>=C2=A0 =C2=A0 =C2=A0* lisp/emacs-lisp/advi=
ce.el (ad-preactivate-advice):
>=C2=A0 =C2=A0 =C2=A0Avoid setting the function definition of nil.
>=C2=A0 =C2=A0 =C2=A0This was happening during bootstrap of org-compat.e=
l,
>=C2=A0 =C2=A0 =C2=A0apparently due to eager macro expansion of code beh=
ind
>=C2=A0 =C2=A0 =C2=A0a (featurep 'xemacs) test.
Really, I think this business of "disallow fset of nil" is a big =
waste
of time and will just lead to more pain than gain.
There are umpteen different ways for the user to shoot himself in the
foot.=C2=A0 This one is not even fatal.
Should we also disallow (fset 'car nil)?=C2=A0 How 'bout (fset '=
;car #'cdr)?
Yes, fset for most primitives should be fo=
rbidden.=C2=A0