Re: improvement for `delete-indentation'

[Richard Stallman]

    >               (delete-region (point) (+ (point) (length fill-prefix))))
    > +         (if (eq ?\\ (char-syntax (preceding-char))) ; assume continued 
line
    > +             (delete-backward-char))
    >           (fixup-whitespace))))

    It might not be a bad idea but I think it deserve more care.  The above
    code can lead to surprises with cases such as

I think it should be controlled by a variable that can be set by the
major mode, because not all languages have this kind of continuation.
The variable might specify which character is the continuation character.