On 12/17/07, Richard Fateman <fateman at cs.berkeley.edu> wrote:
> Please consider the re-definition of nformat, part of the display program,
> in
>
> http:/www.cs.berkeley.edu/~fateman/generic/nnformat.lisp
>
> it should be faster and more general.
After looking over this code, I don't think NFORMAT should emit
strings for non-string arguments such as numbers or CLOS objects.
I think Stavros said the same thing about an earlier version.
At present, float->string format is handled, with some heuristics,
in EXPLODEN, which is called from MAKESTRING.
If we don't like the way floats are formatted, maybe we should
look at those functions.
Expanding NFORMAT to handle CLOS objects makes sense to me,
although I would rather see a cons returned instead of a string,
e.g. ((MLIST) FOO BAR) for an interval instead of "[foo, bar]".
I think that's how NFORMAT works on special cases at present.
FWIW
Robert Dodier