Re: lynx-dev Saner mouse in lynx with ncurses

[Doug Kaufman]

On Sun, 22 Nov 1998, Ilya Zakharevich wrote:

> This patch fixes all the mouse problems I see with lynx as far as
> ncurses are used.  The only thing really missing is a (possibly 
> context-sensitive) menu of possible actions assigned to button-2 of the 
> mouse.  Well, one could also enjoy a menu of possible print/download 
> methods on Shift-clicking a link!  ;-)
> 
> (These changes should be propagated into other branches (PDCURSES etc),
> but I cannot do it myself.  It should be pretty straightforward.)
> ... 
> --- ./src/LYEditmap.c~        Thu Aug  6 07:28:22 1998
> +++ ./src/LYEditmap.c Sun Nov 22 01:09:22 1998
> -/* 100..10E function key definitions in LYStrings.h */
> +/* 100..10F function key definitions in LYStrings.h */
>  LYE_NOP,        LYE_NOP,        LYE_FORW,       LYE_BACK,
>  /* UPARROW      DNARROW         RTARROW         LTARROW     */
>  
> @@ -110,7 +110,7 @@ LYE_NOP,        LYE_TAB,        LYE_BOL,
>  /* F1           Do key          Find key        Select key  */
>  
>  LYE_NOP,        LYE_DELP,       LYE_NOP,        LYE_NOP,
> -/* Insert key   Remove key      DO_NOTHING      ...         */
> +/* Insert key   Remove key      MOUSE_KEY       DO_NOTHING         */
> ...
>  #define REMOVE_KEY   269     /* 0x10D */
> -#define DO_NOTHING   270     /* 0x10E */
> +#define MOUSE_KEY    270     /* 0x10E */
> +#define DO_NOTHING   271     /* 0x10F */

I am a little worried about these changes as far as portability goes.
There are a number of pdcurses, slang, and djgpp_keyhandler specific
mappings in LYKeymap.c. I don't understand exactly how changes in
LYEditmap.c will interact, but changes in LYStrings.h should. Mouse
support has not yet been achieved in the DOS port, but is in the Win32
port. The value for MOUSE_KEY chosen here is already used by the
PDCURSES and DJGPP_KEYHANDLER with SLANG ports. In DJGPP_KEYHANDLER,
0x10E is Alt_Backspace and 0x10F is BackTab. In PDCURSES, 0x10E is
F6 and 0x10F is F7. In the SLANG without DJGPP_KEYHANDLER, 0x10F is
backspace and mapped to LYK_HISTORY. (I hope I got this right. It has
been a while since worked on LYKeymap). I wonder if it might not be
better to expand the array in LYEditmap.c to match that in LYKeymap.c
(I think that this goes up to 0x293) and move DO_NOTHING to the end
of the array, picking a value for MOUSE_KEY not currently used by any
of the programs to represent existing keys. If we bring DO_NOTHING
to the end of the array, we can get rid of the code at the end
of LYStrings.c that distinguishes between the acceptable keys in
DJGPP/Windows versus other ports.

I hope these worries are groundless, but it is probably easier to
bring this up now, rather than after the changes are incorporated in
the code.

                            Doug
__
Doug Kaufman
Internet: address@hidden (preferred)
          address@hidden