Re: [tlug] Why change a linux server's locale?

[Darren Cook <darren@example.com>]

> Our customers include companies with very old systems. I suspect that
> their data is presently is sjis or similar and they want to keep it
> that way. It makes me nervous and I wish someone had mentioned it
> when I was setting up the dev server.

UTF-8 is mostly 3 bytes for Japanese, compared to 2 bytes for Shift-JIS.

To put that in perspective, when summed over all their text databases
that can often steal enough memory and disk space to store two whole
scans of a faxed customer order.

> I think I'll do some local testing and see if how I can break sjis

Katakana "so" was the usual way; we used it to test string escaping had
been correctly implemented. The second byte in Shift-JIS is a backslash. (*)

Darren

*: That is from distant memory... I've thankfully only dealt with UTF8
the past 15-20 years. After all, once the Katakana Digraph Koto
(https://en.wikipedia.org/wiki/Kana_ligature) was added to Unicode 3.0
(in 2002) I knew I'd have every Japanese glyph I'd ever need. :-)