Fix forging L1Batch too early

When scheduling an L1Batch, make sure the previous L1Batch has been
synchronized.  Otherwise, an L1Batch will be forged that may not contain all
the L1UserTxs that are supposed to be included.

Update buckets withdrawals on API

ZKInput with L2Txs compatible with circom circuits

- Til
  - update Til users BJJ key generation for better js tests
    compatibility
- Common
  - PoolL2Tx to L2Tx use AuxToIdx in case that ToIdx is 0
  - Update ZKInputs parameter descriptions
- TxProcessor
  - Fix AccumulatedFees in case that there is no CoordIdx for that token
  - Fix zki.NewExit usage
  - Use same order for AccumulatedFees & FeeIdx & FeePlanTokens
  - Add Nonce usage to ExitLeafs
  - Update TestZKInput6 and check its compatibility with circom Hermez
circuits

Add debug info in batch debug file

Swagger improvements

- Debug Batch File
    - Add details about why a batch is scheduled to be L1Batch or regular
    - Add timestamps in time and blockNumber of important events
    - Use a different naming for the files (batchNum + timestamp)
- Coordinator:
    - To make decisions and store debug info, always use the current block that
      is being mined, instead of the last mined block

Fix repeated items when coordinator is updated

Update node and apply some fixes

- Node
	- Load Coordinator Fee Account from config
		- Sign the AccountCreationMsg to generate the
		  AccountCreationAuth
		- Resolve #465
	- Wait for synchronizer termination before stopping coordinator to avoid
	  getting stuck when closing in the following case:
		- The coordinator stops reading the synchronizer msg channel,
		  and the node gets stuck sending a message to that channel.
- Common
	- Move account creation auth signature code to common.
	- Update RollupConstInputSHAConstantBytes
- Coordinator
	- Set batch status in the debug file
	- Propagate SCVariables on reorg
	- Pipeline: Get SCVariables updates
		- Resolve #457
	- Fix off by 1 error in Pipeline.shouldL1L2Batch() (which shouldn't have
	  caused any problem, but it was not right)
- KVDB
	- Delete future checkpoints after reset
	- In `ResetFromSynchronizer`, remove all checkpoints first, and follow
	  the same logic as `reset()`.
- Cli
	- Add command to generate a BabyJubJub key pair (to be used for the
	  Coordinator Fee Account)
- Node
	- Adjust example config `Coordinator.L1BatchTimeoutPerc` to avoid
	  missing the L1Batch deadline with the following setup:
		- a block is mined every 2 seconds
		- single proof server that takes 2 seconds to calculate a proof
- TxProcessor
	- Close temporary pebble used for the exit tree after usage.
		- Resolve #463

Feature/txselector2

Update account.PublicKey to account.BJJ for naming consistency with the rest of the code

- Upgrade go-merkletree version to include the last changes of Pebble
that fixes the cgo issues (which should fix #453), from:
c2b05f12d7
- TxSelector
	- Remove parameter batchNum for GetL2TxSelection & GetL1L2TxSelection
	- Add checks of ToBJJ & ToEthAddr when ToIdx>255
	- Avoid getting the sender account twice to get the TokenID of a l2tx
	- Add test to check that selected L2Txs are sorted by Nonce
	- Discard L2Tx that return error at ProcessL2Txs
- executed `go mod tidy`

Add max txs limit to tx pool

Sanitize varchars that may not be UTF-8

Feature/txselector0

- Add missing Fees to CoordAccounts after processing PoolL2Txs
- Add Nonces checks for L2Txs (txs with incorrect nonces not included
in the selection)
- Add missing MakeCheckpoint() at the LocalAccountsDB once the
selection is done
- Add TxSelector test of full flow using Til.SetBlockchainMinimumFlow0
checking balances & parameters

- Til
  - tests that were using `til.SetBlockchainMinimumFlow0` have been
updated, as now the L1CoordinatorTxs are placed simulating TxSelector
creation (when needed, not before)
  - Add `EthSk` to `tc.User` to allow to sign `AccoutCreationAuths` at
the tests flows
- TxProcessor
  - for L2Txs get TokenID from tx.FromIdx Account.TokenID instead than
  - update tests vectors for new values
from tx.TokenID (which not always is set)
- Common
  - move TxIDsFromL2Txs & TxIDsFromPoolL2Txs from `coordinator` to
`common` to allow usage from other packages

Change queries that produced undesired meddler logs

Add config flag to enable/disable meddler logs

Added new metrics tests

Update TxSelector to return CoordIdxs used & other

- StateDB
  - Update GetIdxByEthAddrBJJ to return ErrToIdxNotFound when idx not found, so can be checked at upper levels
- TxSelector
  - rm CoordIdxsDB that is no longer needed (also related methods)
  - add `getCoordIdx` method to get the Coordinator Idx for a given TokenID
  - Update coordinator account creation related to new TokenIDs from L2Txs
  - Reorganize GetL1L2TxSelection
  - return CoordIdxs used in the selection
- Update go-merkletree version which avoids marshaling Siblings to json
with 'null' value in case of empty array

- In config, add Debug.MeddlerLogs to enable meddler debug logs.  This only
  affects the node started via cli, tests swill ran with `meddler.Debug = true`
- Remove unused ethereum client functions and config parameters
- Document all the config parameters

Add circuit configuration to node config

- Remove MaxL1CoordTxs parameter from the TxSelector because this parameter
  doesn't exist
- Use ChainID in l1tx byte encoding
- Pass txprocessor configuration to batch builder via an existing parameter

Add cli command to wipe the SQL DB

- Refactor the migrations code so that packr is only called (via an init
  function in `db/utils.go`).
- When loading the migrations, make sure there is at least one migration,
  otherwise panic (this would happen if the node is built incorrectly)

Load ethereum private key

empty array instead of 404

Update ExitProofs-Siblings padding

Load an ethereum keystore when the node is started in coordinator mode.  The
private key corresponding to the forger address must be imported into the
keystore before running the node in coordinator mode.  You can see an examples
in `cli/node/load-sk-example.sh`.

Include addrs in exit responses

Fix next forgers when no bids

Abstract KVDB from StateDB

- KVDB contains the Checkpoint & Resets system
- StateDB uses KVDB and adds all the StateDB related methods