Re: [Emacs-diffs] emacs-25 1f680db: Fix compilation next-error in buffers with selective-display

[Eli Zaretskii]

> From: Stefan Monnier <address@hidden>
> Cc: Eli Zaretskii <address@hidden>
> Date: Sun, 03 Jan 2016 11:02:17 -0500
> 
> > +(defun compilation-beginning-of-line (&optional n)
> > +  "Like `beginning-of-line', but accounts for lines hidden by 
> > `selective-display'."
> > +  (if (or (not (eq selective-display t))
> > +          (null n)
> > +          (= n 1))
> > +      (beginning-of-line n)
> > +    (re-search-forward "[\n\r]" nil 'end (1- n))
> > +    (if (< n 0)
> > +        (beginning-of-line))))
> 
> All this extra cruft to support an obsolete feature we should eliminate
> is the reason why I didn't fix it earlier.

I'm sorry, but I'm firmly opposed to such a "strategy" of removing
features you don't like by leaving the related bugs unsolved.  In the
reality of Emacs maintenance, where all of us are busy people with
lots of other stuff on our plates, the result will always be that the
feature stays in, and bugs stay unsolved, for years.  This bug is an
excellent example.

> There are reams of other places where selective-display can cause errors
> and where similar "simple" fixes could be added.
> 
> The problem is/was not in compile.el but in selective-display (which
> should just die its deserved death).

If you really want to remove selective-display, please submit patches
that remove it entirely, replacing it with alternative features.  Or
talk someone else into doing that.  Then we can discuss if the
replacements are good enough, and if they are, then bye-bye
selective-display.

But as long as the feature stays, it's wrong to refrain from solving
such simple problems, hoping that the issue will go away by itself
somehow.