On 1 Mar 2013 at 22:43, Miro Kropácek wrote: > Well, on the other hand, my version (which was converted to hyp; April > 1994) says 92 (0x5C). > Aha, I've located the problem. The table in Appendix A of the Compendium (on p A.10) is wrong, but the actual description (on p 4.100) is correct. Roger