Name: bb33257 Date: 02/19/98


According to the people at the IBM Arabic Competency Center,
there are some errors in the CharToByteConverters and
ByteToCharConverters for the Arabic code pages.

In ByteToCharCp1256.java:
  0x80 should map to \u0080
  0x8a should map to \u008a
  0x8f should map to \u008f
  0x98 should map to \u0098
  0x9a should map to \u009a
  0x9f should map to \u009f

In ByteToCharCp864.java:
  0x9f should map to \u200c
  0xd7 should map to \ufec3
  0xd8 should map to \ufec7
  0xf1 should map to \ufe7c

In ByteToCharCp420.java:
  0x45 should map to \u200c
  0x77 should map to \uf8f6
  0x80 should map to \ufeb5
  0x8b should map to \ufeb9
  0x8d should map to \ufebd

And, of course, the corresponding conversions in the CharToByte
converters also need to be updated.

In addition, there are probably other conversions in the
CharToByte converters that need to be updated, and IBM wants
support for code page 1089, which is not currently supported.

I have complete tables for all of the code pages in both
directions. I can provide them to whoever works on this bug.
Contact me at ###@###.###.
======================================================================