This seems to getting more complicated than I thought. But if not anybody
beats me to it, feel free to do that(!), I can start looking at it. The
basic functionality does not look that complicated anyway.
2013/3/14 Stephen J. Turnbull <stephen(a)xemacs.org>
Jerry James writes:
> It looks easy to support, so for the sake of compatibility, I say go
> ahead. Should we do their INTEGER_TO_CONS trick on non-bignum builds
> as well?
If this is like the way time values (32-bit numbers that don't fit
into a Lisp fixnum) on 32-bit systems were split into two 16-bit
values, and that is done to support the full range of IDs, yes.
But I wonder if a cons can support the full range. Specifically, are
we sure that there are no 32-bit systems using a 64-bit ID (for
XEmacs-Beta mailing list