Skip to content

Quantachain Testnet Alpha V.0.7.2

Choose a tag to compare

View all tags
@xaexaex xaexaex released this 05 May 17:14
· 351 commits to main since this release
v0.7.2-alpha
105efa5

QuantaChain Testnet — Alpha v0.7.2

Post-quantum secure blockchain using Falcon-512 signatures and SHA3-256 Proof of Work.

v0.7.2 — CONSENSUS-CRITICAL patch. No testnet reset required.
All nodes MUST upgrade. v0.7.1 nodes will diverge from upgraded nodes on any block
containing a TimeLock credit to the miner's address at height > 85,000.
Existing quanta_data/ directories are fully compatible — drop-in upgrade.

This is a pre-release testnet build. Do not use real funds. APIs and chain parameters may change between alpha releases.

Genesis Block

Parameter Value
Network Testnet (QUA7)
Timestamp 1775001600 (2026-04-01 00:00:00 UTC)
Testnet Genesis Hash 00000012d3a2cbb7eb9579330ccdaa4f83ca9e6e016bfe6d2c8a38539cf3733b
Mainnet Genesis Hash 1cdbccdff3db462378f4acbe4553b49040ffcdebf74b5c77e685ba05ccfa8cb0
Difficulty 8,304,130 (Testnet) / 16,777,216 (Mainnet)
Block Time 30 seconds

Testnet Faucet Wallets (Genesis Premine)

Each address below received 1,000,000 QUA at genesis. Faucet account 0 is the active sender used by the faucet API.

Index Address Role
0 0x1683be267318d2ddd8cee8df4a4548dcffb1e088 Faucet Sender (active)
1 0xd528c18ce7a8844e4a4dcd841975b20ae599b020 Faucet Reserve
2 0xfd6e36bfa2b2798d08592802206c943d5513adfb Faucet Reserve
3 0xed15573ad312d41aaef74cff56a8ef28122ec2db Faucet Reserve
4 0xaffd6d4f74c5651110efcf1b9736f7a5cf2ccdbb Faucet Reserve
5 0xbf5ee055f399323fdd0cefe3d4aa923678d46107 Faucet Reserve
6 0x1dc9637b183093d723ea8d1fb18083b06490facb Faucet Reserve
7 0xa2270f30ca1aad922510375508bf68cd95509f29 Faucet Reserve
8 0xe15a689775685ae324559ea9a492fc650354ca0b Faucet Reserve
9 0x005dcff212d27b55e7a74bf745e1349ab44ca25d Faucet Reserve

Treasury

Parameter Value
Treasury Address ms69216b1d10425689704d5ae3b2a4aa17049f59b1
Multisig Scheme 3-of-5 Falcon-512
Block Reward to Treasury 5% per block
Fee to Treasury 20% of transaction fees

Tokenomics

Parameter Value
Year 1 Block Reward 100 QUA
Annual Reward Reduction 15% per year
Minimum Reward Floor 5 QUA (reached ~year 20)
Mining Reward Lock 50% locked for 6 months
Fee Burn 70% of all transaction fees
Fee to Treasury 20% of all transaction fees
Fee to Miner 10% of all transaction fees
Max Block Transactions 1,200
Max Block Size 2 MB
Min Transaction Fee 0.0001 QUA

Quick Start with Docker

Option 1: Docker Desktop (Graphical Interface)

  1. Open Docker Desktop and find xd637/quanta-node:v0.7.1-alpha (or :latest) in your Images.
  2. Click Run.
  3. Under Optional settings, configure:
    • Container name: quanta-node
    • Ports: Map 3000, 7782, 8333, 9090 to themselves.
    • Volumes:
      • Add host path quanta-data to container path /home/quanta/quanta_data
      • Add host path quanta-logs to container path /home/quanta/logs
  4. Click Run.

Option 2: Docker CLI

docker pull xd637/quanta-node:v0.7.2-alpha

docker run -d \
  --name quanta-node \
  -p 3000:3000 -p 8333:8333 -p 7782:7782 -p 9090:9090 \
  -v quanta-data:/home/quanta/quanta_data \
  -v quanta-logs:/home/quanta/logs \
  xd637/quanta-node:v0.7.2-alpha

Option 3: Docker Compose (Recommended)

Upgrading from v0.7.0 — no data wipe needed:

docker compose -f docker-compose.single.yml pull
docker compose -f docker-compose.single.yml up -d

Server Setup (Ubuntu / VPS)

1. Update system and install Docker:

sudo apt update && sudo apt upgrade -y
curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker ubuntu && newgrp docker

2. Open necessary ports:

sudo ufw allow 8333/tcp
sudo ufw allow 7782/tcp
sudo ufw allow 3000/tcp
sudo ufw allow ssh
sudo ufw --force enable

3. Upgrade to v0.7.2 (no data wipe required):

docker pull xd637/quanta-node:latest
docker stop quanta-node && docker rm quanta-node
docker run -d \
  --name quanta-node \
  --restart always \
  --network host \
  -v ~/quanta_data:/home/quanta/quanta_data \
  xd637/quanta-node:latest

4. Check logs:

docker logs quanta-node --tail 30 -f

Manual Build from Source

git clone https://github.com/quantachain/quanta
cd quanta
git checkout v0.7.2-alpha
cargo build --release

./target/release/quanta start -c quanta.toml

Wallet Management

# Native
./target/release/quanta new-wallet --file wallet.qua

# Docker
docker exec -it quanta-node quanta new_wallet --file wallet.qua

# HD Wallet (recommended — 24-word recovery phrase)
docker exec -it quanta-node quanta new_hd_wallet --file hd_wallet.json

Mining

# Native
./target/release/quanta start_mining YOUR_WALLET_ADDRESS --rpc-port 7782

# Docker (background)
docker exec -d quanta-node quanta start_mining YOUR_WALLET_ADDRESS --rpc-port 7782

# Stop
docker exec -it quanta-node quanta stop_mining --rpc-port 7782

Node Status

docker exec -it quanta-node quanta status --rpc-port 7782
docker exec -it quanta-node quanta print_height --rpc-port 7782
docker exec -it quanta-node quanta mining_status --rpc-port 7782

Ports

Port Service
3000 REST API
8333 P2P Network
7782 RPC
9090 Prometheus Metrics

What Changed in Alpha v0.7.2

No testnet reset. No wire format change. Consensus-critical patch — all nodes must upgrade.

Fix — State root determinism (calculate_state_root)

The locked_balances field on each account is a Vec<LockedBalance>. When a block
contains a TimeLockTransfer credit to the miner's own address alongside a coinbase
credit, two LockedBalance entries are pushed to that address's vec — but in different
orders depending on which code path runs:

  • Mining path (create_block_template): coinbase tx processed first → coinbase lock
    pushed first, TimeLock lock pushed second.
  • Validation path (validate_block_consensus): user txs applied first → TimeLock
    lock pushed first, coinbase lock pushed second.

Both vecs contain the same two entries, but SHA3-256 is order-sensitive — the resulting
state root hashes differed between the mining node and every syncing peer, causing:

[ERROR] Invalid state root at block N: expected <mining_hash>, got <validation_hash>

This manifested sporadically (only when a miner received a TimeLockTransfer to their
own wallet in the same block they mined) and was the root cause of the "nodes fail at
varying heights" sync bug reported across the testnet.

Fix: calculate_state_root now sorts locked_balances by (unlock_height, amount)
before iterating. The sort is stable, deterministic on all platforms, and
order-independent — both code paths now produce an identical SHA3-256 digest.

Guard — STATE_ROOT_SORT_FIX_HEIGHT = 85_000

Blocks below height 85,000 skip state root validation — they are already secured by
hardcoded checkpoints and were committed under the old (buggy) ordering rule. Applying
the new sort rule retroactively would fail for any historical block that happened to
have the mismatch, turning a sync fix into a sync break.

From height 85,000 onward, the new deterministic state root is enforced on all nodes.

New Checkpoints (through block 80,000)

Three testnet checkpoints verified live from scan.quantachain.org on 2026-05-05:

Height Hash
60,000 0000010ce22920660ba1e42423ea46e76dc7582963d6f9f220e3930031bd9bc9
70,000 000001fcb0637b06601b4f111b22070e856c8cabf2eaa545c41b938b4478d186
80,000 0000002d80e66bce37596616a9c9c3c1988da6e65811ad132926162c7e000a0e

These protect the chain from deep reorgs below 80k even on nodes that have not yet
reached that height.

What Changed in Alpha v0.7.1

No testnet reset. No wire format change. Drop-in upgrade.

Fix — deep_reorg used wrong validator on peer blocks

add_block_to_main_chain_reorg() was calling validate_block_consensus() — the strict
validator that requires the incoming block's difficulty to exactly match the local LWMA.
During a deep reorg, peer blocks were mined against their LWMA which can differ slightly
from ours (the two chains diverged at a prior block with a different timestamp).

Fix: reorg path now calls validate_block_consensus_reorg(), the 50%-bounds permissive
validator that was already written for this purpose but wasn't being used.

Fix — deep_reorg corrupted cumulative_work counter

After rolling back the chain, the in-memory cumulative_work was still at the old tip's
value. Each new block applied by add_block_to_main_chain_reorg added to this stale
total, producing a cumulative_work value roughly double the correct amount. This caused
the node to always believe it had more work than all peers and skip future syncs.

Fix: deep_reorg now recomputes the correct base work from storage before replaying
new blocks, and resets both the in-memory counter and the sled key to this value.

Fix — single-block tip swap (reorg_to_block) never updated cumulative_work

The 1-deep reorg path correctly swapped the block and rebuilt account state, but never
adjusted the cumulative_work counter. The counter was left at the old tip's value.

Fix: subtracts the old tip's difficulty and adds the incoming tip's difficulty after commit.

Fix — add_block_to_main_chain_reorg had dangling orphan code (compile error)

A previous edit left a if !tx.is_coinbase() { ... } block without its enclosing
for tx in &block.transactions loop. This was a compile-time error in practice.

Fix: restored the complete nonce-clearing loop matching add_block_to_main_chain.

Fix — linear sync treated as reorg (request_start <= bc_height → < bc_height)

When the sync engine requested the next batch of blocks starting exactly at the current
chain height, request_start == bc_height evaluated true for the reorg branch and
triggered a deep_reorg call. This caused O(n²) behaviour during normal linear sync —
every downloaded block triggered a full chain rollback and account-state rebuild.

Fix: condition changed to strictly-less-than so only blocks below the current tip
are treated as a reorg.

Improvement — Storage: no per-block fsync

save_block and save_account_state no longer call db.flush() after every write.
Sled's write-ahead log guarantees crash safety without a per-block fsync. A single
flush_storage() call is issued at the end of each sync batch and after mining a block.
At 18,000 blocks × ~5 ms/fsync this removes ~90 seconds of wasted IO during IBD.

Improvement — O(1) cumulative work lookup

cumulative_work is now stored as a sled key and kept in an in-memory Arc<Mutex<u128>>.
cumulative_work_at(tip) returns the stored value in O(1) for the current tip.
Previously every call scanned all blocks from genesis (O(height) disk reads while
holding the blockchain read lock — the primary cause of seed-node connection timeouts).

Improvement — Account state snapshots every 1000 blocks

add_block_to_main_chain now saves a full account state snapshot at every 1000-block
boundary. rebuild_account_state_up_to() loads the nearest snapshot and replays only
the delta — previously it always replayed from genesis, O(height) on every reorg.

What Changed in Alpha v0.7.0

UPGRADE NOTICE — v0.7.0 required a testnet reset.
The cumulative_work handshake field changed the binary wire format.
v0.6.0 and v0.7.0 nodes are not mutually compatible.

Major Architecture — Headers-First Sync (Bitcoin IBD style)

Two new wire messages — GetHeaders and Headers — allow a syncing node to download
light headers (index, hash, previous_hash, difficulty, cumulative_work) before requesting
full blocks. The sync engine validates headers first, finds the fork point, then requests
only the missing full blocks in ordered batches.

Cumulative work-based peer selection

The handshake now exchanges cumulative_work alongside height. Sync always targets
the peer with the highest cumulative PoW — not the tallest chain.

Atomic deep reorg with rollback

Failed reorgs no longer leave the node at a partial intermediate state. The original
chain is saved before rollback and restored on failure.

Security — Cross-chain replay protection

network_id: u32 added to Transaction. Signatures are cryptographically bound to a
specific network. Testnet = 0, Mainnet = 1.

Security — State root empty-string bypass closed

Blocks with state_root = "" can no longer bypass state root validation.

Security — Reorg path now verifies all transaction signatures

validate_block_consensus_reorg() now runs the parallel Rayon signature pass.

Security — Inbound peer connection cap

listen_for_connections() now enforces max_peers before accepting the TCP stream.

Improvement — Light block gossip

broadcast_block() sends only the block header (~200 B) instead of the full block
(~2 MB). Peers request the full block if they need it.

Security Notice

This release has undergone internal audit only. It has not been formally verified
by a third-party security firm. Do not use for real financial transactions.

Falcon-512 and Kyber-1024 implementations are based on NIST PQC Round 3 finalists.

License

Apache 2.0 — see LICENSE

"Quanta" and "QuantaChain" are trademarks. Forks may not use these names without permission.

Assets 2
Loading