Skip to content

testing.md

Latest commit

 

History

History
173 lines (123 loc) · 4.03 KB

testing.md

File metadata and controls

173 lines (123 loc) · 4.03 KB

Debugging Guide

mega-evme CLI

The official debugging tool for MegaETH transactions.

Installation

git clone https://github.com/megaeth-labs/mega-evm
cd mega-evm/bin/mega-evme
cargo build --release

Replay Transactions

# Basic replay
mega-evme replay <txhash> --rpc https://mainnet.megaeth.com/rpc

# With execution trace
mega-evme replay <txhash> --trace --trace.output trace.json --rpc <endpoint>

# With call tracer
mega-evme replay <txhash> --trace --tracer call --trace.output calls.json

Docs: https://github.com/megaeth-labs/mega-evm/tree/main/bin/mega-evme

Gas Profiling

Opcode-Level Analysis

# 1. Get debug trace
cast run <txhash> --rpc-url <vip-endpoint> > trace.json

# 2. Profile opcodes
python scripts/trace_opcode_gas.py trace.json

Script: https://github.com/megaeth-labs/mega-evm/blob/main/scripts/trace_opcode_gas.py

Example Output

op              count      total         avg      min      max
SSTORE             29      85400      2944.8      100    22100
SLOAD             473     121300       256.4      100     2100
LOG2               25     598601     23944.0    14257    25521
KECCAK256         288      12108        42.0       36       60

Common Issues

"Intrinsic Gas Too Low"

Local gas estimation uses standard EVM costs. MegaEVM differs.

Fix:

# Foundry: skip local simulation
forge script Deploy.s.sol --gas-limit 5000000 --skip-simulation

# Or use higher hardcoded limit

"Nonce Too Low"

With eth_sendRawTransactionSync, this can mean:

  1. Tx already executed (check receipt)
  2. Race condition with pending tx

Debug:

# Check current nonce
cast nonce <address> --rpc-url <endpoint> --block pending

"Block Pruned" on eth_call

Public endpoint only keeps ~15 days of state.

Solutions:

  • Use Alchemy/QuickNode for historical calls
  • Run archive node
  • Use VIP endpoint with longer retention

WebSocket Disconnections

Connection drops after idle period.

Fix: Send keepalive every 30s:

setInterval(() => {
  ws.send(JSON.stringify({
    jsonrpc: '2.0',
    method: 'eth_chainId',
    params: [],
    id: Date.now()
  }));
}, 30000);

Volatile Data Access Limit

Error after using block.timestamp + heavy computation.

Cause: 20M total compute gas cap (retroactive) when any block metadata opcode is accessed. The cap applies to the entire transaction's compute gas, not just gas used after the opcode.

Fix: Keep total compute gas under 20M in transactions that touch block metadata. For heavy computation + time-awareness, split into separate transactions or use the timestamp oracle.

Cast Commands

# Estimate gas
cast estimate --from <addr> --to <addr> --value 0.001ether \
  --rpc-url https://mainnet.megaeth.com/rpc

# Call at specific block
cast call --block <number> <contract> "method(args)" \
  --rpc-url https://mainnet.megaeth.com/rpc

# Get transaction details
cast tx <txhash> --rpc-url https://mainnet.megaeth.com/rpc

# Decode calldata
cast 4byte-decode <calldata>

HAR Analysis

For debugging frontend RPC patterns:

# Export HAR from browser devtools, then:
python parse_eth_rpc_har.py export.har > calls.csv
python method_activity_timeline.py calls.csv

Identifies:

  • Unnecessary RPC calls
  • Batching opportunities
  • Slow methods blocking UX

Monitoring Latency

# Detailed timing breakdown
curl -i -X POST https://mainnet.megaeth.com/rpc \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"eth_blockNumber","params":[]}' \
  -w "dns: %{time_namelookup} | connect: %{time_connect} | total: %{time_total}\n"

Block Explorers

Network Explorer
Mainnet https://mega.etherscan.io
Testnet https://megaeth-testnet-v2.blockscout.com

Etherscan may lag a few blocks behind real-time.

Getting Help

  1. Check transaction on explorer
  2. Replay with mega-evme locally
  3. Profile gas by opcode
  4. Contact MegaETH team with:
    • Transaction hash
    • Error message
    • Reproduction steps