Re: printing bytecode in backtraces

[Richard Stallman]

      In fact, `print-escape-nonascii' seems to
    make no difference:

It only makes a difference when you output into a unibyte buffer.
That was because nonascii unibyte characters in a multibyte buffer
used to cause trouble.

Now they no longer cause trouble, so perhaps we should change that so
that print-escape-nonascii really controls the decision for all kinds
of buffers.  That would make things cleaner, but might break something.
What do people think?

For the moment, I will document this.