bug#17947: 24.3; ruby-mode sets require-final-newline unconditionally

[Eli Zaretskii]

> From: Ethan <ethan.glasser.camp@gmail.com>
> Date: Sat, 5 Jul 2014 14:19:06 -0400
> Cc: 17947@debbugs.gnu.org
> 
> The idea behind the package is that bad whitespace is highlighted, not
> automatically cleaned up. So if the file doesn't have a final newline when
> it is first loaded, no final newline will be added without explicit user
> action -- instead a red "eof" marker is placed where the final newline
> ought to be.
> 
> require-final-newline obviously interferes with this. ethan-wspace could
> therefore set it to nil when it is operating, unconditionally. However, I
> added functionality to complain if require-final-newline was set -- not
> because *users* set it, but because *modes* do, and I wanted to get notice
> when that happened. So if you really think ruby-mode should unconditionally
> set require-final-newline, it's easy to remove the warning, make it
> optional, override require-final-newline on a buffer-local basis, etc.

It seems to me that if the buffer being checked has
require-final-newline set to a non-nil value, ethan-wspace should
simply refrain from checking the final newline, because obviously the
user and/or the mode took control of that issue.  Wouldn't this be
better than bitching at users for setting this option?

And I don't think introducing yet another option would solve it.  If
the user or mode set require-final-newline, IMO that is a clear
declaration that the final newline issue is that user's or mode's
responsibility; no further options are needed to make that clear.
Don't you agree?