SBCL port



On Sat, 2003-02-22 at 12:00, Wolfgang Jenkner wrote:
> I have checked in my SBCL patch without further changes and recompiled
> Maxima with the current CVS version SBCL 0.7.12.52 and also with Clisp
> but not with GCL or CMUCL (if you have one of these two please try to
> recompile Maxima and report success or failure).
> 

Excellent. I verified that GCL and CMUCL still compile and pass "make
check". Good work.


> If you want to compile Maxima with SBCL just do the usual
> 
> ./bootstrap; ./configure --enable-sbcl; make
> 
> As for the current state of the Maxima/SBCL combination, I just copy
> here what I wrote in previous messages:
> 
> SBCL will pass the test suite except where `kill(rules)' is expected
> to work correctly.  It doesn't, but this is Maxima's fault, not SBCL's
> (Maxima assumes that compiled rules are built-in and therefore
> shouldn't be removed, but SBCL compiles everything, so nothing is
> removed).  Note that this causes some tests to fail with `circular
> rule attempted'.
> 

Hmm. That's probably worthy of a bug report. I think there are several
problems with kill.

> Known bugs are various pathname (?) annoyances.

What are those, exactly? I've long been afraid that I haven't treated
pathnames properly in the code I have written. If so, please enlighten
me.

-- 
James Amundson <amundson@fnal.gov>