More on PowerLoom and SBCL

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

- -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

- - -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Folks,

I determined what's going on.

SBCL adopts a "declarations are assertions" stance, which it claims to 
inherit from CMU CL. For optimizations at SAFETY > 0, type declarations 
are checked at runtime. Changing PowerLoom's optimization settings to 
SAFETY 0 solved the problem.

What's interesting to me, though, is that this means that there were a 
bunch of other typechecks in the code at SAFETY 1 that passed! Nice 
work! So you may want to consider doing whatever is necessary in order 
to get "length" on kv-cons to be type-safe. As it stands, the problem 
is that NULL, which is set to :NULL_VALUE, is, of course, not of type 
KV-CONS, and so "length" chokes on it.

Hope this helps,
Paul Snively
- - -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iEYEARECAAYFAkBndLYACgkQbot1wzHBQBVrLQCeLceuxGrwQJoKYIOqJNfWGWiL
4fsAoIi8dAevL+pRTIXas58KIkDuOs5x
=PHs9
- - -----END PGP SIGNATURE-----

- -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iEYEARECAAYFAkBndMwACgkQbot1wzHBQBUvSwCfWxVkNifgx5JJ8aGycMAM6ne5
+fYAni1zkhsDRScHw/NR+JwKD7BNtCU3
=pFsp
- -----END PGP SIGNATURE-----

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iEYEARECAAYFAkBnktAACgkQbot1wzHBQBWW0ACfQrMmN82GHgobdgX1qHu76tRF
43UAniadl3VKySsmmo3wAfuptYRYAyKn
=fylQ
-----END PGP SIGNATURE-----