Network.shutdown() called Peer.shutdown() on each Peer
while holding synchronization lock on this.connectedPeers.
This would cause a problem during Peer.shutdown() as some
other reachable code would also want synchronized access
to this.connectedPeers. Typical symptoms would be log
entries like:
2019-07-02 11:13:05 DEBUG Peer:512 - Message processor for peer 192.144.182.73:9889 failed to terminate
Eventually Network.shutdown() would exit, releasing synchronization
lock and awaking stuck Peer threads, which could then try to access
repository (now closed) causing further log spam.
Now it uses Network.getConnectedPeers to return duplicated
List<Peer>, minimizing lock time on this.connectedPeers.
Also made Peer main thread logging more informative when a IOException
occurs, as most situations are harmless EOF or connection reset by peer.
Refactored duplicate code into Transaction.setInitialApprovalStatus().
This is make sure transactions HAVE a group-approval status
in Synchronizer before Block.isValid is called.
This wasn't a problem for new, unconfirmed, individual transactions
arriving over the wire due to Transaction.importAsUnconfirmed()
doing the right thing.
Also added a groupApprovalTimestamp to BlockChain feature-triggers
to support legacy chains.
Tested by syncing mainnet from scratch.
lastHeight/blockSig/blockTimestamp, etc. moved from PeerData/repository
to Peer as it's transient so no need to store in repository.
Repository now keeps track of when/who added peer, e.g. API/INIT/another peer.
API calls DELETE /peers and DELETE /peers/known also disconnect peer
as well as deleting from repository.
Connection timestamp now reported by API call GET /peers
Some repository-updating code removed from Network/Controller as no
longer needed.
Removed obsolete Controller.hasShorterBlockchain predicate.
Previous version was prone to throwing exceptions during shutdown
sequence, especially in Peer/Network-related threads.
Shutdown now tries to wait for Peer/Network threads to cleanly exit.
API call POST /admin/forcesync now tries to grab blockchain lock,
with timeout, and then repeatedly sync with requested peer until
either fully synced or something unexpected happens.
BlockGenerator will now attempt to generate a new block if none of its
peers have a recent block either (in case of network stall). BlockGenerator
still needs a minimum number of peers before generating though.
Reduce BlockGenerator workload and use of blockchain lock if it
can't generate a block.
Reduce Synchronizer logging output.
Unify calculating timestamp threshold for 'recent' block into Controller.
Don't disconnect peers that fail to send a requested transaction,
as they may no longer have it. e.g. transaction might have expired
or become invalid.
For some other cases, e.g. we have transaction already, move on to
requesting the next transaction instead of giving up on the list.
NOTE: Downloaded update JARs are now expected to have been XORed with 0x5A!
This is to help prevent Windows Firewall from blocking update downloads
based on deep packet inspection.
Download read timeout reduced from 5s to 3s.
Download locations reordered so github entries are at the top as they have
better CDNs.
ApplyUpdate now assumes null response from GET /admin/stop means node
is not running.
ApplyUpdate now checks replacement JAR actually exists before attempting
to overwrite previous version.
ApplyUpdate now tries to use Windows EXE launcher in preference to raw
java command line. (This should improve Windows installer behaviour
in detecting running process and possibly firewall implications too).
Included unit test to cover change.
Modified test blockchain config "test-chain-v2.json" to add maxProxyRelationships.
Added comments to some proxy-related methods in AccountRepository class.
Added transaction [de]serialization test, along with corresponding random transaction generators.
Minor typo fix in Transaction.
Minor clarification in MessageTransactionTransformer.
Added debugging to Account.
Arbitrary transactions now [de]serialize data-type (raw/hash) for v4+ transactions.
Data type also stored in repository. Very small (<=255 byte) data payloads are also stored directly in HSQLDB.
Added ArbitraryDataManager which looks for hash-only data payloads and possibly requests raw data over network
depending on 'policy' (which currently is "fetch everything").
Added networking support for finding, and transferring, arbitrary data payloads.
Minor optimization to message ID generation in Peer.
Minor optimization in Serialization.serializeSizedString()
Bumped version
Controller no longer uses block height to determine whether to sync
but now uses peer's latest block's timestamp and signature.
Also BlockGenerator checks whether it's generating in isolation using
the same peer info (latest block timestamp and signature).
Added API call POST /admin/forcesync peer-address to help get wayward
nodes back on track.
Unified code around, and calling, Transaction.importAsUnconfirmed().
Tidied code around somelock.tryLock() to be more readable.
Controller (post-sync) now broadcasts new chaintip info if
our latest block's signature has changed, not simply the height.
Network.broadcast() only sends out via outbound peer if node has
more than one connection from the same peer. So Controller would
only update one of the peer records with chaintip info.
Controller now updates all connected peers with the ID when it
receives a HEIGHT or HEIGHT_V2 message.
Added node1 thru node7.mcfamily.io to default peers in Network.
Network ignores first "listen port" entry when receiving peers
list from an outbound-connection peer as it already knows
by virtue of having connected to it!
More network message debug logging (hopefully never to be seen).
[some old code left in, but commented out, for a while]
Controller now sets (volatile) requestSync flag when a peer sends new height info.
This allows much quicker response to new blocks which might hopefully improve synchronization
compared with the old periodic sync method.
"Unsolicited" network messages are now added to a BlockingQueue,
and a separate unsolicited message processing thread (one per peer)
deals with this messages in turn.
This allows "reply" network messages to propagate up to the
threads that are waiting for them, preventing deadlocks and
peer disconnections due to lost pings.
Controller tries to do as much new transaction processing
outside of the blockchain lock as possible, and only
broadcasts new transaction's signature if we successfully
import transaction to our unconfirmed pile.
Synchronizer.findSignaturesFromCommonBlock now returns null
to indicate some sort of connection issue (no cool-off)
and an empty list to indicate NO COMMON BLOCK.
That method also tries to work back to genesis block
instead of giving up too early if test block height
becomes negative.
Network.createConnection additionally filters out
candidates if their addresses resolve to the same
IP+port as an existing connection. So now it won't
connect to localhost:1234 if it has an existing
connection with 127.0.0.1:1234.
Network.broadcast only considers unique peers,
i.e. prefers outbound connection if a peer has
corresponding inbound connection.
Added Thread.currentThread().setName() where possible.
Notably: network messages passed up to Controller are now processed in their
own thread, as opposed to peer's thread.
So each message processor in Controller needs to thread-safe.
V2 network protocol asks for unconfirmed transactions, can send round lists
of transaction signatures and ask for individual transactions, to save
bandwidth/processing.
Remove fetching unconfirmed from Synchronizer
Add extra validity/reference/processable checks to
Transaction.isValidUnconfirmed
Update TransactionUtils to use Transaction.importAsUnconfirmed
for unit tests.
Transaction subclass isValid methods have last reference checks removed.
(Now handled by Transaction.hasValidReference)
Some checks in subclass' isValid moved to isProcessable, typically
for transaction types that require group-approval.
Removed fee extraction and last-reference update code from
subclass' process() method to Transaction.processReferencesAndFees.
Other changes to transaction/block processing.
Don't attempt to sync, or generate blocks, if we think we're not up to date.
Notify Controller of newly generated block AFTER releasing blockchain lock.
Loads of changes to synchronizer.
Added missing GET_PEERS handling to Controller.onNetworkMessage.
More detailed peer information (last block sig, last generator sig, last block timestamp, ...)
New HEIGHT_V2 network message to help support above.
More, and improved, logging.
Fix for HSQLDB serialization failure caused by trying to save the same new transaction
received by more than one peer/thread simultaneously.
Added mass (x1000) testing of key conversion and shared secret calculations.
Fix incorrect proxy private key test that has expected result from previous
algorithm.
Added another test HTML/JS file.
Use IPv6 wildcard address "::" for P2P bind address as null
seems to make listen socket bind to 127.0.0.1
Increase target minimum outbound connections from 3 to 10.
Increate maximum connection limit from 10 to 30.
Previously we preferred the outbound connection but actually
this leaves our info about inbound connection peers out of date.
Synchronizer still prefers outbound.
HTML/JS in src/test/resources/proxy-key-example.html updated accordingly.
Add handshake status to output of API call GET /peers
Add/correct @ApiErrors annotations on some API calls.
Add API call POST /admin/orphan (target height as body)
to force blockchain orphaning for when node is wildly out of sync.
Added support for above to BlockChain class.
BlockGenerator now requires a minimum number of peers
before it will generate any new blocks.
See "minBlockchainPeers" in settings.
Controller now requires a minimum number of peers
before it will consider synchronizing.
See "minBlockchainPeers" in settings.
Old "minPeers" entry in settings.json no longer valid!
Networking now allows both an outbound and inbound connection
to a peer although will use the outbound connection in preference.
Networking checks peer ID of inbound connections to detect,
and resolve, peer ID clashes/theft.
GET /addresses/proxykey/{privkey}/{pubkey} now POST /addresses/proxykey
GET /utils/fromBase58 now GET /utils/frombase58
GET /utils/fromBase64 now GET /utils/frombase64
GET /utils/toBase58/{hex} now GET /utils/tobase58/{hex}
GET /utils/toBase64/{hex} now GET /utils/tobase64/{hex}
GET /utils/privateKey/{entropy} now POST /utils/privatekey
GET /utils/publicKey/{privateKey} now POST /utils/publickey
Lots of edits to Transaction subclasses to change/remove 'delete'.
Corresponding extra changes to help reset some transaction fields to pre-process
state during orphaning.
Changed Block, GenesisBlock & Synchronizer to save transactions where appropriate.
Added enhanced GET_SIGNATURES_V2 network message to reduce the number of
block signatures sent over network.
Peers are now version 2 if they send a new-style build version string,
instead of using first digit from build version.
X25519 shared secrets now match those generated by libsodium.
New tests show that shared secrets are the same using either set
of private+public key combinations.
Changed proxy private key generation from using simple SHA256
of shared secret to using SHA256(shared secret + both public keys).
Added a temporary "BouncyCastle25519" shim class to provide missing
key conversion from Ed25519 to X25519.
Added a finer-grained result (enum) to synchronization to replace
stark boolean result. This allows Controller to decide whether
peer can be retried in the next round (e.g. network issue) or
cooled off for an hour (e.g. peer way too far behind)
Fixed bug with wrong UPDATE_GROUP serialized transaction length.
Added logging to BlockMessage when deserialization to help catch
future bugs.
Restrict synchronizer to only consider peers at most 60 blocks behind.
Synchronizer will only process a batch of 200 blocks at a time.
Improve debugging message in Peer.
Checking interval increased from 5 seconds to 5 minutes.
(Reduces repeated fails).
System tray icon should show message just prior to applying downloaded update.
Proxy forging recipient no longer needs a public key on the blockchain
at the point PROXY_FORGING transaction is submitted.
Proxy forging recipient is given a last-reference, if they don't have one,
when they receive their first block rewards.
Split block fees in proxy forging scenario, using same share proportion.
100% proxy sharing is now allowed.
Fixed account flags processing for accounts in genesis block.
AutoUpdate needs separate logfiles for its process as
log4j2's rolling appender doesn't support locking.
So we create AU_LOGGER in AutoUpdate and set it up
using static { } block.
Added support for starting/querying/stopping auto-update
Windows service.
Changed Controller's public connectionUrl field to
getRepositoryUrl() public static method so it
can incorporate repositoryPath from Settings.
Controller has added support for auto-update on startup.
Fix bugs with Settings not using userPath properly.
Removed obsolete standalone binary classes:
* blockgenerator
* txhex
* v1feeder
Removed all @Produces from API resources as response content type is
sorted by Swagger.
Added API /admin/info for generic node info.
Added API /arbitrary/ endpoints.
Moved arbitrary data storage from ArbitraryTransaction to ArbitraryRepository.
V4 arbitrary transaction signature is based on data's hash.
Original commit was d02f282, and commit message was:
Initial auto-update support, network MAGIC change, arbitrary tx improvements
Various fixes to synchronization
Added missing code for processing incoming block summaries in Network.
Fixed block summaries serialization and removed references to BlockData.
Fixed bug in transaction transformation where base transaction length
didn't include reference or fee lengths.
Original commit was ebbab7b
After opening repository connection with RepositoryManager.getRepostory(),
any 'read' from repository (e.g. SELECT) starts the transaction
even though HSQLDB documentation states there are no shared/read locks
in MVCC concurrency model.
The work-around for this is to 'reset' HSQLDB's in-transaction flag
by performing a ROLLBACK (Repository.discardChanges) immediately
after acquiring the blockchain lock (which is used to ringfence
changes that might collide like these).
Also adding an extra check to prevent payments to nonexistent AT
addresses as it touches Transaction.
API /addresses/{address} now returns lastReference taking unconfirmed into account.
Added DELETE /peers/known to remove all known peers from repository.
Added blockchain locking around Transaction methods like isValidUnconfirmed
as they (temporarily) update account lastReference.
Ditto getInvalidTransactions, etc.
Move hard-coded forging tiers to blockchain config.
Tests for granting forging rights.
Added API call to list top block forgers.
Fixed typo with Reward[s]ByHeight class name.
Remove old whispersystems, etc. *25519 and use new v1.61 BouncyCastle.
Fix proxy forging private key derivation from X25519 shared secret.
Also include Javascript test version for comparison.
Fix block rewards for proxy forging.
Add extra useful info to API call GET /admin/forgingaccounts.
Fix API response to POST/DELETE /admin/forgingaccounts when
passed invalid private keys.
Added block rewards and account flags to testchain config.
Tests to cover changes above.
Added /addresses/proxying to find proxy forging mappings.
Added /addresses/proxykey/{genprivkey}/{recipientpubkey} to calculate proxy private key.
New Block.regenerate factory method to create new Blocks
but without having to reprocess ATs, etc.
Added support for proxied generator in Block.calcGeneratorsTarget
BlockGenerator now generates and checks new blocks for various generators,
including proxy generators.
BlockGenerator now uses generator private keys supplied by Settings.
Corresponding changes to Settings to load base58-encoded private keys.
+ minor stuff
Also improved asset tests in that they don't use QORA any more
which takes fees/block rewards out of the picture.
More test assets are issued in the genesis block to
accomplish this:
1m TEST issued to Alice
1m OTHER issued to Bob
1m GOLD issued to Alice
e.g. supply "assetName" in JSON for TRANSFER_ASSET transactions
Also supply have/want asset names, amount asset ID/name, price-pair
and creator address in asset orders.
Show CREATE_ASSET_ORDER amount ID/name & price-pair in correct
format depending on whether transaction was placed before/after
'new' asset pricing took effect. (Orders are always in 'new' form).
Change API call /assets/transfers/{assetid}/{address} to
/assets/transfers/{assetid} with optional "address" query param.
Orders are back to having "amount" and "price".
(No more "unitPrice" or "wantAmount").
Order "amount" is expressed in terms of asset with highest
assetID.
"price" is expressed in (lowest-assetID)/(highest-assetID).
Given an order with two assets, e.g. QORA (0) and GOLD (31),
"amount" is in GOLD (31), "price" is in QORA/GOLD (0/31).
Order's "fulfilled" is in the same asset as "amount".
Yet more tests and debugging.
For simplicity's sake, the change to HSQLDB repository is
assumed to take place when 'new' pricing switch also
occurs.
Don't forget to change "newAssetPricingTimestamp" in
blockchain config JSON file.
Typical log entry would be:
2019-04-05 09:47:08 ERROR WadlGeneratorJAXBGrammarGenerator:401 - Failed to generate the schema for the JAX-B elements
com.sun.xml.internal.bind.v2.runtime.IllegalAnnotationsException: 4 counts of IllegalAnnotationExceptions
at com.sun.xml.internal.bind.v2.runtime.IllegalAnnotationsException$Builder.check(IllegalAnnotationsException.java:91) ~[?:1.8.0_191]
Changed API call GET /assets to NOT return asset "data" fields
as they can be huge. If need be, call GET /assets/info to fetch
a specific asset's data field.
Improve asset trade amount granularity, especially for indivisible
assets, under "new" pricing scheme only.
Added corresponding tests for granularity adjustments.
Fix/unify asset order logging text under "old" and "new"
pricing schemes.
Change asset order related API data models so that old "price" is
now "unitPrice" and add new "return" as in amount of want-asset
to receive if have-asset "amount" was fully matched.
(Affects OrderData, CreateAssetOrderTransactionData)
Some changes to the HSQLDB tables.
Don't forget to add "newAssetPricingTimestamp" to your blockchain config's
"featureTriggers" map.
Better order matching, especially in situations
where inexact fractional representations (e.g. 1/12)
or rounding issues might occur. Also better matching
with indivisible assets.
Essentially change ordering from have-amount & price
to have-amount and want-return, leaving unit price
to be calculated internally to a finer degree (in
some cases to 48 decimal points).
Corresponding unit tests to cover both legacy and new
scenarios. Support for tests to switch between
blockchain configs.
"New" pricing schema is its own 'feature trigger'
independent from general qorav2 switch.
Safety checks added during trading process.
HSQLDB schema changes (will probably need
careful conflict resolution on merge).
Still to do:
API changes
etc.
+ unit test
+ newer unit test harness
but still needs:
BlockChain config support for activating newer "price" arg
New unit test to check old "price" arg usage
Rework existing asset-related unit tests
Check API inputs/output pre/post "price" arg crossover
Non-native-coin payments (e.g. transfer-asset) didn't build map
for checking whether sender had enough balance.
Also fixed raw layout description for TRANSFER_ASSET.
Also fixed ISSUE_ASSET deserialization to use empty string (instead
of null) for v1 transactions.
GET /assets/trades/recent expanded to allow multiple otherAssetIds
When GET /assets/balances is called with address(es) but no assetIDs
then it will return balances for all assets, including zero balances
for assets the addresses don't own.
GET /addresses/{address} no long fakes a default groupID - in fact
defaultGroupId now removed from blockchain config.
Some Eclipse IDE files now hidden/removed from git repo.
Imported Block/BlockChain fixes from "minting" branch to do
with block timestamps.
GET /assets/holders/{assetid}
and
GET /assets/address/{address}
and
GET /assets/balance/{assetid}/{address}
all combined into
GET /assets/balances?address=...&address=...&assetid=...&assetid=...
New GET /assets/trades/recent?assetid=...&assetid=...
that returns most recent two trades for each asset-pair.
GET /assets/orders/{address}/{assetid}/{otherassetid} has
includeClosed and includeFulfilled repurposed as
isClosed (true/false/omitted) and isFulfilled (true/false/omitted).
ALSO, Order.isClosed is now set to true when isFulfilled is set to true
during processing (and correspondingly set to false during orphaning).
AccountBalanceData now includes optional assetName field for use with API
but generally not set for internal use.
Fixed bug in GenesisBlock which wasn't stripping
out ISSUE_ASSET transactions for v1 blockchains.
Updated blockchain.json for v1 qora blockchain to
fall into line with new BlockChain config
unmarshalling code.
Improved TransactionData.equals when comparing
exact same object.
Improved HSQLDBBlockRepository.getHeightFromTimestamp to use OffsetDateTime
object, which includes time zone info, to fix incorrect SQL full-scan
so that DB now does index-scan instead. Also converted use of MAX(column)
to ORDER BY column DESC as MAX() not optimized when MVCC in effect.
Added corresponding INDEX to Blocks on columns (generation, height);
Similar MAX(column) to ORDER BY column DESC improvement for getBlockchainHeight().
Bumped size of HSQLDB TYPE ATTags from VARCHAR(32) to VARCHAR(80)
Improved reflection-based method calls to rethrow underlying exceptions
like DataException, TransformationException, etc. instead of losing
them in general InvocationTargetException.
Re-added atTransaction.toBytes() method so v1 transactions can be verified.
Fixed extraneous additional copy of voter's public key when serializing
in VoteOnPollTransactionTransformer.toBytes()
Fix-up of v1feeder
UPDATE_NAME requires txGroupId to match original txGroupId used in REGISTER_NAME
UPDATE_GROUP requires txGroupId to match original txGroupId used in CREATE_GROUP
Extraneous JAXB unmarshal code removed from various transaction constructors.
Incorrect, old, groupID vs txGroupId checks removed from various transactions.
Fix bug in UPDATE_GROUP fromBase() method using mismatched column indexes.
Fix bug in API's pending transaction fetch not checking admin is of the same
group as the txGroupId. (Fix actually in HSQLDBTransactionRepository).
Similar fix in GroupApprovalTransaction.
GET /assets/orderbook/{assetid}/{otherassetid} renamed to
GET /assets/openorders/{assetid}/{otherassetid}
Replacement /assets/orderbook/{assetid}/{otherassetid} now
returns aggregated orders, with entries containing only
"price" and "unfulfilled" (amount).
Added /assets/orders/{assetid}/{otherassetid}/{address} to return
orders by specific account, for a specific asset-pair.
Block timestamp validity extracted to separate method so that
BlockGenerator can test timestamp and generate blocks at the usual
rate, even for testnets. This still allows testnets to a way to
generate blocks on demand as Block's isValid skips some timestamp
validity checks if testnet.
txGroupId was sometimes incorrectedly checked for approval-less tx types.
GenesisBlock (v4) now supports various transaction types (issue-asset, etc.)
with generated signatures (like genesis transaction signature) and
missing references inserted.
JUnit reverted back to v4 for Eclipse support (for now).
Some dev/testing API calls are now turned off by default in production mode,
see "restrictApi" settings entry, returning NON_PRODUCTION API error.
Corrections to how account's defaultGroupId works, removing "effective groupID"
which overly complicated matters.
In relation to above, DEFAULT_GROUP (0) no longer exists and NO_GROUP(-1) now has
the value 0 instead.
So transactions can no longer have txGroupId of DEFAULT_GROUP, which in turn
required all the erroneous "effective groupID" code.
API call /addresses/{address} now supplies blockchain-wide defaultGroupId if
account doesn't exist or if account's default not set and NO-GROUP not allowed.
API /transactions/pending now offloaded to repository instead of Java-based
processing and filtering.
Transaction approval checks added to Block.isValid
Groups now have min/max approval block delays.
Checks added to incoming unconfirmed, block generator, block.isValid, etc.
'needing approval' and 'meets approval threshold' now split into separate calls.
NB: settings.json no longer part of git repo
log4j2.properties now has debugging entries removed.
log4j2-test.properties (not in repo) takes priority
so using that in development instead.
Unconfirmed transactions no longer wiped on start-up
by default - see Settings
Reworking of {Public,Private,Genesis}Accounts as it seemed
possible to silently lose public key in repository.
The use of AccountData didn't work and so field-specific
repository calls have been made instead
(e.g. setLastReference) that try to opportunistically
store public key too, if available (i.e. caller is
PublicKeyAccount subclass, or better).
Added API call GET /addresses/{address} to return
general account info in one go. (Essentially the
AccountData object as fetched from repository).
Initial work on adding default groupID to accounts,
along with corresponding SET_GROUP transaction type.
In additional, added blockchain-wide default groupID
and flag to allow/disallow no-group/groupless
transactions.
Initial work on group-admin approval of transactions
tied to a specific group via txGroupId.
More work needed on transaction's "effective txGroupId"!
API call /transactions/pending to list transactions
pending group-admin approval. However, this needs more
work (see effective txGroupId above) and potentially
offloading to HSQLDB repository if possible.
Minor CIYAM AT renames to help static reflection initializers.
Block.orphan() no longer adds orphaned transactions back to
unconfirmed pile as they are themselves deleted during
Transaction.orphan(). Maybe the answer is to NOT delete
them during Transaction.orphan() but to add them to
unconfirmed pile at that point? Very old transactions
leftover from major resync would simply expire, whereas
recently transactions leftover from minor resync could
still make it into a new block on synced chain fork.
Changes/tidying/improvements to block generator regarding
removing invalid transactions and dealing with transactions
pending group approval.
Approval threshold added to groups.
Mass refactoring of transaction-related classes to unify
constructors, particularly field ordering, to fall in line
with raw transaction layout.
e.g. constructors now reflect that raw transactions mostly
start with type, timestamp, txGroupId, publicKey, reference
e.g. JAXB afterUnmarshal methods added where needed and corresponding
nasty code in Transaction subclass constructors ripped out.
e.g. TransactionTransformer subclasses contain less duplicated code.
Fixed bug with repository save points thanks to swapping to Deque.
Some fixes to do with missing transaction types being passed to JAXB
TransactionData subclass constructors.
Ripped out obsolete toJSON in TransactionTransformers as this
is all nicely taken care of by Swagger/OpenAPI (thanks @Kc)
Added short sleep() to GET /admin/stop to allow time for HTTP response body to be sent back.
Improved documentation for /peers API resources. Added examples, tidied API output models.
Fixed issue where IPv6 literals with port couldn't be parsed. Now uses RFC5952/RFC3986 style
literal IPv6 addresses with ports, e.g. [::1]:9084
Fixed NPE in Controller.potentiallySynchronize() where peer might not have sent height yet.
Improved Handshake to discard inbound connections if we already have an outbound connection
to a peer with that ID. This prevents us from having two connections to the same peer, one
in each direction.
Network.mergePeers() now runs in a separate thread as acquiring the lock might block.
Network.creationConnection() exits fast based on number of outbound connections, instead of
number of total connections.
Network no longer sends 'local' peer addresses to non-local peers.
e.g. it won't send localhost:9084 to node4.qora.org:9084
Added try-catch to Network.broadcast for when we try to broadcast while shutting down.
Added PeerAddress class to deal with the whole hostname/IPv4/IPv6 address situation.
Reworked PEERS_V2 message type to only send sized-strings instead of separate port,
and potentially IPv6 byte arrays.
Change to HSQLDB database shape.
Corresponding changes to HSQLDBNetworkRepository.
Peers now broadcast height after successful synchronization.
Added support for sending unconfirmed transactions to other peers.
This is done on connect and also after a new unconfirmed transaction is submitted via API.
Fixed synchronizer to handle blocks with transactions correctly.
Fixed network-related PoW to not use class-global SHA256 message digester!
(It was being corrupted by simulataneous access by different threads - whoops)
Surrounded Network.mergePeers with a lock to prevent HSQLDB deadlocks.
Also changed HSQLDB concurrency model to MVCC (only takes effect if DB rebuilt).
Added support for logging other HSQLDB sessions in the event of exception.
(Currently only used by HSQLDBSaver)
Transaction transformer modifications to help deserialize TransactionMessages.
Some pom.xml changes to reduce maven-shade-plugin conflicting classes warnings.
Repository now supports SAVEPOINT and ROLLBACK TO SAVEPOINT.
HSQLDB concurrency/transaction model changed from LOCKS to MVCC to help with
transaction deadlocks/rollbacks.
More XXXs and TODOs added to Block.java for investigation/fix/improvements.
Also used new repository SAVEPOINT feature when validating transactions
instead of rolling back entire transaction. This fixes problem during
synchronization where the rollback would undo previously synchronized,
but not yet committed, blocks!
Transactions orphaned by Block.orphan ARE now added to unconfirmed pile,
unlike before.
Concurrent lock now prevents simultaneous block generation and synchronization,
including synchronization via multiple peers.
Lots of new networking code: peer lists, block signatures, blocks,
blockchain synchronization.
PEERS_V2 message now supports hostnames, IPv6 and port numbers.
Fixed bug with block serialization for transport over network.
DB shape change from v29 to add peer info.
New NetworkRepository to handle above.
Peer handshaking with v2 anti-DoS PoW code.
Handshaking refactored into a state-machine-like enum.
Some peer-related API calls added.
Peers exchange pings, heights, peers.
No actual peer sync yet.
Other changes:
Peer version info taken from Maven build properties/resource file.
AnnotationPostProcessor more resilient when fetching PathItems.
Per-repository session debugging flag that can be toggled at will.
HSQLDBRepository.delete() now returns int so callers can detect
whether anything was actually deleted.
Some renaming to settings.
Moved as much reflection out to class-static initializers as possible.
Renamed some classes to fall in line with transaction type name
to class name conversion, e.g. DEPLOY_AT -> DeployAt
API calls that return lists now take limit, offset and reverse params.
API calls that used to return data & optional list (e.g. blockWithTransactions)
now only return base data. The optional lists can be fetched via
a different API call.
Also: SLF4J now routes logging to log4j2 so start up output cleaned up.
Suppressed extraneous Jersey warning about Providers during start-up injection.
