]> git.wh0rd.org - dump.git/blob - CHANGES
Off by one miscalculation (dumping a one letter directory).
[dump.git] / CHANGES
1 $Id: CHANGES,v 1.72 2000/08/19 23:15:38 stelian Exp $
2
3 Changes between versions 0.4b17 and 0.4b19 (released ?????????????)
4 ===================================================================
5
6 1. Fixed the signal handling in dump (which I broke in 0.4b17)
7 which was causing several strange dump failures (dump
8 hanged or segmentation faults etc).
9
10 2. Specified the default blocksize in the dump man page.
11
12 3. Changed two info messages of restore to be written on stdout
13 instead of stderr in order to leave stderr only for errors
14 or warnings. Thanks to Stephen Atwell
15 <satwell@urbana.css.mot.com> for the suggestion.
16
17 4. Corrected an off by one calculation which prevented
18 dumping a directory having a 1 character name length.
19 Thanks to Bernhard Erdmann <bernhard.erdmann@gmx.de>
20 for reporting the bug.
21
22 Changes between versions 0.4b17 and 0.4b18 (released June 30, 2000)
23 ===================================================================
24
25 1. Fixed a potential buffer overflow in restore. Thanks
26 to Stan Bubrouski <satan@fastdial.net> for reporting
27 the bug.
28
29 2. Fixed a readline-related bug which prevented
30 'cat DUMPFILE | restore -i -f -' from working. Thanks
31 to Charles Karney <karney@users.sourceforge.net>
32 for the bug report.
33
34 3. Changed a few "panic" into "exit", causing restore to
35 be more stable against some attacks (like the last one
36 reported on Bugtraq, although the last version of restore
37 was not vulnerable - just dumped core). Thanks to
38 Andreas Hasenack <andreas@conectiva.com.br> for reporting
39 the bugs to me.
40
41 4. Removed the suid-root bit on dump and restore in the
42 default build (and generated RPMs). It should be safer
43 now. Users who need the suid-root capabilities in order
44 to make network backups should read first the man page
45 of dump and enable the suid bit manually.
46
47 5. Added -ltermcap to the compile parameters for restore
48 when using readline support, in order to make the compile
49 process work on some systems (SuSE ?). Thanks to
50 Patrik Schindler <poc@pocnet.net> for reporting the bug.
51
52 Changes between versions 0.4b16 and 0.4b17 (released June 1st, 2000)
53 ====================================================================
54
55 1. The -F script is called now *only* at the end of a tape,
56 not at the end of the dump. Thanks to Christian Haul
57 <haul@informatik.tu-darmstadt.de> for the bug report.
58
59 Normally, the device name and the current volume number
60 are passed on the command line of the script. However,
61 if you want the old style script (with no arguments
62 passed on the command line), you can enable it in
63 configure with the --enable-oldstylefscript.
64
65 2. Use posix signal handling to preserve dump functionality
66 with libc5. Thanks to Jeff Johnson <jbj@redhat.com> for
67 the patch.
68
69 3. Made the exit code of restore in case of a 'C'ompare
70 command reflect the result of the compare. An exit status
71 of 0 means the dump archive is correct, 1 means tape errors,
72 2 means that some files were modified. Thanks to Kenneth Porter
73 <shiva@well.com> for the suggestion.
74
75 4. Made (finally) quotes work in interactive restore.
76
77 5. Small fixes in order to allow dump to compile with a
78 really old e2fsprogs version. Thanks to Ian Zimmerman
79 <itz@speakeasy.org> for the bug report.
80
81 6. Add GNU readline capabilities to interactive restore.
82 Use configure's --enable-readline flag to enable this feature.
83 Thanks to Patrik Schindler <poc@pocnet.net> for the
84 suggestion.
85
86 7. Do the compare on the fly in restore 'C' mode (this will
87 allow not to exhaust the available /tmp space when
88 ccmparing large files). Thanks to Kenneth Porter
89 <shiva@well.com> for the suggestion.
90
91 Changes between versions 0.4b15 and 0.4b16 (released March 11, 2000)
92 ====================================================================
93
94 1. Fixed some several duplicate 'DUMP: DUMP:' in the
95 output of dump.
96
97 2. Corrected the estimation of blocks to dump. Note that
98 this is still wrong for large directory inodes, due
99 to the size of a BSD directory entry on the tape
100 (estimation is based on the size of the ext2 directory,
101 which is smaller because it doesn't contain the null
102 character at the end).
103
104 3. Made dump report the total number of MB written to
105 tape at the end of run. Thanks to W. Reilly Cooley
106 <wcooley@nakedape.cc> for the patch.
107
108 4. Added the -X option to restore giving the possibility
109 to read the names of the files to be extracted/listed
110 from a text file (in addition of the command line).
111 Thanks to Dejan Muhamedagic <dejan@quant-x.com> for the
112 patch.
113
114 5. Added the device name and the current volume number
115 as arguments to the end of tape script (dump -F option).
116
117 6. Made the multi-volume dump work again (after having
118 broken it in 0.4b15).
119
120 Changes between versions 0.4b14 and 0.4b15 (released March 2, 2000)
121 ===================================================================
122
123 1. Added a prompt command in interactive restore mode. Thanks
124 to Andreas Dilger <adilger@home.com> for the patch.
125
126 2. Fixed a buffer overflow problem in dump (caused by
127 not checking the size of the filesystem parameter).
128 Thanks to Kim Yong-jun <loveyou@hackerslab.org> for
129 reporting this on Bugtraq (and to several dump users
130 who forwarded me his mail).
131
132 3. Added the '-F script' option to dump in order to
133 launch a script at the end of each tape (to be used
134 with a tape changer, or to notify the sysadmin by
135 pager etc.).
136
137 4. Fixed a bug in restore compare code caused by the changes
138 I made in 0.4b14.
139
140 5. Fixed the treatment of options using the old BSD syntax
141 in both dump and restore.
142
143 Changes between versions 0.4b13 and 0.4b14 (released February 10, 2000)
144 =======================================================================
145
146 1. Fixed a bug in dump which may have caused invalid deleted
147 directories to be dumped out if they were deleted after the
148 mapping pass. This could occure on active filesystem and lead
149 to heap corruption (causing dump malfunction in many possible ways).
150 Thanks to Jan Sanislo <oystr@cs.washington.edu> for finding this
151 bug and submitting the patch.
152
153 2. Fixed the handling of the filesystem parameter in dump. A
154 '/mnt/disk' parameter caused the disk contents to be dumped,
155 but a '/mnt/disk/' parameter caused the mountpoint directory
156 to be dumped (generally an empty directory).
157
158 3. Improved the output of dump in order to tell which directory
159 it is currently dumping (when dumping a subtree).
160
161 4. Added the '-e' exclude inode option to dump. Thanks to
162 Isaac Chuang <ike@isl.stanford.edu> for contributing with the patch.
163
164 5. Added a REPORTING-BUGS file in order to provide a guide
165 on how to correctly report a bug in dump/restore.
166
167 6. Fixed a restore problem when restoring a hard link to an inode
168 having the immutable or append-only attribute set. Thanks to
169 Ambrose Li <acli@mingpaoxpress.com> for submitting the patch.
170
171 7. Fixed a compatibility problem between dumps made on little
172 endian machines (the format was correct) and big endian
173 machines (the format was incorrect). This fix break the
174 compatibility with the older dumps made on big endian
175 machines (sparc, m86k, ppc etc). For the first time in
176 linux dump's history (I believe), the dumps made by *BSD,
177 Linux/alpha, Linux/sparc and Linux/x86 are compatible,
178 so interchangeable. Thanks to Rob Cermak
179 <cermak@ahab.rutgers.edu> for submitting the bug and
180 helping me test the fix.
181
182 8. Fixed the way dump reports the remaining percent/time, if
183 the number of blocks actually dumped exceeds the estimated
184 number of blocks. Thanks to Jean-Paul van der Jagt
185 <jeanpaul@dutepp0.et.tudelft.nl> for reporting the bug.
186
187 Changes between versions 0.4b12 and 0.4b13 (released January 21, 2000)
188 ======================================================================
189
190 1. Small Y2K fix in the man pages :). Thanks to Bernhard Sadlowski
191 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug.
192
193 2. Removed the requirement to build the RPM as root from the
194 spec file. Thanks to Christian Weisgerber
195 <naddy@mips.rhein-neckar.de> for submitting this.
196
197 3. Fixed a bug in dump related to the 'filetype' feature of ext2fs,
198 causing dump to block when dumping really huge filesystems.
199 Many thanks to Patrik Schindler <poc@pocnet.net> for
200 helping me find this bug.
201
202 4. Fixed the treatment for an interrupt signal when dump access
203 the remote tape through RSH. Thanks to Christian Weisgerber
204 <naddy@mips.rhein-neckar.de> for providing the patch.
205
206 5. Fixed a bug which was causing dump/restore to display
207 garbage characters instead of the remote host name.
208
209 Changes between versions 0.4b11 and 0.4b12 (released January 8, 2000)
210 =====================================================================
211
212 1. Small fix in the dump man page. Thanks to Thorsten Kukuk
213 <kukuk@suse.de> for submitting the patch.
214
215 2. Fix for the exit code when using the size estimate option of
216 dump. Thanks to Matti Taskinen <mkt@rni.helsinki.fi> for
217 submitting the patch.
218
219 3. Handle EINTR in atomical reads/writes in dump, which was causing
220 dump to fail on some systems. Thanks to Eric Jergensen
221 <eric@dvns.com> for reporting the bug and submitting the patch.
222
223 4. Handle more than 16 characters for the device names in dumpdates.
224 (up to 255 now). Thanks to Rainer Clasen <bj@ncc.cicely.de> for
225 tracking down the problem and proposing the solution.
226
227 5. Fixed a bug in dump which prevented the creation of the
228 dumpdates file when doing a 0-level dump without already
229 having a dumpdates file. Thanks to Patrik Schindler
230 <poc@pocnet.net> for reporting the bug.
231
232 6. Changed the way dump 'S' flag reports the size estimate
233 from number of blocks into bytes (making it compatible
234 with the Solaris version, and simplifying things for
235 amanda users). Thanks to Jason L Tibbitts III
236 <tibbs@math.uh.edu> for reporting the bug.
237
238 7. Fixed a compatibility problem in linux/alpha dump tape format.
239 Now the linux/alpha dump are (again) compatible with the
240 other dump formats. But this breaks compatibility with
241 older dumps made on alpha. Thanks to Mike Tibor
242 <tibor@lib.uaa.alaska.edu> for helping me in finding this bug.
243
244 Changes between versions 0.4b10 and 0.4b11 (released December 5, 1999)
245 ======================================================================
246
247 1. Added a '--enable-kerberos' to configure.
248
249 2. Added a 'S' option to dump which determines the amount of space
250 that is needed to perform the dump without actually doing it, similar
251 to the Sun's ufsdump 'S' option. Patch contributed by Rob Cermak
252 <cermak@ahab.rutgers.edu>.
253
254 3. Added a 'M' multi-volume option to dump and restore which enables
255 dumping to multiple files (useful when dumping to an ext2
256 partition to make several dump files in order to bypass the 2GB
257 file size limitation). The 'f' argument is treated as a prefix and
258 the output files will be named <prefix>001, <prefix>002 etc. With
259 the 'M' flag, restore automatically selects the right file without
260 asking to enter a new tape each time.
261
262 4. Fixed a memory leak which was causing dump to grow very big
263 (270MB when dumping a 10GB filesystem...). Thanks to Jason
264 Fearon <jasonf@netrider.org.au> for reporting the bug.
265
266 Changes between versions 0.4b9 and 0.4b10 (released November 21, 1999)
267 ======================================================================
268
269 1. Make configure test if the system glob routines support
270 extended syntax (ALTDIRFUNC). If not, use the internal glob
271 routines instead of system ones. Thanks to Bernhard Sadlowski
272 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug
273 and helping me resolve this and other minor libc5 compiling
274 glitches.
275
276 2. Fix a problem when dumping a ext2fs with the 'filetype'
277 feature enabled. Thanks to Patrick J. LoPresti
278 <patl@cag.lcs.mit.edu> for reporting the bug and to
279 Theodore Y. Ts'o <tytso@mit.edu> for providing the patch.
280
281 3. Made the nodump flag work on directories. A directory which
282 has the nodump flag gets never dumped, regardless of its
283 contents.
284
285 4. Integrate a patch from Jeremy Fitzhardinge <jeremy@goop.org>
286 which allows dump on an active ext3 filesystem. However, this
287 is a "quick and dirty" patch which enables backup of an ext3
288 filesystem through the ext2 compatibility (by ignoring the
289 NEEDS_RECOVERY bit). The journal file is not recognized and
290 it is dumped (it should not).
291
292 5. Test the superblock compatibility flags when dumping, in order
293 to be sure that we know how to deal with specific features.
294
295 Changes between versions 0.4b8 and 0.4b9 (released November 5, 1999)
296 ====================================================================
297
298 1. Use lchown instead of chown, fixing a possible security problem
299 when restoring symlinks (a malicious user could use this
300 to deliberately corrupt the ownership of important system files).
301 Thanks to Chris Siebenmann <cks@utcc.utoronto.ca> for detecting
302 this and providing the patch.
303
304 Changes between versions 0.4b7 and 0.4b8 (released November 3, 1999)
305 ====================================================================
306
307 1. Put dump sources under CVS, added Id tags in all files so
308 one can use 'ident' on binary files.
309
310 2. Added the dump/restore version in the usage text so one can
311 easily verify the version he is using.
312
313 3. Small patch from Nuno Oliveira <nuno@eq.uc.pt> which fixes
314 a va_start/va_end problem on linux-ppc (always call va_start
315 va_end in pairs each time we use a vararg function).
316
317 4. Added again the DT_* constants because old libc does not
318 contain them :(. Thanks to Eric Maisonobe <virnet@nat.fr>
319 for submitting the bug report.
320
321 5. Use ext2fs_llseek instead of llseek. With recent e2fsprogs
322 this should enable dumping big (huge) filesystems.
323
324 6. Added the RSH environment variable in order to be able to
325 use a rsh replacement like ssh when doing remote backups (and
326 bypass the security limitations of rcmd). Now you can do remote
327 backups without being root (or making dump setuid root).
328
329 7. Modified again the way dumpdates works. For incremental dumps,
330 we need to read dumpdates even if we are not using 'u' option.
331 Thanks to Bdale Garbee <bdale@gag.com> for his ideas on how
332 this should work.
333
334 Changes between versions 0.4b6 and 0.4b7 (released October 8, 1999)
335 ===================================================================
336
337 1. Removed the 'k' flag from the restore 'about' text if kerberos
338 was not compiled in.
339
340 2. Prototyped (f)setflags from e2fsprogs and corrected the calls
341 to them (fsetflags takes a char*, setflags an open fd!).
342
343 3. (f)setflags is called only if the flags aren't empty. If the
344 file is a special file, a warning is printed, because changing
345 flags implies opening the device. Normally, a special file
346 should not have any flag... (Debian bug #29775, patch provided
347 by Abhijit Dasgupta <abhijit@ans.net>).
348
349 4. Made possible to dump a file system not mentioned in /etc/fstab.
350 (Debian bug #11904, patch provided by Eirik Fuller <eirik@netcom.com>).
351
352 5. Changed the default behaviour to not create dumpdates
353 unless 'u' option is specified. Removed the old "debian-patch"
354 which provided the same thing. (Debian bug #38136, #33818).
355
356 6. Removed all those dump*announce, since they were getting old...
357
358 7. Added warning messages if dumpdates does not exist and
359 when an inferior level dump does not exist (except for a level 0
360 dump).
361
362 8. Debugged the glob calls in interactive mode: restore used a
363 dirent struct which was different from the /usr/include/dirent.h
364 one (this used to work, is it a glibc2 change?), so none of the
365 compat glob (which used /usr/include/dirent.h) or the system glob
366 worked. Restore use now the system dirent (and the system
367 DT_* constants), which are compatible with BSD ones.
368
369 9. Added a configure flag (--with-dumpdatespath) to specify
370 the location of dumpdates file. By default, it is
371 /etc/dumpdates.
372
373 10. Added the "AUTHOR" and "AVAILABILITY" sections and
374 included the current date/version in man pages.
375
376 11. Corrected the estimation of remaining time when
377 the operator doesn't change the tapes quickly enough. This
378 was an old bug, I thought I corrected it, and discovered
379 that in fact it was corrected in two different places, so
380 the results canceled each other...
381
382 Changes between versions 0.4b5 and 0.4b6 (released October 1, 1999)
383 ===================================================================
384
385 1. Integrated multiple patches from RedHat, Debian and SuSE:
386
387 - tweak dump/itime.c to not try to read dumpdates if the 'u' option
388 isn't specified.
389 - several fixes in the man pages.
390 - update the default tape device to /dev/st0.
391 - many updates for Linux Alpha (byte ordering, size_t etc).
392 - buffer overruns.
393 - use environment variable for TMPDIR (instead of /tmp).
394 - use sigjmp_buf instead of jmp_buf (RedHat bug #3260).
395 - workaround egcs bug (RedHat bugs #4281 and #2989).
396 - wire $(OPT) throughout Makefile's.
397
398 2. Upgrade the dump revision to 1, making possible to dump filesystems
399 made with e2fsprogs-1.15 or newer. Nothing seems to break...
400
401 3. Fix some compile warnings, prototype all functions.
402
403 4. Use glibc err/glob instead of internal compatibility
404 routines (only if available).
405
406 5. Fix a compile error on Linux 2.2.7 / libc5 (5.4.44) (patch provided
407 by Bernhard Sadlowski <sadlowsk@mathematik.uni-bielefeld.de>).
408
409 Changes between versions 0.4b4 and 0.4b5 (released September 22, 1999)
410 ======================================================================
411
412 1. Integrated the changes from FreeBSD-3.1-RELEASE
413 (mostly bug fixes, buffer overruns, dump has now an "automatic
414 tape length calculation" flag, dump/restore can use kerberos now
415 (this is NOT tested), use environment variables for TAPE and
416 RMT etc.).
417
418 2. Integrated three RedHat patches ("glibc", "kernel" and "bread" patches)
419
420 3. Corrected a bug in restore when using 'C' option with multi-volumes
421 tapes (files splited accros two tapes give "size changed" errors
422 when comparing).
423
424 4. Corrected the long standing bug when dumping multiple tapes.
425 This works for me, needs further testing.
426
427 Changes between versions 0.4b3 and 0.4b4 (released January 17, 1997)
428 ====================================================================
429
430 1. Dump now runs correctly on kernels 2.1.x
431 Fix made by Gerald Peters <gapeters@worldnet.att.net>
432
433 Changes between versions 0.4b2 and 0.4b3
434 ========================================
435
436 1. Use realpath() if available
437
438 2. Report statistics
439
440 Changes between versions 0.4b1 and 0.4b2
441 ========================================
442
443 1. Fixed the bug fix from Greg Lutz (I had made a mistake when integrating
444 the patch)
445
446 2. Fixed restore to make it able to read FreeBSD 2.x dumps again
447
448 3. Fixed configure.in to correctly handle --enable-rmt
449
450 Changes between versions 0.3 and 0.4b1
451 ======================================
452
453 1. Integrated the changes from 4.4BSD-Lite2
454
455 2. Integrated the patches from Debian and Red Hat
456
457 3. Portability changes: use the __u32, __u16, __s32, and __s16 types
458
459 4. Changed dump to use the Ext2fs library to get block addresses. This
460 should solve the endianness problem on SparcLinux.
461
462 5. Created a configure.in file (shamelessly stolen from the e2fsprogs
463 distribution's one) to use autoconf
464
465 6. Fixed a few minor bugs
466
467 Changes between versions 0.2e and 0.2f
468 ======================================
469
470 1. Added the creation of named pipes (fifos) in restore.
471
472 2. Added the -N flag in the restore manual page.
473
474 3. Added the file kernel.patch which contains the llseek() optimization
475 patch for 1.2.x kernels.
476
477 4. Fixed a bug in the restoration of symbolic links: owner and group were
478 not restored.
479
480 5. Integrated some changes from FreeBSD 2.2.
481
482 6. Added a call to ftruncate() after restoring each file to restore
483 correctly files ending by a hole.
484
485 Changes between versions 0.2d and 0.2e
486 ======================================
487
488 1. Fixed a bug in the "set owner/mode" process. Every file was restored
489 with owner = root (0) and group = root/wheel/whatever (0).
490
491 Changes between versions 0.2c and 0.2d
492 ======================================
493
494 1. Dump is now able to backup 2GB+ filesystems.
495
496 2. Dump and restore can now be linked as static binaries.
497
498 Changes between versions 0.2b and 0.2c
499 ======================================
500
501 1. Fixed a bug when dumping ``slow'' (i.e. normal) symbolic links.
502
503 Changes between versions 0.2a and 0.2b
504 ======================================
505
506 1. Really fixed the bug that I should have corrected in 0.2a.
507
508 2. Enabled optimization again.
509
510 Changes between versions 0.2 and 0.2a
511 =====================================
512
513 1. Disabled the optimization during compilation.
514
515 Changes between versions 0.1 and 0.2
516 ====================================
517
518 1. Fixed a bug in fstab.c which caused a null pointer to be stored in
519 the fs_type field (actually, I modified the file fstab.c to make it
520 use the mntent functions).
521
522 2. Dump and restore now use a 4.3 BSD compatible dump format. Backups
523 made by dump should be readable by the BSD restore and backups made
524 by the BSD dump should be readable by restore. Unfortunately, this
525 means that the dump format has changed between version 0.1 and version
526 0.2 :-(
527
528 3. Dump is now able to backup a subtree, it is no longer limited to whole
529 filesystems like the BSD version.
530
531 4. Dump now uses ext2_llseek() so it is able to backup filesystems bigger
532 than 2 GB.
533
534 Changes between versions 0.0 and 0.1
535 ====================================
536
537 1. Now create links rdump and rrestore during the `make install' step.
538
539 2. Linux port specific bugs added to the manual pages
540
541 3. Incorrect estimation of the number of tapes blocks fixed when doing
542 incremental backups.
543
544 4. Better ls-like format in restore in interactive mode.