Re: Possible keypad() bug

[Thomas Dickey]

On Sun, Jan 19, 2020 at 05:17:14AM -0700, address@hidden wrote:
> Possible bug:
> ncursesw and terminfo v:6.1.20180923 (Fedora 30, GNOME v:3.32.2 or Konsole
> v:19.08.3)
> 
> In this version, keypad() no longer allows the ‘escape’ key through to the
> application.
> If keypad() is turned off (no translation), ‘escape’ comes through correctly.
> 
> Is this intentional, or a problem?

I haven't seen that: usually when I do weekly patches, I run the
ncurses test-program with 1-2 menu selections (to validate the
test-packages that I build to validate the patches, etc).

Those menu-selections exit the test-screen with the escape key,
and have keypad enabled.

(There's a short delay of course - always been there).

> Any hint or guidance on this would be welcome.

I've a Fedora 30, and just tried the ncurses-examples with that.
        https://invisible-island.net/datafiles/current/ncurses-examples.tar.gz

Seems to work.  I did
        ./configure --with-screen=ncursesw && make && ./ncurses

and (for example) the a/A menu entry has keypad enabled.
Pressing the escape key stopped that screen after a second.

Doing

        ESCDELAY=0 ./ncurses
and
        ESCDELAY=3000 ./ncurses

changed the delay time as expected.
> 
> I have scanned the change log since that time and do not see any mention of
> changes to keypad(). The docs still describe the “short delay” to determine
> whether an escape sequence follows the escape keycode, after which the
> escape should come through the stream as an independent keycode.
> 
> I am the maintainer of the NcDialogAPI, which relies upon the ‘escape’ key
> when user want to discard the edits within a dialog control and move to the
> next control.
> 
> Mahlon
> http://www.SoftwareSam.us/
> address@hidden
> 
> 
> 

-- 
Thomas E. Dickey <address@hidden>
https://invisible-island.net
ftp://ftp.invisible-island.net

signature.asc
Description: PGP signature