Block: do not auto-activate hardforkByBlockNumber in static BlockHeader constructor

[holgerd77]

This PR reverts on #2081 after a discussion with @acolytec3 on Discord, some extracts:

Example posted by @acolytec3, which was not working any more after the changes:

Reasoning to revert by @holgerd77:

---

I am coming more and more to the conclusion that this in BlockHeader.fromRLPSerializedHeader() was the wrong choice:

// If an RLP serialized block header is provided and no `hardforkByBlockNumber` opt is provided, default true to
    // avoid scenarios where no `opts.common` or `opts.hardforkByBlockNumber` is provided and a serialized blockheader
    // is provided that predates London result in a default base fee being added to the block
    // (resulting in an erroneous block hash since the default `common` hardfork is London and the blockheader constructor
    // adds a default basefee if EIP-1559 is active and no basefee is provided in the `headerData`)
    if (opts.hardforkByTTD === undefined) {
      opts.hardforkByBlockNumber = opts.hardforkByBlockNumber ?? true
    }

This is again one of these implicit-options behaviors which is not clear from the outside and this also breaks the above example. If someone wants a hardforkByBlockNumber behavior, it is very much possible to set the respective option (that's where the option is for). If someone wants a "HF has precedense" behavior, he/she sets the explicit HF (here: London). There one would except that everything stays in the London realm (this expectation is broken here).

On top I am already pretty sure that there are these use cases out there where people (mainly these dev tool guys like HH/Ganache) have some mainnet RLP serialized block with some artificial block number which should behave in a London (e.g.) context. We had such cases various times. This forcing into the block number logic will likely lead to problems for them.

I would somewhat be inclined to revert (unless I might have forgotten the strong reasoning why we introduced).

---

[codecov]

Codecov Report

Merging #2205 (4181609) into master (c6254ce) will decrease coverage by 1.98%.
The diff coverage is n/a.

Flag	Coverage Δ
block	93.03% <ø> (-0.05%)	⬇️
blockchain	88.82% <ø> (ø)
client	87.00% <ø> (ø)
common	98.09% <ø> (ø)
devp2p	92.25% <ø> (ø)
ethash	∅ <ø> (∅)
evm	79.07% <ø> (?)
rlp	∅ <ø> (∅)
statemanager	88.09% <ø> (ø)
trie	90.15% <ø> (ø)
tx	97.98% <ø> (ø)
util	92.33% <ø> (ø)
vm	85.24% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

[acolytec3]

LGTM