forked from TrueCloudLab/neoneo-go
CHANGELOG: release 0.78.3
This commit is contained in:
parent
3f50f90dc5
commit
39a2360448
1 changed files with 59 additions and 0 deletions
59
CHANGELOG.md
59
CHANGELOG.md
|
@ -2,6 +2,65 @@
|
|||
|
||||
This document outlines major changes between releases.
|
||||
|
||||
## 0.78.3 "Calcification" (22 February 2021)
|
||||
|
||||
We're updating NeoGo node for Neo 2 network with this release bringing a
|
||||
number of fixes for various problems. Some were found during our internal RPC
|
||||
compatibility audit and some are just backports from Neo 3 development
|
||||
branch. There are no new features introduced, but it's an important
|
||||
upgrade. Unfortunately we had to change the DB format to fix some problems, so
|
||||
add some time for resynchronization to your node maintenance plan.
|
||||
|
||||
Improvements:
|
||||
* RPC documentation was updated mentioning all known differences with C#
|
||||
implementation that are considered to be not worth fixing (#1760)
|
||||
|
||||
Bufs fixed:
|
||||
* when mempool is full new transaction's hash could still be added into it
|
||||
even if transaction is to be rejected afterwards (#1623)
|
||||
* `gettxout` RPC API was returning data for already spent outputs, although it
|
||||
shouldn't (C# node doesn't do that, #1649)
|
||||
* `getrawtransaction` RPC call returned metadata for pooled (not yet accepted)
|
||||
transactions (#1648)
|
||||
* `gettransactionheight` RPC call returned zero height for pooled (not yet
|
||||
accepted) transactions (#1648)
|
||||
* VM CLI set breakpoints after target instructions, although it should set
|
||||
them before execution (#1648)
|
||||
* a peer could be disconnected and not reconnected again properly in some
|
||||
cases (#1648)
|
||||
* missing write timeout could lead to broadcasting stalls (#1648)
|
||||
* CN transaction requests could try to ask for more transactions from peers
|
||||
than it is allowed to by the protocol (#1648)
|
||||
* CN transaction requests could be trimmed unnecessarily (#1648)
|
||||
* FP Fixed8 values (GAS usually) could be unmarshalled from JSON/YAML
|
||||
incorrectly in some cases (#1761)
|
||||
* previous proposal reuse could lead to empty blocks accepted even if there
|
||||
are transactions in the mempool (#1761)
|
||||
* UTXO asset issuer wasn't properly stored by the node (#1760)
|
||||
* `getassetstate` RPC call answer used wrong field name for "frozen" data
|
||||
(#1760)
|
||||
* UTXO asset type was serialized as numeric code instead of a string in
|
||||
`getrawtransaction` and getassetstate RPC call results (#1760)
|
||||
* `getblockheader` RPC call didn't support block numbers in parameters (#1760)
|
||||
* `getblocksystemfee` RPC call returned block's system fee while it should
|
||||
return cumulative system fee from genesis block to the one requested
|
||||
(#1760)
|
||||
* system fee calculation was wrong for some types of transactions (#1760)
|
||||
* `gettxout` RPC call implementation was returning asset hash in wrong byte
|
||||
order (#1760)
|
||||
* `getpeers` RPC call implementation returned port numbers wrapped into
|
||||
strings (#1760)
|
||||
* `getenrollments` RPC call implementation returned duplicate wrong records
|
||||
along with proper ones (#1760)
|
||||
* NEP5 tracking code wasn't deleting entries with zero balance leading to
|
||||
excessive getnep5balances request outputs for some addresses (#1760)
|
||||
* NEP5 tracking code wasn't able to handle tokens overflowing int64 in
|
||||
transfer and balance amounts (like pnWETH, #1760)
|
||||
* `confirmations` value in `getblock` RPC call response was shifted by two
|
||||
(#1760)
|
||||
* some fields with default values were omitted for `getrawtransaction` and
|
||||
`getblock` RPC call answers with Register and Publish transactions (#1760)
|
||||
|
||||
## 0.78.2 "Colorization" (03 December 2020)
|
||||
|
||||
Release 0.78.2 adds some minor features to the node while still being
|
||||
|
|
Loading…
Reference in a new issue