Charset.forName("PCK").newEncoder.canEncode(c) incorrectly returns true
for some 0x80 <= c <= 0xff,
and incorrectly returns false for Half-width katakana.

charset=x-PCK canEncode=true '\u0080' - '\u00a1'
charset=x-PCK canEncode=true '\u00a4'
charset=x-PCK canEncode=true '\u00a9' - '\u00ab'
charset=x-PCK canEncode=true '\u00ad' - '\u00af'
charset=x-PCK canEncode=true '\u00b2' - '\u00b3'
charset=x-PCK canEncode=true '\u00b5'
charset=x-PCK canEncode=true '\u00b7' - '\u00d6'
charset=x-PCK canEncode=true '\u00d8' - '\u00f6'
charset=x-PCK canEncode=true '\u00f8' - '\u00fe'
charset=x-PCK canEncode=false '\u203e'
charset=x-PCK canEncode=false '\uff61' - '\uff9f'

(This is a part of the fix for 5065777)

###@###.### 2004-06-22