Tag: testing

Fork testing smart contracts runs your tests on a copy of mainnet state. Learn when it helps, when it gives false confidence, and how to pin the block.

Flaky e2e tests have five causes: timing races, shared state, order dependence, environment limits and unstable selectors. Each one reproduced and fixed.

Sanity testing vs smoke testing explained in one table, with a worked checkout discount example and a clear answer on when you need both or just one.

How to run tests in parallel in CI/CD with workers, shards and job matrices. A GitHub Actions example, a risk table and the shared-state bugs that appear.

How to write a reentrancy test in Solidity with Foundry. Build the attacker contract, watch the test fail, then apply checks-effects-interactions and a guard.

A k6 soak test script, how long to run it and what to watch, shown on a small service that leaks on purpose so you can tell a real leak from normal noise.

Puppeteer vs Playwright for E2E tests: the same login test in both, a decision table, and when Puppeteer still fits. Start with Playwright for new suites.

A working k6 spike test example with ramping-vus stages, thresholds on error rate and p95, annotated output and the mistakes that make spike results useless.

Six Solidity unit test patterns with Foundry code: revert, event, access control, rounding, fuzz and state-transition tests, plus the bug each one catches.

Smoke testing vs regression testing in one page: scope, size, triggers and a CI/CD pipeline example that runs smoke on every merge and regression nightly.

The edge cases that break NFT contracts: stale approvals, mint limits, reentrancy, royalties, metadata and ownership, with Hardhat tests for each.

Build realistic DeFi test data: fee-on-transfer, rebasing and blacklistable token mocks, forked mainnet state and failure scenarios, with Hardhat code.

Use gas ceilings, gas reports and snapshot tests to find and lock in gas savings in smart contracts, with measured Hardhat examples.

How to test upgradeable contracts: storage layout checks, proxy upgrade simulation, initializer tests, gas checks and a pre-upgrade checklist.

Budget for web3 testing: audits, QA salaries, infrastructure and gas, with sourced ranges and a sample annual budget for a mid-sized project.

How to test the contract-to-frontend flow in a dapp: a local Hardhat chain, an injected wallet provider, Playwright tests and a CI run, with a DEX swap example.

Web3 testing differs from web2 in environment, state, async behavior, security, performance and auth. Practical fixes for each, with Hardhat examples.

Four load testing types explained with ready-to-run k6 scripts: load, stress, soak and spike. What each tests, when to use it and which tool to pick.

Should end-to-end tests live in the project repository or a separate one? Pros, cons, folder structures and a rule of thumb for choosing.

Six decisions that determine whether your automated test suite scales: structure, data, environments, reporting, CI and ownership.

How to cut automation run time with parallelisation in CI/CD, avoid flaky and duplicated tests, and keep the suite maintainable.

What a QA engineer does day to day, the skills required, how the role differs from tester and SDET, and where the career can go.

Screenshot, UI, API and performance tests compared: what each checks, when to add it, and example tools for each.

Stop UI tests breaking on CSS changes: data-testid attributes, a selector strategy that starts with roles, and how to get developers to help.

Smoke, sanity and regression testing compared in one table: purpose, scope, timing and automation, plus how to place each in a CI/CD pipeline.

Set up end-to-end tests with Playwright and Jest: install, config, a real first test, and when to use @playwright/test instead. Updated for current tooling.

Do you need manual testing first? Which language and framework to pick, what to automate first and where to learn, from my own start in test automation.

The three mistakes I see when interviewing QA candidates: language basics, team fit, and handling unknown problems.