Avoid applying fork choice during snap-sync

[MegaRedHand]

We're currently calling apply_fork_choice in our forkChoiceUpdate handler during a snap-sync. We should instead return a "SYNCING" early.

We already had this functionality, but it was commented out in #5196:

ethrex/crates/networking/rpc/engine/fork_choice.rs

Lines 222 to 237 in 1bebff3

	/* Revert #4985
	if context.syncer.sync_mode() == SyncMode::Snap {
	// Don't trigger a sync if the block is already canonical
	if context
	.storage
	.is_canonical_sync(fork_choice_state.head_block_hash)?
	{
	// Disable snapsync mode so we can process incoming payloads
	context.syncer.disable_snap();
	} else {
	context
	.syncer
	.sync_to_head(fork_choice_state.head_block_hash);
	return Ok((None, PayloadStatus::syncing().into()));
	}
	} */

[MegaRedHand]

Update: this only mitigates the issue. It can still happen that we process an FCU that jumps to before our latest state. We should address that.

Edit: I'll keep this alive since it's a different problem