David Elworthy <[EMAIL PROTECTED]> writes:
> Hugs 1.4 will process Kanji (Japanese) characters on Japanese Windows 95
> when input is read from a file, but not when it comes from stdin. I
> believe this is because the terminal characteristics are set to read one
> character at a time (using the VMIN setting), which prevents the IME
> (input method editor) from being able to intercept them and present
> possible Kanji for them. The IME generally needs 2+ characters before it
> can do this.
>
> I don't know enough about Japanese processing to know what to do about
> this: we might be able to find out something from our Japanese
> colleagues. However, if you have any experience in the matter, we'd
> appreciate suggestions.
Gosh, I know nothing about Japanese processing.
The standard Haskell IO library provides the function hSetBuffering to
let you fiddle with buffering of Handles (including stdin). We
haven't implemented it yet - but we've already had one request for it.
Let me know if your Japanese colleagues think it would solve your problems.
(If you have GHC installed, you could try it out with GHC and see if
it helps.)
Alastair
