Re: master c8a2af3037 1/3: Add new function function-alias-p

[Stefan Monnier <monnier@iro.umontreal.ca>]

>>> An the semantics would be kinda odd -- in the non-alias case, we'd still
>>> return nil, right?
>> No, I was thinking of returning the content of `symbol-function`.
>> So, it basically returns a more explicit info than `indirect-function`.
> My feeling is that people that want to get at the function object will
> use `indirect-function'.  Callers that want both the intermediate chain
> and the final object are pretty rare, I think.

Could be.  I was thinking of cases like `interactive-form`, `commandp`,
and `documentation` where we usually want to fetch the info off of the
function object, but we need to check some symbol property along the way
in case it shadows the function's own value.

These functions currently start with `indirect-function` to rule out the
presence of cycles and then proceed to loop through the aliases (if
any).


        Stefan