gh-116622: Kill Android Signal Catcher thread before running tests

[mhsmith]

This is an alternative attempt to fix the problem described in #123981, by killing the Android Signal Catcher completely. But it ended up being quite large and complex, and depends on several internal details which Android might change in the future.

Also, when I tried to test the "SignalCatcher TID %d still exists" message by commenting out the first tgkill call, I found that about 80% of the time, the 0.1 second delay between installing the signal handler and sending the signal somehow caused the handler to be bypassed, and the signal to kill the process. I have no idea why that would happen, and that makes me feel even less confident about this approach.

As #123981 says, the original problem only occurs in a fairly rare combination of circumstances, has a clearly-identifiable log message, and some easy workarounds, so it isn't worth pursuing this.

• Issue: Add support for Android as a target platform #116622