Commit graph

1752 commits

Author SHA1 Message Date
Evgenii Stratonikov
357bb4ce41 core: get rid of global variables in tests
It can lead to unnecessary race conditions and is just
a bad practice.
2020-03-02 17:04:08 +03:00
Evgenii Stratonikov
8ca94e23c8 core: replace makeBlocks() with addBlocks() in tests
This simplifies tests a bit.
2020-03-02 17:04:08 +03:00
Roman Khimov
a2b9d85c80
Merge pull request #681 from nspcc-dev/feature/tx_batching
network: batch transactions on broadcast
2020-03-02 15:54:19 +03:00
Evgenii Stratonikov
cbf26bac83 mempool: do not allocate new slice for verified transactions
Because transactions a iterated in an increasing order,
we can filter slice in-place.
2020-03-02 10:38:27 +03:00
Evgenii Stratonikov
0ef7a76e84 network: batch transactions on broadcast 2020-03-02 10:37:27 +03:00
Roman Khimov
ff551d2015
Merge pull request #700 from nspcc-dev/fix/forloop
compiler: do not short-circuit in complex conditions
2020-02-28 20:14:26 +03:00
Roman Khimov
36197056b1
Merge pull request #691 from nspcc-dev/network-conn-overflow-and-shutdown
Network conn overflow and shutdown
2020-02-28 18:16:29 +03:00
Evgenii Stratonikov
b461a6ab63 compiler: do not short-circuit in complex conditions
Current implementation of short-circuting is just plain wrong
as it uses `last` or `before-last` labels which meaning depend
on context. It doesn't even handle simple assignements like
`a := x == 1 && y == 2`.

This commit makes all jumps in such conditions local
and adds tests.

Closes #699, #700.
2020-02-28 17:44:46 +03:00
Roman Khimov
e213e69a33 network: rework shutdown sequence of Server
Close transport and disconnect peers right in the Shutdown(), so that no new
connections would be accepted and so that all the peers would be disconnected
correctly (avoiding the same deadlock as in e2116e4c3f).
2020-02-28 16:22:04 +03:00
Roman Khimov
77624a8847 network: add Close() to discoverer, shut it down on exit 2020-02-28 16:22:04 +03:00
Roman Khimov
f2397607b3
Merge pull request #698 from nspcc-dev/fix/sendtoaddress
rpc: rename (*Client).SendToAddress to TransferAsset
2020-02-28 15:05:13 +03:00
Evgenii Stratonikov
624a14f730 rpc: rename (*Client).SendToAddress to TransferAsset
SendToAddress name is similar to the `sendtoaddress` RPC call,
this can cause confusion.

Closes #686.
2020-02-28 14:16:39 +03:00
Roman Khimov
26c4e83ddf
Merge pull request #697 from nspcc-dev/claim-verification-in-mempool
Claim TX verification for mempool
2020-02-27 15:33:53 +03:00
Roman Khimov
3e2b490025 mempool: rework inputs verification, check Claim txes
Use more efficient check for Input and also check Claims to avoid double
claiming.
2020-02-27 13:43:35 +03:00
Roman Khimov
f329de73e8 mempool: fix Remove bug (dropping wrong tx!)
Uncovered by TestMemPoolAddRemoveWithInputsAndClaims (to be added later)
2020-02-27 13:43:35 +03:00
Roman Khimov
7d59fa0066
Merge pull request #696 from nspcc-dev/tx-verification-fixes
Claim and enrollment TX verification fixes
2020-02-27 12:45:36 +03:00
Roman Khimov
00d199ed85 core: fix NetworkFee calculation
Add doesn't change the variable state. Thanks to GolangCI for catching this.
2020-02-27 12:32:56 +03:00
Roman Khimov
7bea6e043e core: add double claim verification check 2020-02-27 12:32:56 +03:00
Roman Khimov
29586f2aa7 core: add duplicate Claim verification check
Refactor HaveInputsDuplicate() out of the core and Blockchain, it doesn't
depend on the Blockchain state. Make it more efficient.
2020-02-27 12:32:56 +03:00
Roman Khimov
9b1ad40273 core: drop GetScriptHashesForVerifyingClaim, fix Claim's ScriptHash getter
Claim transactions should _add_ claims scripthashes to the standard list, not
replace them. And this code is actually very reusable.
2020-02-27 12:32:56 +03:00
Roman Khimov
9f7018503a core/transaction: use new transaction.InOut for References
We don't need a map here, use simpler structures.
2020-02-27 12:00:03 +03:00
Roman Khimov
80de208a68 transaction: unbind GroupInputsByPrevHash from Transaction
Which allows to use it for ClaimTX. Make it also a bit more efficient as maps
are expensive.
2020-02-27 12:00:03 +03:00
Roman Khimov
71a18b1727
Merge pull request #688 from nspcc-dev/fix/jmplabels
compiler: use uint16 for labels
2020-02-27 00:28:46 +03:00
Roman Khimov
b1b184f77f
Merge pull request #692 from nspcc-dev/persist-timer-fix
core: reschedule the next persist in a second from previous one
2020-02-26 10:34:40 +03:00
Roman Khimov
2f2a4afe21
Merge pull request #693 from nspcc-dev/memorystore-seek-lock
storage: add locking into (*MemoryStore).Seek
2020-02-26 10:15:40 +03:00
Roman Khimov
c258adb532 core: move spent coin management out of the inner storeBlock loop
prevHash == input.PrevHash, so make less DB accesses and more real work. Fix
some bugs along the way:
 * spentCoins structure may already be present in the DB when persisting TX,
   there is nothing wrong with that and we shouldn't overwrite it
 * it's only used for NEO and only to check for claim validity. Thus, when
   processing claim tx the corresponding spentCoins should always be present
   in the DB
2020-02-25 17:21:09 +03:00
Roman Khimov
36c6b6af14 core: put unspents get/put one loop above in storeBlock
Everywhere in this code prevHash == input.PrevHash, thus we can easily move
some common code out of the loop saving on DB accesses and
serialization/deserialization.
2020-02-24 18:53:10 +03:00
Roman Khimov
5c9681508b transaction: strip off a layer of redirection from ClaimTX.Claims
We don't need a pointer here and this change makes this field compatible with
Transaction.Inputs which is useful in many scenarios.
2020-02-24 18:22:27 +03:00
Roman Khimov
ed9e9e8590 storage: add locking into (*MemoryStore).Seek
It was missing there.
2020-02-24 17:51:50 +03:00
Roman Khimov
b49808b766 core: reschedule the next persist in a second from previous one
It makes little sense queueing up several persistence goroutines (or actually
even running them concurrently).
2020-02-24 17:17:25 +03:00
Roman Khimov
e2116e4c3f network: don't deadlock on connection overflow
(*Peer).Disconnect send an unregister signal to this goroutine, so invoking it
from here is not a good idea, run it asynchronously.
2020-02-24 15:56:49 +03:00
Roman Khimov
d5e6028c13 core: add validator public key into the list of hashes to verify
That should be done for Enrollment transactions.
2020-02-21 20:04:16 +03:00
Roman Khimov
5d4dfbfa1a
Merge pull request #684 from nspcc-dev/feature/getblocksysfee
rpc: implement getblocksysfee RPC
2020-02-21 18:48:56 +03:00
Anna Shaleva
76a0a6e7e8 rpc: implement getblocksysfee RPC
Closes #341
2020-02-21 18:45:52 +03:00
Evgenii Stratonikov
177b725dc1 compiler: make writeJumps return error for bad jumps
The script is invalid anyway so it is better to notify user.
2020-02-21 17:45:45 +03:00
Evgenii Stratonikov
5e229d84d4 compiler: use uint16 for label numbers
As noted in #687 this will make compiler a bit more predictable.
2020-02-21 17:45:45 +03:00
Roman Khimov
56f87cd44e
Merge pull request #676 from nspcc-dev/dedup-rpc-types
Split up RPC into separate packages

Closes #510.
2020-02-21 16:22:36 +03:00
Roman Khimov
8da20055d6 rpc: drop duplicating structures used in tests
We have proper results now, so use those. The only left is Invoke, but that
depends on another issue at the moment.
2020-02-21 15:23:11 +03:00
Roman Khimov
6a3be6081c smartcontract: support "ByteArray" string for ByteArrayType
We actually do emit it ourselves in String()
2020-02-21 15:23:11 +03:00
Roman Khimov
3fa9de764b rpc/client: only return the Result from calls, handle Error internally
Adjust structures accordingly and throw away most of them, they're useless.
2020-02-21 15:23:11 +03:00
Roman Khimov
28a26d2cae rpc/client: look into data first, then HTTP error code
In case of error our own server responds with an HTTP error and proper
JSON-RPC error in the body, so look there first as it has more specific data.
2020-02-21 15:12:04 +03:00
Roman Khimov
877b987ecf rpc: omitempty some fields in TransactionOutputRaw
These might be undefined for mempool transactions, thus they should be defined
as omitempty.
2020-02-21 15:12:04 +03:00
Roman Khimov
bba8ac15ff rpc: introduce proper type for sendrawtransaction, drop useless types
sendrawtransaction just returns a bool, sendtoaddress returns a proper
transaction and that should be the same as the one we have in
TransactionOutputRaw.
2020-02-21 15:12:04 +03:00
Roman Khimov
465476201c rpc/response: drop duplicating AccountState structure 2020-02-21 15:12:04 +03:00
Evgenii Stratonikov
b50704fd3b rpc: move server-related code to a separate package 2020-02-21 15:12:04 +03:00
Evgenii Stratonikov
d24c6d1d9e rpc: move client-related code to a separate package
This includes Client struct with RPC methods and
BalanceGetter implementation with NeoSCAN.
2020-02-21 15:12:04 +03:00
Evgenii Stratonikov
b6bc4e580a rpc: move functions for Tx building in request package/ 2020-02-21 15:12:04 +03:00
Evgenii Stratonikov
98a1e2170a rpc: move ContractDetails to request/ package 2020-02-21 15:12:04 +03:00
Evgenii Stratonikov
63751a6e6b rpc: merge neoscan-related definitions into one file
It will make refactoring easier.
2020-02-21 15:12:04 +03:00
Roman Khimov
f330f2f40b rpc: separate out request and response structures
Mostly as is, no real effort done yet to optimize them, so there are still a
lot of duplicates there, but at least we sort them out into different smaller
packages.
2020-02-21 15:12:04 +03:00