Erroneous result when transcoding to other output
encoding than utf-8 or utf-16 using
int_daisy_unicodeTranscoder. Use for example a doc that
contains the diff chars between windows-1252 or
iso-8859-1; and output to these two encodings; output
is bit-by-bit identical.
Logged In: YES
user_id=523358
Note that the UCharReplacer functionality of this
transformer works ok, and the bug does not surface as long
as output encoding is set to utf8 or utf16.
The problem is caused by the currently used version of StAX
(Woodstox 2.0.x).
Probable solution is a) check if Woodstox 3.x solves the
problem, if not b) try the SJSXP outputwriter (although the
latest version of that seems not to represent non-bmp chars
correctly as per the jdk5 update to drop char and use int or
char[] for representation of non-bmp codepoints, need to
double check that).