Stephen J. Turnbull wrote:
Raymond Toy writes:
> Now in *junk* do M-: and eval
>
> (add-local-hook 'post-command-hook 'dummy t)
>
> What I see is that post-command-hook now has the value
>
> (t dummy)
This is the correct result, I believe. The semantics of a hook
variable is that the local value of the hook variable is retrieved,
and the functions in it are run. But if the hook is the value t, then
that means to run the global hooks there. In other words, the t
separates the prepended local hooks from the appended ones.
> Without append, add-local-hook just creates (dummy), which looks right
> to me.
With a different (nonexistent) hook variable, I get (dummy t) which is
consistent with the above semantics. If you actually do get just
(dummy) from `(add-local-hook 'post-command-hook 'dummy)', that's a
bug. So I'll check later to see if I can reproduce.
I did some more investigation with an older version of slime.
Everything is working there, and the hook is (dummy t), and everything
works fine. So it's something about slime. It's weird though. With
(dummy t), slime is broken. I get some error about (natnump t), and the
backtrace shows some OPAQUE object. If I remove t, slime works, but
then my default paren-hightlight hook isn't run, so I'm missing the very
useful paren highlighting.
Sorry for the noise. I'll have to dig some more to see what's going on.
Ray
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta