I think it would be a good idea to prepend reset() to each regression
test.
Right now, reset() spews information that I doubt we want.
Either:
(1) make reset() silent (easy)
(2) introduce an icky option variable to control the output of reset(),
or
(3) invent a uniform mechanism for controlling warnings and other
messages.
For (3), there is a an experimental function in linearalgebra. This
function
should be improved (I think it is confusing), but maybe it's a start:
(defun $inform (level pck msg &rest arg)
(if (member level (member ($get '$INFOLEVEL pck) `($debug $verbose
$silent)))
(apply 'mtell `(,msg , at arg))))
(%i2) qput(infolevel, verbose, system)$
Since verbose = verbose, print the message:
(%i3) inform('verbose, system, "We're pumping mud");
&We're pumping mud
(%o3) false
Since verbose > silent, print the message:
(%i4) inform('silent, system, "We're pumping mud");
&We're pumping mud
(%o4) false
Because verbose < debug, don't print the message:
(%i5) inform('debug, system, "We're pumping mud");
(%o5) false
This might not be what we want; maybe silent should always suppress
output:
(%i3) qput(infolevel, silent, system)$
(%i5) inform('silent, system, "We're pumping mud");
&We're pumping mud
Barton
PS With
(%i9) build_info();
Maxima version: 5.9.2.19cvs
Maxima build date: 9:47 1/26/2006
host type: i686-pc-mingw32
lisp-implementation-type: GNU Common Lisp (GCL)
lisp-implementation-version: GCL 2.6.7
I only get the one known rtest_trig #68 bug. I can run the testsuite
multiple times and I still only get this one bug.