Re: logtape.c stats don't account for unused "prefetched" block numbers

On Mon, 2020-09-14 at 15:54 -0700, Peter Geoghegan wrote:
> Maybe the LogicalTapeRewindForRead() inconsistency you mention could
> be fixed, which would enable the simplification you suggested. What
> do
> you think?

Yes, it was apparently an oversight. Patch attached.

RC1 was just stamped, are we in a sensitive time or is it still
possible to backport this to REL_13_STABLE?

If not, that's fine, I'll just commit it to master. It's a little less
important after 9878b643, which reduced the overpartitioning issue.

Regards,
    Jeff Davis