Re: defsystem: initially-do & finally-do ignored?

[Marco A. <ma...@gm...>]

Hi

Yes.  It is a buglet in preprocess-component-definition.  The
following version is correct.



(defun preprocess-component-definition (definition-body)
  `(list* ,@(loop for slot in *component-evaluated-slots*
=09          for value =3D (getf definition-body slot)
=09          when value
                    do (remf definition-body slot)
                    and nconc `(,slot ,value))
=09  ,@(loop for slot in *component-form-slots*
=09=09  for form =3D (getf definition-body slot)
=09          do (remf definition-body slot)
                  nconc `(,slot (lambda () ,form)))
=09  ',definition-body))

Sam,  I am working from gmail and probably you will have to ok this email.

Also,  I am travelling right now.  Could you make sure the new version
is in CVS?

Thanks

Marco




On 3/9/06, Sam Steingold <sd...@po...> wrote:
> The following message is a courtesy copy of an article
> that has been posted to gmane.lisp.clocc.general as well.
>
> I just noticed that initially-do and finally-do are ignored:
> $ cd clocc/src/cllib
> $ make system
> no longer call system-source-size, at does not create and compile auto.li=
sp
>
> I think the latest change is the culprit:
>
> revision 1.102
> date: 2005/09/05 18:39:21;  author: marcoxa;  state: Exp;  lines: +425 -1=
53
> Made several changes to ensure that Sam Steingold modifications to
> control evaluation of definition slots do not interfere with
> subcomponents.
>
> Cleaned up the code in several places.
>
> Added RUN-SHELL-COMMAND (modified from UFFI and ASDF to make it more port=
able).
> The original versions were by Dan Barlow and Kevin Rosenberg.
>
>
> --
> Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 4 (Sten=
tz)
> http://www.honestreporting.com http://www.savegushkatif.org
> http://www.mideasttruth.com http://ffii.org http://www.jihadwatch.org
> There are many reasons not to use Lisp - but no good ones.
>


--
--
Marco Antoniotti
New York, NY, USA