Unicode support

Marcin 'Qrczak' Kowalczyk qrczak@knm.org.pl
10 Oct 2001 11:16:13 GMT


Tue, 9 Oct 2001 14:59:09 -0700, John Meacham <john@repetae.net> pisze:

> I think a cannonical way to get at iconvs ('man 3 iconv' for info.)
> functionality in one of the standard librarys would be great. perhaps
> I will have a go at it. even if the underlying platform does not have
> iconv then some basic conversions (utf8, utf16, latin1, [Char]) could
> easily be provided with the same API and minimal implementation effort.

I've already done some of this: <http://www.sf.net/projects/qforeign/>.
There is a framework for conversions - I'm not entirely happy with
it but I don't know how to do it nicer.

I implemented a binding to iconv, UTF-8, some 8-bit encodings. It
should be suitable for associating Handles with a conversion: in
particular the conversion performed by iconv can be done without
intermediate Haskell lists but directly on some handle buffers,
provided that we can get a Ptr to it (it's tricky in the case of
ghc's ByteArrays).

There is also a Char module replacement with extra character
properties.

It works with ghc and parts of it work with nhc98.

-- 
 __("<  Marcin Kowalczyk * qrczak@knm.org.pl http://qrczak.ids.net.pl/
 \__/
  ^^                      SYGNATURA ZASTĘPCZA
QRCZAK