427 			 * these are all ok
428 			 */
429 			break;
430 		case FTS_DP:
431 			/*
432 			 * already saw this directory. If the user wants file
433 			 * access times reset, we use this to restore the
434 			 * access time for this directory since this is the
435 			 * last time we will see it in this file subtree
436 			 * remember to force the time (this is -t on a read
437 			 * directory, not a created directory).
438 			 */
439 #			ifdef NET2_FTS
440 			if (!tflag || (get_atdir(ftent->fts_statb.st_dev,
441 			    ftent->fts_statb.st_ino, &mtime, &atime) < 0))
442 #			else
443 			if (!tflag || (get_atdir(ftent->fts_statp->st_dev,
444 			    ftent->fts_statp->st_ino, &mtime, &atime) < 0))
445 #			endif
446 				continue;
447 			set_ftime(ftent->fts_path, mtime, atime, 1);
448 			continue;
449 		case FTS_DC:
450 			/*
451 			 * fts claims a file system cycle
452 			 */
453 			tty_warn(1,"File system cycle found at %s",
454 			    ftent->fts_path);
455 			continue;
456 		case FTS_DNR:
457 #			ifdef NET2_FTS
458 			syswarn(1, errno,
459 #			else
460 			syswarn(1, ftent->fts_errno,
461 #			endif
462 			    "Unable to read directory %s", ftent->fts_path);
463 			continue;
464 		case FTS_ERR:
465 #			ifdef NET2_FTS
466 			syswarn(1, errno,
467 #			else
468 			syswarn(1, ftent->fts_errno,
469 #			endif
470 			    "File system traversal error");
471 			continue;
472 		case FTS_NS:
473 		case FTS_NSOK:
474 #			ifdef NET2_FTS
475 			syswarn(1, errno,
476 #			else
477 			syswarn(1, ftent->fts_errno,
478 #			endif
479 			    "Unable to access %s", ftent->fts_path);
480 			continue;
481 		}
482 
483 		/*
484 		 * ok got a file tree node to process. copy info into arcn
485 		 * structure (initialize as required)
486 		 */
487 		arcn->skip = 0;
488 		arcn->pad = 0;
489 		arcn->ln_nlen = 0;
490 		arcn->ln_name[0] = '\0';
491 #		ifdef NET2_FTS
492 		arcn->sb = ftent->fts_statb;
493 #		else
494 		arcn->sb = *(ftent->fts_statp);
495 #		endif
496 
497 		/*
498 		 * file type based set up and copy into the arcn struct
499 		 * SIDE NOTE:
500 		 * we try to reset the access time on all files and directories
501 		 * we may read when the -t flag is specified. files are reset
502 		 * when we close them after copying. we reset the directories
503 		 * when we are done with their file tree (we also clean up at
504 		 * end in case we cut short a file tree traversal). However
505 		 * there is no way to reset access times on symlinks.
506 		 */
507 		switch(S_IFMT & arcn->sb.st_mode) {
508 		case S_IFDIR:
509 			arcn->type = PAX_DIR;
510 			if (!tflag)
511 				break;
512 			add_atdir(ftent->fts_path, arcn->sb.st_dev,
513 			    arcn->sb.st_ino, arcn->sb.st_mtime,
514 			    arcn->sb.st_atime);
515 			break;
516 		case S_IFCHR:
517 			arcn->type = PAX_CHR;
518 			break;
519 		case S_IFBLK:
520 			arcn->type = PAX_BLK;
521 			break;
522 		case S_IFREG:
523 			/*
524 			 * only regular files with have data to store on the
525 			 * archive. all others will store a zero length skip.
526 			 * the skip field is used by pax for actual data it has
527 			 * to read (or skip over).
528 			 */
529 			arcn->type = PAX_REG;
530 			arcn->skip = arcn->sb.st_size;
531 			break;
532 		case S_IFLNK:
533 			arcn->type = PAX_SLK;
534 			/*
535 			 * have to read the symlink path from the file
536 			 */
537 			if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
538 			    PAXPATHLEN)) < 0) {
539 				syswarn(1, errno, "Unable to read symlink %s",
540 				    ftent->fts_path);
541 				continue;
542 			}
543 			/*
544 			 * set link name length, watch out readlink does not
545 			 * allways null terminate the link path
546 			 */
547 			arcn->ln_name[cnt] = '\0';
548 			arcn->ln_nlen = cnt;
549 			break;
550 		case S_IFSOCK:
551 			/*
552 			 * under BSD storing a socket is senseless but we will
553 			 * let the format specific write function make the
554 			 * decision of what to do with it.
555 			 */
556 			arcn->type = PAX_SCK;
557 			break;
558 		case S_IFIFO:
559 			arcn->type = PAX_FIF;
560 			break;
561 		}
562 		break;
563 	}
564 
565 	/*
566 	 * copy file name, set file name length
567 	 */
568 	arcn->nlen = l_strncpy(arcn->name, ftent->fts_path, PAXPATHLEN+1);
569 	arcn->name[arcn->nlen] = '\0';
570 	arcn->org_name = ftent->fts_path;
571 	return(0);
572 }
573