Re: [Maxima-bugs] [ maxima-Bugs-574635 ] machine-mantissa-precision hard-wired

>>>>> "Barton" == noreply  <noreply@sourceforge.net> writes:

    Barton> Initial Comment:
    Barton> In simp.lisp, the value of machine-mantissa-precision is hard-wired to 24:

    Barton> (eval-when (compile load)
    Barton>    (defvar machine-mantissa-precision 24 );;is this correct 
    Barton> )

    Barton> This is correct for single-floats.  I'd suggest changing the code
    Barton> to something like

    Barton> (eval-when (compile load)
    Barton>      (defvar machine-mantissa-precision (float-digits 1d0)))

This looks good.  I'll try adding it, and deleting the reference in
float.lisp that you mention.

    Barton> Notes:

    Barton> 1. For some background, see

    Barton>           http://www.math.utexas.edu/pipermail/maxima/2001/000175.html

Could you add these to the bug tracker?

My comments to message 175:

3.  Isn't this because you are trying to call the Lisp function ROOM
    directly so that the "?" is required?  Section 2.2 of the manual
    says this is how you can call Lisp functions from the
    command-line.  Note that with CMUCL this doesn't work---you have
    to type "?xroom()" for any x that is not a space.  This is an ANSI
    CL compatibility bug that I don't know how to work around in
    mdebug.lisp.

4.  Seems to be already added.

7.  Looks like a misplaced declaration in numeric/fft.lisp.  I'll try
    to fix this.  Alternatively, I have a simple FFT in Lisp that can
    be used.


Thanks,

Ray