pure texinfo files: do we need to do anything special about encoding?

See #254.  We read these directly using `fileread`:
```
  if any(strcmpi(ext, {'.texinfo' '.texi' '.txi' '.tex'}))
```
Anything special need be done to ensure we get a utf-8 internal string directly from that?

Low priority: I don't think folks are using doctest extensively with `.texinfo` files.  I also don't know how prevalent non-ascii texinfo files are.