- Make "len" parameter of do_punch_hole an OFF_T.
- Clear sparse_past_write in sparse_end(), otherwise when write_sparse()
is called for the next file, do_punch_hole() will be called with a pos
that's not actually the current position in file, causing it to fail.
Previously files were hashed in blocks of CSUM_CHUNK (64) bytes. This
causes significant overhead. The CSUM_CHUNK define cannot be changed as
md5.c depends on it, but there is no obvious reason to use it in
file_checksum(). By using CHUNK_SIZE (32 kB) instead, in some test
cases throughput more than doubles.
- Add the zlibx (external-code compatible) compression name.
- Re-enable zlib support with the external library so it can be
tried as a fallback if zlibx isn't available.
- Add --compress-choice=STR (aka -zz=STR) option.
- Make --cc=STR an alias for --checksum-choice=STR.
- Hook up the new compression negotiation logic.
- Add/improve --debug=CSUM2 messages.
- Add an "xxh64" alias for "xxhash" name because we should be
getting a few more xxhash variations in the future.
- Tweak the matching code to handle entries that have multiple
names.
- Tweak some of the vars/defines.
Avoid a newline issue during the output of --DEBUG=CSUM info from
both the server and the client -- we need to output the full message
with its newline as much as possible.
Originally created by Marc Bevand and placed in the public domain.
Enable/disabled via the same --enable-simd configure switch as
the rolling checksum optimizations.
Additionally restructures build switches and defines from SSE2 to SIMD,
to allow potential reuse should patches become available with SIMD
instructions for other processor architectures.
(Some minor tweaks of Jorrit's patch to avoid requiring GNU make and to
avoid C++ comments in .c files.)
- Add checksum negotiation to the protocol so that we can easily add new
checksum algorithms and each will be used when both sides support it.
- Increase the size of the compat_flags value in the protocol from a
byte to an int.
