Vincent Tunru @vinnl@mastodon.social

Like and boost for a chance at getting frisked next time you're at the airport, or tag a friend who deserves a tax audit.

https://www.socialcooling.com

This looks great: property-based end-to-end testing.

Define some general conditions that should hold (e.g. the value of this input field should always be shown here), and it will perform arbitrary actions in your UI to see whether they break the condition (e.g. clicking this button replaces the displayed value by [object Object]).

https://quickstrom.io

The resemblance between this endangered animal and the browser named after it is uncanny.

I've been using git worktree a couple of times recently. I don't think I'll ever go back to stashing.

I have just been informed that apparently Web Replay is being developed as a separate product rather then as part of the Firefox DevTools: https://webreplay.io

Just found out that the Web Replay experiment of the Firefox DevTools was discontinued.

Such a shame; it really would have been a killer feature: https://developer.mozilla.org/en-US/docs/Mozilla/Projects/WebReplay

Darn it. I listened back some of the clips I recorded for https://commonvoice.mozilla.org. Turns out I've got the fake "look at me I'm trying to pretend I'm posh English even though I'm clearly not" accent.

This is also a good way to introduce tests to an untested code base: ignore all existing files for the coverage report, then set the threshold to 100%.

Arbitrary thresholds are a great way to invoke Goodhart's law. Code coverage should be an assistant, not a stick to hit you with.

I really like this quote:

> a lot of the value of code coverage data is to highlight not what’s covered, but what’s not covered.

Be liberal about marking code not worth writing tests for skipped for the coverage report, and set your coverage threshold to 100%. The report can then remind you about parts of your code you meant to but forgot to write a test for.

https://testing.googleblog.com/2020/08/code-coverage-best-practices.html

We're going to release the things I've been working on for the past couple of months officially in a bit, and I'm more excited than I expected.

It's been open source and publicly accessible for a while, but now it's in an 🌟announcement🌟. Shiny!

My day.

Just realised that I never really shared that I wrote a small npm package to send #WebMentions a while ago. It's automatically tested against the full https://webmention.rocks test suite, work in both Node and the browser, and includes type definitions.

This work was supported by @NGIZero and is currently used in production on https://plaudit.pub.

You can find it here: https://www.npmjs.com/package/webmen

I just can't get over how great StreetComplete is. It looks fantastic, is incredibly easy to use, and actually makes it fun to contribute data about the Real World for the benefit of everyone.

If you run Android, be sure to give it a try: https://github.com/westnordost/StreetComplete#download #OpenStreetMap #StreetComplete

SemVer seems clear enough, but in practice often leads to questions like "is this bugfix a breaking change?" or "shouldn't a major new feature be a major release?"

I wrote a post to collect and answer those questions:

https://vincenttunru.com/semver-explained

The tl;dr is: don't think about what your version "means", but about what conclusions you want your package's consumers to draw. That change of perspective can help resolve many common dilemma's.