Re: master 72b1379f079: Increase `emacs-lisp-docstring-fill-column` to 72

[Po Lu <luangruo@yahoo.com>]

Eli Zaretskii <eliz@gnu.org> writes:

> Please excuse me for being blunt, but this is a tempest in a teapot.
> To see why, please collect the statistics of the line length of our
> doc strings: you will see that we already have a lot of them that
> exceed even the new value of 72.  The reason is simple: it is very
> rare to fill doc strings using fill commands.  I, for one, almost
> never do that, instead "filling" them by hand to make each line, and
> the doc string in general, read reasonably.

I only do so when the filling commands fail to produce properly
formatted messages, such as when a table or long symbols are present
which the filling commands wrap pedantically for being only 1 or two
characters past the fill column.

> So I think this setting should affect almost no one, and making a
> showcase out of this change just misses the entire point of the
> "December's debacle".  We have more important stuff to discuss than
> this insignificant detail.

Before responding, I tried to measure the impact of this change by
running:

  (mapatoms (lambda (atom)
	    (when (functionp atom)
	      (and (documentation atom)
		   (insert (documentation atom) "\n")))))

in a buffer, with the fill column set to 72, and typing C-x h M-q.  Upon
restoring the fill column to 64 and enabling the fill column indicator,
the number of lines and words beyond it increased visibly.  So this is
not so insignificant a detail.