Stavros Macrakis wrote:
> On Maxima 5.11.0 GCL 2.6.8, internal errors look like this:
>
>
>
>
> Maxima encountered an internal error:
> ***the Lisp error message***
> This indicates a bug in Maxima; use bug_report() to report bugs.
> To debug in Lisp, set *debugger-hook* to nil.
> <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>
> The last line is not helpful to most users, but I admit that I'd
> forget the variable name myself if it weren't there, though better
> might be:
>
>
> Maxima encountered an internal error:
> ***the Lisp error message***
> This indicates a bug in Maxima; use bug_report() to report bugs.
> Lisp programmers: try debugging_advice().
> <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>
> Where debugging_advice could print out something useful.
>
>
Like what? Would it say something about setting *debugger-hook* to
NIL? Or maybe running "maxima -g" to enable the Lisp debugger?
Something else? I always run with -g so that I can at least get a
backtrace from the error. As it is now, I'd be just as happy if nothing
were printed out because such an error doesn't really tell me anything
useful. (Yes, this would be annoying to other users.)
> It appears that the string "Error in ... [or a callee]" is part of the
> 'condition' object passed by GCL to Maxima. I don't know GCL well
> enough to know how to control this. Do we have to do something silly
> like string manipulation on the result to splice it out?
>
>
Is that something we really want to do? People do use other Lisps
besides GCL. CMUCL says "Argument X is not a number: ((MTIMES SIMP)
-1.0 X)"
Ray