bug#22018: 25.0.50; REGRESSION wrt `isearch-new-word'

[Drew Adams <drew.adams@oracle.com>]

> > You essentially did a global replace of `isearch-word' and
> > `isearch-new-word' with `isearch-regexp-function' and
> > `isearch-new-regexp-function'.
> >
> > You provided a `define-obsolete-variable-alias' for `isearch-word',
> > which is a global variable.  But you did nothing to accommodate
> > `isearch-new-word'.
> 
> So adding an alias for isearch-new-word should (in principle)
> fix the issue?

I'm not sure that would be the right fix.  I suggested doing this
inside the `let':

(setq ...
      isearch-regexp-function  (if (boundp 'isearch-new-word)
                                   isearch-new-word
                                 isearch-new-regexp-function)
      ...)

This variable, unlike `isearch-word', exists only locally, bound
by `let'.  An alias (e.g., `defvaralias') would make it a global
variable, no?

That was why I made the above suggestion - to avoid creating a
global variable for this.  But feel free to fix it however you
think is best.