growing too much if the sender is much faster than the receiver

where a rsync receive process could hang around after the connection
has gone away if a fatal error occurs (such as someone killing the
other end)

this stops "rsync -v" causing a segv.

generation

  daemon
- update spec file

- distinguish downloads and uploads in the syslog messages
- add stats on uploads in syslog

then we supply a zero filled buffer at the end and rely on the
checksum to cause a retry. This is really the best we can do as there
is no correct semantics for copying a changing file!

provide hours of fun for those people who like to tune their systems
to the utmost degree.

starts so we know the syslog facility for the "starting" message

- some more paranoid buffer size checks
- separate open syslog call
- handle systems without LOG_NDELAY

chance of a buffer overflow

wildcards) to a rsync server. For example you can do:

rsync -avz samba::'ftp/pub/samba/README ftp/pub/samba/*.gz' .

a bit more efficient (less system calls)

necessary on the server

pages

- documented the RSYNC_PASSWORD environment variable

This is useful for mirroring a web site when you don't want users to
mirror everything.

fixed SIGCLD -> SIGCHLD

fixed formatting of man page (thanks to Karsten)