Re: applying changes in .emacs

[Eli Zaretskii]

> From: don provan <dprovan@comcast.net>
> Date: Mon, 29 Aug 2005 00:35:25 -0700
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> > There's no command to do that, since quite a few fragments in a
> > typical .emacs file cannot be re-evaluated without causing bad side
> > effects.
> 
> Wow. I would have said .emacs code is broken if it cannot be
> re-evaluated without causing bad side effects, so I must be missing
> something.

You are not missing anything, I think.  Simply put, typical .emacs
code is written assuming that it is evaluated only once, when Emacs
starts.  While that might be a bad assumption for general-purpose Lisp
code, what one puts in .emacs is not general-purpose code, and in any
case, that's how people tend to do it.

> Can you give some examples and explain why they can't be avoided?
> About the only thing I'm aware of doing is adding duplicate entries
> to some lists, and even that I could avoid if I thought the effect
> would amount to something bad.

Adding things to a list is one typical example (adding a hook function
is a frequently seen variant of this).  Doing something if some symbol
is not bound or if some feature is not available (meaning a package is
not loaded) is another.  There's more.