"-b --suffix '' --delete --backup-dir /path/" combination does not act as expected

jw schultz jw at pegasys.ws
Fri Aug 1 06:37:59 EST 2003

On Thu, Jul 31, 2003 at 11:36:57AM -0700, Wayne Davison wrote:
> On Mon, Jul 28, 2003 at 03:55:28PM -0700, jw schultz wrote:
> > Any other developer's thoughts?
> The original logic looks faulty to me in another way:  if a backup
> directory was specified, the default suffix turns into nothing.  The
> other code currently handles this in a really weird way by setting a
> "suffix_specified" flag rather than making sure that the backup suffix
> is set right.  Thus, this code in receiver.c is not dealing with the
> backup suffix correctly when a directory was specified but a suffix was
> not.
> I'm going to check-in some better code into options.c that will improve
> this situation.  It will also pre-compute the length of the backup
> suffix so that we don't have to keep calling strlen() on a constant-
> length string.
> As for the actual patch to receiver.c, I think your counter-patch looks
> pretty good, but I think it can be simplifed a bit more.  I'll have more
> time to look at this later.

Have at it!

The only real reason for my counter-patch was the state of
that code.  One shouldn't have to puzzle over that few lines
of code to comprehend what it is doing unless it is critical
fast-path.  I'd pretty much refactored the conditional by
the time i was sure what it was doing to even evealuate the
patch.  If i have to do that the code is more difficult to
support that in needs to be.  A refactoring of outer the
loop seemes in order; with a lot more whitespace.

	J.W. Schultz            Pegasystems Technologies
	email address:		jw at pegasys.ws

		Remember Cernan and Schmitt

More information about the rsync mailing list