>>>And another note: there's no such thing as EndOfMemo character - the memo block has a header which denotes its status and length. Fox never analyzes the content of the memo, and doesn't really care what sort of butes you stuff inside, else we'd never be able to Save To Memo, Append Memo From (file) etc, i.e. it wouldn't allow arbitrary binary content.
>>
>>There is an end of memo word -- 0x1a1a. This is a hold over from dbase.
>>
>>The following site has detailed descriptions:
>>
>>http://www.e-bachmann.dk/docs/xbase.html
>
>It may be useful for some emulation mode, but try this:
>create curs xx (m m)
>appe blank
>repl m with repl(chr(26), 200)
>?len(m)
>?m
>
>I've tried this in FPD2.6a, and it gives me 200, and eight lines of right arrows, which means it doesn't care about the content, and the double eof marker is not treated as end-of-memo. It may mean that in tables which have the type byte in the header designating them as dBase tables.

Yes, I was just observing that the characters are still appended, even though not used. The current scheme is superior to the old dBase scheme (shocking!).