NoLocks // 1.9.8b10 adoption of record locks. Feedback

[jda]

>
>YES.
>
>This was made on request of Erne :-)
>
>If you set all your cursors to NoLocks, then you get EXACTLY the same
>behavior as in current 1.9.7
>

I'm afraid I ran into an exception for this (literally and 
figurtively). I made one cursor that returned one field of all 
records i a database. A second cursor then iterated through the first 
cursor, getting all fields for each record and modifying them. This 
worked before, but returned a nil for the second cursor with b10. I 
then added kNoLocks (= 1) to each SQLSelect -- still got a nil 
cursor. I solve the problem by changing the logic of the method 
(which was flawed, anyway) so that I only needed one cursor. Now it 
works again.

Jon