1 NEWS for rsync 2.6.2 (30 Apr 2004)
2 Protocol: 28 (unchanged)
7 - Fixed a major bug in the sorting of the filenames when --relative
8 is used for some sources (probably just "/", but don't depend on
9 that). This fix ensures that we ask for the right file-list item
10 when requesting changes from the sender.
12 - Rsync now checks the return value of the close() function to
13 better report disk-full problems on an NFS file system.
15 - Restored the old daemon-server behavior of logging error messages
16 rather than returning them to the user. (A better long-term fix
17 will be sought in the future.)
19 - An obscure uninitialized-variable bug was fixed in the uid/gid
20 code. (This bug probably had no ill effects.)
22 - Got rid of the configure check for sys/sysctl.h (it wasn't used
23 and was causing a problem on some systems). Also improved the
24 broken-largefile-locking test to try to avoid failure due to an
27 - Fixed a compile problem on systems that don't define
30 - Fixed a compile problem in the popt source for compilers that
31 don't support __attribute__.
35 - Improved the testsuite's "merge" test to work on OSF1.
37 - Two new diffs were added to the patches dir.
40 NEWS for rsync 2.6.1 (26 Apr 2004)
41 Protocol: 28 (changed)
46 - Paths sent to an rsync daemon are more thoroughly sanitized when
47 chroot is not used. If you're running a non-read-only rsync
48 daemon with chroot disabled, *please upgrade*, ESPECIALLY if the
49 user privs you run rsync under is anything above "nobody".
53 - Lower memory use, more optimal transfer of data over the socket,
54 and lower CPU usage (see the INTERNAL section for details).
56 - The RSYNC_PROXY environment variable can now contain a
57 "USER:PASS@" prefix before the "HOST:PORT" information.
60 - The --progress output now mentions how far along in the transfer
61 we are, including both a count of files transferred and a
62 percentage of the total file-count that we've processed. It also
63 shows better current-rate-of-transfer and remaining-transfer-time
66 - The configure script now accepts --with-rsyncd-conf=PATH to
67 override the default value of the /etc/rsyncd.conf file.
69 - Added a couple extra diffs in the "patches" dir, removed the ones
70 that got applied, and rebuilt the rest.
72 - Documentation changes now attempt to describe some often mis-
73 understood features more clearly.
77 - When -x (--one-file-system) is combined with -L (--copy-links) or
78 --copy-unsafe-links, no symlinked files are skipped, even if the
79 referent file is on a different filesystem.
81 - The --link-dest code now works properly for a non-root user when
82 (1) the UIDs of the source and destination differ and -o was
83 specified, or (2) when the group of the source can't be used on
84 the destination and -g was specified.
86 - Fixed a bug in the handling of -H (hard-links) that might cause
87 the expanded PATH/NAME value of the current item to get
88 overwritten (due to an expanded-name caching bug).
90 - We now reset the "new data has been sent" flag at the start of
91 each file we send. This makes sure that an interrupted transfer
92 with the --partial option set doesn't keep a shorter temp file
93 than the current basis file when no new data has been transfered
94 over the wire for that file.
96 - Fixed a byte-order problem in --batch-mode on big-endian machines.
99 - Fixed configure bug when running "./configure --disable-ipv6".
101 - Fixed "make test" bug when build dir is not the source dir.
103 - When using --cvs-exclude, the exclude items we get from a
104 per-directory's .cvsignore file once again only affect that one
105 directory (not all following directories too). The items are also
106 now properly word-split and parsed without any +/- prefix parsing.
108 - When specifying the USER@HOST: prefix for a file, the USER part
109 can now contain an '@', if needed (i.e. the last '@' is used to
110 find the HOST, not the first).
112 - Fixed some bugs in the handling of group IDs for non-root users:
113 (1) It properly handles a group that the sender didn't have a name
114 for (it would previously skip changing the group on any files in
115 that group). (2) If --numeric-ids is used, rsync no longer
116 attempts to set groups that the user doesn't have the permission
119 - Fixed the "refuse options" setting in the rsyncd.conf file.
121 - Improved the -x (--one-file-system) flag's handling of any mount-
122 point directories we encounter. It is both more optimal (in that
123 it no longer does a useless scan of the contents of the mount-
124 point dirs) and also fixes a bug where a remapped mount of the
125 original filesystem could get discovered in a subdir we should be
128 - Rsync no longer discards a double-slash at the start of a filename
129 when trying to open the file. It also no longer constructs names
130 that start with a double slash (unless the user supplied them).
132 - Path-specifying options to a daemon should now work the same with
133 or without chroot turned on. Previously, such a option (such as
134 --link-dest) would get its absolute path munged into a relative
135 one if chroot was not on, making that setting fairly useless.
136 Rsync now transforms the path into one that is based on the
137 module's base dir when chroot is not enabled.
139 - Fixed compilation problem on Tru64 Unix (having to do with
140 sockaddr.sa_len and sockaddr.sin_len).
142 - Fixed a compatibility problem interacting with older rsync
143 versions that might send us an empty --suffix value without
144 telling us that --backup-dir was specified.
146 - The "hosts allow" option for a daemon-over-remote-shell process
147 now has improved support for IPv6 addresses and a fix for systems
148 that have a length field in their socket structs.
150 - Fixed the ability to request an empty backup --suffix when sending
151 files to an rsync daemon.
155 - Most of the I/O is now buffered, which results in a pretty large
156 speedup when running under MS Windows. (Craig Barratt)
158 - Optimizations to the name-handling/comparing code have made some
159 significant reductions in user-CPU time for large file sets.
161 - Some cleanup of the variable types make the code more consistent.
163 - Reduced memory requirements of hard link preservation.
166 - Implemented a new algorithm for hard-link handling that speeds up
167 the code significantly. (J.W. Schultz and Wayne Davison)
169 - The --hard-link option now uses the first existing file in the
170 group of linked files as the basis for the transfer. This
171 prevents the sub-optimal transfer of a file's data when a new
172 hardlink is added on the sending side and it sorts alphabetically
173 earlier in the list than the files that are already present on the
176 - Dropped support for protocol versions less than 20 (2.3.0 released
177 15 Mar 1999) and activated warnings for protocols less than 25
178 (2.5.0 released 23 Aug 2001). (Wayne Davison and J.W. Schultz,
181 - More optimal data transmission for --hard-links (protocol 28).
183 - More optimal data transmission for --checksum (protocol 28).
185 - Less memory is used when --checksum is specified.
187 - Less memory is used in the file list (a per-file savings).
189 - The generator is now better about not modifying the file list
190 during the transfer in order to avoid a copy-on-write memory
191 bifurcation (on systems where fork() uses shared memory).
192 Previously, rsync's shared memory would slowly become unshared,
193 resulting in real memory usage nearly doubling on the receiving
194 side by the end of the transfer. Now, as long as permissions
195 are being preserved, the shared memory should remain that way
196 for the entire transfer.
198 - Changed hardlink info and file_struct + strings to use allocation
199 pools. This reduces memory use for large file-sets and permits
200 freeing memory to the OS. (J.W. Schultz)
202 - The 2 pipes used between the receiver and generator processes
203 (which are forked on the same machine) were reduced to 1 pipe and
204 the protocol improved so that (1) it is now impossible to have the
205 "redo" pipe fill up and hang rsync, and (2) trailing messages from
206 the receiver don't get lost on their way through the generator
207 over to the sender (which mainly affected hard-link messages and
208 verbose --stats output).
210 - Improved the internal uid/gid code to be more portable and a
211 little more optimized.
213 - The device numbers sent when using --devices are now sent as
214 separate major/minor values with 32-bit accuracy (protocol 28).
215 Previously, the copied devices were sent as a single 32-bit
216 number. This will make inter-operation of 64-bit binaries more
217 compatible with their 32-bit brethren (with both ends of the
218 connection are using protocol 28). Note that optimizations in the
219 binary protocol for sending the device numbers often results in
220 fewer bytes being used than before, even though more precision is
223 - Some cleanup of the exclude/include structures and its code made
224 things clearer (internally), simpler, and more efficient.
226 - The reading & writing of the file-list in batch-mode is now
227 handled by the same code that sends & receives the list over the
228 wire. This makes it much easier to maintain. (Note that the
229 batch code is still considered to be experimental.)