Patch: Promoting characters to 31 bits (for discussion)
Stephen J. Turnbull writes:
I don't see that loss of one bit of precision in integers
hurts;
I know of only one place where it matters: buffer sizes. The maximum
integer value is the maximum buffer size. Doesn't seem much of a
loss for the average user.
First, in src/data.c and src/print.c I changed "switch
(XTYPE(obj))
..." to "if (LRECORDP(obj)) ... else if (CHARP(obj) ..." so that these
control structures don't depend on the representation of Lisp
character and Lisp integer types. I think the efficiency hit should
be negligible, but I'd like a qualified opinion.
It basically means the type extraction happens for each
comparison instead of once. For Ftype_of, I don't care.
For the print functions, I worry.