Wayne Davison
86fc7e6880
Got rid of a superfluous newline.
2005-11-10 16:47:52 +00:00
Wayne Davison
1f86fcf5dc
The combination of --update and --progress now outputs progress
...
information during the checksum scan of the already-sent data.
2005-07-30 16:13:19 +00:00
Wayne Davison
6cc1198288
Support new --append option.
2005-07-28 01:46:12 +00:00
Wayne Davison
0394e34a69
Moved the end_progress() call from match.c to sender.c so that we
...
report progress on 0-length files when pushing files (the receiver
already called it, so we already produced progress on a 0-length
file when pulling).
2005-03-05 17:51:23 +00:00
Wayne Davison
37a5644576
One arg to matched() is supposed to be an int32.
2005-02-14 22:48:28 +00:00
Wayne Davison
7cacd47edd
Changed the size_t variables into int32 variables.
2005-02-14 08:12:22 +00:00
Wayne Davison
a36ffd3910
Changed the checks for --inplace to use the new updating_basis_file
...
variable.
2005-01-17 23:11:13 +00:00
Wayne Davison
a06b419d42
Use an int32 for the each block-size variable.
2005-01-01 21:08:14 +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
7560c17adc
We call map_ptr() with a data range than includes any unmatched data
...
(which we might need to reference again) in addition to the current
rolling-checksum block (this prevents the unmatched data from being
lost when we slide the buffer and read more data).
2004-08-03 08:05:23 +00:00
Wayne Davison
a3221d2ac1
My version of Mark Curtis's --inplace option.
2004-07-16 20:06:24 +00:00
Wayne Davison
b9f592fbf5
My modified version of Chris Shoemaker's improved batch-file handling.
2004-07-15 02:20:08 +00:00
Wayne Davison
a04d77bcbc
Use want_i instead of last_i+1.
2004-07-07 08:38:40 +00:00
Wayne Davison
066a844c4e
Changed a "for" loop into a "do ... while" loop.
2004-07-04 08:07:23 +00:00
Wayne Davison
5e252dea4b
Optimized away a loop in hash_search().
2004-07-02 23:35:30 +00:00
Wayne Davison
ba582f753a
Changed sum_init() to take a seed value as an arg instead of always
...
using checksum_init. This fixes an authentication problem in server
mode (as pointed out by Craig Barratt).
2004-05-21 08:27:04 +00:00
Wayne Davison
edecdad54d
If we got a read-error on a file, make sure that the whole-file
...
checksum we send to the receiver is wrong (so they won't save
the bogus file).
2004-05-13 06:46:20 +00:00
Wayne Davison
0e36d9da42
More variable-size cleanup.
2004-01-03 19:28:03 +00:00
Wayne Davison
da9d12f5d9
Made the types used in the sum_buf and sum_struct structures consistent
...
with the variables in the code that manipulates these values.
2004-01-03 18:37:41 +00:00
Wayne Davison
64c3523a6f
Got rid of /* dw */ comments.
2004-01-02 07:42:20 +00:00
Wayne Davison
58cadc8608
Merged in the security fixes from 2.5.7.
2003-12-06 21:07:27 +00:00
Wayne Davison
c95dcb3935
Cast various s->blength references inside of MIN() macros to OFF_T
...
so that compilers don't warn about signed/unsigned comparisons.
2003-08-22 21:26:08 +00:00
Wayne Davison
1e4f48d6c4
Whitespace tweaks.
2003-08-22 05:51:42 +00:00
Wayne Davison
16417f8b9d
Only call show_progress() and end_progress() if do_progress is set.
2003-07-08 16:49:10 +00:00
J.W. Schultz
fc0257c9fd
Now support dynamic per-file checksum2 size.
...
This pushes protocol version to 27
2003-04-10 01:13:30 +00:00
J.W. Schultz
bc63ae3f13
Removing vestigial support for protocol versions prior to 15.
2003-03-31 17:28:34 +00:00
Martin Pool
e35080cede
long int and signedness fix from Nelson Beebe
2002-04-09 06:11:06 +00:00
Martin Pool
538ba24fd7
Doc match_sums.
2002-04-08 08:33:13 +00:00
Martin Pool
f5f95a38c4
Document the perversely-named matched() function, which is called even
...
if we have not matched!
2002-04-08 08:23:50 +00:00
Andrew Tridgell
0e9480317d
the signed/unsigned change seems to have caused a logic bug on some
...
systems (only those without large file support perhaps?)
this fixes it
2002-02-03 01:38:39 +00:00
Martin Pool
a261989cda
More signedness fixes; should be harmless.
2002-01-25 23:07:33 +00:00
Martin Pool
6902ed178b
Merge across rsync+ patch; add a little documentation to the manpage. More documentation would be better.
2001-08-14 02:04:47 +00:00
Andrew Tridgell
5e12ce1186
fix segv bug in --progress handling
2000-01-25 13:16:42 +00:00
Andrew Tridgell
166aa72332
patch from David Murn to make sure the final 100% is always printed
...
when using --progress
2000-01-25 06:39:33 +00:00
Andrew Tridgell
5f808dfbd7
fix a problem with files > 2GB
...
(thanks to T.J.Adye@rl.ac.uk )
2000-01-23 12:30:34 +00:00
Andrew Tridgell
536541d52b
moved the block length mismatch code to another part of the loop.
1999-01-05 06:43:59 +00:00
Andrew Tridgell
496d9272c1
don't try to match checksums of two blocks which are of unequal
...
size. This explains the high false_alarms rate that I saw for one of
the sample data files used in my thesis.
The bug was harmless as the strong checksum easily caught all the
false matches but it's been bugging me as I couldn't explain it :)
1999-01-05 06:31:58 +00:00
Andrew Tridgell
34d3eed462
fixed a bug in the adjacent target optimisation
1999-01-05 01:57:13 +00:00
Andrew Tridgell
923fa97808
an optimization that tries to make rsync choose adjacent matches if
...
multiple matching blocks are available. This make the run-length
coding of the output more efficient.
1999-01-05 01:15:32 +00:00
Andrew Tridgell
4440b8aa3f
no longer use mmap() in rsync because of the risk of a SIGBUS when
...
another program (such as a mailer) truncates a file.
To offset the speed loss I have rewritten the map_ptr() code to make
much better use of read().
1998-12-30 14:48:45 +00:00
Andrew Tridgell
1309d90dde
fixed a bug handling files larger than 2GB
1998-11-16 03:53:43 +00:00
Andrew Tridgell
b536f47e3c
- don't show "created directory" message unless verbose is selected
...
- check for null buf in show_progress
1998-07-02 02:08:55 +00:00
Andrew Tridgell
eb86d661d7
added --progress option which shows the progress of transfers. This
...
gives bored users something to watch.
1998-07-02 00:47:13 +00:00
Andrew Tridgell
100e5241b0
the tag table should be of type int* not tag*.
...
This bug resulted in rsync being much less efficient that it could be
for files with more than 64k blocks. With the adaptive block size code
giving a maximum block size of 16k this means that files larger than
1GB were handled very inefficiently. The transfer was still accurate,
just slow.
1998-06-03 02:47:52 +00:00
Andrew Tridgell
a800434a82
added --stats option for verbose stats on the file transfer
1998-06-01 03:42:14 +00:00
Andrew Tridgell
4c36ddbeec
heaps of cleanup in the io code.
...
we no longer use non-blocking IO, instead it uses select a lot more,
being careful to always allow for reading whenever a valid read fd is
available and chcking timeouts.
also split the file io calls into fileio.c
1998-05-27 12:37:22 +00:00
Andrew Tridgell
f900f5fe71
removing an unused variable
1998-05-22 12:06:25 +00:00
Paul Mackerras
5914bf15d2
Update to use the new zlib-1.1.2 code.
...
The compressed token code now handles the null (-2)
token from the match logic.
1998-05-22 06:58:52 +00:00
Andrew Tridgell
45f133b976
this fixes two problems:
...
1) handle 64 bit file offsets in the token code. I wonder how large
bit files worked up till now?
2) send a null token when we have passed over a large lump of data
without finding a token match. This reduces the number of IOs
considerably as it removes the need for seeks/reads on the checksum
calculation and literal send code. This is not enabled yet for the
compressed case as the deflate token code can't handle it yet.
1998-05-22 01:53:02 +00:00
Andrew Tridgell
9486289ce4
some large systematic changes to make socket conversion easier. The
...
biggest one is the use of rprintf() instead of fprintf()
1998-05-07 14:50:22 +00:00
