Hi all, I might have found a way to avoid unkilled subprocesses when they are stuck in the open syscall. When the parent shell is killed, we can force it to open and close all the named pipes that it still have. It might be a bit harsh, but at least there is no unkilled processes anymore.
It might also fix your case, Chet. Though I haven't checked. > There's a case very similar to that in the test suite, which is supposed to > test a degenerate case (programmer error). What should the shell do if a > script creates a FIFO and doesn't do anything with it? Bash tries to detect > these sort of stray FIFOs and remove them. I'm currently checking if it's working correctly on daily uses. And if yes, I'll submit the patch to see if it can be merged as is. Sincerely, Clément
