bug#5251: 23.1; problem in decode_eol of coding.c

[Eli Zaretskii]

> From: Andrew Hyatt <ahyatt@gmail.com>
> Date: Tue, 16 Feb 2016 22:41:32 -0500
> Cc: 5251@debbugs.gnu.org
> 
> 
> I can verify that this issue still happens in Emacs 25.  This simple fix
> seems to have slipped by unnoticed many years ago, hopefully someone
> that is qualified to comment on the patch will see it now.

Thanks for the reminder, I fixed this now on the emacs-25 branch.
(The proposed patch was not entirely correct, although it was in the
right direction, and happened to fix this particular case.)

It is indeed a shame that such a simple bug was left unresolved for
such a long time, but better late than never.

> BTW, this seems like a nice thing to unit test with ert, if a fix does
> take place.

Yes, that would be very welcome, thanks.  Looks like coding-tests.el
is a good place to add such a test.  Please be sure to include lone ^M
characters, as well as those followed by a newline, so that we are
sure lone CRs are not removed.

Thanks.