Add logging to try to diagnose SIGSEGV issues

[kcooney]

• Print message when tests start end complete
• Print summary of test runs
• Archive JVM crash logs and test logs
• Update InputValidationTest to use @RunWith(Enclosed.class)
• Add a delay before closing NetworkTableInstance in IsolatedPreferences

[cuttestkittensrule]

lgtm; We may want to include stderr on non-debug runs, but other than that, I don't have any comments. This also isn't the worst thing since all runs are debug by default here, so stderr will usually be included

[cuttestkittensrule]

fyi, the artifacts for the two SISEGV only have the hs_err_pid*.log, and not the html summary

[kcooney]

fyi, the artifacts for the two SISEGV only have the hs_err_pid*.log, and not the html summary

My guess is the SIGSEGV killed the process that was processing the test results.

[kcooney]

Looking at one of the crash logs, team2813 isn't anywhere on the stack frames.

I see multiple references to edu.wpi.first.networktables.NetworkTableInstance$ListenerStorage::lambda$startThread in the logs.

I'm guessing this is some kind of race condition between these threads and our calls to NetworkTableInstance.close()

[kcooney]

It's a race condition related to queued up network table events and the call to NetworkTableInstance.close(). I added a delay.