Further to my own prediction, a deliberate excerpt is not the reason for my feeds failing to validate. Nope, it is a dastardly invisible character, as pointed out by Jan Boddez in a comment. This problem has bitten me before, and I always forget it. It would be very cool if I could automagically test for the presence of this kind of stuff; StackOverflow suggests that iconv --from-code=UTF-8
would do the needful, so that could be something, especially if I can trigger something like that when I save a new blog post.
There’s more ➢