This is the mail archive of the guile@cygnus.com mailing list for the guile project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

hairy types (Re: emacs with guile instead of elisp)

> it is unlikely that a random long-aligned pointer would turn out to be
> SCM-compatible.  most likely, it would alias w/ a "real" SCM and cause
> problems.
> 
> redesigning SCM type space partitioning (tagging) is possible, but
> almost too hairy to consider[1].
> 
> if we try to go around SCM, we introduce another (probably) expensive
> step for type resolution for every object reference.  and it may not
> even be possible, depending on where the compiler (linker?) maps the
> string pool and whether or not SCM uses all bits of a long.

If anyone does decide to redesign the SCM type space then I put my
dibs in now for making SCM_BOOL_F equal zero. Since C and C++ have already
set a blindingly simple standard for booleans, since booleans are
one of the most commonly used types and since for now and the foreseeable
future guile is going to involve interfacing scheme to C, why not
take this tiny step towards sanity?

	- Tel