7:18 a.m.
>>>> "David" == David A Cobb
<superbiskit(a)home.com> writes:
David> If I were to put a couple of "(message" lines into
David> autoload.el, would the batch-mode run send these to stderr?
David> Or is there a better built-in to use?
It takes a little fooling around to get messages to go to the
terminal. Use the ancient prin1 and princ forms. (They differ in how
they quote Lisp data.)
What does your environment look like (environment varialbes, PATH, and
the Windows registry)? XEmacs should ignore that stuff during build,
but often we get caught looking at the wrong external information.
David> I'm enclosing my 'update-elc.el' file; could you perhaps
David> diff it against yours? Mine is the way it came off the
David> 'net or as munged by configure.
It's OK. configure does not touch any Lisp. Late in the make it gets
compiled. Three lisp files (finder-inf.el, auto-autoloads.el and
custom-load.el) are built from scratch. Otherwise it should be pristine.
[The rest is pretty OT.]
David> I'll look, but it still seems strange.
Lisp is not your average programming language. It has bizarre idioms.
David> Even if those /if/ forms are nested inside an enclosing
David> function it would only produce one result, no? The several
David> /if/ forms may return several values but I don't think they
David> will get where they're desired as it's written.
*scratch* is your friend. Try it yourself. C-x b *scratch* RET
(if t (list "A" "B") (list "C" "D")) LF
(if nil (list "E" "F") (list "G" "H")) LF
(append (if t (list "A" "B") (list "C" "D"))
(if nil (list "E" "F") (list "G" "H")))
LF
Also try substituting `list' and `vector' for `append'. Note that
although LF is magic, it doesn't know anything about what you do, it
reads the buffer to find out what you want. So feel free to erase,
mung, edit, copy, use abbrev-mode, keyboard macros, or even cut and
paste from this message.
David> If I correctly understand that, should it not be a /(cond /
David> form?
Try substituting `cond' in the `append' form above. You'll get an
error. Then take each (if ...) form and wrap it with extra parens in
the cond ((if ...)). See what happens.
So each if form, each cond form, produces _one_ value. It might be a
list containing many things. But Lisp grabs a "handle" to it, and you
can pass that "handle" around very efficiently, like a pointer to a
struct in C or an object in Smalltalk.
Anyway, this should all be moved to comp.lisp or comp.emacs or
comp.emacs.xemacs.