Patrick Smith wrote: > > Well, the problem is that you can't simply kill a thread--it shares > > memory with other threads that it could be leaving in an inconsistent > > state. Imagine that it was, say, holding a lock when it was forceably > > killed. Now any other thread that tries to acquire that lock will > > block forever. > > > > You really do need the thread's cooperation so that it only exits when > > everything is in a kosher state. Can you dig into the incredibly long > > function and change it to do the flag-checking and exiting safely? > > > > I second the notion to consider using subprocesses instead of threads; > > that's almost always a good idea unless you're really sharing a lot of > > complex data structures. > > > > Hi, > Thanks again for your replies. > > The way this was originally working was actually using a subprocess, > however, the people who this program is for were having problems that when > the application was closed, the subprocess was still around in the system,
Have the main process kill the subprocess when it exits; this should be fine unless you use kill -9 or the windows equivalent (e.g terminating from the task manager), in which case you can kill the subproc manually as well. > Also, the people who I am doing this for need something ASAP and for now are > fine with the cancel button completely closing the entire program. Is there > at least a nice way to do this? maybe os.abort or os._exit depending on what you want and how they behave on your platform? -- http://mail.python.org/mailman/listinfo/python-list
