segmentation fault with Maxima/ECL



 
> GCL does not.  If the most important platforms do not support > 63 function
> arguments, we should probably (alas) inherit that limitation in Maxima,
> explicitly prohibiting > 63 arguments in mapply so that Maxima code will be
> portable between platforms.

I disagree: What would be the point of supporting multiple lisps, if we
stick to the least common set of features?

OTOH some compatibility mode with strict checking for such issues
might be nice, but IMHO it's not of high importance either ...

> In almost all cases where people are using apply with a large number of
> arguments, the function is associative, and xreduce does the right thing.
> We should be encouraging the use of xreduce.

Yes.

Harald