lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: lynx-dev coloring with character-set=utf-8


From: Klaus Weide
Subject: Re: lynx-dev coloring with character-set=utf-8
Date: Tue, 27 Jul 1999 22:04:08 -0500 (CDT)

On 28 Jul 1999, Christian Weisgerber wrote:
> Klaus Weide <address@hidden> wrote:
> 
> > > [When the display character set is set to UTF-8] then in both xterm and
> > > Linux console, links are not highlighted in blue any more, and headlines
> > > are not underlined any more. Only when a link is left via the cursor 
> > > up/down
> > > keys, it gets highlighted in blue (was red before).
> > 
> > Can anyone else reproduce this?
> 
> Yessir. This bug has been around for a long time and I'm under the
> impression that I reported it.
> 
> lynx 2.8.2rel.1 on FreeBSD 4.0/i386 with the native (ancient) ncurses.
> 
> On the other hand, lynx 2.8rel.2 on Debian 2.0/i386 and 2.8.1rel.2
> on Red Hat 6.0/alpha are not affected.

[ and linux is all I can easily test with ncurses. ]

> > The following helps under linux to see what lynx actually writes
> > to the screen:  attach to the running lynx process with
> > 
> >   strace -ewrite -s1000 -p <pid>
> 
> Tracing lynx simply confirms that no highlighting codes are sent.
> This also affects underlining (EM, STRONG, B, I, etc.).

That's all very strange.  As long as no non-ASCII characters are
in the text, the calls to curses functions should be the same
whether display character set is set to UTF-8 or not.[*]  So I
don't have a clue why a different (n)curses or a different platform
should change anything.  Someone with such a setup will have to
debug this.  (If you have this problem and can compile lynx with
debugging and can use gdb or similar, I can give some tips what
to look at, to start narrowing it down.)

[*] Afaik.  One difference though: Lynx does more screen refreshing
for UTF-8 output.  Near end of display_page() in GridText.c:

    if (HTCJK != NOCJK || text->T.output_utf8) {
        /*
         *  For non-multibyte curses.
         */
        lynx_force_repaint();
    }

Does taking this out change things?   

I really don't know why it should, if lynx_force_repaint() were 
wrong for some curses versions then there should be more problems
e.g. with ^L.

   Klaus

   


reply via email to

[Prev in Thread] Current Thread [Next in Thread]