Open Hoverbear opened 8 years ago
Partitioning would be to just split the available peers per simulation and run them independently. Partition and merge functions in the simulator will need to cover this.
EDIT: The most important thing to note here is that the tests would be deterministic without I/O or platform specific code. With no side affects - the tests become proofs.
It may be useful to have the debug builds of Raft feature a testing harness. This could be used to simulate failure, delays, etc. Up in the air about what it might even look like, but I assume we could use some
#[cfg(feature = "harness")]
or some manner or speaking.Possible features: