[cells-gtk-devel] cells-gtk and sbcl

Peter Denno peter.denno at nist.gov
Mon Mar 14 14:53:35 UTC 2005


On Monday 14 March 2005 08:48, rm at fabula.de wrote:
> On Sun, Mar 13, 2005 at 09:24:08PM -0500, Peter Denno wrote:
> > Hi,
> >
> > cells-gtk does not currently run sbcl.

What I should have said was that I (the maintainer) have no knowledge of it 
running on SBCL. The one time I tried it, it didn't work, and I expected that 
there would be plenty of problems.

>
> Ahhrgh, conflicting information - from c.l.l:
> |  Fred Gilham <gilham at snapdragon.csl.sri.com> writes:
> |
> |  ...
> |
> |  Cells-gtk builds very easily using sbcl, though, and is interesting in
> |  its own way.
>
> whom to belive?
> In my follow-up i mentioned some problems (missing package :lisp for
> example)
>
> and Peter Scott replied:
> | I got the same problem, but I just changed :lisp to :common-lisp and
> | went from there. I still had some issues (a warning about the lack of
> | libcellsgtk.so and some places where it apparently tried to redefine
> | constants, but nothing major) but I was able to get gtk-ffi.lisp
> | compiled with SBCL 0.8.16 on x86.

OK, so it compiles....

> |
> > AFAIK no one has looked into it yet.
> > I  hope to look into making it run sbcl, but I will not have time until
> > May. Maybe someone else will get to it before then.
>
> Well, i hade a quick glance at it (and spotted some problems) - i might
> have a closer look at it soon (but i'm about to get sick, so i can't
> promise any time- frame). Some problems found so far:
>
>  - uses package :LISP (should most likely be :COMMON-LISP)
>
>  - does attempt to redefine some constants (but redifines to the same value
>    so not too big a problem - still, it should be fixed).

Send it along...

>
>  - has _hardcoded_ pathes in the  test for library locations!

Really? Where is that?

>
>  - the test '(directory "/path/glob*")' doesn't seem to work ...

I don't know what this refers to.

>
>  - complains about undefined alien functions (tons of'em).

Which sounds to me like it isn't going to work.

> Except the last problem i have already fixes ..,
>
>  HTH Ralf Mattes
>
> > cells-gtk compiles easily on cmucl.

He didn't say it runs. But if you just enjoy compiling it ... ;^)

If you do get it to run, please let me/us know.


-- 
Best Regards,
 - Peter



More information about the cells-gtk-devel mailing list