There were a couple instances where a division was used on an epoch
to derive the corresponding sync committee period instead of calling the
`compute_sync_committee_period` function.
These instances were changed to also use the function.
This renames the `sync_committee_aggregate` field of `LightClientUpdate`
to `sync_aggregate` for consistency with the terminology in the rest of
the spec.
Some tests are currently restricted to a single phase using @with_phases
even though they could likely run unchanged in later phases. This patch
changes the default for such tests to also run in later phases. If the
beacon chain changes enough in later phases to break these tests, this
highlights that the tests need to be adjusted or extended accordingly.
Building merkle proofs is required functionality for implementing light
client sync. Although the spec currently only defines a function to
verify merkle proofs (`is_valid_merkle_branch`) there are still a few
PySpec unit tests that produce merkle proofs. This patch adds a new
generator to extract test vectors from those static unit tests, so that
light client implementations can validate their merkle proof logic.
The `test_next_sync_committee_tree` currently only supports the minimal
preset, as it incorrectly initializes the `next_sync_committee`. On the
mainnet preset, `SYNC_COMMITTEE_SIZE` is 512, but the default states use
only 256 validators, leading to an IndexError during the test execution.
`next_sync_committee` is already initialized correctly prior to the test
run using the spec's `get_next_sync_committee` function, which fills up
extra committee slots with duplicate validators in this scenario. This
makes it unnecessary to manually initialize the `next_sync_committee`.
Removed the incorrect initialization to allow testing on mainnet preset.
There are three defined unit tests for the light client sync protocol.
They all follow a similar structure. However, there is an inconcistency
how they pass the slot to compute_aggregate_sync_committee_signature.
In one instance it is passed as `block.slot`. In the other two cases
it is passed as `block_header.slot`. As the `block_header` is created
from the `block`, they share the same value. This patch makes the way
how the slot is passed consistent across all of the test cases.
