pax/ftree.c

   1 /*      $OpenBSD: ftree.c,v 1.28 2008/05/06 06:54:28 henning Exp $      */
   2 /*      $NetBSD: ftree.c,v 1.4 1995/03/21 09:07:21 cgd Exp $    */
   3
   4 /*-
   5  * Copyright (c) 1992 Keith Muller.
   6  * Copyright (c) 1992, 1993
   7  *      The Regents of the University of California.  All rights reserved.
   8  *
   9  * This code is derived from software contributed to Berkeley by
  10  * Keith Muller of the University of California, San Diego.
  11  *
  12  * Redistribution and use in source and binary forms, with or without
  13  * modification, are permitted provided that the following conditions
  14  * are met:
  15  * 1. Redistributions of source code must retain the above copyright
  16  *    notice, this list of conditions and the following disclaimer.
  17  * 2. Redistributions in binary form must reproduce the above copyright
  18  *    notice, this list of conditions and the following disclaimer in the
  19  *    documentation and/or other materials provided with the distribution.
  20  * 3. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  */
  36
  37 #ifndef lint
  38 #if 0
  39 static const char sccsid[] = "@(#)ftree.c       8.2 (Berkeley) 4/18/94";
  40 #else
  41 static const char rcsid[] = "$OpenBSD: ftree.c,v 1.28 2008/05/06 06:54:28 henning Exp $";
  42 #endif
  43 #endif /* not lint */
  44
  45 #include <sys/types.h>
  46 #include <sys/time.h>
  47 #include <sys/stat.h>
  48 #include <sys/param.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;                /* current 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(void);
  83 static char *getpathname(char *, int);
  84
  85 /*
  86  * ftree_start()
  87  *      initialize the options passed to fts_open() during this run of pax
  88  *      options are based on the selection of pax options by the user
  89  *      fts_start() also calls fts_arg() to open the first valid file arg. We
  90  *      also attempt to reset directory access times when -t (tflag) is set.
  91  * Return:
  92  *      0 if there is at least one valid file arg to process, -1 otherwise
  93  */
  94
  95 int
  96 ftree_start(void)
  97 {
  98         /*
  99          * set up the operation mode of fts, open the first file arg. We must
 100          * use FTS_NOCHDIR, as the user may have to open multiple archives and
 101          * if fts did a chdir off into the boondocks, we may create an archive
 102          * volume in an place where the user did not expect to.
 103          */
 104         ftsopts = FTS_NOCHDIR;
 105
 106         /*
 107          * optional user flags that effect file traversal
 108          * -H command line symlink follow only (half follow)
 109          * -L follow sylinks (logical)
 110          * -P do not follow sylinks (physical). This is the default.
 111          * -X do not cross over mount points
 112          * -t preserve access times on files read.
 113          * -n select only the first member of a file tree when a match is found
 114          * -d do not extract subtrees rooted at a directory arg.
 115          */
 116         if (Lflag)
 117                 ftsopts |= FTS_LOGICAL;
 118         else
 119                 ftsopts |= FTS_PHYSICAL;
 120         if (Hflag)
 121                 ftsopts |= FTS_COMFOLLOW;
 122         if (Xflag)
 123                 ftsopts |= FTS_XDEV;
 124
 125         if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
 126                 paxwarn(1, "Unable to allocate memory for file name buffer");
 127                 return(-1);
 128         }
 129
 130         if (ftree_arg() < 0)
 131                 return(-1);
 132         if (tflag && (atdir_start() < 0))
 133                 return(-1);
 134         return(0);
 135 }
 136
 137 /*
 138  * ftree_add()
 139  *      add the arg to the linked list of files to process. Each will be
 140  *      processed by fts one at a time
 141  * Return:
 142  *      0 if added to the linked list, -1 if failed
 143  */
 144
 145 int
 146 ftree_add(char *str, int chflg)
 147 {
 148         FTREE *ft;
 149         int len;
 150
 151         /*
 152          * simple check for bad args
 153          */
 154         if ((str == NULL) || (*str == '\0')) {
 155                 paxwarn(0, "Invalid file name argument");
 156                 return(-1);
 157         }
 158
 159         /*
 160          * allocate FTREE node and add to the end of the linked list (args are
 161          * processed in the same order they were passed to pax). Get rid of any
 162          * trailing / the user may pass us. (watch out for / by itself).
 163          */
 164         if ((ft = (FTREE *)malloc(sizeof(FTREE))) == NULL) {
 165                 paxwarn(0, "Unable to allocate memory for filename");
 166                 return(-1);
 167         }
 168
 169         if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
 170                 str[len] = '\0';
 171         ft->fname = str;
 172         ft->refcnt = 0;
 173         ft->newercnt = 0;
 174         ft->chflg = chflg;
 175         ft->fow = NULL;
 176         if (fthead == NULL) {
 177                 fttail = fthead = ft;
 178                 return(0);
 179         }
 180         fttail->fow = ft;
 181         fttail = ft;
 182         return(0);
 183 }
 184
 185 /*
 186  * ftree_sel()
 187  *      this entry has been selected by pax. bump up reference count and handle
 188  *      -n and -d processing.
 189  */
 190
 191 void
 192 ftree_sel(ARCHD *arcn)
 193 {
 194         /*
 195          * set reference bit for this pattern. This linked list is only used
 196          * when file trees are supplied pax as args. The list is not used when
 197          * the trees are read from stdin.
 198          */
 199         if (ftcur != NULL)
 200                 ftcur->refcnt = 1;
 201
 202         /*
 203          * if -n we are done with this arg, force a skip to the next arg when
 204          * pax asks for the next file in next_file().
 205          * if -d we tell fts only to match the directory (if the arg is a dir)
 206          * and not the entire file tree rooted at that point.
 207          */
 208         if (nflag)
 209                 ftree_skip = 1;
 210
 211         if (!dflag || (arcn->type != PAX_DIR))
 212                 return;
 213
 214         if (ftent != NULL)
 215                 (void)fts_set(ftsp, ftent, FTS_SKIP);
 216 }
 217
 218 /*
 219  * ftree_notsel()
 220  *      this entry has not been selected by pax.
 221  */
 222
 223 void
 224 ftree_notsel()
 225 {
 226         if (ftent != NULL)
 227                 (void)fts_set(ftsp, ftent, FTS_SKIP);
 228 }
 229
 230 /*
 231  * ftree_skipped_newer()
 232  *      file has been skipped because a newer file exists and -u/-D given
 233  */
 234
 235 void
 236 ftree_skipped_newer(ARCHD *arcn)
 237 {
 238         /* skipped due to -u/-D, mark accordingly */
 239         if (ftcur != NULL)
 240                 ftcur->newercnt = 1;
 241 }
 242
 243 /*
 244  * ftree_chk()
 245  *      called at end on pax execution. Prints all those file args that did not
 246  *      have a selected member (reference count still 0)
 247  */
 248
 249 void
 250 ftree_chk(void)
 251 {
 252         FTREE *ft;
 253         int wban = 0;
 254
 255         /*
 256          * make sure all dir access times were reset.
 257          */
 258         if (tflag)
 259                 atdir_end();
 260
 261         /*
 262          * walk down list and check reference count. Print out those members
 263          * that never had a match
 264          */
 265         for (ft = fthead; ft != NULL; ft = ft->fow) {
 266                 if ((ft->refcnt > 0) || ft->newercnt > 0 || ft->chflg)
 267                         continue;
 268                 if (wban == 0) {
 269                         paxwarn(1,"WARNING! These file names were not selected:");
 270                         ++wban;
 271                 }
 272                 (void)fprintf(stderr, "%s\n", ft->fname);
 273         }
 274 }
 275
 276 /*
 277  * ftree_arg()
 278  *      Get the next file arg for fts to process. Can be from either the linked
 279  *      list or read from stdin when the user did not them as args to pax. Each
 280  *      arg is processed until the first successful fts_open().
 281  * Return:
 282  *      0 when the next arg is ready to go, -1 if out of file args (or EOF on
 283  *      stdin).
 284  */
 285
 286 static int
 287 ftree_arg(void)
 288 {
 289
 290         /*
 291          * close off the current file tree
 292          */
 293         if (ftsp != NULL) {
 294                 (void)fts_close(ftsp);
 295                 ftsp = NULL;
 296         }
 297
 298         /*
 299          * keep looping until we get a valid file tree to process. Stop when we
 300          * reach the end of the list (or get an eof on stdin)
 301          */
 302         for (;;) {
 303                 if (fthead == NULL) {
 304                         /*
 305                          * the user didn't supply any args, get the file trees
 306                          * to process from stdin;
 307                          */
 308                         if (getpathname(farray[0], PAXPATHLEN+1) == NULL)
 309                                 return(-1);
 310                 } else {
 311                         /*
 312                          * the user supplied the file args as arguments to pax
 313                          */
 314                         if (ftcur == NULL)
 315                                 ftcur = fthead;
 316                         else if ((ftcur = ftcur->fow) == NULL)
 317                                 return(-1);
 318                         if (ftcur->chflg) {
 319                                 /* First fchdir() back... */
 320                                 if (fdochdir(cwdfd) < 0) {
 321                                         syswarn(1, errno,
 322                                           "Can't fchdir to starting directory");
 323                                         return(-1);
 324                                 }
 325                                 if (dochdir(ftcur->fname) < 0) {
 326                                         syswarn(1, errno, "Can't chdir to %s",
 327                                             ftcur->fname);
 328                                         return(-1);
 329                                 }
 330                                 continue;
 331                         } else
 332                                 farray[0] = ftcur->fname;
 333                 }
 334
 335                 /*
 336                  * watch it, fts wants the file arg stored in a array of char
 337                  * ptrs, with the last one a null. we use a two element array
 338                  * and set farray[0] to point at the buffer with the file name
 339                  * in it. We cannot pass all the file args to fts at one shot
 340                  * as we need to keep a handle on which file arg generates what
 341                  * files (the -n and -d flags need this). If the open is
 342                  * successful, return a 0.
 343                  */
 344                 if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
 345                         break;
 346         }
 347         return(0);
 348 }
 349
 350 /*
 351  * next_file()
 352  *      supplies the next file to process in the supplied archd structure.
 353  * Return:
 354  *      0 when contents of arcn have been set with the next file, -1 when done.
 355  */
 356
 357 int
 358 next_file(ARCHD *arcn)
 359 {
 360         int cnt;
 361         time_t atime;
 362         time_t mtime;
 363
 364         /*
 365          * ftree_sel() might have set the ftree_skip flag if the user has the
 366          * -n option and a file was selected from this file arg tree. (-n says
 367          * only one member is matched for each pattern) ftree_skip being 1
 368          * forces us to go to the next arg now.
 369          */
 370         if (ftree_skip) {
 371                 /*
 372                  * clear and go to next arg
 373                  */
 374                 ftree_skip = 0;
 375                 if (ftree_arg() < 0)
 376                         return(-1);
 377         }
 378
 379         /*
 380          * loop until we get a valid file to process
 381          */
 382         for (;;) {
 383                 if ((ftent = fts_read(ftsp)) == NULL) {
 384                         if (errno)
 385                                 syswarn(1, errno, "next_file");
 386                         /*
 387                          * out of files in this tree, go to next arg, if none
 388                          * we are done
 389                          */
 390                         if (ftree_arg() < 0)
 391                                 return(-1);
 392                         continue;
 393                 }
 394
 395                 /*
 396                  * handle each type of fts_read() flag
 397                  */
 398                 switch (ftent->fts_info) {
 399                 case FTS_D:
 400                 case FTS_DEFAULT:
 401                 case FTS_F:
 402                 case FTS_SL:
 403                         /*
 404                          * these are all ok
 405                          */
 406                         break;
 407                 case FTS_SLNONE:        /* was same as above cases except Unix
 408                                            conformance requires this error check */
 409                         if (Hflag || Lflag) {       /* -H or -L was specified */
 410                                 if (ftent->fts_errno)
 411                                         paxwarn(1, "%s: %s",
 412                                                 ftent->fts_name, strerror(ftent->fts_errno));
 413                         }
 414                         break;
 415                 case FTS_DP:
 416                         /*
 417                          * already saw this directory. If the user wants file
 418                          * access times reset, we use this to restore the
 419                          * access time for this directory since this is the
 420                          * last time we will see it in this file subtree
 421                          * remember to force the time (this is -t on a read
 422                          * directory, not a created directory).
 423                          */
 424                         if (!tflag || (get_atdir(ftent->fts_statp->st_dev,
 425                             ftent->fts_statp->st_ino, &mtime, &atime) < 0))
 426                                 continue;
 427                         set_ftime(ftent->fts_path, mtime, atime, 1);
 428                         continue;
 429                 case FTS_DC:
 430                         /*
 431                          * fts claims a file system cycle
 432                          */
 433                         paxwarn(1,"File system cycle found at %s",ftent->fts_path);
 434                         continue;
 435                 case FTS_DNR:
 436                         syswarn(1, ftent->fts_errno,
 437                             "Unable to read directory %s", ftent->fts_path);
 438                         continue;
 439                 case FTS_ERR:
 440                         syswarn(1, ftent->fts_errno,
 441                             "File system traversal error");
 442                         continue;
 443                 case FTS_NS:
 444                 case FTS_NSOK:
 445                         syswarn(1, ftent->fts_errno,
 446                             "Unable to access %s", ftent->fts_path);
 447                         continue;
 448                 }
 449
 450                 /*
 451                  * ok got a file tree node to process. copy info into arcn
 452                  * structure (initialize as required)
 453                  */
 454                 arcn->skip = 0;
 455                 arcn->pad = 0;
 456                 arcn->ln_nlen = 0;
 457                 arcn->ln_name[0] = '\0';
 458                 memcpy(&arcn->sb, ftent->fts_statp, sizeof(arcn->sb));
 459
 460                 /*
 461                  * file type based set up and copy into the arcn struct
 462                  * SIDE NOTE:
 463                  * we try to reset the access time on all files and directories
 464                  * we may read when the -t flag is specified. files are reset
 465                  * when we close them after copying. we reset the directories
 466                  * when we are done with their file tree (we also clean up at
 467                  * end in case we cut short a file tree traversal). However
 468                  * there is no way to reset access times on symlinks.
 469                  */
 470                 switch (S_IFMT & arcn->sb.st_mode) {
 471                 case S_IFDIR:
 472                         arcn->type = PAX_DIR;
 473                         if (!tflag)
 474                                 break;
 475                         add_atdir(ftent->fts_path, arcn->sb.st_dev,
 476                             arcn->sb.st_ino, arcn->sb.st_mtime,
 477                             arcn->sb.st_atime);
 478                         break;
 479                 case S_IFCHR:
 480                         arcn->type = PAX_CHR;
 481                         break;
 482                 case S_IFBLK:
 483                         arcn->type = PAX_BLK;
 484                         break;
 485                 case S_IFREG:
 486                         /*
 487                          * only regular files with have data to store on the
 488                          * archive. all others will store a zero length skip.
 489                          * the skip field is used by pax for actual data it has
 490                          * to read (or skip over).
 491                          */
 492                         arcn->type = PAX_REG;
 493                         arcn->skip = arcn->sb.st_size;
 494                         break;
 495                 case S_IFLNK:
 496                         arcn->type = PAX_SLK;
 497                         /*
 498                          * have to read the symlink path from the file
 499                          */
 500                         if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
 501                             PAXPATHLEN)) < 0) {
 502                                 syswarn(1, errno, "Unable to read symlink %s",
 503                                     ftent->fts_path);
 504                                 continue;
 505                         }
 506                         /*
 507                          * set link name length, watch out readlink does not
 508                          * always NUL terminate the link path
 509                          */
 510                         arcn->ln_name[cnt] = '\0';
 511                         arcn->ln_nlen = cnt;
 512                         break;
 513                 case S_IFSOCK:
 514                         /*
 515                          * under BSD storing a socket is senseless but we will
 516                          * let the format specific write function make the
 517                          * decision of what to do with it.
 518                          */
 519                         arcn->type = PAX_SCK;
 520                         break;
 521                 case S_IFIFO:
 522                         arcn->type = PAX_FIF;
 523                         break;
 524                 }
 525                 break;
 526         }
 527
 528         /*
 529          * copy file name, set file name length
 530          */
 531         arcn->nlen = strlcpy(arcn->name, ftent->fts_path, sizeof(arcn->name));
 532         if (arcn->nlen >= sizeof(arcn->name))
 533                 arcn->nlen = sizeof(arcn->name) - 1; /* XXX truncate? */
 534         arcn->org_name = ftent->fts_path;
 535         return(0);
 536 }
 537
 538 /*
 539  * getpathname()
 540  *      Reads a pathname from stdin, handling NUL- or newline-termination.
 541  * Return:
 542  *      NULL at end of file, otherwise the NUL-terminated buffer.
 543  */
 544
 545 static char *
 546 getpathname(char *buf, int buflen)
 547 {
 548         char *bp, *ep;
 549         int ch, term;
 550
 551         if (zeroflag) {
 552                 /*
 553                  * Read a NUL-terminated pathname, being especially
 554                  * paranoid about proper termination and pathname length.
 555                  */
 556                 for (bp = buf, ep = buf + buflen; bp < ep; bp++) {
 557                         if ((ch = getchar()) == EOF) {
 558                                 if (bp != buf)
 559                                         paxwarn(1, "Ignoring unterminated "
 560                                             "pathname at EOF");
 561                                 return(NULL);
 562                         }
 563                         if ((*bp = ch) == '\0')
 564                                 return(buf);
 565                 }
 566                 /* Too long - skip this path */
 567                 *--bp = '\0';
 568                 term = '\0';
 569         } else {
 570                 if (fgets(buf, buflen, stdin) == NULL)
 571                         return(NULL);
 572                 if ((bp = strchr(buf, '\n')) != NULL || feof(stdin)) {
 573                         if (bp != NULL)
 574                                 *bp = '\0';
 575                         return(buf);
 576                 }
 577                 /* Too long - skip this path */
 578                 term = '\n';
 579         }
 580         while ((ch = getchar()) != term && ch != EOF)
 581                 ;
 582         paxwarn(1, "Ignoring too-long pathname: %s", buf);
 583         return(NULL);
 584 }