bug#15746: 24.3; [PATCH] bookmark should confirm when overwrite

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

> Occationally I have overwritten bookmarks with regrets. So maybe
> something along the following lines is needed.

It is not needed, IMO.

The point of `bookmark-set' is to set a bookmark, just like the
point of `setq' (unlike `defvar') is to set a variable value,
whether or not the variable already has a value.

As far as I am concerned, it would be OK to add a separate new
command for the behavior you want (without binding it to
`C-x r m').

Or to add a user option that `bookmark-set' tests, to optionally
(not by default) do what you want.  But the default behavior
of `bookmark-set' should not be changed this way, IMO.

(FWIW, even if you do do the wrong thing occasionally, you need
not necessarily regret it.  Your updated bookmark is not saved 
immediately, depending on your value of `bookmark-save-flag'.
If you are worried about your occasional `bookmark-set' mistakes, consider setting `bookmark-save-flag' to `nil'.  Of course, you
might also want to then check such a bookmark before explicitly
using `bookmark-save'.)


[(OT) Replying to a yes-or-no question is not a *user*
error, even if the misnamed `user-error' is appropriate here
(which it is).  Just one more indication why "user-error" is
the wrong name for its intended behavior (= any error other
than a coding error).]