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