[Openmcl-devel] 32-bit x86 port status

[mikel evins]

On May 1, 2008, at 9:06 PM, R. Matthew Emerson wrote:
> I've been working on a port of CCL to the 32-bit x86 architecture for
> a while now.  For the most part, this work has involved staring at a
> gdb prompt for extended periods; progress has therefore been hard to
> see up to this point.
>
> I'm therefore glad to be able to report that the lisp now starts up
> and maps in a (cross-dumped) boot image, which subsequently loads in
> the rest of the (cross-compiled) fasls that make up the system.  This
> results in a working (for some fairly lenient definition of working)
> toplevel.
>
> $ ./dx86cl x86-boot32.image
> ;Loading level-1.dx32fsl
> ;Loading ./l1-fasls/l1-cl-package.dx32fsl
> [...]
> ;Loading ./bin/ccl-export-syms.dx32fsl
> ;Loading ./l1-fasls/version.dx32fsl
> ;Loading /Users/rme/ccl/library/lispequ.dx32fsl
> ? (lisp-implementation-version)
> "Version 1.1-r8833:9322MS (DarwinX8632)"
> ? (defun fact (n) (if (= n 1) 1 (* n (fact (1- n)))))
> FACT
> ? (fact 5)
> 120
> ? (fact 50)
> 30414093201713378043612608166064768844377641568960512000000000000
> ?
>
> This is single-threaded, the GC doesn't work, and there are many other
> flaws.  It's not suitable for even experimental use yet---it has a
> tendency to dump you into gdb at the slightest provocation---but a
> great many things *do* work.
>
> Next up is getting the GC operational, and getting the lisp to compile
> itself.

Wow, this is pretty much terribly exciting.

Thanks for this!

--me