[Sbcl-devel] merged patch: Better MACHINE-VERSION for Linux/PPC?

[William H. N. <wil...@ai...>]

On Thu, Jun 26, 2003 at 10:11:10AM -0500, William Harold Newman wrote:
> > (defun get-machine-version ()
> >   #!+linux
> >   (with-open-file (stream "/proc/cpuinfo" :if-does-not-exist nil)
> >     (loop with line while (setf line (read-line stream nil))
> >           when (eql (search "cpu" line) 0)
> >           return (string-trim " " (subseq line (1+ (position #\: line))))))
> >   #!-linux
> >   nil)
> > 
> > (defvar *machine-version* :unset)
> > 
> > (defun machine-version ()
> >   "Returns a string describing the version of the local machine."
> >   (if (eq *machine-version* :unset)
> >       (setf *machine-version* (get-machine-version))
> >       *machine-version*))
> 
> I don't plan to do anything with the PPC code unless someone sends me
> a nice PPC machine to play with, or at least a killer PPC shell
> account, since I only have X86s and I dislike checking in code I can't
> test directly. However, other maintainers have PPCs; and I myself plan
> to borrow the basic idea for the x86 implementation of
> MACHINE-VERSION.

After offers on #lisp to test it, I decided to change the PPC code as
well as the X86 code, and now both are in CVS.

> Thank you.

again

-- 
William Harold Newman <wil...@ai...>
Saying that taste is just personal preference is a good way to prevent
disputes. The trouble is, it's not true. You feel this when you start
to design things. -- <http://www.paulgraham.com/taste.html>
PGP key fingerprint 85 CE 1C BA 79 8D 51 8C  B9 25 FB EE E0 C3 E5 7C