Implementing "extremes"

[John Darrington]

As part of my (re)implementation of EXAMINE, I have a decision to
make.  I can store extreme values in one of two ways:

1. Keep them in dynamically allocated memory (which is what the
   current implementation does).

2. Use a casereader.

My first thought was that the casereader would be the prefered
option.  However, whilst it's more robust, it will require an extra
data pass and therefore slow down the command.  Keeping them on the
heap will be faster, but could result in out-of-memory conditions if
huge list of extremes are requested.

So I'm tending towards option  1 at the moment, since in the most
normal uses it will be faster, and could only cause a problem if
somebody does something silly like:

EXAMINE x
        /STATISTICS = EXTREME (1000000000000000).


Does anyone disagree with this line of thought?

J'



-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for public key.

signature.asc
Description: Digital signature