>BTW, WRT spotting multi-byte UTF-8 encoding, I don't think that's a
>goer. Valid UTF-8 and valid GB2312 can share the same sequences,
>especially if it's just the odd `£' or `拢` in ASCII text.
It was just a suggestion, not one I was particularly crazy about ... but
not all arbitrary 8-bit sequences are valid UTF-8. And it looks like
for GB2312 (using the EUC-CN encoding, right?) it would be harder, but
there are certainly invalid sequences for GB2312. Although I do not
think this is a business we should be in; pick your locale properly
or explicitly specify a character set in the draft.
Nmh-workers mailing list