+# Use these to convert to/from UTF-8 bytes. I8 is the encoding that
+# corresponds to UTF-8 with start bytes, continuation bytes, and invariant
+# bytes. UTF-EBCDIC is derived from this by a mapping which causes things
+# like the start byte C5 to map to something else, as C5 is actually an 'E' in
+# EBCDIC so can't be a real start byte, as it must be an invariant; and it
+# maps 0x45 (an ASCII 'E') to C5.