OprimeDeveloper (@Crowdin_eth) header
OprimeDeveloper (@Crowdin_eth) profile picture
Joined February 2025
Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)Photo by OprimeDeveloper (@Crowdin_eth)
I had a really tough day today, but this truly made me so happy. Sometimes dreams really do come true. ❤️ Thank you so much @moonpay ❤️
1
2
103
Two SQL checks I run after any test that writes data: Dupes: GROUP BY email HAVING COUNT(*) > 1 Orphans: LEFT JOIN users u ON u.id = o.user_id WHERE u.id IS NULL The UI can say "Saved ✅" while the DB quietly disagrees. #QA #SQL #SoftwareTesting
🤖 Made with AI
1
Fight day in Doha 🇬🇪 @TopuriaAleks made weight and meets Santiago Luna on the UFC Qatar prelims tonight, chasing a third straight UFC win. Big brother @Topuriailia was right beside him at the weigh-ins. From sharing one bicycle as kids to this 🙌 #UFCQatar
1
1
73
Lots of devs are mourning the hand-coding era this week. Those years aren't wasted: reading code line by line is exactly what you need when an agent hands you a 400-line diff. Less typing, more reviewing. The craft moved from the keyboard to the review. #vibecoding #AI #coding
🤖 Made with AI
1
10
pytest habit that saves me time every day: - pytest --lf reruns only the tests that failed last run - pytest -x stops at the first failure - pytest -k "login and not slow" picks tests by name While fixing: pytest --lf -x. Before pushing: the full suite. #Python #pytest #QA
6
PostgreSQL 19 is out. Two picks from a QA seat: - ON CONFLICT DO SELECT: "get or create" in one round trip - WAIT FOR LSN: read-your-writes on replicas, fewer flaky "row not found" checks Plus REPACK (CONCURRENTLY) for bloat, no locks. @PostgreSQL #SQL postgresql.org/about/news/po…
10
OpenAI's Performance Factory is QA run by agents: a perf alert fires, Codex finds the root cause, proposes a fix an engineer reviews, then verifies prod is back to baseline. 1,700+ investigations in ~3 months. @OpenAIDevs #Codex #QA openai.com/index/how-openai-…
1
1
47
UFC Qatar this Saturday 🇬🇪 Aleksandre Topuria vs Santiago "BorderBoy" Luna in Doha, with @TopuriaIlia in his corner. Luna is coming off his first career loss; Aleksandre goes in as the favorite, hunting a finish. Let's go Aleksandre 👊 @ufc #UFCQatar #UFC #MMA #Georgia
1
47
XO Cash is a clean stack: created by @exodus, powered by @moonpay, built on @m0, live on @solana. A stablecoin that moves like crypto and spends like cash is exactly the kind of UX that brings everyday users onchain. Curious to try it 👀 teapawt.org/moonpay/status/2123416…
it’s time to learn about the XO Cash stablecoin 🪙 created by @exodus ⚡️ powered by MoonPay ⚒️ built on @m0 ⛓️ live on @solana moves like crypto, spends like cash: programmablemoney.net/stable…
1
26
Selenium habit that fixes a lot of flaky tests: drop time.sleep() and don't mix implicit with explicit waits. Use WebDriverWait with expected_conditions like element_to_be_clickable, so the test waits exactly as long as the page needs instead of a guessed number of seconds.
7
Seeing devs delete whole test suites so AI agents can ship faster. As a QA engineer, I get pruning low-value tests, but tests are how you know the agent built what you meant. Cut the brittle ones, keep the checks on payments, auth and data. Speed without a safety net isn't speed.
6
Playwright tip for faster E2E tests: don't click through the UI just to set up state. Create test data via the API (the request fixture), save login once with storageState, and let the UI test cover only the flow it's actually about. Fewer steps, fewer flakes.
14
Fight week in Doha: Aleksandre Topuria 🇬🇪 takes on Santiago Luna at UFC Qatar this Saturday, Nov 21. Big brother Ilia flew in to corner him, and Merab Dvalishvili is in town too. It's Aleksandre's first fight in about a year, so a statement win would mean a lot. Rooting for him 👊
80
JUnit 5 habit for data-driven tests: skip copy-pasting the same test with new inputs. Use ParameterizedTest with CsvSource (or MethodSource for complex objects). One method, many cases, and each row shows up as its own result, so a failure points straight at the bad input.
5
MoonPay Day at Solana Breakpoint London, quick recap: - @KeithGrossman and @HalseyHuth on stage - PayBox mode live at the venue, with more from @PayBox coming soon - @EntendreAI is now an official @OpenAI connector - Their take: the future of AI x crypto is built on @solana
27
Vibecoding rule I keep coming back to: let the AI write the code, but write the test yourself first. If you can't express the expected behavior as an assertion, you're not ready to prompt for it. The test is the spec the model can't argue with.
3
Hey @moonpay , send me $1,500 and in exchange I'll get a tattoo of your name on my arm 🥹🥰♥️
POV: your life just changed
6
SQL tip for data checks in tests: don't stop at COUNT(*). Run expected EXCEPT actual, then actual EXCEPT expected. Both empty = the data really matches, and any leftover rows show exactly what's off. (Oracle calls it MINUS.)
2
Reconciling transactions across entities is usually the most painful part of finance ops, so being able to do it right inside a ChatGPT conversation is a big step. Curious how permissions work: read-only by default? teapawt.org/moonpay/status/2122455…
BREAKING: @EntendreAI is now an official @OpenAI connector! 💬 bring your financial data into the conversation 🧠 review transactions and reconcile across entities, all from ChatGPT connect with a click: chatgpt.com/plugins/plugin_a…
4
Small pytest habit that pays off: put setup/teardown in a fixture with yield instead of try/finally in every test. Use scope="session" for expensive stuff like a browser or DB connection, and pytest.mark.parametrize instead of copy-pasted tests.
6