>>>> "SJT" == Stephen J Turnbull
<turnbull(a)sk.tsukuba.ac.jp> writes: 
>>>> "Nick" == Nick V Pakoulin
<npak(a)ispras.ru> writes: 
  Nick> How to make XEmacs use koi8-r fonts
instead of iso8859-5 fonts?
 It seems that I should be more prescise ;))
 The observations are that XEmacs-MULE uses iso8859-5 fonts to display
 cyrillics.  As I get it, XEmacs encodes buffer with cyrillic stuff into
 ISO8859-5 character set and then display it using iso8859-5 fonts. Right? 
 If this reasoning is right, then let's name ISO8859-5 be "basic cyrillic
 encoding".  
 Then the question is: is it possible to make KOI8-R to be "basic cyrillic
 encoding" in MULE XEmacs? Ok, now I see the problem -- KOI-8 does not have
 some characters that one can find in oher cyrillic languages (like Ukranian).
 May be it is possible to make KOI*-R to be another cyrillic charset, to be
 independent from iso8859?
  SJT> Build without Mule, use koi8-r file and process encodings, and set the
  SJT> font to koi8-r using X resources or command line arguments.
  Ok, but the question is for MULE XEmacs.
  Nick> And many people complain that they have to install 8859-5 fonts for
  Nick> XEmacs solely while all other apps do well with koi8.
  SJT> If you have an example of a multilingual app that works with koi8-r, I'd
  SJT> like to know about it.
  KDE2(qt-2.2.1) and KDE2 apps.
  XEmacs-MULE does understand koi8, but it needs iso8859-5 fonts to display
  it.  This is inconvenient.
  SJT> Does GNU Emacs do it correctly?  If so, it should be possible to port.
  SJT> Are there any tricks involved (eg, like using unibyte buffers)?
  Yes, GNU Emacs uses koi8-r fonts to display russian text.  For more details
  need some hackung :))
  
  Nick> Is there easy and straightforward way to use koi8-r fonts in Mule
  Nick> xemacs?
  SJT> Only if you are willing to limit yourself to 8859-5 portions of the KOI8
  SJT> character set.  The problem is that XEmacs basically assumes that KOI8
  SJT> is a remapping of ISO Cyrillic, and translates other characters to a
  SJT> dummy replacement.  In that case it should be possible to derive a font
  SJT> mapping CCL program from the code for KOI8 file encoding.
  SJT> Another possibility would be to create a KOI8-R charset (XEmacs
  SJT> currently doesn't have one), and install it in private space.  Like the
  SJT> "no Mule" solution, this would require that you use KOI8-R file
  SJT> encodings; you would not be able to easily save those buffers in
  SJT> ISO-8559-5.  You would also find it difficult to save multilingual files
  SJT> mixing KOI8 with any other character set (except ASCII, of course).
  SJT> I would guess that would be acceptable (probably everybody already uses
  SJT> KOI8 in their files); is it?
 Well, most russian texts/files use three encodings: koi8-r, windows-1251,
 alternativnaya(rare).  Nobody uses iso8859-5 to write in russian.
  SJT> Can you tell me where I could get free KOI8 fonts?  The only one I have
  SJT> is -etl-fixed-medium-r-normal--16-160-72-72-c-80-koi8-1 from the GNU
  SJT> intlfonts collection.
http://sunsite.doc.ic.ac.uk/Mirrors/ftp.redhat.com/pub/redhat/redhat-6.2/...
  SJT> Also, how are the KOI8-R fonts specified under X and Windows?  (I can't
  SJT> do much about the latter, I don't know anything about Windows, but I can
  SJT> at least document it for when somebody who does Windows wants to work on
  SJT> it.)
 Don't understand your question :(
  SJT> It's going to be a while before I can work much on this, but let me know
  SJT> about the GNU Emacs, what registry(s) are used for KOI8-R in X11, and
  SJT> whether saving files in KOI8-R is acceptable.
 Ok.
  SJT> I may have more questions to make things more robust and allow
  SJT> multilingual use, but answers to these should be enough to allow Mule to
  SJT> work with pure KOI8-R files and display them on screen.
Nick.