cmucl -nositeint



>>>>> "Robert" == Robert Dodier <robert.dodier at gmail.com> writes:

    Robert> On 12/6/06, Dan Gildea <gildea at cs.rochester.edu> wrote:
    >> We have a site-init file that defines a top-level command line interface,
    >> and I was getting an error on starting maxima from redefining a symbol in this package,
    >> until I added -nositeinit to the call to cmucl in the maxima wrapper script:

    Robert> Hmm, aren't command line arguments for maxima passed through
    Robert> to the Lisp command line? Doesn't "maxima -nositeinit" have the
    Robert> intended effect?

It seems like they're not passed on.  In fact, I'm not exactly sure
what maxima does with -nositeinit---it causes a Lisp error for me.
Probably because the script isn't passing -nositeinit quite right.

    Robert> I'd rather not change the script if you can get what you want by
    Robert> just entering the extra argument (unless someone wants to argue
    Robert> that the per-site initialization file should always be ignored).

I don't think maxima should do that.  It's up to the site manager to
do what is necessary.  FWIW, I think a top-level command line
interface in siteinit is the wrong place.  It should be in the user's
init file.

In any case, we should fix it so that extra command-line args can be
passed to the underlying lisp correctly.  Either by passing any
unrecognized args or by adding a new arg that maxima recognizes and
sends to the underlying lisp.

Ray