(moved from hard_link_cluster()) that checks the inode/dev of an
  existing file and either does nothing (when properly linked), or
  removes the existing file before calling hard_link_one().
- Changed hard_link_check() so that it checks if the master item is
  marked with FINISHED_LINK (in its hlindex int), and if so, it
  finishes the current file by calling maybe_hard_link() (and
  marking it as finished too).  Otherwise it marks the item as
  skipped by setting hlindex to SKIPPED_LINK.
- Fixed the outputting of an error in hard_link_one() so that the
  message is always output when the caller did not request "terse"
  processing.
- Changed hard_link_cluster() so that it marks the master item with
  FINISHED_LINK and then links only the skipped items (marking them
  with FINISHED_LINK too).

- Got rid of a duplicate error after the hard_link_one() call.

to copy its contents.

single-file transfer, we need to make sure that the file's
name is the generic-transfer name and not a local name.

belong.

from the latest GPL file on the web.

failing.

$fromdir.

  when sending a single file.
- Made a couple variable improvements in link_idev_data().
- Got rid of the non-NULL check of hlink_list in init_hard_links()
  because is enough to just check file->link_u.links these days.

  logfile_reopen().
- Changed log_close() into logfile_close().
- Improved the logic in logit().

is used and we're pushing files to a remote system.

- When we're handling --only-write-batch on the client side, we now
  send all index/header info to the receiver (not just to the batch
  file, like we do with the xfer data) so that the connection cannot
  timeout.

when we're on the server side and --only-write-batch was specified.

"log file".

syslog (rather than trying to output an error on stderr,
which would usually be lost, and then exiting).

for directories.

the challenge string base64-encoded (instead of forcing the caller
to handle this).  Also improved a couple comments.

- Don't define SUPPORT_HARD_LINKS unless HAVE_LINK is defined.

  (with appropriate negation).
- If write_batch is < 0, we write out the file-transfer data to
  just the batch file, not the socket.

  (with appropriate negation).
- If we get a transfer when write_batch is < 0, discard it.

(with appropriate negation).

  now that write_batch is no longer forced to 0 for the server).
- If write_batch is < 0, force "dry_run = 1" (but only after we've
  finished any appropriate write-batch file creation).

- Handle the new --only-write-batch=FILE option.

changing --write-batch).

- Don't log an auth-failed error -- auth_server() now handles that.

- Pass the hostname in to auth_server().
- Generate a unique error for each failure type in auth_server() so
  that the log-file contains why the authorization failed.
- Don't use sscanf() to parse the client's auth-challenge response.