[PATCH] Say explicitly that eq is useful for chars; xref number comparison, lispref

[Stephen J. Turnbull]

Aidan Kehoe writes:

 > Indeed, it’s not. The docs I cross-reference (from Jerry?) call immediate
 > integers in a build with bignum support "integers" though, not fixnums, and

That's OK in an context where repeated usage of bigint can make the
distinction clear.  In a context where bignums aren't mentioned, it
will be ambiguous.

 > suggest reasonable contexts where one might still want to use #'eq. In
 > particular, it's perfectly cromulent to say (eq X Y) where X is a literal
 > integer and Y is a bignum, since bignums with values less than (/ (lsh 1
 > INT_VALBITS) 2) should not escape to Lisp.

Feel free to add that useful information. ;-)

 >  > It's only true for fixnums, which do. It's true for characters, which
 >  > also have an immediate representation. You should a mention that this is
 >  > an implementation detail, although it's very unlikely to change. Unless
 >  > Emacs promises that this is true?
 > 
 > Their Lisp reference does, yes. 

OK, no need to mention implementation details then.