fix: wasn't able to decode Electra Era BeaconBlock + invalid request_hash() generated

[KolbyML]

What was wrong?

I was trying to generate E2HS files from ElectraBeaconBlock's and

• e2store era reader was failing to decode ElectraBeaconBlock
• We were generating the request_hash() wrong.

How was it fixed?

• I added Electra to get_beacon_fork, in the future we should use NetworkSpec instead, but NetworkSpec currently doesn't store CL fork data (it will in the future) so we can update that code then
• apparently for request_hash() you are supposed to convert the whole list of X objects to ssz, not do it individually

After I added these fixes I was able to generate E2HS files from ElectraBeaconBlocks.

I added a test case for requests_hash() to prevent regressions.

[morph-dev]

... but NetworkSpec currently doesn't store CL fork data (it will in the future) so we can update that code then

This is not true. We use CL fork data in several places already. In this case, we could do something like this:

network_spec().is_ethereum_fork_active_at_timestamp(
    EthereumHardfork::Prague,
    network_spec().slot_to_timestamp(slot),
)

Probably better option would be to add is_ethereum_fork_active_at_slot directly inside NetworkSpec.

I don't think that this should block this PR, I just wanted to point out that it's possible at the moment, just not very elegant.

[morph-dev]

I tried, and this specific test would pass with the old code, so it doesn't really test the regression that well :)

I think it would be better if you would use the real mainnet data (the one that it failed for). You would of course mention the slot / block number, so it doesn't appear just like random data.

[KolbyML]

I don't think real test data matters in this case, but sure I can use real testdata

[KolbyML]

... but NetworkSpec currently doesn't store CL fork data (it will in the future) so we can update that code then

This is not true. We use CL fork data in several places already. In this case, we could do something like this:

network_spec().is_ethereum_fork_active_at_timestamp(
    EthereumHardfork::Prague,
    network_spec().slot_to_timestamp(slot),
)

Probably better option would be to add is_ethereum_fork_active_at_slot directly inside NetworkSpec.

I don't think that this should block this PR, I just wanted to point out that it's possible at the moment, just not very elegant.

Not all CL forks happened at the same time as EL forks, so the idea above only works for the last 2 CL forks, but there is no guarantee it will work in the future as well. To do this properly we need to explicitly load the CL fork data into the NetworkSpec