On Jan 14, 2008 6:34 AM, Werner LEMBERG <[EMAIL PROTECTED]> wrote:
>
> > > [...] if you want to access, say, U+2018, it is first checked
> > > whether LaTeX itself provides a definition for this character
> > > (defined with \DeclareUnicodeCharacter in a DFU file -- this is
> > > output in the log file, BTW), and then CJKutf8.sty looks up the
> > > CJK definition.  With other words: Your definitions in the FDX
> > > file are never seen by LaTeX!
/../
> > Is it possible to force latex to read the definition in the fdx file
> > and overwrite whatever it had previously read from a .dfu file?  I'm
> > afraid I still know far too little about latex and tex programming
> > to see exactly what issues are at stake here.
>
> Well, \DeclareUnicodeCharacter acts globally (and is allowed in the
> preamble only).  However, we need to override it locally.
>
> > Is the setting or "fetching" of the glyph for horizontal CJK also
> > not under control of the CJK package but rather done by normal latex
> > in this case?
>
> CJKutf8.sty redefines some underlying macros of \CJKchar, regardless
> of the writing direction ([EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL 
> PROTECTED], [EMAIL PROTECTED]).  I
> could imagine to add a macro which switches to and from the original
> definitions of those four macros.  Another possibility is to activate
> and deactivate the particular macros defined with
> \DeclareUnicodeCharacter for those 6 characters.

Thank you, this is all interesting information to me, since I did not
understand exactly what is happening in CJKutf8.sty. So there are
definitely possibilities, and I will try to code something to fulfill
my needs here.

It seems to me that the problem is wider than just 6 characters, as
there may be many characters that need such treatment.

> Of course, a much simpler solution is to not use CJKutf8.sty...

Now this I fail to understand properly. Maybe I misunderstand the
purpose of CJKutf8.sty?
I thought that when writing CJK, one had to choose an encoding to go
with it. For Japanese one has a choice between, say, EUC-JP or
Shift-JIS. These will allow input of Japanese and English characters
only, but not other languages. Since I wish to mix Japanese with
languages for which there is no support in EUC-JP or Shift-JIS, I
assumed that Unicode was the way to go, and that for this CJKutf8 was
proper style to use. Is this not the case?

Best regards, Gernot

_______________________________________________
Cjk maillist  -  [email protected]
https://lists.ffii.org/mailman/listinfo/cjk

Reply via email to