Re: LYNX-DEV ac98: 1.) configure,

[Foteos Macrides]

"T.E.Dickey" <address@hidden> wrote:
>> >Numbering probably starts at 0 for internal KEY_F(n) codes (I'll have to 
>> >look at
>> >the terminfo description to be sure, but I've been bitten by that before).
>> 
>>      KEY_F(15) also doesn't work, but I tried KEY_REDO and that's
>> DO_KEY (i.e., the VTxxx Do key works, too, now with Unix curses in the
>> devel code).  It looks like you need those cases only for wgetch(), not
>> getch(), at least with the native solaris2 curses.  Also note that you
>> don't need any of that stuff for the devel code built with slang.
>
>In ncurses, at least, it would be KEY_F(16) - I tested that.
>
>I'm not sure about Solaris curses: I implemented the change this morning and
>tested with ncurses on XFree86 xterm (which does provide an extra set of
>function-keys, so I have a working F16).  I just logged in to check my
>test-builds on slcc, so I can check in the change.

        Here's what I added at the bottom of the keypad() mode switch()
section:
        
#ifdef KEY_REDO
        case KEY_REDO:             /* VTxxx Do */
           c = DO_KEY;
           break;
#endif /* KEY_REDO */
#ifdef KEY_FIND
        case KEY_FIND:
           c = FIND_KEY;           /* VTxxx Find */
           break;
#endif /* KEY_FIND */
#ifdef KEY_SELECT
        case KEY_SELECT:
           c = SELECT_KEY;         /* VTxxx Select */
           break;
#endif /* KEY_SELECT */
#ifdef KEY_IC
        case KEY_IC:
           c = INSERT_KEY;         /* VTxxx Insert */
           break;
#endif /* KEY_IC */

I don't know why a KEY_foo isn't needed for REMOVE_KEY in the devel code,
but since it isn't, I have no way to track it down on Scott's SPARC and
add a case for it.


>Also am not sure about the VMS problem: the symptoms point to something
>have gotten trashed that's only picked up when 'exit()' runs (e.g., a
>signal handler or a file-pointer).  Since I don't test on VMS, I didn't
>see that - I didn't see it months ago when I did a test there.

        I posted a message about that.  Since the problem is identical
in the devel code and 2.7.1fm, and that stderr redirection trick is
done identically in the two code sets, it's a good bet.  The problem
seems specific to VAXC or VAXen.  It works fine with DECC on VMS/AXP.
There are other problems for VMS in the devel code which the old
VAXC compilers don't report (but they don't matter if this problem
isn't solved :).

        For that problem with HTmmdecode(), it's still only used
when HTCJK == JAPANESE, so for now I put back the v2.7.1 HTmmdecode(),
and #ifdef NODEFINED out Klaus' "generalized" version.

                                Fote

=========================================================================
 Foteos Macrides            Worcester Foundation for Biomedical Research
 address@hidden         222 Maple Avenue, Shrewsbury, MA 01545
=========================================================================