On Wed, 15 Jun 2011 10:33:21 -0400
David McBride
The problem is the "forever" in the main thread. It never has a way to know that the two threads have died, but the stm knows that the other two channels have disappeared, so they no longer block, and this causes a busy loop.
So check this out. I would have rather done it with a state monad to count the number of threads I spawn and then wait for the appropriate number of messages to arrive, but this way works too.
This is one of those cases where datatypes are awesome. Now the thread passes back either a line to be printed, or it tells the parent thread that it has nothing left to print. That way the main thread knows exactly when to die.
Thanks a lot. I will study your solution carefully.
Also, if you are going to be making command line scripts, I highly recommend the cmdargs package on hackage, as it is pretty cool for doing commandline arguments in a safe way.
Yes, you are right of course. But here I wanted to have just a minimal example. -- Manfred