Commit Graph

2700 Commits

Author SHA1 Message Date
Wayne Davison
aa0ea373cd Mention the enhancement that was made (quite a while ago) to the
daemon's wildcard-expansion limit (formerly 1000 items, now memory
limited).
2004-09-30 10:46:43 +00:00
Wayne Davison
6c3fda83ba Refined the text of a few of the news items. 2004-09-30 09:58:39 +00:00
Wayne Davison
a3571c6cce - Mention the fix for lost output with 2>&1.
- Prepare the file for the 2.6.3 final release.
2004-09-30 09:37:38 +00:00
Wayne Davison
6fcedb7dbe Mention the (anticipated) 2.6.3 release date. 2004-09-30 09:36:29 +00:00
Wayne Davison
18882701d2 Set our stderr output to blocking I/O to avoid any loss of output. 2004-09-29 17:58:07 +00:00
Wayne Davison
30c041f9ad Changed the errors concerning the secrets file to output with FLOG
instead of FERROR.
2004-09-24 17:04:05 +00:00
Wayne Davison
be7cf82299 - Make sure that match_address() always restores the "tok" string,
even on error.
- Turned the various FERROR messages into (the more proper) FLOG.
2004-09-24 16:50:07 +00:00
Wayne Davison
fde045cd77 Turned the various FERROR messages into (the more proper) FLOG. 2004-09-24 16:39:41 +00:00
Wayne Davison
183150b741 Added some more --inplace info (i.e. it implies --partial and conflicts
with 3 basis-file-affecting options).
2004-09-23 21:15:04 +00:00
Wayne Davison
a2570930e8 Made the refused-option message clearer, like Paul suggested. 2004-09-23 17:39:05 +00:00
Wayne Davison
fdb6716c0f Create a FIFO using mkfifo. 2004-09-23 16:34:43 +00:00
Wayne Davison
a20a88d235 Assume that gettimeofday() takes two arguments when cross-compiling. 2004-09-23 16:00:44 +00:00
Wayne Davison
48d3ff94c9 Only output major/minor info for a character/block device, not for
a FIFO or a socket.
2004-09-23 05:33:48 +00:00
Wayne Davison
a33857da09 Mention that --partial-dir now sets an exclude that helps to avoid
losing the partial data when deleting, and avoid copying any partial
data from the sender.
2004-09-22 04:14:43 +00:00
Wayne Davison
13791b1eeb If the partial-dir value is relative, add a directory-exclude for it
to the end of the user's exclude list.
2004-09-22 04:12:13 +00:00
Wayne Davison
9a5e37fca8 In get_exclude_tok(), if XFLG_DIRECTORY was passed in the xflags,
set MATCHFLG_DIRECTORY in the mflags we return.
2004-09-22 04:11:15 +00:00
Wayne Davison
3e976df0fb Added XFLG_DIRECTORY define. 2004-09-22 04:10:10 +00:00
Wayne Davison
42afed9c1a Preparing for release of 2.6.3pre2 v2.6.3pre2 2004-09-21 16:10:35 +00:00
Wayne Davison
37c36e2692 Mention the daemon-socket-connection change. 2004-09-21 15:47:20 +00:00
Wayne Davison
7fbc7031f4 Mentioned a couple more changes. 2004-09-21 15:31:33 +00:00
Wayne Davison
9f004a9ea9 The delete_one() function no longer needs to handle the case where
it thinks it is removing a directory and it is really removing a
symlink.
2004-09-21 09:24:06 +00:00
Wayne Davison
23f4587f2b - Turned readlink_stat() into a static function.
- Made readlink_stat() no longer honor keep_dirlinks.
- The make_file() function now checks keep_dirlinks late in the
  function so that it only transforms a symlink to a local dir
  into a directory if the receiver also has a directory by that
  name.  This makes the use of --delete with --keep-dirlinks
  work much better.
2004-09-21 09:24:02 +00:00
Wayne Davison
4d8f5b0ae7 - Fixed a problem with the $bakdir value.
- Made the files have better contents to copy.
- Also test --backup without --backup-dir.
2004-09-20 19:50:04 +00:00
Wayne Davison
89389a29ef Output a backup message when verbose > 1 and we did a copy prior
to an --inplace update.
2004-09-20 19:47:59 +00:00
Wayne Davison
29fe3961ab Output the same backup-message prefix when verbose > 1 regardless of
the setting of --backup-dir.
2004-09-20 19:46:45 +00:00
Wayne Davison
4e8a085ac9 Use $diffopt instead of -u. 2004-09-20 05:17:57 +00:00
Wayne Davison
fb22c2774d Got rid of a superfluous static buffer. 2004-09-20 05:01:38 +00:00
Wayne Davison
7d059d4c37 A simple test of the backup functionality. 2004-09-20 04:59:01 +00:00
Wayne Davison
9715c5899a Improved a comment. 2004-09-20 04:17:42 +00:00
Wayne Davison
cc07f21211 Mention the new RSYNC_PARTIAL_DIR environment variable. 2004-09-18 17:37:54 +00:00
Wayne Davison
b4d1e854ef Document the new RSYNC_PARTIAL_DIR environment variable. 2004-09-18 17:35:20 +00:00
Wayne Davison
075aa18fd4 Look for the RSYNC_PARTIAL_DIR environment variable when --partial
was specified (and --partial-dir was not).
2004-09-18 17:34:56 +00:00
Wayne Davison
e0204f5621 Mention that older rsync versions had a problem with --link-dest and
how to work around it.
2004-09-18 01:49:19 +00:00
Wayne Davison
a9ac4411e5 Mention the latest changes. 2004-09-17 16:53:51 +00:00
Wayne Davison
50b31539c2 - Added the ability to parse a literal IPv6 address in an "rsync:" URL
(e.g. rsync://[2001:638:500:101::21]:873/module/dir).
- Improved a couple --files-from error messages.
2004-09-17 16:50:53 +00:00
Wayne Davison
56194bcd95 When outputting the flist info (in a debug-level of verbosity) we
now mention the UID of the file when we are the sender (as well as
when we are root).
2004-09-17 16:39:34 +00:00
Wayne Davison
eb8ffa9040 Actually, since the close calls shouldn't fail (now that listener
is properly set to -1 on close), we don't really need to play the
save-errno game after all.
2004-09-16 17:22:31 +00:00
Wayne Davison
a7a1cc2c75 Make sure that /etc and /bin actually are readable before we try
to list them.
2004-09-16 17:16:36 +00:00
Wayne Davison
ab217f7ffa - Set "listener" to -1 after we close it so that the error-handler
doesn't try to re-close it.
- Set blocking I/O before the second (final) connect() call.
2004-09-16 17:09:46 +00:00
Wayne Davison
a20c9893e4 Don't try to optimize-away the sending of the --delete option if
--delete-after was specified (since we don't know what the protocol
version will be yet).
2004-09-08 07:33:06 +00:00
Wayne Davison
3bb400ca14 If --backup was used with --inplace, we don't limit the basis-file
matches (as we normally would) because the receiver is using the
backup-file as the basis-file.
2004-09-07 21:45:19 +00:00
Wayne Davison
cd6aa5b5c0 When --backup is used with --inplace, we make a copy of the destination
file into its backup spot while generating the checksums.
2004-09-07 21:44:02 +00:00
Wayne Davison
dc55d7bdab If we're making backups with --inplace, use the backup file as the
basis file while still updating the real destination file inplace.
2004-09-07 21:34:26 +00:00
Wayne Davison
8b115ac8dc Turn off make_backups during the redo phase, just like the receiver. 2004-09-07 21:32:36 +00:00
Wayne Davison
c94e4afbfa Moved the code that determines the backup filename into a new function
named get_backup_name().
2004-09-07 21:29:26 +00:00
Wayne Davison
6566d205e2 Made full_write() non-static. 2004-09-07 21:26:26 +00:00
Wayne Davison
e484f0cc04 W mustn't backup an inplace file in finish_transfer(). 2004-09-07 20:37:36 +00:00
Wayne Davison
bd397b8cba Reject the use of --compare-dest or --link-dest with --inplace
(it will take extra code to suppor this).
2004-09-07 20:36:36 +00:00
Wayne Davison
f8c8ef9eac When we say we're skipping a non-regular file, actually skip it. 2004-09-07 19:49:09 +00:00
Wayne Davison
72c19bb3de A minor optimization to the partial-dir code. 2004-09-07 17:03:51 +00:00