When a parent block is rejected, also reject its children

[teor2345]

Motivation

After Zebra rejects a queued non-finalized block, it keeps on processing that block's children, as if their parent was valid.

This might cause panics or hangs, or break consensus rules in the non-finalized state.

I'm not sure if we should merge this fix, leave it for later, or turn it into a ticket.

Equivalent zcashd Behaviour

zcashd implements a similar "drop all children of invalid blocks" behaviour:

Pruned nodes may have entries in setBlockIndexCandidates for
which block files have been deleted. Remove those as candidates
for the most work chain if we come across them; we can't switch
to a chain unless we have all the non-active-chain parent blocks.

https://github.com/zcash/zcash/blob/c34162d6ddfaa8d6276aa0628cee838a51ad9df7/src/main.cpp#L3905

See also BLOCK_FAILED_CHILD:

BLOCK_FAILED_CHILD = 64, //! descends from failed block

https://github.com/zcash/zcash/blob/785803382a27c85301bc75b34576c3704966fe87/src/chain.h#L176

And its uses:
https://github.com/zcash/zcash/blob/c34162d6ddfaa8d6276aa0628cee838a51ad9df7/src/main.cpp#L3919
https://github.com/zcash/zcash/blob/c34162d6ddfaa8d6276aa0628cee838a51ad9df7/src/main.cpp#L4144

Designs

Non-finalized state validation:
https://github.com/ZcashFoundation/zebra/blob/main/book/src/dev/rfcs/0005-state-updates.md#pubsuper-fn-queue_and_commit_non_finalized_blocksmut-self-new-arcblock---tokiosynconeshotreceiverblockhash

Solution

• When a parent block errors, reject all descendant blocks

Review

Do you think we should merge this fix, leave it for later, or turn it into a ticket?

We might add tests for this error case as part of upcoming PRs.
We could hold this PR until it has tests.

Reviewer Checklist

• Code implements Specs and Designs
• Tests for Expected Behaviour
• Tests for Errors