Philipp Fritsche

If you're working on a tool/script that uses @github's APIs, try github.com/ph-fritsche/gi… to write concise and meaningful tests without touching live data.

@TAbrodi That implies that there is no use for a product manager on a team of good software engineers. Hot take.

Good software engineers are technically good and think like product managers.

@TestingLib 🐙

Toolbox will be immensely helpful in finding implementations that work across multiple environments 😃

@meijer_s It is possible :)

I wish it was possible to 'inline' assignments like these. I don't want 'match' to be available outside the while scope.

@meijer_s I prefer the comment over the variable declaration.

When properly documenting our code, TS feels like the less obvious choice. I'm strongly reconsidering my current preferences.

@meijer_s The only recipe which requires just two cloves of garlic is that for two cloves of garlic. And you better use three.

Recipe says two cloves of garlic. What size are we talking about?

@kettanaito Semantic Release as GitHub action: github.com/ph-fritsche/ac…

3. Semantic release github.com/semantic-relea… I used SR in the past but I grew to dislike it. If you're not sure why, you haven't tried reverting a release after CI crashed mid-way but SR already pushed the release commit and tags to the remote. I had big issues with its defaults.

If you are shipping to NPM, please set up an automated release process. Begging for library releases is ridiculous. Here are some of the tools I recommend for automated releases 👇

@OtsoKivekas @0xgaut @Otto_Lehto That sounds like something an AI which wrote the US constitution would say

@0xgaut @Otto_Lehto Nah, boring

someone used an AI detector on the US Constitution and the results are concerning. Explain this, OpenAI!

@meijer_s Semantic Release as Github Action: github.com/ph-fritsche/ac…

What do you use to automate releases of your js sdks (to npm, or private registry)?

@kentcdodds Well, ...

Articles with a straightforward yes or no question as the title that don't start with "Yes" or "No" are really annoying.

@florinpop17 Just CSS borders 🙂 codesandbox.io/s/inverted-bor…

Hey CSS nerds! Here's a challenge for you! How would you do this in CSS? 👀 A follower asked me but tbh, I never did such a thing. Those inverted borders seem tricky 😆

Do not just add act() calls until React shuts up! ph-fritsche.github.io/blog/post/reac… #react #testing

@TAbrodi This completely depends on the bridge/software project. A pedestrian bridge over a creek? Just start building, fix it later if you need to. Do you know the complexity of the software upfront? What might happen if it fails? Can you afford to apply a fix later?

It doesn't make sense to compare building bridges with writing software! Here is why: Building bridges requires a significant and detailed plan upfront: the waterfall approach. That approach is literally the one that's killing software projects.

@meijer_s sentimentalversioning.org

semver doesn't think big enough. Not all features are minor.

@meijer_s For this GitHub release action I dogfeed the build to create a release: github.com/ph-fritsche/ac…

Wrapped up a quick test to confirm that our bundled files look like what we'd expect. Do you do something similar? Do you test your bundles at all?

@meijer_s I use a shared script for building libraries. I add tests there for issues with the resulting builds. github.com/ph-fritsche/sc…

@kentcdodds @__grunet @marcysutton I like the idea to run tests twice in keyboard- and pointer-mode. I'm not so sure about the idea to replace `.click()` per spy. Might be better to integrate some utility like `.navigateTo()` into user-event.

Bringing Together Keyboard-only and Click-based UI Tests with keyboard-testing-library by @__grunet on @marcysutton's site testingaccessibility.com/introducing-ke… *Very* cool new utility for testing 👍

Calling a render prop or using it as a #react element can look similar. But it's not the same. ph-fritsche.github.io/blog/post/rend…

@sebsilbermann So it should be `git-praise`🤪

@ph_fritsche You clearly haven't worked with my code. Those lines never end up as bugs

git-blame isn't for finding out who made the change but why the change was made. Should be named git-why to remove the negative connotation.