Re: glibc UCS4_to_BIG5-HKSCS mapping should contain compatibility points

[Anthony Fok]

Hello Ulrich,

On Sat, Feb 23, 2002 at 03:29:23PM -0800, Ulrich Drepper wrote:
> Anthony Fok <address@hidden> writes:
> > Thanks to a reminder from Ben Collins earlier today, the code now
> > consistently uses unsigned char.  Attached is a new patch.
> > (glibc-2.2.5-big5hkscs.patch2.bz2).  Thanks!
>
> The patch does not work or the test suite is not adjusted.  Always run
> the test suite before submitting a patch.  And please add a ChangeLog
> entry.

Thanks for accept the zh_HK locale patch.  As for big5hkscs.c, after
some discussion with Andrew Fung (ITSD, HKSARG) and with Roger So, we
agreed that it is time to move to the HKSCS:2001 standard, but for now,
in the Big5-HKSCS -> Unicode direction, we don't map beyond BMP (i.e.
don't go to CJK Ext B U+20000 and above) until the rest of the system
(XFree86, fonts, etc.) have caught up to the new standard.

We will be working on this new big5hkscs.c soon, and will send you the
result when it is available.  Thanks for your note about the test
suite; I will update that too.  Do we need to update
localedata/charmaps/BIG5-HKSCS too?  If so, do the Big5-code entries in
localedata/charmaps/BIG5-HKSCS need to be unique?  For example:

        <UF308>      /x88/x45     <CJK>
        <U2010C>     /x88/x45     <CJK>
and
        <U3551>      /x89/xE8     <CJK>
        <UF426>      /x89/xE8     <CJK>

In big5hkscs.c, the following mappings would be used:

    from_ucs4: U+F308 -> "\x88\x45", U+2010C -> "\x88\x45";
      to_ucs4: "\x88\x45" -> U+F308

    from_ucs4: U+3551 -> "\x89\xE8", U+F426 -> "\x89\xE8";
      to_ucs4: "\x89\xE8" -> U+3551

Thanks for your help!

Cheers,

Anthony

-- 
Anthony Fok Tung-Ling
ThizLinux Laboratory   <address@hidden> http://www.thizlinux.com/
Debian Chinese Project <address@hidden>       http://www.debian.org/intl/zh/
Come visit Our Lady of Victory Camp!           http://www.olvc.ab.ca/