NetBSD-current now integrates multibyte LC_CTYPE locale support, from the
Citrus XPG4DL codebase.

It currently supports both stateless and stateful multibyte encodings,
for example:

        - ISO2022 variant encodings, including iso-2022-jp and X11 compound text
        - EUC variant encodings, including euc-jp, euc-kr, euc-cn
        - euc-tw encoding
        - MS kanji encoding, commonly used in Japanese MS-DOS arena
        - Big5, used commonly in Hongkong
        - UTF-2 and UTF-8

We already have support for singlebyte encodings, like ISO-8859-1 or KOI8-R.

Now we should be able to start attacking other workareas, like wide
character stdio functions (like fgetwc), wide character curses/regex
functions, other locale types like LC_TIME and LC_COLLATE, more
locale definition files, and so forth.

Note:

To avoid code footprint bloat, we disabled multibyte LC_CTYPE locale
support for statically linked binaries. Statically linked binaries
can still enjoy singlebyte LC_CTYPE locale support (like "fr_FR.ISO8859-1").
In the future, when NetBSD supports dlopen() from within statically linked
binaries, we may be able to support multibyte LC_CTYPE locale from statically
linked binaries.

To understand goals of Citrus project, please take a look at:

        http://citrus.bsdclub.org/index-en.html

itojun

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]