My version of Mark Curtis's --inplace option.
[rsync/rsync.git] / NEWS
... / ...
CommitLineData
1NEWS for rsync 2.6.3 (UNRELEASED)
2Protocol: 28 (unchanged)
3Changes since 2.6.2:
4
5 BUG FIXES:
6
7 - Fixed a crash bug that might appear when --delete was used and
8 multiple source directories were specified.
9
10 - Fixed the 32-bit truncation of the file length when generating the
11 checksums.
12
13 - The --backup code no longer attempts to create some directories
14 over and over again (generating warnings along the way).
15
16 - Fixed a bug in the reading of the secrets file (by the daemon) and
17 the password file (by the client): the files no longer need to be
18 terminated by a newline for their content to be read in.
19
20 - If a file has a read error on the sending side, the receiver will
21 no longer keep the resulting file unless the --partial option was
22 specified. (Note: both sides must be running 2.6.3 for this to
23 work -- older receivers always keep the file, and older senders
24 don't tell the receiver that the file was not read correctly.)
25
26 - If a file gets resent in a single transfer and the --backup option
27 is enabled, rsync no longer performs a duplicate backup (losing the
28 original file in the process).
29
30 - Fixed an age-old crash problem with --read-batch on a local copy
31 (rsync was improperly assuming --whole-file for the local copy).
32
33 - Files specified in the daemon's "exclude" or "exclude from" config
34 items are now excluded from being uploaded (assuming that the module
35 allows uploading at all) in addition to the old download exclusion.
36
37 - Got rid of a potential hang in the receiver when near the end of a
38 phase.
39
40 - When using --backup without a --backup-dir, rsync no longer preserves
41 the modify time on directories. This avoids confusing NFS.
42
43 - When --copy-links (-L) is specified, we now output a separate error
44 for a symlink that has no referent instead of claiming that a file
45 "vanished".
46
47 - Error messages from the daemon server's option-parsing (such as
48 refused options) now get sent back to the client (the server used
49 to just exit because the socket wasn't in the right state to send
50 the message).
51
52 - Most errors that occur during a daemon transfer are now returned to
53 the user in addition to being logged (some messages are intended to
54 be daemon-only).
55
56 - Fixed a bug in the daemon authentication code when using one of the
57 batch-processing options.
58
59 - We try to work around some buggy IPv6 implementations that fail to
60 implement IPV6_V6ONLY. This should fix the "address in use" error
61 that some daemons get when running on an OS with a buggy IPv6
62 implementation. Also, if the new code gets this error, we might
63 suggest that the user specify --ipv4 or --ipv6 (if we think it will
64 help).
65
66 ENHANCEMENTS:
67
68 - Added --keep-dirlinks (-K), which allows you to symlink a directory
69 onto another partition on the receiving side and have rsync treat it
70 as matching a normal directory from the sender.
71
72 - Added the "write only" option to the daemon's config file.
73
74 - Added long-option names for -4 and -6 (namely --ipv4 and --ipv6)
75 and documented all these options in the man page.
76
77 - Improved the handling of the --bwlimit option so that it's less
78 bursty, more accurate, and works properly over a larger range of
79 values.
80
81 - The rsync daemon-over-ssh code now looks for SSH_CONNECTION and
82 SSH2_CLIENT in addition to SSH_CLIENT to figure out the IP address.
83
84 - Added the --checksum-seed=N option for advanced users.
85
86 - Batch writing/reading has a brand-new implementation that is simpler,
87 fixes a few weird problems with the old code (such as no longer
88 sprinkling the batch files into different dirs or even onto different
89 systems), and is much less intrusive into the code (making it easier
90 to maintain for the future). The new code generates just one data
91 file instead of three, which makes it possible to read the batch via
92 stdin. Also, the old requirement of using the same fixed checksum-
93 seed for all batch processing has been removed. (Chris Shoemaker)
94
95 INTERNAL:
96
97 - Some cleanup in the exclude code has saved some per-exclude memory
98 and made the code easier to maintain.
99
100 - Improved the argv-overflow checking for a remote command that has
101 a lot of args.
102
103 - Use rsyserr() in the various places that were still calling
104 rprintf() with strerror() as an arg.
105
106 - If an rsync daemon is listening on multiple sockets (to handle both
107 IPv4 and IPv6 to a single port), we now close all the unneeded file
108 handles after we accept a connection (we used to close just one of
109 them).
110
111 - Optimized away a loop in hash_search().
112
113 BUILD CHANGES:
114
115 - Added a "gen" target to rebuild most of the generated files,
116 including configure, config.h.in, the man pages, and proto.h.
117
118 - If "make proto" doesn't find some changes in the prototypes, the
119 proto.h file is left untouched (its time-stamp used to always be
120 updated).
121
122 - The variable $STRIP (that is optionally set by the install-strip
123 target's rule) was changed to $INSTALL_STRIP because some systems
124 have $STRIP set in the environment.
125
126 - Fixed a build problem when SUPPORT_HARD_LINKS isn't defined.
127
128 DEVELOPER RELATED:
129
130 - The scripts in the testsuite dir were cleaned up a bit.
131
132 - Some new diffs were added to the patches dir, and some accepted
133 ones were removed.
134