Building https://xitter.cf/t.co/32lRRmtL0V system at @awscloud. All opinions are mine.
Berlin, Germany
Joined June 2012
- Tweets6K
- Following197
- Followers446
- Likes2.9K
Pinned Tweet
How browser testing tools work:
Webdriver – created a spec, let browser vendors implement it on their own.
Cypress – create a browser extension, use it to send commands and manipulate the page via JavaScript
Playwright – patch browser engines to add a special automation protocol
Prediction
The current skills in markdown files is a temporary solution to work around tooling limitations. In the future, they will become JSDOC annotations like /** @skill */ on top of the respective files
"Code itself is the best documentation" - they always say. So as the skills
New programing language with more explicit constructs and predictable behaviors
But what's the deal with agents? Humans would benefit from more readable code too, or?
Counterintuitive take: if a tool supports multiple frameworks, that’s bad for you. Developer attention is limited, and it gets diverted to features and fixes you don’t need
(image is for reference only, there are many such cases)
Hello chat,
Does anyone have a hint how to stop confusing resetAllMocks, restroreAllMocks and clearAllMocks API?
When you see this line in the console
node --max-old-space-size=8192
this is a red flag.
There is no legitimate use-cases for increasing heap memory. The reason is always some bloated code shipping unnecessary dependencies
Code freeze is basically a public declaration of defeat: you no longer control the code, the code controls you
just-boris retweeted
Every test framework I've ever used that chooses to captures logs (stderr, console.log, etc.) has managed to make my life harder by doing so. Please, if you're making a testing framework, do not capture logs.
just-boris retweeted
I used to be against a lot of heavy unit testing of internals. I felt they often cost more in dev time than they were worth compared to e2e. Now that AI makes these tests effectively free, and agents use them for localised feedback, this equation has shifted back a lot for me.
Little trick with React error boundaries which might save your app
Instead of component prefer using `withErrorBoundary` wrapper, so even the component body with hooks will be covered as well
Using AI as an auto-complete tool is superior to just chatting.
When you write code, you will get suggestions, but even if you got none, you have still written some meaningful code
With chatting, if suggestions did not work, you cannot do anything useful with that chat history
Maybe I am slow, but only today I realised that Remix team rewrites their framework without React.
Look, they build their own virtual DOM: github.com/remix-run/remix/b…
Very promising, much hyped
just-boris retweeted
Unpopular opinion:
Unless you’re just prototyping, you should aim to understand as close to 100% of production code generated by LLMs.
Yes, all of it.
Effective mental models are still important for humans to sustainably maintain and evolve a codebase via prompting alone.
Hello @code, using a single curly bracket without its closing pair is an offense for the developer community
where do I file a complaint?
just-boris retweeted
Why I don't trust MCP output (and neither should you)
Not a cool research, just some small notes and thoughts
blog.slonser.info/posts/smug…
