[Bug 280697] editors/fileobj fails to start due to a curses error

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=280697

Sean Farley <scf@FreeBSD.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |scf@FreeBSD.org
           Severity|Affects Only Me             |Affects Some People

--- Comment #2 from Sean Farley <scf@FreeBSD.org> ---
This looks like an issue related to the latest ncurses (v6.5) import since this
used to work with 14-STABLE and still does on 13.3-RELEASE-p6.  I think this
issue belongs to base, but I am hesitant to move it over as I am only an
alumni.  :)

When Python is built, is relies upon keyname() to determine which KEY_*
variables to incorporate. Previously, keyname() would return KEY_RESIZE when
the terminal was resized, but now, it returns NULL.

This is point where Python makes the determination.  If keyname() returns a
NULL, then the KEY will not be included. 
https://github.com/python/cpython/blob/v3.11.10/Modules/_cursesmodule.c#L4908

I did not dive into ncurses code much, but this has mention of KEY_RESIZE: 
https://invisible-island.net/ncurses/NEWS.html#t20240120

Here is a simple program to show what keyname() can return:

#include <ncurses.h>

int main(void) {
    int ch;

    initscr();
    cbreak();
    noecho();

    printw("KEY_RESIZE: %s\n", keyname(KEY_RESIZE));
    while ((ch = getch()) != ERR) {
        printw("Got: %s\n", keyname(ch));
        refresh();
    }

    return (endwin());
}

-- 
You are receiving this mail because:
You are the assignee for the bug.