Change getByBlockNumber to be getByBlockHeader

[siladu]

Following #4743

We should update all the usages we can to be getByBlockHeader and phase out getByBlockNumber.

getByBlockNumber can cause issues in a post-shanghai (TimestampSchedule-enabled) chain.

This has been intentionally separated out of #4743 so it doesn't make the PR huge, see this spike for an idea of what the changes might look like: #4687

Another related item for clean-up:
#4743 (comment)

Tasks

• Simple replacements of getByBlockNumber to getByBlockHeader
• Create getForNextBlock (used in AbstractBlockCreator, PoWMinerExecutor and likely in BFT)
• Investigate FlexiblePrivacyPrecompiledContract usage in RunnerBuilder.createPrivateTransactionObserver()
• Handle Dao fork and classic fork in ProtocolScheduleBuilder
• BFT sequence number usage
• Change streamMilestoneBlocks to a findSchedule(Predicate) [need to fix usage in TransactionPoolFactory, RunnerBuilder]
• Consider deleting getByBlockNumber usage in CombinedProtocolScheduleFactory
• Fix tests to getByBlockHeader or create a test ProtocolSchedule that has getByBlockNumber inc streamMilestoneBlocks for checking milestones
• ScheduleBasedBlockHeaderFunctions getBlockHeaderFunctions usage of getByBlockNumber
• CheckpointSyncDownloadPipelineFactory getBlockHeaderFunctions usage of getByBlockNumber
• Replace HeaderBasedProtocolSchedule with ProtocolSchedule

[gfukushima]

Partial work of the trivial replaces done here: https://github.com/gfukushima/besu/tree/replace-getByBlockNumber-by-getByBlockHeader. Missing a few fixes for the tests.

[siladu]

Bug found during goerli sync: #5064