Cells, a dataflow extension to CLOS

Michael Raskin | 29 Jan 2010 08:28

Cells-devel: allow ECL support

		Hello.
	I have some interest in simplifying installation of Cells-GTK on as
wide range of platforms as possible (from the "expansive" point-of-view:
I have a perfectly fine installation of Cells-GTK on SBCL on Linux on my
notebook).
	The attached patch enables ECL support in Ramarren cells fork. It
actually does nothing except noting that ECL uses :clos package for MOP.
I hope it will be included in master branch somewhen soon..
	(cells::test-cells) gives no Cells errors (dribble errors do not count,
right?).

	Thanks
Michael Raskin

diff -U3 -r Ramarren-cells-beb09a0/cells.asd /home/raskin/.sbcl/site/Ramarren-cells-beb09a0/cells.asd
--- Ramarren-cells-beb09a0/cells.asd	2010-01-29 09:49:41.000000000 +0300
+++ /home/raskin/.sbcl/site/Ramarren-cells-beb09a0/cells.asd	2010-01-20 20:20:11.000000000 +0300
 <at>  <at>  -1,6 +1,6  <at>  <at> 
 ;;;; -*- Mode: Lisp; Syntax: ANSI-Common-Lisp; Base: 10 -*-

-#+(or allegro lispworks cmu mcl ccl clisp cormanlisp sbcl scl ecl)
+#+(or allegro lispworks cmu mcl ccl clisp cormanlisp sbcl scl)
 (progn
 (declaim (optimize (debug 3) (speed 3) (safety 1) (compilation-speed 0)))

diff -U3 -r Ramarren-cells-beb09a0/defpackage.lisp /home/raskin/.sbcl/site/Ramarren-cells-beb09a0/defpackage.lisp
--- Ramarren-cells-beb09a0/defpackage.lisp	2010-01-29 09:51:58.000000000 +0300
+++ /home/raskin/.sbcl/site/Ramarren-cells-beb09a0/defpackage.lisp	2010-01-20
20:20:11.000000000 +0300

(Continue reading)

headers

Ramarren | 29 Jan 2010 09:40

Re: Cells-devel: allow ECL support

Hello,

I pushed your patch to GitHub. I also tried running cells-gtk3 demo
application on ECL, and it ran after I eliminated remains of UFFI
dependency, since cffi-uffi-compat doesn't load. I pushed those fixes
also (although I'm not sure if ALLOC-COL-TYPE-BUFFER shouldn't be just
deleted, since it seems to be either exported or used), and changed (/
3 0) in test application, since ECL just refused to compile it.

The only problem left is that cl-glut sometimes throws a package error
(something with exporting SPECIAL), but it works on RETRY.

Regards,
Jakub Higersberger

2010/1/29 Michael Raskin <f9cef2aa <at> yandex.ru>:
>                Hello.
>        I have some interest in simplifying installation of Cells-GTK on as
> wide range of platforms as possible (from the "expansive" point-of-view:
> I have a perfectly fine installation of Cells-GTK on SBCL on Linux on my
> notebook).
>        The attached patch enables ECL support in Ramarren cells fork. It
> actually does nothing except noting that ECL uses :clos package for MOP.
> I hope it will be included in master branch somewhen soon..
>        (cells::test-cells) gives no Cells errors (dribble errors do not count,
> right?).
>
>        Thanks
> Michael Raskin
>

(Continue reading)

headers

Michael Raskin | 29 Jan 2010 09:59

Re: Cells-devel: allow ECL support

On 01/29/2010 11:40 AM, Ramarren wrote:
> Hello,
> 
> I pushed your patch to GitHub. I also tried running cells-gtk3 demo
> application on ECL, and it ran after I eliminated remains of UFFI
> dependency, since cffi-uffi-compat doesn't load. I pushed those fixes

I am currently convincing CFFI maintainers to accept either of two
two-line fixes for that UFFI dependency..

> also (although I'm not sure if ALLOC-COL-TYPE-BUFFER shouldn't be just
> deleted, since it seems to be either exported or used), and changed (/
> 3 0) in test application, since ECL just refused to compile it.

I guess you can just replace it with eval. Although that may lead to ECL
wanting GCC.

> The only problem left is that cl-glut sometimes throws a package error
> (something with exporting SPECIAL), but it works on RETRY.

I currently fail to understand what they are doing wrong. I fail to make
sense of Common Lisp standard meaning of their shadow statement.