UTF-8 in lynx

[Klaus Weide <kweide@enteract.com>]

Hello,

I have made some changes recently to improve behavior of Lynx with
"display character set" == "UNICODE (UTF-8)".  Some of them have been
already added to Lynx 2.8.3dev.13 at <http://www.slcc.edu/lynx/current/>,
and another patch is currently in <http://enteract.com/~kweide/lynx/>
(expected to become part of 2.8.3dev.14 soon, unless TD finds something
to object - he usually doesn't).

I would appreciate some testing from this UTF-8 crowd...
(either apply the diff to 2.8.3dev.13, or wait for 2.8.3dev.14)

Specifically, Bruno had some problem with previous versions which I have
not been able reproduce, an interaction between curses refreshing and
background and UTF-8 characters (or something like that...).  I would
appreciate confirmation that I have not re-introduced this or a similar
problem with my changes.

Don't expect too much - there are no fundamental changes.  The UTF-8
display mode of Lynx doesn't work together with CJK charsets, it makes
the assumption that all characters have the same width of 1 cell, etc.
And it tries to make the best out of running under a non-multibyte-aware
display library, which can't be perfect (or even very good).  My
changes are mostly about removing or minimizing various display
glitches.  One day, when there is a (generally available) UTF-8-aware
curses, most of them will be unnecessary - but I don't see that day
around the corner yet.

For best results with UTF-8 display: compile with slang library, and
explicitly define SLANG_MBCS_HACK macro.  (I usually do that last one
by compiling as `make SITE_DEFS="-DSLANG_MBCS_HACK'.)
For worst results (with more remaining glitches), configure
--enable-color-style.


Feedback appreciated to this list or <lynx-dev@sig.net>

   Klaus

-
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/lists/