2009/7/30 Christophe Rhodes <csr21@...>:
> I'm slightly intrigued by this patch: I just wonder why the
> optimization is only performed for 1D simple arrays (as opposed to 1D
> arrays in general, where the element type is known at compile-time).
> It strikes me that I'm more likely to write
> (coerce x 'vector) ; or 'bit-vector
> than to remember to put simple- on the front, and we should be able to
> perform the same optimization at no extra cost, no? Something like
> (if (typep x `(,(if simplep 'simple-array 'array) ,eltype (*)))
> (replace ...))
> I think even if we don't do the above, we should special-case
> base-string, vector, bit-vector and string as well as simple-string,
> because they're particularly tempting (short) type names.
A very good point. I've commited a somewhat extended version as 220.127.116.11, though
it does not deal with VECTOR yet -- but it does with STRING, and (ARRAY X (*)).