- Got rid of some awkward spacing in the long_options table.
[rsync/rsync.git] / NEWS
... / ...
CommitLineData
1NEWS for rsync 2.6.4 (UNRELEASED)
2Protocol: 28 (unchanged)
3Changes since 2.6.3:
4
5 BUG FIXES:
6
7 - Restore the list-clearing behavior of "!" in a .cvsignore file (2.6.3
8 was only treating it as a special token in an rsync include/exclude
9 file).
10
11 - An OS that has a binary mode for its files (such as cygwin0 needed
12 setmode(fd, O_BINARY) called on the temp-file we opened with
13 mkstemp().
14
15 - We now check if the OS doesn't support using mknod() for creating
16 FIFOs and sockets, and compile-in using mkfifo() and socket() when
17 necessary.
18
19 ENHANCEMENTS:
20
21 - In _exit_cleanup(): when we are exiting with a partially-received
22 file, we now flush any data in the write-cache before closing the
23 partial file.
24
25 BUILD CHANGES:
26
27 - Support an OS that uses mkdev() instead of makedev().
28
29\f
30NEWS for rsync 2.6.3 (30 Sep 2004)
31Protocol: 28 (unchanged)
32Changes since 2.6.2:
33
34 SECURITY FIXES:
35
36 - A bug in the sanitize_path routine (which affects a non-chrooted
37 rsync daemon) could allow a user to craft a pathname that would get
38 transformed into an absolute path for certain options (but not for
39 file-transfer names). If you're running an rsync daemon with chroot
40 disabled, *please upgrade*, ESPECIALLY if the user privs you run
41 rsync under is anything above "nobody".
42
43 OUTPUT CHANGES (ATTN: those using a script to parse the verbose output):
44
45 - Please note that the 2-line footer (output when verbose) now uses the
46 term "sent" instead of "wrote" and "received" instead of "read". If
47 you are not parsing the numeric values out of this footer, a script
48 would be better off using the empty line prior to the footer as the
49 indicator that the verbose output is over.
50
51 - The output from the --stats option was similarly affected to change
52 "written" to "sent" and "read" to "received".
53
54 - Rsync ensures that a filename that contains a newline gets mentioned
55 with each newline transformed into a question mark (which prevents a
56 filename from causing an empty line to be output).
57
58 - The "backed up ..." message that is output when at least 2 --verbose
59 options are specified is now the same both with and without the
60 --backup-dir option.
61
62 BUG FIXES:
63
64 - Fixed a crash bug that might appear when --delete was used and
65 multiple source directories were specified.
66
67 - Fixed a 32-bit truncation of the file length when generating the
68 checksums.
69
70 - The --backup code no longer attempts to create some directories
71 over and over again (generating warnings along the way).
72
73 - Fixed a bug in the reading of the secrets file (by the daemon) and
74 the password file (by the client): the files no longer need to be
75 terminated by a newline for their content to be read in.
76
77 - If a file has a read error on the sending side or the reconstructed
78 data doesn't match the expected checksum (perhaps due to the basis
79 file changing during the transfer), the receiver will no longer
80 retain the resulting file unless the --partial option was specified.
81 (Note: for the read-error detection to work, neither side can be
82 older than 2.6.3 -- older receivers will always retain the file, and
83 older senders don't tell the receiver that the file had a read
84 error.)
85
86 - If a file gets resent in a single transfer and the --backup option
87 is enabled, rsync no longer performs a duplicate backup (it used to
88 overwrite the original file in the backup area).
89
90 - Files specified in the daemon's "exclude" or "exclude from" config
91 items are now excluded from being uploaded (assuming that the module
92 allows uploading at all) in addition to the old download exclusion.
93
94 - Got rid of a potential hang in the receiver when near the end of a
95 phase.
96
97 - When using --backup without a --backup-dir, rsync no longer preserves
98 the modify time on directories. This avoids confusing NFS.
99
100 - When --copy-links (-L) is specified, we now output a separate error
101 for a symlink that has no referent instead of claiming that a file
102 "vanished".
103
104 - The --copy-links (-L) option no longer has the side-effect of telling
105 the receiving side to follow symlinks. See the --keep-dirlinks
106 option (mentioned below) for a way to specify that behavior.
107
108 - Error messages from the daemon server's option-parsing (such as
109 refused options) are now successfully transferred back to the client
110 (the server used to fail to send the message because the socket
111 wasn't in the right state for the message to get through).
112
113 - Most transfer errors that occur during a daemon transfer are now
114 returned to the user in addition to being logged (some messages are
115 intended to be daemon-only and are not affected by this).
116
117 - Fixed a bug in the daemon authentication code when using one of the
118 batch-processing options.
119
120 - We try to work around some buggy IPv6 implementations that fail to
121 implement IPV6_V6ONLY. This should fix the "address in use" error
122 that some daemons get when running on an OS with a buggy IPv6
123 implementation. Also, if the new code gets this error, we might
124 suggest that the user specify --ipv4 or --ipv6 (if we think it will
125 help).
126
127 - When the remote rsync dies, make a better effort to recover any error
128 messages it may have sent before dying (the local rsync used to just
129 die with a socket-write error).
130
131 - When using --delete and a --backup-dir that contains files that are
132 hard-linked to their destination equivalents, rsync now makes sure
133 that removed files really get removed (avoids a really weird rename()
134 behavior).
135
136 - Avoid a bogus run-time complaint about a lack of 64-bit integers when
137 the int64 type is defined as an off_t and it actually has 64-bits.
138
139 - Added a configure check for open64() without mkstemp64() so that we
140 can avoid using mkstemp() when such a combination is encountered.
141 This bypasses a problem writing out large temp files on OSes such as
142 AIX and HP-UX.
143
144 - Fixed an age-old crash problem with --read-batch on a local copy
145 (rsync was improperly assuming --whole-file for the local copy).
146
147 - When --dry-run (-n) is used and the destination directory does not
148 exist, rsync now produces a correct report of files that would be
149 sent instead of dying with a chdir() error.
150
151 - Fixed a bug that could cause a slow-to-connect rsync daemon to die
152 with an error instead of waiting for the connection to finish.
153
154 - Fixed an ssh interaction that could cause output to be lost when the
155 user chose to combine the output of rsync's stdout and stderr (e.g.
156 using the "2>&1").
157
158 ENHANCEMENTS:
159
160 - Added the --partial-dir=DIR option that lets you specify where to
161 (temporarily) put a partially transferred file (instead of over-
162 writing the destination file). E.g. --partial-dir=.rsync-partial
163 Also added support for the RSYNC_PARTIAL_DIR environment variable
164 that, when found, transforms a regular --partial option (such as
165 the convenient -P option) into one that also specifies a directory.
166
167 - Added --keep-dirlinks (-K), which allows you to symlink a directory
168 onto another partition on the receiving side and have rsync treat it
169 as matching a normal directory from the sender.
170
171 - Added the --inplace option that tells rsync to write each destination
172 file without using a temporary file. The matching of existing data
173 in the destination file can be severely limited by this, but there
174 are also cases where this is more efficient (such as appending data).
175 Use only when needed (see the man page for more details).
176
177 - Added the "write only" option for the daemon's config file.
178
179 - Added long-option names for -4 and -6 (namely --ipv4 and --ipv6)
180 and documented all these options in the man page.
181
182 - Improved the handling of the --bwlimit option so that it's less
183 bursty, more accurate, and works properly over a larger range of
184 values.
185
186 - The rsync daemon-over-ssh code now looks for SSH_CONNECTION and
187 SSH2_CLIENT in addition to SSH_CLIENT to figure out the IP address.
188
189 - Added the --checksum-seed=N option for advanced users.
190
191 - Batch writing/reading has a brand-new implementation that is simpler,
192 fixes a few weird problems with the old code (such as no longer
193 sprinkling the batch files into different dirs or even onto different
194 systems), and is much less intrusive into the code (making it easier
195 to maintain for the future). The new code generates just one data
196 file instead of three, which makes it possible to read the batch on
197 stdin via a remote shell. Also, the old requirement of forcing the
198 same fixed checksum-seed for all batch processing has been removed.
199
200 - If an rsync daemon has a module set with "list = no" (which hides its
201 presence in the list of available modules), a user that fails to
202 authenticate gets the same "unknown module" error that they would get
203 if the module were actually unknown (while still logging the real
204 error to the daemon's log file). This prevents fishing for module
205 names.
206
207 - The daemon's "refuse options" config item now allows you to match
208 option names using wildcards and/or the single-letter option names.
209
210 - Each transferred file now gets its permissions and modified-time
211 updated before the temp-file gets moved into place. Previously, the
212 finished file would have a very brief window where its permissions
213 disallowed all group and world access.
214
215 - Added the ability to parse a literal IPv6 address in an "rsync:" URL
216 (e.g. rsync://[2001:638:500:101::21]:873/module/dir).
217
218 - The daemon's wildcard expanding code can now handle more than 1000
219 filenames (it's now limited by memory instead of having a hard-wired
220 limit).
221
222 INTERNAL:
223
224 - Some cleanup in the exclude code has saved some per-exclude memory
225 and made the code easier to maintain.
226
227 - Improved the argv-overflow checking for a remote command that has a
228 lot of args.
229
230 - Use rsyserr() in the various places that were still calling rprintf()
231 with strerror() as an arg.
232
233 - If an rsync daemon is listening on multiple sockets (to handle both
234 IPv4 and IPv6 to a single port), we now close all the unneeded file
235 handles after we accept a connection (we used to close just one of
236 them).
237
238 - Optimized the handling of larger block sizes (rsync used to slow to a
239 crawl if the block size got too large).
240
241 - Optimized away a loop in hash_search().
242
243 - Some improvements to the sanitize_path() and clean_fname() functions
244 makes them more efficient and produce better results (while still
245 being compatible with the file-name cleaning that gets done on both
246 sides when sending the file-list).
247
248 - Got rid of alloc_sanitize_path() after adding a destination-buffer
249 arg to sanitize_path() made it possible to put all the former's
250 functionality into the latter.
251
252 - The file-list that is output when at least 4 verbose options are
253 specified reports the uid value on the sender even when rsync is
254 not running as root (since we might be sending to a root receiver).
255
256 BUILD CHANGES:
257
258 - Added a "gen" target to rebuild most of the generated files,
259 including configure, config.h.in, the man pages, and proto.h.
260
261 - If "make proto" doesn't find some changes in the prototypes, the
262 proto.h file is left untouched (its time-stamp used to always be
263 updated).
264
265 - The variable $STRIP (that is optionally set by the install-strip
266 target's rule) was changed to $INSTALL_STRIP because some systems
267 have $STRIP already set in the environment.
268
269 - Fixed a build problem when SUPPORT_HARD_LINKS isn't defined.
270
271 - When cross-compiling, the gettimeofday() function is now assumed to
272 be a modern version that takes two-args (since we can't test it).
273
274 DEVELOPER RELATED:
275
276 - The scripts in the testsuite dir were cleaned up a bit and a few
277 new tests added.
278
279 - Some new diffs were added to the patches dir, and some accepted
280 ones were removed.
281