In this commit, the RPC API is adapted to run on top of the new Go API.
This required lots of changes to many packages, but has a few side
benefits:
- Ethereum and LightEthereum can now be used as a contract backend.
- Some duplicated code could be removed (there is added duplication in
other places though)
- It is now much easier to see which operations are unsupported with the
light client. Package les previously relied on the full node RPC API
backend, which masked several issues because the RPC API performed
direct access to the chain database.
Changes to packages in detail:
accounts/abi/bind:
- Contract call boilerplate has moved to package core.
cmd/utils:
- les now inherits the event.TypeMux from the Node instance
contracts/release:
- The ReleaseService now uses Ethereum and LightEthereum as backend.
core:
- MissingNumber is exposed so it can be used in package eth.
- GetTransaction now returns the index as an int, for convenience
reasons.
- ApplyCallMessage has been added as the new one and only
implementation of read-only contract calls.
- TxPool exposes NonceAt instead of the more general accessor for the
ManagedState.
core/types:
- Signer.SignECDSA is gone (it was basically unused).
- WithSignature doesn't return an error anymore (all implementations panic for
invalid length). I made this change to avoid code duplication in the API.
eth:
- EthApiBackend is gone. In its place, Ethereum gains many new methods
which implement a large portion of the new Go API. It does not
yet support event subscriptions and log filtering.
- Some accessors for internal objects are gone.
- ethapi.PrivateDebugAPI and ethapi.TxPoolDebugAPI are now created in
package eth for dependency reasons.
eth/downloader:
- Progress returns a pointer to simplify callers.
eth/filters:
- The Backend interface is simpler and uses the stable Go API where
possible. The new BlockReceipts method saves one DB read because
BlockReceipts also takes a block number argument.
- ChainDB is no longer passed via the Backend interface.
- EventSystem now relies on HeaderByHash for reorgs in light client mode
instead of reading from the chain database.
eth/gasprice:
- The LightPriceOracle now uses ethereum.ChainReader instead of
ethapi.Backend.
ethclient:
- TransactionByHash is adapted for the last-minute API change which
adds the isPending return value.
internal/ethapi:
- PublicTxPoolAPI is now called TxPoolDebugAPI, moved to its own file
and talks to the transaction pool instead of using the main Backend.
- The API no longer accesses the chain database directly. All access
is mediated through Backend methods.
- The backend is now split into three interfaces.
Implementing Backend is mandatory but does not require the pending
state. The other two (PendingState, TransactionInclusionBlock) are
optional and discovered at runtime.
les:
- LesApiBackend is gone, LightEthereum gets all the methods.
- Weird accessors copied from package eth are now gone as well.
light:
- TxPool.Stats now returns a queued count of zero. It implements the
ethapi.TxPool interface and can be used with TxPoolDebugAPI.
This commit implements EIP158 part 1, 2, 3 & 4
1. If an account is empty it's no longer written to the trie. An empty
account is defined as (balance=0, nonce=0, storage=0, code=0).
2. Delete an empty account if it's touched
3. An empty account is redefined as either non-existent or empty.
4. Zero value calls and zero value suicides no longer consume the 25k
reation costs.
params: moved core/config to params
Signed-off-by: Jeffrey Wilcke <jeffrey@ethereum.org>
* core: Add metrics collection for transaction events; replace/discard for pending and future queues, as well as invalid transactions
* core: change namespace for txpool metrics
* core: define more metrics (not yet used)
* core: implement more tx metrics for when transactions are dropped
* core: minor formatting tweeks (will squash later)
* core: remove superfluous meter, fix missing pending nofunds
* core, metrics: switch txpool meters to counters
This commit replaces the deep-copy based state revert mechanism with a
linear complexity journal. This commit also hides several internal
StateDB methods to limit the number of ways in which calling code can
use the journal incorrectly.
As usual consultation and bug fixes to the initial implementation were
provided by @karalabe, @obscuren and @Arachnid. Thank you!
Shutting down geth prints hundreds of annoying error messages in some
cases. The errors appear because the Stop method of eth.ProtocolManager,
miner.Miner and core.TxPool is asynchronous. Left over peer sessions
generate events which are processed after Stop even though the database
has already been closed.
The fix is to make Stop synchronous using sync.WaitGroup.
For eth.ProtocolManager, in order to make use of WaitGroup safe, we need
a way to stop new peer sessions from being added while waiting on the
WaitGroup. The eth protocol Run function now selects on a signaling
channel and adds to the WaitGroup only if ProtocolManager is not
shutting down.
For miner.worker and core.TxPool the number of goroutines is static,
WaitGroup can be used in the usual way without additional
synchronisation.
Added chain configuration options and write out during genesis database
insertion. If no "config" was found, nothing is written to the database.
Configurations are written on a per genesis base. This means
that any chain (which is identified by it's genesis hash) can have their
own chain settings.
* Removed some strange code that didn't apply state reverting properly
* Refactored code setting from vm & state transition to the executioner
* Updated tests
