Re: emacs-26 c87d04e: Avoid aborts in 'md5'

[Eli Zaretskii]

> Cc: address@hidden, address@hidden
> From: Paul Eggert <address@hidden>
> Date: Wed, 28 Feb 2018 11:14:36 -0800
> 
> > With the current code, I agree.
> 
> OK, proposed patch attached.

That's not what I meant.  I meant that as long as the code is frozen
in its current form, nothing is gained by omitting 'else'.  But the
change I made is more future-proof.

As I said, we always check STRINGP before using SBYTES or SSDATA, so I
see no reason why we shouldn't in this case.