test(debug_getRawReceipts): use post-Byzantium block for get-block-n test

[mattsse]

Summary

Block 0x3 is pre-Byzantium (byzantiumBlock=9 in the test chain's genesis), which means receipts at that block use the legacy state root encoding instead of EIP-658 status code encoding.

This causes clients that only support EIP-658 status encoding (like reth) to fail the debug_getRawReceipts/get-block-n test, even though they correctly implement the EIP-658 receipt format.

Changes

Change the test to use block 0xa (10) which is post-Byzantium, ensuring the test validates EIP-658 status encoding which is what all modern clients use post-mainnet-Byzantium.

Context

• Byzantium activated on mainnet at block 4,370,000 (Oct 2017)
• Pre-Byzantium receipt encoding with state root is extremely rare in practice
• Reth deliberately doesn't store pre-Byzantium state roots as discussed in paradigmxyz/reth#853

This change maintains test coverage while aligning with modern client implementations.

[MysticRyuujin]

On the one hand, I understand the reluctance to support legacy stuff, but on the other hand, this is literally part of the blockchain's history and it does not change the fact that if clients request legacy receipts from Reth or Erigon they get a different response with different values. This is not fixing anything, it's just sweeping the API response differences under the rug.

[fjl]

What needs to happen here, is that we split the test suite into profiles, where one of the profiles is compatibility with legacy formats.

[canepat]

What needs to happen here, is that we split the test suite into profiles, where one of the profiles is compatibility with legacy formats.

Either this or the profiles could be e.g. archive_node, full_node, pruned_node and this test at block 0x03 becomes part of the archive_node profile