CHANGES

   1 $Id: CHANGES,v 1.68 2000/06/30 09:12:14 stelian Exp $
   2
   3 Changes between versions 0.4b17 and 0.4b18 (released June 30, 2000)
   4 ===================================================================
   5
   6 1.      Fixed a potential buffer overflow in restore. Thanks
   7         to Stan Bubrouski <satan@fastdial.net> for reporting
   8         the bug.
   9
  10 2.      Fixed a readline-related bug which prevented
  11         'cat DUMPFILE | restore -i -f -' from working. Thanks
  12         to Charles Karney <karney@users.sourceforge.net>
  13         for the bug report.
  14
  15 3.      Changed a few "panic" into "exit", causing restore to
  16         be more stable against some attacks (like the last one
  17         reported on Bugtraq, although the last version of restore
  18         was not vulnerable - just dumped core). Thanks to
  19         Andreas Hasenack <andreas@conectiva.com.br> for reporting
  20         the bugs to me.
  21
  22 4.      Removed the suid-root bit on dump and restore in the
  23         default build (and generated RPMs). It should be safer
  24         now. Users who need the suid-root capabilities in order
  25         to make network backups should read first the man page
  26         of dump and enable the suid bit manually.
  27
  28 5.      Added -ltermcap to the compile parameters for restore
  29         when using readline support, in order to make the compile
  30         process work on some systems (SuSE ?). Thanks to
  31         Patrik Schindler <poc@pocnet.net> for reporting the bug.
  32
  33 Changes between versions 0.4b16 and 0.4b17 (released June 1st, 2000)
  34 ====================================================================
  35
  36 1.      The -F script is called now *only* at the end of a tape,
  37         not at the end of the dump. Thanks to Christian Haul
  38         <haul@informatik.tu-darmstadt.de> for the bug report.
  39
  40         Normally, the device name and the current volume number
  41         are passed on the command line of the script. However,
  42         if you want the old style script (with no arguments
  43         passed on the command line), you can enable it in
  44         configure with the --enable-oldstylefscript.
  45
  46 2.      Use posix signal handling to preserve dump functionality
  47         with libc5. Thanks to Jeff Johnson <jbj@redhat.com> for
  48         the patch.
  49
  50 3.      Made the exit code of restore in case of a 'C'ompare
  51         command reflect the result of the compare. An exit status
  52         of 0 means the dump archive is correct, 1 means tape errors,
  53         2 means that some files were modified. Thanks to Kenneth Porter
  54         <shiva@well.com> for the suggestion.
  55
  56 4.      Made (finally) quotes work in interactive restore.
  57
  58 5.      Small fixes in order to allow dump to compile with a
  59         really old e2fsprogs version. Thanks to Ian Zimmerman
  60         <itz@speakeasy.org> for the bug report.
  61
  62 6.      Add GNU readline capabilities to interactive restore.
  63         Use configure's --enable-readline flag to enable this feature.
  64         Thanks to Patrik Schindler <poc@pocnet.net> for the
  65         suggestion.
  66
  67 7.      Do the compare on the fly in restore 'C' mode (this will
  68         allow not to exhaust the available /tmp space when
  69         ccmparing large files). Thanks to Kenneth Porter
  70         <shiva@well.com> for the suggestion.
  71
  72 Changes between versions 0.4b15 and 0.4b16 (released March 11, 2000)
  73 ====================================================================
  74
  75 1.      Fixed some several duplicate 'DUMP: DUMP:' in the
  76         output of dump.
  77
  78 2.      Corrected the estimation of blocks to dump. Note that
  79         this is still wrong for large directory inodes, due
  80         to the size of a BSD directory entry on the tape
  81         (estimation is based on the size of the ext2 directory,
  82         which is smaller because it doesn't contain the null
  83         character at the end).
  84
  85 3.      Made dump report the total number of MB written to
  86         tape at the end of run. Thanks to W. Reilly Cooley
  87         <wcooley@nakedape.cc> for the patch.
  88
  89 4.      Added the -X option to restore giving the possibility
  90         to read the names of the files to be extracted/listed
  91         from a text file (in addition of the command line).
  92         Thanks to Dejan Muhamedagic <dejan@quant-x.com> for the
  93         patch.
  94
  95 5.      Added the device name and the current volume number
  96         as arguments to the end of tape script (dump -F option).
  97
  98 6.      Made the multi-volume dump work again (after having
  99         broken it in 0.4b15).
 100
 101 Changes between versions 0.4b14 and 0.4b15 (released March 2, 2000)
 102 ===================================================================
 103
 104 1.      Added a prompt command in interactive restore mode. Thanks
 105         to Andreas Dilger <adilger@home.com> for the patch.
 106
 107 2.      Fixed a buffer overflow problem in dump (caused by
 108         not checking the size of the filesystem parameter).
 109         Thanks to Kim Yong-jun <loveyou@hackerslab.org> for
 110         reporting this on Bugtraq (and to several dump users
 111         who forwarded me his mail).
 112
 113 3.      Added the '-F script' option to dump in order to
 114         launch a script at the end of each tape (to be used
 115         with a tape changer, or to notify the sysadmin by
 116         pager etc.).
 117
 118 4.      Fixed a bug in restore compare code caused by the changes
 119         I made in 0.4b14.
 120
 121 5.      Fixed the treatment of options using the old BSD syntax
 122         in both dump and restore.
 123
 124 Changes between versions 0.4b13 and 0.4b14 (released February 10, 2000)
 125 =======================================================================
 126
 127 1.      Fixed a bug in dump which may have caused invalid deleted
 128         directories to be dumped out if they were deleted after the
 129         mapping pass. This could occure on active filesystem and lead
 130         to heap corruption (causing dump malfunction in many possible ways).
 131         Thanks to Jan Sanislo <oystr@cs.washington.edu> for finding this
 132         bug and submitting the patch.
 133
 134 2.      Fixed the handling of the filesystem parameter in dump. A
 135         '/mnt/disk' parameter caused the disk contents to be dumped,
 136         but a '/mnt/disk/' parameter caused the mountpoint directory
 137         to be dumped (generally an empty directory).
 138
 139 3.      Improved the output of dump in order to tell which directory
 140         it is currently dumping (when dumping a subtree).
 141
 142 4.      Added the '-e' exclude inode option to dump. Thanks to
 143         Isaac Chuang <ike@isl.stanford.edu> for contributing with the patch.
 144
 145 5.      Added a REPORTING-BUGS file in order to provide a guide
 146         on how to correctly report a bug in dump/restore.
 147
 148 6.      Fixed a restore problem when restoring a hard link to an inode
 149         having the immutable or append-only attribute set. Thanks to
 150         Ambrose Li <acli@mingpaoxpress.com> for submitting the patch.
 151
 152 7.      Fixed a compatibility problem between dumps made on little
 153         endian machines (the format was correct) and big endian
 154         machines (the format was incorrect). This fix break the
 155         compatibility with the older dumps made on big endian
 156         machines (sparc, m86k, ppc etc). For the first time in
 157         linux dump's history (I believe), the dumps made by *BSD,
 158         Linux/alpha, Linux/sparc and Linux/x86 are compatible,
 159         so interchangeable. Thanks to Rob Cermak
 160         <cermak@ahab.rutgers.edu> for submitting the bug and
 161         helping me test the fix.
 162
 163 8.      Fixed the way dump reports the remaining percent/time, if
 164         the number of blocks actually dumped exceeds the estimated
 165         number of blocks. Thanks to Jean-Paul van der Jagt
 166         <jeanpaul@dutepp0.et.tudelft.nl> for reporting the bug.
 167
 168 Changes between versions 0.4b12 and 0.4b13 (released January 21, 2000)
 169 ======================================================================
 170
 171 1.      Small Y2K fix in the man pages :). Thanks to Bernhard Sadlowski
 172         <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug.
 173
 174 2.      Removed the requirement to build the RPM as root from the
 175         spec file. Thanks to Christian Weisgerber
 176         <naddy@mips.rhein-neckar.de> for submitting this.
 177
 178 3.      Fixed a bug in dump related to the 'filetype' feature of ext2fs,
 179         causing dump to block when dumping really huge filesystems.
 180         Many thanks to Patrik Schindler <poc@pocnet.net> for
 181         helping me find this bug.
 182
 183 4.      Fixed the treatment for an interrupt signal when dump access
 184         the remote tape through RSH. Thanks to Christian Weisgerber
 185         <naddy@mips.rhein-neckar.de> for providing the patch.
 186
 187 5.      Fixed a bug which was causing dump/restore to display
 188         garbage characters instead of the remote host name.
 189
 190 Changes between versions 0.4b11 and 0.4b12 (released January 8, 2000)
 191 =====================================================================
 192
 193 1.      Small fix in the dump man page. Thanks to Thorsten Kukuk
 194         <kukuk@suse.de> for submitting the patch.
 195
 196 2.      Fix for the exit code when using the size estimate option of
 197         dump. Thanks to Matti Taskinen <mkt@rni.helsinki.fi> for
 198         submitting the patch.
 199
 200 3.      Handle EINTR in atomical reads/writes in dump, which was causing
 201         dump to fail on some systems. Thanks to Eric Jergensen
 202         <eric@dvns.com> for reporting the bug and submitting the patch.
 203
 204 4.      Handle more than 16 characters for the device names in dumpdates.
 205         (up to 255 now). Thanks to Rainer Clasen <bj@ncc.cicely.de> for
 206         tracking down the problem and proposing the solution.
 207
 208 5.      Fixed a bug in dump which prevented the creation of the
 209         dumpdates file when doing a 0-level dump without already
 210         having a dumpdates file. Thanks to Patrik Schindler
 211         <poc@pocnet.net> for reporting the bug.
 212
 213 6.      Changed the way dump 'S' flag reports the size estimate
 214         from number of blocks into bytes (making it compatible
 215         with the Solaris version, and simplifying things for
 216         amanda users). Thanks to Jason L Tibbitts III
 217         <tibbs@math.uh.edu> for reporting the bug.
 218
 219 7.      Fixed a compatibility problem in linux/alpha dump tape format.
 220         Now the linux/alpha dump are (again) compatible with the
 221         other dump formats. But this breaks compatibility with
 222         older dumps made on alpha. Thanks to Mike Tibor
 223         <tibor@lib.uaa.alaska.edu> for helping me in finding this bug.
 224
 225 Changes between versions 0.4b10 and 0.4b11 (released December 5, 1999)
 226 ======================================================================
 227
 228 1.      Added a '--enable-kerberos' to configure.
 229
 230 2.      Added a 'S' option to dump which determines the amount of space
 231         that is needed to perform the dump without actually doing it, similar
 232         to the Sun's ufsdump 'S' option. Patch contributed by Rob Cermak
 233         <cermak@ahab.rutgers.edu>.
 234
 235 3.      Added a 'M' multi-volume option to dump and restore which enables
 236         dumping to multiple files (useful when dumping to an ext2
 237         partition to make several dump files in order to bypass the 2GB
 238         file size limitation). The 'f' argument is treated as a prefix and
 239         the output files will be named <prefix>001, <prefix>002 etc. With
 240         the 'M' flag, restore automatically selects the right file without
 241         asking to enter a new tape each time.
 242
 243 4.      Fixed a memory leak which was causing dump to grow very big
 244         (270MB when dumping a 10GB filesystem...). Thanks to Jason
 245         Fearon <jasonf@netrider.org.au> for reporting the bug.
 246
 247 Changes between versions 0.4b9 and 0.4b10 (released November 21, 1999)
 248 ======================================================================
 249
 250 1.      Make configure test if the system glob routines support
 251         extended syntax (ALTDIRFUNC). If not, use the internal glob
 252         routines instead of system ones. Thanks to Bernhard Sadlowski
 253         <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug
 254         and helping me resolve this and other minor libc5 compiling
 255         glitches.
 256
 257 2.      Fix a problem when dumping a ext2fs with the 'filetype'
 258         feature enabled. Thanks to Patrick J. LoPresti
 259         <patl@cag.lcs.mit.edu> for reporting the bug and to
 260         Theodore Y. Ts'o <tytso@mit.edu> for providing the patch.
 261
 262 3.      Made the nodump flag work on directories. A directory which
 263         has the nodump flag gets never dumped, regardless of its
 264         contents.
 265
 266 4.      Integrate a patch from Jeremy Fitzhardinge <jeremy@goop.org>
 267         which allows dump on an active ext3 filesystem. However, this
 268         is a "quick and dirty" patch which enables backup of an ext3
 269         filesystem through the ext2 compatibility (by ignoring the
 270         NEEDS_RECOVERY bit). The journal file is not recognized and
 271         it is dumped (it should not).
 272
 273 5.      Test the superblock compatibility flags when dumping, in order
 274         to be sure that we know how to deal with specific features.
 275
 276 Changes between versions 0.4b8 and 0.4b9 (released November 5, 1999)
 277 ====================================================================
 278
 279 1.      Use lchown instead of chown, fixing a possible security problem
 280         when restoring symlinks (a malicious user could use this
 281         to deliberately corrupt the ownership of important system files).
 282         Thanks to Chris Siebenmann <cks@utcc.utoronto.ca> for detecting
 283         this and providing the patch.
 284
 285 Changes between versions 0.4b7 and 0.4b8 (released November 3, 1999)
 286 ====================================================================
 287
 288 1.      Put dump sources under CVS, added Id tags in all files so
 289         one can use 'ident' on binary files.
 290
 291 2.      Added the dump/restore version in the usage text so one can
 292         easily verify the version he is using.
 293
 294 3.      Small patch from Nuno Oliveira <nuno@eq.uc.pt> which fixes
 295         a va_start/va_end problem on linux-ppc (always call va_start
 296         va_end in pairs each time we use a vararg function).
 297
 298 4.      Added again the DT_* constants because old libc does not
 299         contain them :(. Thanks to Eric Maisonobe <virnet@nat.fr>
 300         for submitting the bug report.
 301
 302 5.      Use ext2fs_llseek instead of llseek. With recent e2fsprogs
 303         this should enable dumping big (huge) filesystems.
 304
 305 6.      Added the RSH environment variable in order to be able to
 306         use a rsh replacement like ssh when doing remote backups (and
 307         bypass the security limitations of rcmd). Now you can do remote
 308         backups without being root (or making dump setuid root).
 309
 310 7.      Modified again the way dumpdates works. For incremental dumps,
 311         we need to read dumpdates even if we are not using 'u' option.
 312         Thanks to Bdale Garbee <bdale@gag.com> for his ideas on how
 313         this should work.
 314
 315 Changes between versions 0.4b6 and 0.4b7 (released October 8, 1999)
 316 ===================================================================
 317
 318 1.      Removed the 'k' flag from the restore 'about' text if kerberos
 319         was not compiled in.
 320
 321 2.      Prototyped (f)setflags from e2fsprogs and corrected the calls
 322         to them (fsetflags takes a char*, setflags an open fd!).
 323
 324 3.      (f)setflags is called only if the flags aren't empty. If the
 325         file is a special file, a warning is printed, because changing
 326         flags implies opening the device. Normally, a special file
 327         should not have any flag... (Debian bug #29775, patch provided
 328         by Abhijit Dasgupta <abhijit@ans.net>).
 329
 330 4.      Made possible to dump a file system not mentioned in /etc/fstab.
 331         (Debian bug #11904, patch provided by Eirik Fuller <eirik@netcom.com>).
 332
 333 5.      Changed the default behaviour to not create dumpdates
 334         unless 'u' option is specified. Removed the old "debian-patch"
 335         which provided the same thing. (Debian bug #38136, #33818).
 336
 337 6.      Removed all those dump*announce, since they were getting old...
 338
 339 7.      Added warning messages if dumpdates does not exist and
 340         when an inferior level dump does not exist (except for a level 0
 341         dump).
 342
 343 8.      Debugged the glob calls in interactive mode: restore used a
 344         dirent struct which was different from the /usr/include/dirent.h
 345         one (this used to work, is it a glibc2 change?), so none of the
 346         compat glob (which used /usr/include/dirent.h) or the system glob
 347         worked. Restore use now the system dirent (and the system
 348         DT_* constants), which are compatible with BSD ones.
 349
 350 9.      Added a configure flag (--with-dumpdatespath) to specify
 351         the location of dumpdates file. By default, it is
 352         /etc/dumpdates.
 353
 354 10.     Added the "AUTHOR" and "AVAILABILITY" sections and
 355         included the current date/version in man pages.
 356
 357 11.     Corrected the estimation of remaining time when
 358         the operator doesn't change the tapes quickly enough. This
 359         was an old bug, I thought I corrected it, and discovered
 360         that in fact it was corrected in two different places, so
 361         the results canceled each other...
 362
 363 Changes between versions 0.4b5 and 0.4b6 (released October 1, 1999)
 364 ===================================================================
 365
 366 1.      Integrated multiple patches from RedHat, Debian and SuSE:
 367
 368         - tweak dump/itime.c to not try to read dumpdates if the 'u' option
 369           isn't specified.
 370         - several fixes in the man pages.
 371         - update the default tape device to /dev/st0.
 372         - many updates for Linux Alpha (byte ordering, size_t etc).
 373         - buffer overruns.
 374         - use environment variable for TMPDIR (instead of /tmp).
 375         - use sigjmp_buf instead of jmp_buf (RedHat bug #3260).
 376         - workaround egcs bug (RedHat bugs #4281 and #2989).
 377         - wire $(OPT) throughout Makefile's.
 378
 379 2.      Upgrade the dump revision to 1, making possible to dump filesystems
 380         made with e2fsprogs-1.15 or newer. Nothing seems to break...
 381
 382 3.      Fix some compile warnings, prototype all functions.
 383
 384 4.      Use glibc err/glob instead of internal compatibility
 385         routines (only if available).
 386
 387 5.      Fix a compile error on Linux 2.2.7 / libc5 (5.4.44) (patch provided
 388         by Bernhard Sadlowski <sadlowsk@mathematik.uni-bielefeld.de>).
 389
 390 Changes between versions 0.4b4 and 0.4b5 (released September 22, 1999)
 391 ======================================================================
 392
 393 1.      Integrated the changes from FreeBSD-3.1-RELEASE
 394         (mostly bug fixes, buffer overruns, dump has now an "automatic
 395         tape length calculation" flag, dump/restore can use kerberos now
 396         (this is NOT tested), use environment variables for TAPE and
 397         RMT etc.).
 398
 399 2.      Integrated three RedHat patches ("glibc", "kernel" and "bread" patches)
 400
 401 3.      Corrected a bug in restore when using 'C' option with multi-volumes
 402         tapes (files splited accros two tapes give "size changed" errors
 403         when comparing).
 404
 405 4.      Corrected the long standing bug when dumping multiple tapes.
 406         This works for me, needs further testing.
 407
 408 Changes between versions 0.4b3 and 0.4b4 (released January 17, 1997)
 409 ====================================================================
 410
 411 1.      Dump now runs correctly on kernels 2.1.x
 412         Fix made by Gerald Peters <gapeters@worldnet.att.net>
 413
 414 Changes between versions 0.4b2 and 0.4b3
 415 ========================================
 416
 417 1.      Use realpath() if available
 418
 419 2.      Report statistics
 420
 421 Changes between versions 0.4b1 and 0.4b2
 422 ========================================
 423
 424 1.      Fixed the bug fix from Greg Lutz (I had made a mistake when integrating
 425         the patch)
 426
 427 2.      Fixed restore to make it able to read FreeBSD 2.x dumps again
 428
 429 3.      Fixed configure.in to correctly handle --enable-rmt
 430
 431 Changes between versions 0.3 and 0.4b1
 432 ======================================
 433
 434 1.      Integrated the changes from 4.4BSD-Lite2
 435
 436 2.      Integrated the patches from Debian and Red Hat
 437
 438 3.      Portability changes: use the __u32, __u16, __s32, and __s16 types
 439
 440 4.      Changed dump to use the Ext2fs library to get block addresses.  This
 441         should solve the endianness problem on SparcLinux.
 442
 443 5.      Created a configure.in file (shamelessly stolen from the e2fsprogs
 444         distribution's one) to use autoconf
 445
 446 6.      Fixed a few minor bugs
 447
 448 Changes between versions 0.2e and 0.2f
 449 ======================================
 450
 451 1.      Added the creation of named pipes (fifos) in restore.
 452
 453 2.      Added the -N flag in the restore manual page.
 454
 455 3.      Added the file kernel.patch which contains the llseek() optimization
 456         patch for 1.2.x kernels.
 457
 458 4.      Fixed a bug in the restoration of symbolic links: owner and group were
 459         not restored.
 460
 461 5.      Integrated some changes from FreeBSD 2.2.
 462
 463 6.      Added a call to ftruncate() after restoring each file to restore
 464         correctly files ending by a hole.
 465
 466 Changes between versions 0.2d and 0.2e
 467 ======================================
 468
 469 1.      Fixed a bug in the "set owner/mode" process.  Every file was restored
 470         with owner = root (0) and group = root/wheel/whatever (0).
 471
 472 Changes between versions 0.2c and 0.2d
 473 ======================================
 474
 475 1.      Dump is now able to backup 2GB+ filesystems.
 476
 477 2.      Dump and restore can now be linked as static binaries.
 478
 479 Changes between versions 0.2b and 0.2c
 480 ======================================
 481
 482 1.      Fixed a bug when dumping ``slow'' (i.e. normal) symbolic links.
 483
 484 Changes between versions 0.2a and 0.2b
 485 ======================================
 486
 487 1.      Really fixed the bug that I should have corrected in 0.2a.
 488
 489 2.      Enabled optimization again.
 490
 491 Changes between versions 0.2 and 0.2a
 492 =====================================
 493
 494 1.      Disabled the optimization during compilation.
 495
 496 Changes between versions 0.1 and 0.2
 497 ====================================
 498
 499 1.      Fixed a bug in fstab.c which caused a null pointer to be stored in
 500         the fs_type field (actually, I modified the file fstab.c to make it
 501         use the mntent functions).
 502
 503 2.      Dump and restore now use a 4.3 BSD compatible dump format.  Backups
 504         made by dump should be readable by the BSD restore and backups made
 505         by the BSD dump should be readable by restore.  Unfortunately, this
 506         means that the dump format has changed between version 0.1 and version
 507         0.2 :-(
 508
 509 3.      Dump is now able to backup a subtree, it is no longer limited to whole
 510         filesystems like the BSD version.
 511
 512 4.      Dump now uses ext2_llseek() so it is able to backup filesystems bigger
 513         than 2 GB.
 514
 515 Changes between versions 0.0 and 0.1
 516 ====================================
 517
 518 1.      Now create links rdump and rrestore during the `make install' step.
 519
 520 2.      Linux port specific bugs added to the manual pages
 521
 522 3.      Incorrect estimation of the number of tapes blocks fixed when doing
 523         incremental backups.
 524
 525 4.      Better ls-like format in restore in interactive mode.