Re: readline support

[Daniel R Risacher <risacher@worldnet.att.net>]

>>>>> "Jim" == Jim Blandy <jimb@red-bean.com> writes:

    Jim> As groovy as the readline support is, I'm thinking about
    Jim> taking it out, because:

    Jim> 1) it breaks the unpublished Emacs interface (because I'm
    Jim> lame)

    Jim> 2) it breaks the Tcl/Tk interface (again, because I'm lame)

    Jim> 3) it is confusing in the way you mention (this is not
    Jim> trivial to fix)

    Jim> 4) you can run Guile under cmuscheme anyway, and it works
    Jim> fine

I object to anyone taking credit for my lameness.  I only implemented
a readline soft-port since I was too lame to even *attempt* integrating
it into the repl.  I think the fact that you *did* work it into the repl
definitively proves that you *aren't* lame, even if there are some
nits left over.

Could the boot code decide which repl to use on the basis of
(isatty?)?  This might fix #1 and #2; I dunno about these.

Ideally, the add-history call should go on an after-read-hook, instead
of in the soft-port -- I only put it in the soft-port because I didn't
want to touch the repl.  This could fix #3.

    Jim> If I can see some quick, groovy workaround, then I'll take
    Jim> it, but it's not a high priority at the moment.

Yeah.  Perhaps the readline-repl should be extracted and stuffed into
readline.scm

    Jim> Certainly, the readline function will still be available at
    Jim> the Scheme level, but it just won't be part of the standard
    Jim> read-eval-print loop.


Dan Risacher

-- 
I wonder if man will walk on the moon in my lifetime?
magnus@alum.mit.edu