Refactor getL1L2TxSelection, which fixes some problems for certain
combinations of txs.

Update to last Poseidon (go-iden3-crypto) & go-merkletree versions,
update the affected test vectors, and check ZKInputs compatibility with
last version of circuits.

Update to last Poseidon version (which includes the changes of the
reference implementation from
26ddaa91db)

Compatible with update at circomlib
(cf853c1cc9)

Fix TxSelector for unselected tx update internal Nonce.
Also updated TxSelector config parameter.

Updated:
	batchbuilder
	common
	coordinator
	db/statedb
	eth
	log
	node
	priceupdater
	prover
	synchronizer
	test/*
	txprocessor
	txselector

Pending (once
https://github.com/hermeznetwork/hermez-node/tree/feature/serveapicli is
merged to master):
	Update golangci-lint version to v1.37.1
	api
	apitypes
	cli
	config
	db/historydb
	db/l2db

- Add config parameter `Coordinator.L2DB.MinPriceUSD` which allows rejecting
  txs to the pool that have a fee lower than the minimum.
- In pool tx insertion, checking the number of pending txs atomically with the
  insertion to avoid data races leading to more than MaxTxs pending txs in the
  pool.

- Add AmountF new parameter to ZKInputs
- Update ZKInputs test vectors with float40 checked with circom circuits
- Small fix at eth/rollup.go with lenL1L2TxsBytes with new length of
Float40

Migrate all packages to use Float40 & Add more test vectors at common

- KVDB/StateDB
        - Pass config parameters in a Config type instead of using many
          arguments in constructor.
	- Add new parameter `NoLast` which disables having an opened DB with a
	  checkpoint to the last batchNum for thread-safe reads.  Last will be
	  disabled in the StateDB used by the TxSelector and BatchBuilder.
	- Add new parameter `NoGapsCheck` which skips checking gaps in the list
	  of checkpoints and returning errors if there are gaps.  Gaps check
	  will be disabled in the StateDB used by the TxSelector and
	  BatchBuilder, because we expect to have gaps when there are multiple
	  coordinators forging (slots not forged by our coordinator will leave
	  gaps).

Fixes the case where there is a PoolTxs of type Exit with fee 0 for a
TokenID. Before this commit the Coordinator was not sending the
accumulated fee (which has value 0) to the Coordinator account for that
TokenID, with this commit the Coordinator always sends the accumulated
fee even when the value is 0 to match the hermez circom circuits
behaviour.

Also added a test to check the values and that also sends the proof to a
real proof server to check that can generate a valid proof.

PoolL2Tx.Info contains information about the status & State of the
transaction. As for example, if the Tx has not been selected in the last
batch due not enough Balance at the Sender account, this reason would
appear at this parameter.
This will help the client (wallet, batchexplorer, etc) to reason why a
L2Tx is not selected in the forged batches.

- Add tests connecting TxSelector, BatchBuilder, ZKInputs, ProofServer
- Added test to check that the signatures of the PoolL2Txs from the L2DB
pool can be verified, to check that the parameters of the PoolL2Tx match
the original parameters signed before inserting them into the L2DB

- TxProcessor move txCompressedDataEmpty inside the if of tp.zki!=nil
- Abstract generation of transactions for ZKInput tests to avoid code
repetition
  - used at txprocessor & test/zkproof tests

Test TxProcessor zkInputs against a real proof server