Wayne Davison
486f8cd1ab
Changed the prerelease-protocol support to use an idiom that will not
...
slow down the initial version handshake between the two sides.
2007-04-23 19:03:29 +00:00
Wayne Davison
875a13b42c
No need to reset quiet, since we tell parse_arguments() that it
...
should prevent quiet from getting set.
2007-04-23 18:40:00 +00:00
Wayne Davison
719522b9bd
Tweaked --version in a few minor ways, including the mentioning of
...
the SUBPROTOCOL_VERSION if it is non-zero.
2007-04-21 19:46:19 +00:00
Wayne Davison
753849fdbc
Batch files need to be written with the new subprotocol value
...
for protocol 30.
2007-04-21 19:32:46 +00:00
Wayne Davison
bb25779de7
Mention how many bits our local timestamps support.
2007-04-21 19:31:54 +00:00
Wayne Davison
348d54d6d1
Mention the latest changes that have been made.
2007-04-21 19:27:04 +00:00
Wayne Davison
4471d9e570
Added a way to exchange a protocol's sub-protocol value. This will be
...
0 when a protocol goes final, and non-zero for an intermediary CVS
version of a protocol that is in flux. If we find that we're talking
to a CVS version with a different sub-protocol #, we automatically
drop back to the prior protocol that the sides have in common.
2007-04-21 18:40:42 +00:00
Wayne Davison
19d4cac996
We now allow verbose messages to be sent when forwarding incremental-recursion
...
file-list data from the receiver to the generator.
2007-04-20 22:40:01 +00:00
Wayne Davison
e4c877cf70
Added a message queue for the receiver->generator messages to handle the case
...
where the message pipe is being used to forward the file-list data.
2007-04-20 22:39:58 +00:00
Wayne Davison
f14b3ef430
Verify that the modtime value is not being truncated, and complain if it is.
2007-04-20 22:39:55 +00:00
Wayne Davison
f3c93b1761
Check the size of a time_t.
2007-04-20 21:39:08 +00:00
Wayne Davison
a72ba0cfac
Call the new {read,write}_varlong30() names for the long-int
...
functions that are protocol sensitive.
2007-04-20 08:17:06 +00:00
Wayne Davison
ba59bd68fc
- Call the new {read,write}_varlong30() names for the long-int
...
functions that are protocol sensitive.
- Changed the sending of the time value in protocol 30 to support
64-bit values without bloating positive 32-bit values.
2007-04-20 08:17:03 +00:00
Wayne Davison
1c60d21947
Got rid of the NVAL*() defines.
2007-04-20 08:16:59 +00:00
Wayne Davison
473feecff3
Added {read,write}_varlong30() inline functions to use the
...
right long-int function based on protocol version.
2007-04-20 08:16:56 +00:00
Wayne Davison
351e23ad8e
- Changed the {read,write}_longint() functions into separate
...
{read,write}_varlong() and {read,write}_longint() functions:
the former handles variable-length transmissions, and the
latter is backward-compatible with older rsync versions.
- Improved the {read,write}_var{int,long}() functions to be
simpler and to use a different byte order.
- The {read_write}_varlong() functions also allow the caller
to specify the minimum number of bytes to send. This allows
time values to be sent efficiently with a 4-byte minimum,
while length values continue to be sent with a (more optimal
for them) 3-byte minimum.
2007-04-20 08:16:53 +00:00
Wayne Davison
f31514adb7
Changed the *_abbbrevint() functions to *_varint().
2007-04-20 02:17:58 +00:00
Wayne Davison
2fcc265c2f
Got rid of protocol-29 check.
2007-04-18 03:30:53 +00:00
Wayne Davison
b58e40332f
If ssh access is not enabled, use a shell function to test that
...
we can copy using a spawned rsync process.
2007-04-18 03:23:26 +00:00
Wayne Davison
a217c45359
Got rid of ACL's uid/gid iterators in favor of a single function
...
that converts the uids & gids in a loop.
2007-04-18 02:33:02 +00:00
Wayne Davison
764782662d
Can eliminate the testtmp.* ignore rule now.
2007-04-14 07:11:32 +00:00
Wayne Davison
4b24f4a2b5
Changed scratchbase from using a "testtmp." name prefix to using a
...
"testtmp/" name prefix (i.e. using a subdir). This allows me to create
a testtmp as a symlink to a different file system to assist with testing
ACLs and xattrs.
2007-04-14 07:10:52 +00:00
Wayne Davison
c64ff141b8
Improved the description of the --checksum option.
2007-04-12 02:44:41 +00:00
Wayne Davison
1b896f8d1e
The check for GCC <= 2 was in the wrong part of an #ifdef sequence.
2007-04-12 02:37:31 +00:00
Wayne Davison
01894cf01f
Allow a local-server copy to use -X and -A even if a --protocol-version
...
is forced, which makes some testing idioms easier.
2007-04-07 19:22:20 +00:00
Wayne Davison
16edf86595
The improved --xattrs option is landing on the trunk.
2007-04-07 17:22:25 +00:00
Wayne Davison
1c53e5171c
Tweaked an error message and a comment.
2007-04-07 17:22:09 +00:00
Wayne Davison
63d83e9477
Improved push_dir() error reporting.
2007-03-25 17:01:48 +00:00
Wayne Davison
5463453489
The print now uses a more straight-forward map() instead of a join().
2007-03-21 13:51:54 +00:00
Wayne Davison
108133689d
Use relative directories, and note that the merge code is only
...
working because it is not yet using --inc-recursion.
2007-03-18 20:41:51 +00:00
Wayne Davison
46ad63b7cf
Added a test of a multi-file copy from a nested directory using a
...
relative source path.
2007-03-18 20:40:01 +00:00
Wayne Davison
dbb20f6bc0
One more tweak.
2007-03-18 20:05:24 +00:00
Wayne Davison
25c2a6ac0e
Fixed a multi-file, no-dir, relative copy, such as:
...
rsync -av lib/*.c /var/tmp/lib
2007-03-18 17:35:39 +00:00
Wayne Davison
161fba6903
Put the checksum seed at the end of the checksum2 buffer in md5 mode.
2007-03-18 06:28:51 +00:00
Wayne Davison
55edf18c42
Made a char* arg const.
2007-03-18 06:20:07 +00:00
Wayne Davison
a0456b9c46
Protocol 30 now uses MD5 checksums instead of MD4.
2007-03-18 06:00:53 +00:00
Wayne Davison
1a515b494b
In option descriptions, the short form of --copy-dirlinks is -k, not -K.
2007-03-15 23:17:26 +00:00
Wayne Davison
33cd9ad54b
I managed to duplicate the hang, so hopefully it's fixed now.
...
Let's restore the full test suite and find out.
2007-03-14 00:22:05 +00:00
Wayne Davison
896f046f79
Fixed more potential hang cases for incremental recursion.
2007-03-14 00:20:44 +00:00
Wayne Davison
b9d5eea5c6
Next step: allow protocol 30 to be used, but disallow the use
...
of incremental recursion. If this still avoids the hanging, I
will need to try to track down how inc_recurse can deadlock.
2007-03-14 00:02:39 +00:00
Wayne Davison
5275029d11
Added an explicit option that will disallow the use of
...
incremental recursive mode (and an option to negate that).
2007-03-14 00:00:01 +00:00
Wayne Davison
271aeaa49e
Next step: see if the hang returns with a partial restoration of
...
protocol 30.
2007-03-12 18:05:03 +00:00
Wayne Davison
91763a9c2f
It looks like some of the compile-farm hosts are hanging on this test,
...
so let's try some remote-diagnostics (let's try protocol 29 first).
2007-03-12 04:56:15 +00:00
Wayne Davison
18438f0bac
Changed the sending of device major/minor numbers in protocol 30 to
...
take fewer protocol bytes and use one less bit in the XMIT_* flags.
2007-03-12 04:33:30 +00:00
Wayne Davison
057f649fc9
When a hard-linked device is duplicating the data from the prior
...
device, we need to allocate 2 extra option slots for the device
numbers.
2007-03-11 22:02:26 +00:00
Wayne Davison
88467ec444
Don't enable ACLs on darwin until we figure out how they work.
2007-03-11 14:25:00 +00:00
Wayne Davison
1b897d50a9
Leave out two unneeded sys*_fd() functions due to their using
...
a function & struct that is not available in the rsync code.
2007-03-11 07:31:17 +00:00
Wayne Davison
d4a013edb7
We now compile on systems where ENOTSUP is not defined.
2007-03-11 07:21:07 +00:00
Wayne Davison
545584cb9a
Leave SUPPORT_ACLS undefined instead of giving it a 0 value.
2007-03-11 07:13:11 +00:00
Wayne Davison
e5abce8e0d
Silence a compiler warning about a printf() %d mismatch.
2007-03-11 06:50:48 +00:00