Skip to content

Releases: aptos-labs/aptos-core

Aptos CLI Release v3.5.0

07 Jul 11:28
b1b6b05
Compare
Choose a tag to compare
aptos-cli-v3.5.0

Release CLI version 3.5.0 (#13925)

[Testnet] Aptos Node Release v1.16.1

02 Jul 17:54
d8ac339
Compare
Choose a tag to compare
Pre-release

Release Notes

Release Tag: aptos-node-v1.16.1-rc

CLI Version this release is compatible with: v3.5.0+

Validator Update Required? Y

Fullnode Update Required? Y

Aptos Improvement Proposals (AIPs)

Check out all of our AIPs and discussions here on GitHub.

  • None included in this release

Aptos Blockchain

General

  • Upgraded rust to version 1.78.0.

Consensus

  • Small optimizations to consensus broadcast to: (i) prioritize distant peers first; (ii) parallelize message serialization more effectively.

Networking

  • Small performance improvements for message deserialization at the networking layer.

Move

VM

  • Updated type size calculation to account for all nodes in non-substituted types. Also introduced TypeBuilder struct to construct types with maximum depth and count.
  • Added a procedural macro bytecode_spec so that inline specifications to the Move bytecode instructions can be defined in Rust.
  • Added runtime environment outside of VM. The env can be used cached and possibly shared across VM instances.

Gas

  • Fixed gas charging in crypto_algebra::serialize() API implementation.

Resolved Issues

  • Upgrade Rust to v1.78.0

[Testnet] Aptos Node Release v1.16.0

01 Jul 21:49
daea1c7
Compare
Choose a tag to compare
aptos-node-v1.16.0-rc

Change proof queue data structure (#13878)

Aptos Framework upgrade v1.15.0

01 Jul 19:37
351918d
Compare
Choose a tag to compare
aptos-framework-v1.15.0

add proposal for VM type size limiting (#13854)

[Mainnet][Testnet] Aptos Node Release v1.15.2

27 Jun 16:32
Compare
Choose a tag to compare

What's Changed

  • cherry-pick 56ca463 by @areshand in #13836
  • [Bug] Network Health Checker. Fixes a infinite blocking call in Network Health Checker when attempting to disconnect from a peer that failed the Health check. The situation arises when a peer connection is blocked writing to a socket unable to process disconnection request from the Health Checker, while the Health Checker is waiting for peer connection termination acknowledgement. The fix also includes a timeout for writing to an outbound socket.

Check out everything included in v1.15 release here!

Full Changelog: adhoc-20240626...aptos-node-v1.15.2

[Testnet] Aptos Node Release v1.15.1

25 Jun 23:12
cea6cf8
Compare
Choose a tag to compare

What's Changed

Full Changelog: aptos-node-v1.15.0...aptos-node-v1.15.1

[Mainnet][Testnet] Adhoc Release 20240626

27 Jun 01:22
cea6cf8
Compare
Choose a tag to compare

Commit hash: f6ee8fe244367b5b306fc4a2dabebe78663f6829

Includes all changes in aptos-node-v1.15.2 + security fix

Upgrade required for Validator nodes, optional for Fullnodes (fullnodes can use aptos-node-v1.15.2 instead)

[Testnet] Aptos Node Release v1.15.0

20 Jun 17:00
Compare
Choose a tag to compare

Aptos Improvement Proposals (AIPs)

Check out all of our AIPs and discussions here on GitHub

Breaking Changes

  • [AIP-88] - BlockEpilogue to replace StateCheckpoint transaction
    • Ecosystem Impact: StateCheckpoint transaction will stop being produced, and BlockEpilogue will be produced instead. For users that handle receiving unknown transactions, no issues should occur. For users that need to handle all transaction, they would need to add handling for the new transaction type.
    • Feature Flag: gated with execution onchain config

Aptos Blockchain

  • #13133 BlockSTM: Parallelize Resource Group Accesses. Different resources in a resource group won't have a conflict for parallelism purposes in most cases. Improves the throughput of resource group based workload by up to 2x.
  • #13531 Mempool: Reduces the number of duplicate transactions due to PFN duplication. In most cases, the upstream VFN will have already seen the txn via the primary VFN's quorum store batch.

Move

VM

  • Correct type size computation: #13028. Note that the feature flag will be enabled in 1.16, and this release rolls out the APIs first with the same type size computations.

Resolved Issues

Bug Fixes

  • Multisig V2 bug fixAborts the multisig transaction if the provided payload does not match the payload stored onchain. This resolves [this filed issue](#12929).
    • Feature flag: ABORT_IF_MULTISIG_PAYLOAD_MISMATCH
  • Network Health Checker. Fixes a infinite blocking call in Network Health Checker when attempting to disconnect from a peer that failed the Health check. The situation arises when a peer connection is blocked writing to a socket unable to process disconnection request from the Health Checker, while the Health Checker is waiting for peer connection termination acknowledgement. The fix also includes a timeout for writing to an outbound socket.

[Mainnet][Testnet] Aptos Node Release v1.14.1

17 Jun 16:33
Compare
Choose a tag to compare

Who needs to upgrade:

  • Validator: Yes, by Jun 20th
  • Fullnodes: Yes, by Jun 26th

What's Changed

See full release notes for v1.14 here.

Full Changelog: aptos-node-v1.14.0...aptos-node-v1.14.1

[Mainnet][Testnet] Aptos Node Release v1.14.0

10 Jun 21:45
Compare
Choose a tag to compare

Release Hash: 17cfcf9

CLI Version this release is compatible with: v3.4.1+

Aptos Improvement Proposals (AIPs)

Check out all of our AIPs and discussions here on GitHub.

Breaking Changes

  • [AIP-70] Parallelize Fungible Assets - opt-in parallel fungible balance
    • Ecosystem Impact: In order to track fungible asset balances correctly (for the accounts that opt-in to the new feature), now both old and new field needs to be summed up: FungibleStore.balance + ConcurrentFungibleBalance.balance. If using indexer, this change will make it transparently handled by provided processors. fungible_asset::balance view function will continue providing correct balance.
    • Dependencies: None
    • Feature Flag: CONCURRENT_FUNGIBLE_BALANCE

New features and enhancements

Aptos Blockchain

State sync

  • Fixed warning log from latency monitor complaining about “unable to read the latest block timestamp”

Consensus

  • Implemented a new “order votes” mechanism to reduce the consensus latency. Earlier, it takes for 4 hops of latency for a block to be ordered after the block is proposed. The new order votes mechanism reduces the latency to 3 hops. To achieve this, when a validator obtains the QC on a block r, the validator immediately signs the QC and broadcasts a new type of vote called “order vote”. This results in obtaining a 2-chain QC on the block r one hop earlier. The feature is added behind the new order_vote_enabled flag in onchain consensus config. The flag is currently disabled and will be enabled later on with an AIP approval.

Move

Framework

  • Add CONCURRENT_FUNGIBLE_BALANCE

VM

  • Add aggregator_v2::is_at_least API which can be used to check for minimal balance among other things.
  • Add checks to the extended checker for safe usage of randomness.
  • Consolidate production configs for VM