Re: lynx-dev a problem of Japanese with 2.8.5pre.5 compiled with ncursesw

[Thomas Dickey]

On Wed, 18 Feb 2004, Atsuhito KOHDA wrote:

> From: Thomas Dickey <address@hidden>
> Subject: Re: lynx-dev a problem of Japanese with 2.8.5pre.5 compiled with 
> ncursesw
> Date: Thu, 12 Feb 2004 19:26:16 -0500
>
> > > The problem is very difficult to explain in English
> > > but, for example, when you visit
> > > http://w1.avis.ne.jp/~nabetaro/text/
> > > and select ?$B!VLO7?JWNr!W then you would see a line something
> > > like ?$B!V$*2[;R$N$* ?$B$1;~Be!W which should be ?$B!V$*2[;R$N$*$^$1;~Be!W
> >
> > I'm not seeing that - probably because I don't know what display charset,
> > etc., to set it to.  (I tried 7-bit and ISO-8859-1 and EUC-JP).
>
> Sorry, I overlooked this mail.  Display charset is EUC-JP
> in my system but it is difficult for me to explain the
> situation precisely in English so I attached screenshots
> of lynx.
>
> - wrong.png is a screenshot of the problem, one letter lost.
>
> - correct.png is a screenshot after pressing "r" key and
> so correctly displays the page.

I think this is a bug that has been fixed in current ncurses.  I have
(actually several flavors) lynx built with current ncursesw and am looking
at the http://w1.avis.ne.jp/~nabetaro/text/ page, following the link to
http://w1.avis.ne.jp/~nabetaro/text/model1.html I see a page which looks
like your picture - the correct one.  This is using mlterm with the
encoding set to EUC-JP.

Perhaps in a week or so, the Debian packages for ncurses will catch up
to the 5.4 release (last week).

When I was looking before, I was using uxterm, which wouldn't have the
same fonts, etc.  I can see that lynx doesn't render _that_ into UTF-8
the same way the EUC-JP is handled in mlterm (so that's something more
to investigate - but a different matter).

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden