Custom Query (1030 matches)

% openmcl
Welcome to Clozure Common Lisp Version 1.6-dev-r14031M (LinuxX8664)!
? (defun one (B)
  (DECLARE (TYPE (INTEGER 51357426816569 68500595286128) B))
  (LOGAND B -2))
ONE
? (defun two (B)
  (LOGAND B -2))
TWO
? (- (one 67660763903986) (two 67660763903986))
-63262717391456

The following should be done at some point.

(in-package "CCL")

(defun remove-extra-symbol-binding-indices ()
 (let* ((n 0))
   (do-all-symbols (s n)
     (let* ((bits (%symbol-bits s))
            (idx (symbol-binding-index s)))
       (unless (zerop idx)
         (unless (and
                  (not (or (logbitp $sym_vbit_const bits)
                           (logbitp $sym_vbit_global bits)))
                  (or (logbitp $sym_vbit_special bits)
                      (let* ((name (symbol-name s))
                             (len (length name)))
                        (and (> len 1)
                             (let* ((initial (schar name 0))
                                    (final (schar name (1- len))))
                               (and (eql initial final)
                                    (or (and (eql initial #\*))
                                        (eql initial #\%))))))))
           (incf n)
           ;; Pretend that the symbol is DEFGLOBALED, to
           ;; persuade ENSURE-BINDING-INDEX to remove
           ;; the symbol from its inverse mapping.
           (%symbol-bits s (logior $sym_vbit_global bits))
           (ensure-binding-index s)
           (%symbol-bits s bits)))))))

Doing:

1) update to the current trunk, especially including the changes to nfcomp.lisp
  in r13745.  Rebuild the image with these changes in effect.

2) In a new image, note that

? (ccl::next-binding-index)

returns a value likely > 13000, then call REMOVE-EXTRA-SYMBOL-BINDING-INDICES
and do a full/clean rebuild.  Quit and load the (new) new image.

3) In that (new) new image, (ccl::next-binding-index) should return ~900.

See e-mail thread below for suggested improvement.

Begin forwarded message:

From: "Mark H. David" <mhd@yv.org>
Date: July 28, 2010 6:37:11 AM EDT
To: Gary Byers <gb@clozure.com>
Cc: openmcl-devel@clozure.com
Subject: Re: [Openmcl-devel] save-application

Yes, (save-application ... :prepend-kernel t) does the trick.

> C:\Program Files\ccl>wx86cl.exe
> Welcome to Clozure Common Lisp Version 1.4-r13122  (WindowsX8632)!
> ? (ccl::save-application "\\foo.exe" :PREPEND-KERNEL t)
>
> C:\Program Files\ccl>\foo.exe
> Welcome to Clozure Common Lisp Version 1.4-r13122  (WindowsX8632)!
> ? (print "Hello World!")
>
> "Hello World!"

Doc could be improved.  The documentation for function save-application (http://openmcl.clozure.com/manual/chapter4.9.html#Saving-Applications) no information about prepend-kernel nor any other of these keyword args:

 (purify t)  impurify  (mode #o644)  prepend-kernel

The text of this chapter leading up to the function does, in retrospect, explain the wonders of prepending the kernel in a somewhat chatty and roundabout style while managing to never actually refer to the PREPEND-KERNEL arg by name.

Thanks!

Mark

Gary Byers wrote:
>
>
> On Tue, 27 Jul 2010, Mark H. David wrote:
>
> > Anyone know why this doesn't work?
>
> You can use the :PREPEND-KERNEL option to SAVE-APPLICATION
> to create a self-contained executable file; without that
> option, it just writes a heap image that can be mapped into
> memory by the lisp kernel.
>
> Giving a file a ".exe" extension doesn't turn it into a valid
> Windows executable file (with the right signatures and structure.)
>
> > 
> > C:\Program Files\ccl>wx86cl.exe
> > Welcome to Clozure Common Lisp Version 1.4-r13122  (WindowsX8632)!
> > ? (ccl::save-application "\\foo.exe")
> > 
> > C:\Program Files\ccl>\foo
> > Access is denied.
>
> Some variant of "File isn't in executable format" might be a clearer
> error message, though it's hard to say "an OS should spend N cycles
> checking to see if alleged executable files are in the right format
> so that better error messages can be produced in this case" - the case
> is probably not exactly common, and the N cycles might add up.
>
> It's likely unrelated to what you describe above, but some versions of
> Windows may dislike writing files to / executing files in the root directory.
> (I don't know what versions of Windows limit this or exactly what the
> limits are, but I'm fairly sure that some limits exist.)
>
> > 
> > C:\Program Files\ccl>
> > _______________________________________________
> > Openmcl-devel mailing list
> > Openmcl-devel@clozure.com
> > http://clozure.com/mailman/listinfo/openmcl-devel

_______________________________________________
Openmcl-devel mailing list
Openmcl-devel@clozure.com
http://clozure.com/mailman/listinfo/openmcl-devel