Re: [PROPOSAL DRAFT]: editor and sensible-editor

[Chris Waters <xtifr@dsp.net>]

ccwf@bacchus.com (Charles C. Fu) writes:

> Personally, I would prefer to change the policy since it makes EDITOR
> preferred over VISUAL, which is the reverse of the behavior on my
> other systems.

Some historical perspective might be useful here.

The use of EDITOR is an ancient *NIX tradition that dates back before
vi was introduced.  (Might have been used to invoke TECO, I'm not
sure.:) When vi was introduced, the VISUAL variable was also
introduced.  At first, programs that needed to invoke an editor would
let you invoke *either* the visual editor (${VISUAL:-vi}) or the
standard editor (${EDITOR:-ed}).  An example of a program that still
works this way is (or should be) mailx.  But most people didn't want
to bother with ed, so soon the idea of having two commands to invoke
different editors went away, but, in the mean time, many people had
taken to setting VISUAL to their editor of choice, and ignoring
EDITOR, so it wasn't possible to get rid of VISUAL, so a lot of
programs took to using ${VISUAL:-${EDITOR:-vi}} (or the C
equivalent).  Examples of programs like this probably include every
more recent console MUAs (i.e. all except mailx).  :-)

I would prefer to change policy to something that matches the
traditional behavior a little better.  There is a little ambiguity
about what constitutes "traditional behavior" though, but I certainly
think that ignoring VISUAL in policy is a Bad Idea.  Ideally, we
should find a way to bless the traditional behavior of mailx, elm, rn,
trn, etc., rather than forcing these well-known programs to be
modified in ways that may disconcert people.

-- 
Chris Waters   xtifr@dsp.net | I have a truly elegant proof of the
      or    xtifr@debian.org | above, but it is too long to fit into
http://www.dsp.net/xtifr     | this .signature file.