oracle-sql quandary (2)

hello again;

> i am no somewhat more than idly considering whether clsql is viable
> in mcl.
>

was "supposed" to be "i am [now] somewhat more...", but who knows.

another situation looks suspicious.

3. the uffi record access translation try to divine the correct formulation for
record accessors. there's this function primitive.lisp#convert-from-uffi-type,
which tries to decide things like, how to translate a '(* unsigned-char) field
reference. only thing is, the formulation was not recognized, and, in addition,
there seems to be only the precedent, to return :pointer field designator
(although that is for allocation contexts only), which really shouldn't work to
dereference a string field.

i've patched it to recognize this, to specify :string instead of :pointer, but
i'm suspicious.

has anyone actually used this stuff (that is, uffi) for mcl?

thanks,
...