xref: /freebsd/bin/pax/ftree.c (revision 380a989b3223d455375b4fae70fd0b9bdd43bafb)
1 /*-
2  * Copyright (c) 1992 Keith Muller.
3  * Copyright (c) 1992, 1993
4  *	The Regents of the University of California.  All rights reserved.
5  *
6  * This code is derived from software contributed to Berkeley by
7  * Keith Muller of the University of California, San Diego.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. All advertising materials mentioning features or use of this software
18  *    must display the following acknowledgement:
19  *	This product includes software developed by the University of
20  *	California, Berkeley and its contributors.
21  * 4. Neither the name of the University nor the names of its contributors
22  *    may be used to endorse or promote products derived from this software
23  *    without specific prior written permission.
24  *
25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35  * SUCH DAMAGE.
36  */
37 
38 #ifndef lint
39 #if 0
40 static char sccsid[] = "@(#)ftree.c	8.2 (Berkeley) 4/18/94";
41 #endif
42 static const char rcsid[] =
43 	"$Id: ftree.c,v 1.10 1998/05/15 06:27:42 charnier Exp $";
44 #endif /* not lint */
45 
46 #include <sys/types.h>
47 #include <sys/time.h>
48 #include <sys/stat.h>
49 #include <unistd.h>
50 #include <string.h>
51 #include <stdio.h>
52 #include <errno.h>
53 #include <stdlib.h>
54 #include <fts.h>
55 #include "pax.h"
56 #include "ftree.h"
57 #include "extern.h"
58 
59 /*
60  * routines to interface with the fts library function.
61  *
62  * file args supplied to pax are stored on a single linked list (of type FTREE)
63  * and given to fts to be processed one at a time. pax "selects" files from
64  * the expansion of each arg into the corresponding file tree (if the arg is a
65  * directory, otherwise the node itself is just passed to pax). The selection
66  * is modified by the -n and -u flags. The user is informed when a specific
67  * file arg does not generate any selected files. -n keeps expanding the file
68  * tree arg until one of its files is selected, then skips to the next file
69  * arg. when the user does not supply the file trees as command line args to
70  * pax, they are read from stdin
71  */
72 
73 static FTS *ftsp = NULL;		/* curent FTS handle */
74 static int ftsopts;			/* options to be used on fts_open */
75 static char *farray[2];			/* array for passing each arg to fts */
76 static FTREE *fthead = NULL;		/* head of linked list of file args */
77 static FTREE *fttail = NULL;		/* tail of linked list of file args */
78 static FTREE *ftcur = NULL;		/* current file arg being processed */
79 static FTSENT *ftent = NULL;		/* current file tree entry */
80 static int ftree_skip;			/* when set skip to next file arg */
81 
82 static int ftree_arg __P((void));
83 
84 /*
85  * ftree_start()
86  *	initialize the options passed to fts_open() during this run of pax
87  *	options are based on the selection of pax options by the user
88  *	fts_start() also calls fts_arg() to open the first valid file arg. We
89  *	also attempt to reset directory access times when -t (tflag) is set.
90  * Return:
91  *	0 if there is at least one valid file arg to process, -1 otherwise
92  */
93 
94 #if __STDC__
95 int
96 ftree_start(void)
97 #else
98 int
99 ftree_start()
100 #endif
101 {
102 	/*
103 	 * set up the operation mode of fts, open the first file arg. We must
104 	 * use FTS_NOCHDIR, as the user may have to open multiple archives and
105 	 * if fts did a chdir off into the boondocks, we may create an archive
106 	 * volume in an place where the user did not expect to.
107 	 */
108 	ftsopts = FTS_NOCHDIR;
109 
110 	/*
111 	 * optional user flags that effect file traversal
112 	 * -H command line symlink follow only (half follow)
113 	 * -L follow sylinks (logical)
114 	 * -P do not follow sylinks (physical). This is the default.
115 	 * -X do not cross over mount points
116 	 * -t preserve access times on files read.
117 	 * -n select only the first member of a file tree when a match is found
118 	 * -d do not extract subtrees rooted at a directory arg.
119 	 */
120 	if (Lflag)
121 		ftsopts |= FTS_LOGICAL;
122 	else
123 		ftsopts |= FTS_PHYSICAL;
124 	if (Hflag)
125 #	ifdef NET2_FTS
126 		pax_warn(0, "The -H flag is not supported on this version");
127 #	else
128 		ftsopts |= FTS_COMFOLLOW;
129 #	endif
130 	if (Xflag)
131 		ftsopts |= FTS_XDEV;
132 
133 	if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
134 		pax_warn(1, "Unable to allocate memory for file name buffer");
135 		return(-1);
136 	}
137 
138 	if (ftree_arg() < 0)
139 		return(-1);
140 	if (tflag && (atdir_start() < 0))
141 		return(-1);
142 	return(0);
143 }
144 
145 /*
146  * ftree_add()
147  *	add the arg to the linked list of files to process. Each will be
148  *	processed by fts one at a time
149  * Return:
150  *	0 if added to the linked list, -1 if failed
151  */
152 
153 #if __STDC__
154 int
155 ftree_add(register char *str)
156 #else
157 int
158 ftree_add(str)
159 	register char *str;
160 #endif
161 {
162 	register FTREE *ft;
163 	register int len;
164 
165 	/*
166 	 * simple check for bad args
167 	 */
168 	if ((str == NULL) || (*str == '\0')) {
169 		pax_warn(0, "Invalid file name arguement");
170 		return(-1);
171 	}
172 
173 	/*
174 	 * allocate FTREE node and add to the end of the linked list (args are
175 	 * processed in the same order they were passed to pax). Get rid of any
176 	 * trailing / the user may pass us. (watch out for / by itself).
177 	 */
178 	if ((ft = (FTREE *)malloc(sizeof(FTREE))) == NULL) {
179 		pax_warn(0, "Unable to allocate memory for filename");
180 		return(-1);
181 	}
182 
183 	if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
184 		str[len] = '\0';
185 	ft->fname = str;
186 	ft->refcnt = 0;
187 	ft->fow = NULL;
188 	if (fthead == NULL) {
189 		fttail = fthead = ft;
190 		return(0);
191 	}
192 	fttail->fow = ft;
193 	fttail = ft;
194 	return(0);
195 }
196 
197 /*
198  * ftree_sel()
199  *	this entry has been selected by pax. bump up reference count and handle
200  *	-n and -d processing.
201  */
202 
203 #if __STDC__
204 void
205 ftree_sel(register ARCHD *arcn)
206 #else
207 void
208 ftree_sel(arcn)
209 	register ARCHD *arcn;
210 #endif
211 {
212 	/*
213 	 * set reference bit for this pattern. This linked list is only used
214 	 * when file trees are supplied pax as args. The list is not used when
215 	 * the trees are read from stdin.
216 	 */
217 	if (ftcur != NULL)
218 		ftcur->refcnt = 1;
219 
220 	/*
221 	 * if -n we are done with this arg, force a skip to the next arg when
222 	 * pax asks for the next file in next_file().
223 	 * if -d we tell fts only to match the directory (if the arg is a dir)
224 	 * and not the entire file tree rooted at that point.
225 	 */
226 	if (nflag)
227 		ftree_skip = 1;
228 
229 	if (!dflag || (arcn->type != PAX_DIR))
230 		return;
231 
232 	if (ftent != NULL)
233 		(void)fts_set(ftsp, ftent, FTS_SKIP);
234 }
235 
236 /*
237  * ftree_chk()
238  *	called at end on pax execution. Prints all those file args that did not
239  *	have a selected member (reference count still 0)
240  */
241 
242 #if __STDC__
243 void
244 ftree_chk(void)
245 #else
246 void
247 ftree_chk()
248 #endif
249 {
250 	register FTREE *ft;
251 	register int wban = 0;
252 
253 	/*
254 	 * make sure all dir access times were reset.
255 	 */
256 	if (tflag)
257 		atdir_end();
258 
259 	/*
260 	 * walk down list and check reference count. Print out those members
261 	 * that never had a match
262 	 */
263 	for (ft = fthead; ft != NULL; ft = ft->fow) {
264 		if (ft->refcnt > 0)
265 			continue;
266 		if (wban == 0) {
267 			pax_warn(1,"WARNING! These file names were not selected:");
268 			++wban;
269 		}
270 		(void)fprintf(stderr, "%s\n", ft->fname);
271 	}
272 }
273 
274 /*
275  * ftree_arg()
276  *	Get the next file arg for fts to process. Can be from either the linked
277  *	list or read from stdin when the user did not them as args to pax. Each
278  *	arg is processed until the first successful fts_open().
279  * Return:
280  *	0 when the next arg is ready to go, -1 if out of file args (or EOF on
281  *	stdin).
282  */
283 
284 #if __STDC__
285 static int
286 ftree_arg(void)
287 #else
288 static int
289 ftree_arg()
290 #endif
291 {
292 	register char *pt;
293 
294 	/*
295 	 * close off the current file tree
296 	 */
297 	if (ftsp != NULL) {
298 		(void)fts_close(ftsp);
299 		ftsp = NULL;
300 	}
301 
302 	/*
303 	 * keep looping until we get a valid file tree to process. Stop when we
304 	 * reach the end of the list (or get an eof on stdin)
305 	 */
306 	for(;;) {
307 		if (fthead == NULL) {
308 			/*
309 			 * the user didn't supply any args, get the file trees
310 			 * to process from stdin;
311 			 */
312 			if (fgets(farray[0], PAXPATHLEN+1, stdin) == NULL)
313 				return(-1);
314 			if ((pt = strchr(farray[0], '\n')) != NULL)
315 				*pt = '\0';
316 		} else {
317 			/*
318 			 * the user supplied the file args as arguements to pax
319 			 */
320 			if (ftcur == NULL)
321 				ftcur = fthead;
322 			else if ((ftcur = ftcur->fow) == NULL)
323 				return(-1);
324 			farray[0] = ftcur->fname;
325 		}
326 
327 		/*
328 		 * watch it, fts wants the file arg stored in a array of char
329 		 * ptrs, with the last one a null. we use a two element array
330 		 * and set farray[0] to point at the buffer with the file name
331 		 * in it. We cannnot pass all the file args to fts at one shot
332 		 * as we need to keep a handle on which file arg generates what
333 		 * files (the -n and -d flags need this). If the open is
334 		 * successful, return a 0.
335 		 */
336 		if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
337 			break;
338 	}
339 	return(0);
340 }
341 
342 /*
343  * next_file()
344  *	supplies the next file to process in the supplied archd structure.
345  * Return:
346  *	0 when contents of arcn have been set with the next file, -1 when done.
347  */
348 
349 #if __STDC__
350 int
351 next_file(register ARCHD *arcn)
352 #else
353 int
354 next_file(arcn)
355 	register ARCHD *arcn;
356 #endif
357 {
358 	register int cnt;
359 	time_t atime;
360 	time_t mtime;
361 
362 	/*
363 	 * ftree_sel() might have set the ftree_skip flag if the user has the
364 	 * -n option and a file was selected from this file arg tree. (-n says
365 	 * only one member is matched for each pattern) ftree_skip being 1
366 	 * forces us to go to the next arg now.
367 	 */
368 	if (ftree_skip) {
369 		/*
370 		 * clear and go to next arg
371 		 */
372 		ftree_skip = 0;
373 		if (ftree_arg() < 0)
374 			return(-1);
375 	}
376 
377 	/*
378 	 * loop until we get a valid file to process
379 	 */
380 	for(;;) {
381 		if ((ftent = fts_read(ftsp)) == NULL) {
382 			/*
383 			 * out of files in this tree, go to next arg, if none
384 			 * we are done
385 			 */
386 			if (ftree_arg() < 0)
387 				return(-1);
388 			continue;
389 		}
390 
391 		/*
392 		 * handle each type of fts_read() flag
393 		 */
394 		switch(ftent->fts_info) {
395 		case FTS_D:
396 		case FTS_DEFAULT:
397 		case FTS_F:
398 		case FTS_SL:
399 		case FTS_SLNONE:
400 			/*
401 			 * these are all ok
402 			 */
403 			break;
404 		case FTS_DP:
405 			/*
406 			 * already saw this directory. If the user wants file
407 			 * access times reset, we use this to restore the
408 			 * access time for this directory since this is the
409 			 * last time we will see it in this file subtree
410 			 * remember to force the time (this is -t on a read
411 			 * directory, not a created directory).
412 			 */
413 #			ifdef NET2_FTS
414 			if (!tflag || (get_atdir(ftent->fts_statb.st_dev,
415 			    ftent->fts_statb.st_ino, &mtime, &atime) < 0))
416 #			else
417 			if (!tflag || (get_atdir(ftent->fts_statp->st_dev,
418 			    ftent->fts_statp->st_ino, &mtime, &atime) < 0))
419 #			endif
420 				continue;
421 			set_ftime(ftent->fts_path, mtime, atime, 1);
422 			continue;
423 		case FTS_DC:
424 			/*
425 			 * fts claims a file system cycle
426 			 */
427 			pax_warn(1,"File system cycle found at %s",ftent->fts_path);
428 			continue;
429 		case FTS_DNR:
430 #			ifdef NET2_FTS
431 			sys_warn(1, errno,
432 #			else
433 			sys_warn(1, ftent->fts_errno,
434 #			endif
435 			    "Unable to read directory %s", ftent->fts_path);
436 			continue;
437 		case FTS_ERR:
438 #			ifdef NET2_FTS
439 			sys_warn(1, errno,
440 #			else
441 			sys_warn(1, ftent->fts_errno,
442 #			endif
443 			    "File system traversal error");
444 			continue;
445 		case FTS_NS:
446 		case FTS_NSOK:
447 #			ifdef NET2_FTS
448 			sys_warn(1, errno,
449 #			else
450 			sys_warn(1, ftent->fts_errno,
451 #			endif
452 			    "Unable to access %s", ftent->fts_path);
453 			continue;
454 		}
455 
456 		/*
457 		 * ok got a file tree node to process. copy info into arcn
458 		 * structure (initialize as required)
459 		 */
460 		arcn->skip = 0;
461 		arcn->pad = 0;
462 		arcn->ln_nlen = 0;
463 		arcn->ln_name[0] = '\0';
464 #		ifdef NET2_FTS
465 		arcn->sb = ftent->fts_statb;
466 #		else
467 		arcn->sb = *(ftent->fts_statp);
468 #		endif
469 
470 		/*
471 		 * file type based set up and copy into the arcn struct
472 		 * SIDE NOTE:
473 		 * we try to reset the access time on all files and directories
474 		 * we may read when the -t flag is specified. files are reset
475 		 * when we close them after copying. we reset the directories
476 		 * when we are done with their file tree (we also clean up at
477 		 * end in case we cut short a file tree traversal). However
478 		 * there is no way to reset access times on symlinks.
479 		 */
480 		switch(S_IFMT & arcn->sb.st_mode) {
481 		case S_IFDIR:
482 			arcn->type = PAX_DIR;
483 			if (!tflag)
484 				break;
485 			add_atdir(ftent->fts_path, arcn->sb.st_dev,
486 			    arcn->sb.st_ino, arcn->sb.st_mtime,
487 			    arcn->sb.st_atime);
488 			break;
489 		case S_IFCHR:
490 			arcn->type = PAX_CHR;
491 			break;
492 		case S_IFBLK:
493 			arcn->type = PAX_BLK;
494 			break;
495 		case S_IFREG:
496 			/*
497 			 * only regular files with have data to store on the
498 			 * archive. all others will store a zero length skip.
499 			 * the skip field is used by pax for actual data it has
500 			 * to read (or skip over).
501 			 */
502 			arcn->type = PAX_REG;
503 			arcn->skip = arcn->sb.st_size;
504 			break;
505 		case S_IFLNK:
506 			arcn->type = PAX_SLK;
507 			/*
508 			 * have to read the symlink path from the file
509 			 */
510 			if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
511 			    PAXPATHLEN - 1)) < 0) {
512 				sys_warn(1, errno, "Unable to read symlink %s",
513 				    ftent->fts_path);
514 				continue;
515 			}
516 			/*
517 			 * set link name length, watch out readlink does not
518 			 * allways NUL terminate the link path
519 			 */
520 			arcn->ln_name[cnt] = '\0';
521 			arcn->ln_nlen = cnt;
522 			break;
523 		case S_IFSOCK:
524 			/*
525 			 * under BSD storing a socket is senseless but we will
526 			 * let the format specific write function make the
527 			 * decision of what to do with it.
528 			 */
529 			arcn->type = PAX_SCK;
530 			break;
531 		case S_IFIFO:
532 			arcn->type = PAX_FIF;
533 			break;
534 		}
535 		break;
536 	}
537 
538 	/*
539 	 * copy file name, set file name length
540 	 */
541 	arcn->nlen = l_strncpy(arcn->name, ftent->fts_path, PAXPATHLEN+1);
542 	arcn->name[arcn->nlen] = '\0';
543 	arcn->org_name = ftent->fts_path;
544 	return(0);
545 }
546