On Wed, 29 Mar 2000, Mike Castle wrote:
> How about getting rid of slang?
>
> As was mentioned on the list before, slang supports regular curses
> interface these days,
It tries to, but it doesn't go all the way.
Some functions/macros of the provided interface are clearly just
no-ops, for example, from my (1.3.9) slcurses.h:
#define hline(x,y) SLcurses_nil ()
#define vline(x,y) SLcurses_nil ()
#define wdeleteln(win) SLcurses_nil()
#define nonl SLcurses_nil
#define echo SLcurses_nil
#define noecho SLcurses_nil
#define nl() SLcurses_nil()
#define tigetstr(x) NULL
> and there is even a bit of code that hangs around
> slang just for the sole purpose of supporting lynx.
If you're referring to (again, slcurses.h):
/* This is a temporary hack until lynx is fixed to not include this file. */
#ifndef LYCURSES_H
- that 'fix' hasn't been necessary for years. Lynx does not 'include
this file'. It may have done so for a short time, when slang support
was first added. I don't know why John Davis still keeps that hack
around, it isn't needed for any moderately recent lynx code.
> And I was under the
> impression that that is the new recommended way of using slang.
My impression is that Tom is under a different impression...
Anyway, you could try for yourself whether slang now provides a
curses emulation that is complete enough for Lynx's needs, at least
if you have a system where library and header files are under your
control. Remove/rename/hide your existing (n)curses header and lib
files, and sneak in the slang files in their place...
(for example, 'ln -s /usr/include/slcurses.h /usr/include/curses.h',
and similar for the libs). Try to configure and compile Lynx so that
it thinks it's using curses. Theoretically this should work...
Klaus
