I was thinking of deploying a lisp based application(clisp guys).
Anyway, if you have an image setup and you have the executable lisp
interpreter and have it load the image on startup. Let us say that you
want to upgrade some of the functions in the original image, because of
bugs or whatever. Could you in theory override the functions in the
previous image by loading new scripts or possibly another image.
What is the best way to write pluggable software in lisp.
Has this worked for anyone?
Ramza Brown <············@gmail.com> writes:
> I was thinking of deploying a lisp based application(clisp
> guys). Anyway, if you have an image setup and you have the executable
> lisp interpreter and have it load the image on startup. Let us say
> that you want to upgrade some of the functions in the original image,
> because of bugs or whatever. Could you in theory override the
> functions in the previous image by loading new scripts or possibly
> another image.
Not another image, since clisp can only load an image at launch time
(and I don't know any implementation that can merge two images).
But indeed it's easy to boot the current image, then compile and load
*.lisp or load *.fas files (the "patches"), then save the new image.
clisp -M image-1.0.mem -x '(progn (load (compile-file "patches.lisp")) \
(ext:saveinitmem "image-1.1.mem" ...))'
clisp -M image-1.1.mem
> What is the best way to write pluggable software in lisp.
Using LOAD. Lisp is all plug-ins.
(dolist (plugin (directory "PLUGINS:*.lisp")) (load plugin))
> Has this worked for anyone?
Yes.
Actually, saving images is so easy that it can be used to save
snapshoots of long computations to be able to restart them from the
last snapshoot instead of from the start.
--
__Pascal Bourguignon__ http://www.informatimago.com/
In deep sleep hear sound,
Cat vomit hairball somewhere.
Will find in morning.