updating_basis set in wrong place in recv_files?

[Matt McCutchen]

On Sun, 2007-10-28 at 12:41 -0700, Wayne Davison wrote:
> On Sun, Oct 28, 2007 at 11:02:35AM -0400, Matt McCutchen wrote:
> > Shouldn't updating_basis be enabled for FNAMECMP_BACKUP as well as
> > FNAMECMP_FNAME?
> 
> No, we use the backup file as the basis file when --inplace is combined
> with --backup.

But according to your own comment on the definition of updating_basis,
it is true when "We're either updating the basis file or an identical
copy", and indeed this is the relevant condition for optimizing writes.

Matt