Re: [Chicken-users] compiling an interpeter script, but remaining interpretable

[felix winkelmann]

On 5/18/06, Graham Fawcett <address@hidden> wrote:
> Hi folks,
>
> I just wanted to check my thinking before I began a fool's errand that
> would be hard to back out of. I have a small application which runs as
> a interpreter script. The script, script.scm, depends on script-lib.scm,
> which in turn depends on misc.scm.
>
> I'd like to compile the script, but for debugging purposes I'd still
> like to be able to run it in an interpreter. To that end, I've
> considered adding declarations like this to my source files:
>
> ;; script.scm
> (eval-when (eval) (use script-lib))
> (eval-when (load) (declare (uses script-lib)))
>
> ;; script-lib.scm
> (eval-when (eval) (use misc))
> (eval-when (load) (declare (unit script-lib) (uses misc)))
>
> ;; misc.scm
> (eval-when (load) (declare (unit misc)))
>
> It's feasible, I think, but it seems rather cumbersome. Is there a
> simpler idiom for keeping the script interpretable, but still
> preparing it for compilation? I feel like I'm missing something obvious...

You can also do something like this:

#+csi <expr eval'd when in the interpreter>

#+(not csi) <expr compiled>

"#+X Y" is the same as `(cond-expand (X Y)(else))'


cheers,
felix