Re: Proposal: immediate strings

On 05/24/2012 12:14 AM, Stefan Monnier wrote:
> we could just grab a second bit from `size'
> for the `immbit'.

Only if MOST_POSITIVE_FIXNUM <= min (SIZE_MAX, PTRDIFF_MAX) / 2.
This inequality is commonly true, but it's false when EMACS_INT
is wider than both size_t and ptrdiff_t.  The method I suggested
should work even on these platforms.