Re: No KEY_RESIZE if nodelay(,TRUE)...

[Thomas Dickey]

On Wed, Jan 02, 2002 at 03:27:15AM -0500, John Anthony Kazos Jr. wrote:
> I have written a program that does absolutely nothing but initialize
> ncurses and getch() while (true), displaying the current getmaxyx() upon
> receipt of KEY_RESIZE. It works perfectly with nodelay(stdscr, FALSE). It
> fails with nodelay(stdscr, TRUE). It fails with nodelay(stdscr, FALSE)
> with nodelay(stscr, TRUE); getch(); nodelay(stdscr, FALSE). It fails also
> if I use timeout(50) or similar rather than nodelay(stdscr, TRUE).

It's not clear what "fails" means in this context, since test/view.c
uses nodelay(stdscr, TRUE), and seems to run satisfactorily.
 
> Another problem (with nodelay(stdscr, FALSE)): some signals are getting
> lost with xterm. I changed "if (CurrChar == KEY_RESIZE)" to "if (CurrChar
> == KEY_RESIZE || CurrChar == 1)" (^A) after this occured, to test it. If I
> resize the terminal slowly (vertically; I didn't test horizontally), it's
> just fine. But if I bring it down quickly, frequently I'll end up with
> several lines of text at the top from before I invoked the executable, and
> the size displayed out of date. Hitting ^A properly clears the screen and
> displays the size (nothing is called but getmaxyx).

what's the version of ncurses you're running?  (I noticed/fixed that problem
last week, but it was broken in September from some other fixes).
 
> What's going on with these two things? I'm trying to write a program that
> uses ncurses and handles a single TCP client connection, and because using
> ncurses destroys the possibility of using select(), I'm trying to 1) getch
> until ERR and 2) read from the socket with a time out of 1/20 of a second,
> in a loop.
> 
> 
> 
> _______________________________________________
> Bug-ncurses mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/bug-ncurses

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