>I use the xsl:output tag with method="html" encoding="ISO-8859-9". And
the
>international chars show up like "İ" in both implementations. I also
>use xsl:output tag with method="text" and the difference is that the 1.2.2
>version produced raw data in the encoding whereas the 2.1.0 produces "
İ"
> still. Is this correct behaivour?
Xalan should produce "raw data" *IF* the character you're trying to write
out is supported in the encoding you specified. If it isn't, HTML and XML
_must_ fall back on numeric character escapes in order to correctly
represent the data.
Unfortunately, plaintext has no standard equivalent of numeric character
escapes... or more accurately, has many incompatable conventions, depending
on what you intend to do with the text, on which platform. If the character
isn't supported by the encoding, there really isn't a lot Xalan can do to
recover. I believe the XML spec actually says that an XSLT processor should
report that case as an error. (In Mike Kay's book, this is at the bottom of
page 260.)
Basically: If there isn't a representation of the character in that
encoding, there _IS_ no correct "raw data" for Text output. Pick another
character, or another encoding, as appropriate.
