Hey guys,
After experiencing some unicode trouble with HSSF today, namely setting the cell
encoding before setting it's text (or lack thereof...), I figured since Java is
unicode based it's not too friendly to require users to set unicode flags explicitly
in order for unicode to work. I traced the problem down to UnicodeString.serialize(),
and found some rather strange code that seems to do nothing (decomposing and creating
a string, try and catch blocks that are identical...anyone have any ideas?), and
thought we should have the serializer (or perhaps the constructor?) detect the case
where the String contains non-ascii/iso-latin chars and set the encoding
automatically. This can be as simple as a
if (str.equals(new String(str.getBytes("iso8859_1"),"iso8859_1"))) // string can be
compressed
...
what do u say? any implications I didn't think of?
-Amichai
