Restrict allowed slashes for a validator to 1 per block height

[brentstone]

We should probably only processes a single slash per block height for a given validator.

Upon revival of the SE2 net, several validators were caught for misbehaving, and in almost all cases, the protocol has enqueued multiple slashes all at the same block height for seemingly the same single misbehavior. Here's an excerpt from the namadac slashes logs:

Enqueued slashes for future processing

To be processed in epoch 6
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1qxdlvt04892jx2m5y3dsqagrlh5grreuqgfmtftm
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1qxdlvt04892jx2m5y3dsqagrlh5grreuqgfmtftm
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1qxdlvt04892jx2m5y3dsqagrlh5grreuqgfmtftm
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1qxdlvt04892jx2m5y3dsqagrlh5grreuqgfmtftm
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1qxdlvt04892jx2m5y3dsqagrlh5grreuqgfmtftm

To be processed in epoch 6
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1q84tqxcm0l792cvtx8ufe6erx00fwwjtggv925l4
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1q84tqxcm0l792cvtx8ufe6erx00fwwjtggv925l4
Infraction epoch 0, block height 5, type Duplicate vote, validator tnam1q84tqxcm0l792cvtx8ufe6erx00fwwjtggv925l4

It doesn't make much sense to account for all of these infractions in computing their cubic slash rates. I think we should de-duplicate such instances before enqueueing the slashes.