Update time format in logs and test failure messages

[sgdxbc]

The log format is updated with nanosecond-scale resolution. Sample:

[FINER  ] [2025-01-15 13:30:54.139487329] [dslabs.framework.Node] MessageReceive(client1 -> pingserver, PingRequest(ping=PingApplication.Ping(value=Hello, World!)))
[FINER  ] [2025-01-15 13:30:54.142355292] [dslabs.framework.Node] TimerReceive(-> client1, PingTimer(ping=PingApplication.Ping(value=Hello, World!)))
[FINER  ] [2025-01-15 13:30:54.150642300] [dslabs.framework.Node] MessageReceive(client1 -> pingserver, PingRequest(ping=PingApplication.Ping(value=Hello, World!)))
[FINER  ] [2025-01-15 13:30:54.151288087] [dslabs.framework.Node] MessageReceive(pingserver -> client1, PongReply(pong=PingApplication.Pong(value=Hello, World!)))
[FINER  ] [2025-01-15 13:30:54.153208765] [dslabs.framework.Node] MessageReceive(pingserver -> client1, PongReply(pong=PingApplication.Pong(value=Hello, World!)))

The failure message for maximum waiting time exceeded, while already in nanosecond-scale resolution, has been updated to keep consistency. Sample:

java.lang.AssertionError: client3 waited too long, 8576 ms (3000 ms allowed), started waiting at 2025-01-15 13:27:20.554216249

The previously second-scale resolution in log message is not enough for identifying livelock-related bugs, because the timeouts are in millisecond scale.

Besides, I am planning to make a message diagram style log visualizer for run tests, which can benefit from finer-grained time in the logs.

(The original idea is to update with microsecond-scale resolution, but apparently there's no format conversion for that according to Formatter.)

[emichael]

Much better, thanks! I'm slightly inclined to cut this off a microseconds as nanos is never really useful, but this is certainly better than only having seconds.