Detritivore Biome

@noahsussman

TOAD ≈ 🐸Arbiter of Arbitrarily-Wide Structured Data Blobs🧚‍♂️Time Bugs Guy🐞My opinions aren’t shared by anyone. 🤷🏻‍♀️ he/him

forest floor of Internet jungle
Vrijeme pridruživanja: svibanj 2007.

Tweetovi

Blokirali ste korisnika/cu @noahsussman

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @noahsussman

  1. Prikvačeni tweet
    8. ruj 2019.

    I said I would write more about breaking into a dev role so here goes. One of the things that absolutely destroys entry-level programming applicants is the fear that they won't be able to do "the job" once hired. This fear is unreasonable and unfounded and I'll tell you why.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    prije 9 sati
    Poništi
  3. proslijedio/la je Tweet
    4. velj
    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    prije 16 sati

    "You asked us to compile an export of your photos and videos, so we sort of did that and also sent some of your stuff (including your videos!) to random other people as a special, fun, community-bonding bonus that is definitely not a privacy breach! You're welcome!"

    Poništi
  5. proslijedio/la je Tweet
    3. velj

    Everyone knows about Exploratory but hardly anyone has publicly documented the technical underpinnings of the practice. Anyway here's the shell script I use to figure out which files get committed to the most. 🤷🏻‍♀️

    Poništi
  6. prije 9 sati
    Poništi
  7. proslijedio/la je Tweet
    prije 22 sata
    Odgovor korisnicima

    Currently yes, current framework less than a year old and running nightly, we have plans to do review our suite over the next few months and get a portion of the tests running in our CI pipeline.

    Poništi
  8. proslijedio/la je Tweet
    Odgovor korisnicima

    Don't do it. If you do, record the number of the retries. But don't do it. The system is telling you that you have something to learn, listen to it. Find the route cause and address it. Is it ok for the system to fail 1 in 100 users? Probably not.

    Poništi
  9. proslijedio/la je Tweet
    Odgovor korisnicima

    We do retry once but based on clear failure conditions.

    Poništi
  10. proslijedio/la je Tweet
    prije 12 sati
    Odgovor korisnicima

    Retry policies usually masks flakiness. It can be useful to allow you to make some progress, but it's a lot better when you stop and investigate the causes of flakiness. Sometimes it's external systems, sometimes it's the test infrastructure, sometimes it's both.

    Poništi
  11. proslijedio/la je Tweet
    prije 12 sati
    Odgovor korisnicima

    As always, context matters, and the same tools can be used for good or evil. FWIW, the team I'm in moved from ~100/8000 flaky tests to ~10/9000 flaky tests in the past year. Would have I been better off leaving the team/company when we were at 100/8000? I highly doubt it.

    Poništi
  12. proslijedio/la je Tweet
    prije 12 sati

    I do agree totally with ; it’s a sign and we need to listen and deal with it. Don’t try to put a green flag in a “war zone” :)

    Poništi
  13. proslijedio/la je Tweet

    The most egregious symptom of the "test retry" mentality I've seen is it disguising an issue where the app under test misrendered pages under certain circumstances... occasionally. Testers lumping that bug into "flakiness" meant they shipped something that could have been better.

    Poništi
  14. proslijedio/la je Tweet

    Is there any really big Amazon sponsored/pushed open source project out there? We know Facebook, Google and Netflix keeps blessing the community with amazing products.

    Poništi
  15. proslijedio/la je Tweet
    4. velj
    Odgovor korisnicima

    I maintain e2e tests if useful, but not retrying as a default. Not only for external system dependencies. And no, I dont find it hard - they stand out like they have wailing siren and flashing lights 😁

    Poništi
  16. proslijedio/la je Tweet
    4. velj
    Odgovor korisnicima

    When running in a pipeline? No. As part of a nightly run or a separate test suite? It might be an acceptable workaround in some cases.

    Poništi
  17. proslijedio/la je Tweet
    4. velj
    Odgovor korisnicima

    I quit my last job in part because I was trying to test an app but the platform the app ran on was unreliable, in multiple ways, and I had no agency to change that.

    Poništi
  18. proslijedio/la je Tweet
    4. velj

    agree. I have exposed bugs when I stopped using retry. -- if tests are flaky I quaranteen them and run them flaky tag and observe to find issues.

    Poništi
  19. proslijedio/la je Tweet
    4. velj
    Odgovor korisnicima

    Retrying e2e tests is too expensive, I stopped a decade ago. Flaky tests can be made reliable; or they can't. If your job is e2e tests and systems you rely on are flaky, seriously consider getting a new job.

    Poništi
  20. proslijedio/la je Tweet
    4. velj

    Many people think this must be conspiracy. But sadly, this underinvestment in quality is still the median software project experience.

    Prikaži ovu nit
    Poništi
  21. proslijedio/la je Tweet
    4. velj

    1/ Tech and : We now have credible reporting out about the app that wrecked Iowa. This was a preventable disaster caused by willful ignorance about development practices. Bottom line: The Democratic Party funded their friends rather than following best practices.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·