AW: [clisp-list] Re: ffi question: c-array

[Hoehle, Joerg-C. <Joe...@t-...>]

Hi,

Yaroslav Kavenchuk wrote:
>> [3]> (ffi:def-c-type t2 (ffi:c-array ffi:character s1))
>S1 is not evaluated.

>try
>(eval `(ffi:def-c-type t2 (ffi:c-array ffi:character ,s1)))
That will probably not work with other forms using this type in compiled mode.

>or
>(ffi:def-c-type t2 (ffi:c-array ffi:character #.s1))
I generally dislike use of #. but that's just a personal bias.

Use macros when you need an extra round of evaluation.

(eval-when (:compile..etc.)
  (defconstant s2 100))

(macrolet ((my-type (n) `(def-c-type t2 (ffi:c-array ffi:character ,n))))
  (my-type s2))

While this looks like overkill or convoluted, it has none of the above drawbacks; it is fully compatible with compilation, doesn't give a hit on the EVAL/EVIL counter and is as top-level as the original (def-c-type ...) form.

Regards,
	Jorg Hohle.