Re: 23.0.50; find-func: can no longer find adviced subrs

[Davis Herring]

> the problem appears to stem from the fact that symbol-function on
> an adviced subr now returns the compiled advice form, rather than
> the #<subr subrname>
> -- this causes find-function-noselect to fail

This should not be surprising: the symbol's function definition now -is-
that advice form and not a subr.

> the problem also shows up when you do describe-function on an
> adviced subr -- the help buffer text wrongly identifies it as an
> interactive function at the top of the help text.

The advice form is an interactive function, and is correctly described as
such.  One can argue that the help system should be more aware of advice,
but it is not necessarily trivial to distinguish between advice and an
arbitrary lambda stored in a symbol's function cell.  Perhaps
`describe-function' should consult the `ad-advice-info' property on its
argument when constructing its first line of output.  `find-function'
could do something similar.

> Other parts of emacs still know it's a subr -- the bototm of the
> help text says "this subr is adviced ..."

That text is generated by the advice system and knows what the original
function is.

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.