Shall we declare Tkinter dead? No-one seems interested in it ;o)

[jepler epler]

All the code below is untested, but I'm familiar with the techniques
from pure tcl/tk programming.

You can have a piece of code be called when the window-manager "close"
function is invoked:
	blah = Toplevel(...)
	blah.wm_protocol("delete", func)
func might set a global variable, raise an exception, or manipulate a
semaphore.

You could also use the wait_window method:
	blah = Toplevel(...)
	blah.wait_window()
wait_window essentially executes the tk mainloop (potentially recursively)
until the window is destroyed.

Finally, you could use the wait_variable method:
	blah = Toplevel(...)
	var  = Variable()
	blah.wait_variable(var)
arrange to set var to a new value when the window is closed.

Whether these work properly in the presence of "microthreads", I admit I'm
unsure.

On Sun, 29 Oct 2000 21:42:59 -0500, Mike Fletcher
 <mfletch at tpresence.com> wrote:
>Side note: are all the mouse interactions under Tkinter atomic C functions?
>For instance, dragging or resizing the window stops all other micro-threads
>until the drag/resize stops.

This may depend on the platform where you use tk, and when under X11
on the details of your window manager.  I use X11 and a window manager
called "icewm".  When I enable "solid" moves/resizes, a tkinter app is
able to respond during the operation.

Jeff
PS if we had to jettison tkinter in favor of some other gui, I wish that
someone would set up gtk+pygtk+glade for me to run on win32.  I looked
at wxpython once, but it was all greek to me, and I couldn't get it to
run under X11 at all, just win32.  The point being, I think it's probably
largely a religious argument.  Tkinter works, and it gets installed by
default everywhere, and I'd be the first to admit the pragmatism of
"doing in Rome as the Romans do."