Unhandled exception occurs if application is resized

[JaySmithWpg]

If the application's terminal is resized while it is waiting for players to buzz in, the following exception occurs:

Traceback (most recent call last):
  File "./jeopardy.py", line 306, in <module>
    curses.wrapper(main)
  File "/usr/lib/python2.7/curses/wrapper.py", line 43, in wrapper
    return func(stdscr, *args, **kwds)
  File "./jeopardy.py", line 194, in main
    scores, answer_server)
  File "./jeopardy.py", line 129, in run_questions_menu
    answer_server
  File "./jeopardy.py", line 165, in run_question
    answered_questions, player_names, scores)
  File "./jeopardy.py", line 223, in run_buzzin_attempts
    buzzed_in_player_id = wait_4_buzz(players_allowed)
  File "/home/jay/hacker-jeopardy/wait_4_buzz.py", line 23, in wait_4_buzz
    data = s.recv(256)
socket.error: [Errno 4] Interrupted system call

[markjenkinsskullspace]

Jeremy and I have never done a resize while operating in front of a live studio audience so as the operator for HJ-4 I don't consider this a practical problem -- I set up gnome-terminal with a large font, no menu bar, no scroll bar and full screen -- if I have to peek at anything else I just alt-tab or minimize.

Tweaking size while operating isn't going to provide the user with the experience they're going to expect anyway, even if this bug is fixed its just going to create more or less black space on the right and bottom.

Doing something much more adaptive to actually use the space would be a wishlist item, because opitmal column widths (and spacing between columns) would have to be re-selected and also height of grid rows and spacing between them.

Currently, the constants for these things are mostly not even calculated at startup, the exception being column width, everything else is a hard coded constant. Before doing a show a operator needs to determine what character cell resolution they'd like to work with (dictated by preferred font size and length of questions) and then adjust the constants accordingly.

[JaySmithWpg]

Agreed, it's a minor bug. I just have a zero tolerance policy for bugs that I may have introduced, and I'll always document them when I find them. It's not a show-stopper by any means though.
Unfortunately, the fix doesn't seem trivial - the socket listener really doesn't like being interrupted by the screen being resized. Short of moving it to anther thread (not worth it), I don't have a fix in mind. This will likely just remain a known issue.