linuxconsole-dev

Re: Cyrillic and Linux Console (fwd)

[James S. <jsi...@ac...>]

Dominik have a idea on how to handle this problem?

---------- Forwarded message ----------
Date: Thu, 10 Aug 2000 12:00:28 +0300 (UKD)
From: Denis Vlasenko <vd...@vd...>
To: jsi...@ac...
Subject: Re: Cyrillic and Linux Console

> > > This is not 8-bit clean. I patched console.c in kernel
> > > source tree and now  printf("-> \x9B <-")  does display character
> > > with code 0x9B. I think it won't break anything, because
> > > any decent program will print ESC [ sequence instead of
> > > 0x9B. What do you think?
>
> Do you have the patch. I will give it a try.

Ok, it is attached. I hope I used diff correctly...

> > > However, other programs like Midnight Commander do not
> > > display that character, althought other Cyrillic chars are
> > > displayed. They don't want to mess up with CSI escape seq.
> > >
> > > How can I tell them that it is now safe to output 0x9B?
> > > In /etc/termcap? In terminfo database? Or I should patch
> > > ncurses?
>
> That sounds like you might have to patch ncurses for linux. I have to try
> out the patch to see the results.

I don't have ncurses source at the moment.

> > > Problems with input:
> > > --------------------
> > > I haven't a faintest idea how to make it work.
> > > Now my box does nothing when I press key bound to
> > > emit 0x9B. Maybe it thinks it's a Meta-Escape?
>
> Can you give me details on what you are exactly doing so I
> can repeat the results.

Ok. You can find everything in the attached tarball.

Have fun,

Re: Cyrillic and Linux Console (fwd)

[James S. <jsi...@ac...>]

Here is more info if you need it.

---------- Forwarded message ----------
Date: Mon, 14 Aug 2000 12:45:54 +0300 (UKD)
From: Denis Vlasenko <vd...@vd...>
To: jsi...@ac...
Subject: Re: Cyrillic and Linux Console

> > Since I didn't get any reply, I wonder did you receive it?
> > Also I'm afraid there wasn't much documentation, for example,
> > I had to say that 0x9B can be generated with my keymap like
> > this: press ScrollLock to switch to Cyrillic, then press
> > uppercase S.
>
> I got it and have been playing with it. Thanks for that info about
> generating that font. I haven't tried your patch yet but I have been
> playing with it with the new console code. Are the fonts 16 bits wide
> instead of the usually 8 bits?

My patch makes console 8-bit clean: all characters with codes 32-255
are printable. Previous code excluded 0x9B. 0x9B was made equivalent
to "ESC [". Maybe that emulates real VT100 behavior. I'd rather drop
this kind of "compatibility". Nobody uses this "feature".

Re my font: it is a 256-char font with 8x16 character cell.
Optimized for VGA 9x16 character cell mode, where 9th column is
added by VGA hardware (replicates 8th one for chars 0xC0-0xDF, always
empty for all other chars).
Codepage 866 - Cyrillic encoding, DOS standard. Cyrillic chars are
made thinner than Latin ones. That way it is easy to distinguish
Cyrillic and Latin chars which otherwise look the same
(A,B,C,E,H,K,M,O,P,T,X).

However, my question is why after my fix ncurses/SLang based
applications still do not display 0x9B character? They don't allow
me to type it in either. What should be fixed? Terminfo database?
Ncurses/SLang sources?

Do you have any suggestions?
--
Denis Vlasenko
Email: vda...@in...