propogate those changes through the code.  This is so that there can
be holes in the list of charsets without breaking old tables.


configure.in:
  - changed pattern for getting number from charsets Index file
mysys/charset.c:
  - changed from using a TYPELIB to a CS_ID struct, so both the name
    and the number of a charset is stored in available_charsets
sql/share/charsets/Index:
  - made the number a real part of the Index file, not just a comment
sql/share/charsets/README:
  - order is no longer significant, but each charset must be paired
    with its number