bug#20168: 25.0.50; (elisp) `Accessing Mouse', for `posn-object-x-y'

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

> > The doc string is OK.  The manual is misleading/incorrect, in this
> > regard: it speaks about a case where "the POSITION is a buffer
> > position".  This is wrong, as a buffer position in Emacs means a natural
> > number 1,2... (or a marker).
> 
> Why is it wrong?

It is wrong to say that the value is a buffer position, because a
buffer position is a number, not a position object (which is a cons).

> > Maybe what should be said in this case is
> > that "POSITION represents a buffer position".  Its value cannot be
> > simply a buffer position, which is a number or a marker.
> 
> Why not?  And what's the difference between the current text and your
> suggestion?

The difference is the difference between "is" and "represents".
The value *is* not a buffer position (which is a number).  It *is*
a position object (which is a cons).  It represents a buffer position
in this case.

It's a minor point, but it was enough to confuse me into trying
to pass an actual buffer position.  Nothing in the current
description says that POSITION must be a position object (i.e.,
something for which `posnp' returns non-nil).