More fun with nullmaps

This is a multi-part message in MIME format...
--
To: mi400@midrange.com
From: jamesl@hb.quik.com
X-Advert: http://emumail.com
Reply-To: jamesl@hb.quik.com
Date: Mon, 12 Nov 2001 12:58:41 AST
X-Mailer: EMUmail
Subject: More fun with nullmaps

Ladies and gentlemen, fellow geeks, geekettes, and gurus:

So far, I've
managed to determine that, in an input buffer for a file opened
for all
modes (i.e., Input, Output, Update, and Delete bits are set), there's a
gap
between te record and the nullmap, and I've empirically determined that
it's
eleven bytes, plus one byte per keyfield, plus one byte per key byte. So

far, though, I can't explain the gap, except that it appears to have nothing
to
do with variable length, since it occurs in files with no variable length

fields, and none of the bits for including a "variable length access
component"
are set.

Unfortunately, none of the obvious places (i.e., in
the input buffer with the
usual gap, in the output buffer with no gap [like
Mr. Gaunt's demo program], in
the output buffer with the usual gap) for
sticking a nullmap seem to work for
writing new records to the file. They
all seem to be ignored.

Any suggestions, either for explaining the gap, or
for where the nullmap should
go when writing new records under these
conditions?

Apologies for any funky
formatting

--
J.Lampert