Re: Pushing the `gnus-range-*' functions down into the C layer

[Stephen J. Turnbull]

Lars Magne Ingebrigtsen writes:
 > Ted Zlatanov <address@hidden> writes:

 > > 64-bit int support in Emacs should not require a 64-bit system.
 > > If they are natively available it's faster, that's all.  It's not
 > > like you'd be packing and unpacking the bits directly.

 > Yeah.  Bignum support in Emacs would totally rock.

Indeed, it rocks in XEmacs and SXEmacs. :-)

 > No more manual conversions to floating point numbers to do safe-ish
 > computations on numbers that might be too big would be needed.
 > 
 > But I suspect that it's a kinda big project.  :-)

Not really.  Somewhat bigger than "linking in libxml2", but not that
much.  For all computations except buffer and string support, it's
already been done in XEmacs and SXEmacs, and the original
implementation has needed very little in the way of bugfixing -- I
suspect it's a real SMOP.  The hard/tedious part would be large buffer
support on 32-bit systems, because for efficiency you probably don't
want to use bignums, but rather 64-bit fixnums.  (But nobody has
actually tried, because the real problem with big buffers is
variable-width representations of characters.)