Re: Editing the 0x80..0x90 characters

[Eli Zaretskii <eliz@is.elta.co.il>]

On Wed, 8 May 2002, Stefan Monnier wrote:

> Currently those coding systems
> decode into (or encode from) ascii + latin-iso8859-5 + eight-bit-control
> + eight-bit-graphic.  The idea is to change it to decode into
> ascii + latin-iso8859-5 + mule-unicode.  So I don't see how the problem
> is mixed character sets is changed.

I thought that the purpose of that part of the change was to make the 
situation better, not just modify the charsets into which the text is 
decoded.  Sorry if I misunderstood.

> I don't understand anything about cpXXX charsets, so I'm not sure
> how to fix the problems you mentioned earlier.

All you need is to put a plist on them, like this:

  (setplist 'cp850-decode-table
            '(charset latin-iso8859-1 language "Latin-1" offset 160))

(See codepage.el for more details.)  For each cpNNN coding system, Emacs 
needs to know the target charset and the associated language 
environment.  That's because on the systems that use cpNNN, there's a 
system call to get the installed codepage, and Emacs uses that to set up 
the correct language environment automatically.

> What do you think of the patch below ?

Can you say a few words about its purpose?  I could figure that out from 
the code, but having some description as guidelines would help.  TIA.

> I'm not sure what the koi8-u stuff is about.  I suspect it's also
> meant for ukrainian so the new language environment should maybe
> be "Ukrainian" rather than "Cyrillic-KOI8-U" since that's what
> mule-cmds.el seems to expect for the "uk" locale.

KOI8-U is indeed for the Ukrainian language.  But I don't understand the 
last part: what is the importance of calling the language environment 
"Ukrainian" rather than "Cyrillic-KOI8-U"?  Is that just to make it 
easier for Ukrainian users to understand it's for them, or is there 
something else involved?