🚀 go-pugleaf

>>>>> "PS" == Paul Smith
>>>>> "kj" == Kyle Jones

    kj> After the trouble we went through to get mail-yank-hooks
    kj> standardized, if VM went off and started using another
    kj> variable Barry Warsaw would kill me. :-)

    PS> Actually, he wouldn't :)

Paul's right, I'd never do *that*.  But I might not invite you to any
more gigs (just kidding Kyle :-).

    PS> FSF Emacs 19 has introduced a new hook, called
    PS> mail-citation-hook, which is supposed to be used exclusively
    PS> for mail citation packages.

I guess you didn't get the mail that RMS and I exchanged about this.
Sorry about that Kyle (I can't seem to find the correspondance now
either, though).  Anyway, RMS is right when he said that
mail-yank-hooks is wrong for 2 reasons: it should have been called
mail-yank-hook, and it doesn't even act like a real hook.  But he
understood the need to keep some backwards compatibility so he still
supports the m-y-h interface, but indicates its obsolete.

    PS> If this hook is entirely empty (nil), a default action is
    PS> taken instead of no action.")

If you look in mail-yank-original in FSFmacs19's sendmail.el, you'll
see that the "default action" it takes when mail-citation-hook is nil
is to call run-hooks on mail-yank-hooks.  That's why the m-y-h
interface still works.  If m-c-h is non-nil, it is run-hooks'd
instead.  So I *think* the proper thing to do now (if you want to be
compatible with FSF ;-) is something like:

    (if mail-citation-hook
        (run-hooks 'mail-citation-hook)
      (run-hooks 'mail-yank-hooks))))

I think RMS's decision on this is fine. The *really* important part of
the interface is that the original mail headers aren't cleansed out
and that point and mark define the region to cite.

Oh, probably also defvar mail-citation-hook to nil and mail-yank-hooks
to some default action.  Look at sendmail.el for how RMS does it now,
but I think it would be great if people could just add:

	(add-hooks 'mail-citation-hook 'sc-cite-original)

to get supercite to work with any mail/news agent.

    PS> SC 3, which is included in the FSF Emacs 19 distribution, has
    PS> already been documented to use this, and RMAIL and (I hear)
    PS> GNUS (also distributed with Emacs 19 now) both use it as well.

Actually FSF19 still uses a semi-hacked version of 2.3. I haven't even
looked at FSF19's diffs to the manual yet. SC3 is in beta testing now
and seems pretty stable.  I just need to rewrite the manual, which
really is the only place that any of these citation hooks is
referenced in supercite anyway!

    PS> After all those years of being on the cutting edge vs. RMAIL
    PS> w.r.t.  citation hooks, VM has now been leaped over! :)

But it was VM's foresight that proved the concept worked, ensuring
citation bliss for all mankind into the forseeable future.  Or
citation torture, depending on your P.O.V. :-)

-Barry