Rick McGowan writes:
> A long long time ago we decided to drop support for most mappings.

With much discussion, contention, and huffing, I might add. Let's not
reopen this discussion. I know Rick doesn't want to. I certainly
don't, and I expect Jenkins would prefer to avoid it too. ;-)

> There are many reasons, not least of which is that some of them are
> moving targets, and we don't have resources to support them. For
> "Shift JIS" there are many problems because there isn't just one
> "Shift JIS"; each vendor has their own flavor. The vendors should
> track their own mappings to Unicode, where they differ from JIS.

The best thing to do is pick a flavor of Shift JIS and stay with
it. At this point I recommend CP-932, only because it is probably the
most ubiquitous. While I don't work with the Japanese encodings that
often, I expect there are even different versions of CP-932 depending
on which version of Windows you are using. This is certainly true for
CP-950. This does, of course, depend on your application.

You can find many mapping tables for these at IBM's ICU site,

http://oss.software.ibm.com/icu/

In particular, 

http://oss.software.ibm.com/cvs/icu/charset/data/xml/

You can also find tables in the GNU iconv source, and numerous other
places, e.g., Mozilla.

    -tree

-- 
Tom Emerson                                          Basis Technology Corp.
Software Architect                                 http://www.basistech.com
  "Beware the lollipop of mediocrity: lick it once and you suck forever"

Reply via email to