miniblog.

I've heard of a 'burger menu' when a button has three horizontal lines on it, but today I heard 'kebab button'. This is similar, a button with '···'. Clearly UI designers are hungry.

Doing another iteration on my diagnostics display. I'm reasonably happy with the bold highlighting within the error message. I'm not sure about the colour on Warning and Error though. It gives the output some visual structure, but arguably the message itself is more important.

Overall it seems like the Rust for Linux is making huge strides. Greg KH is super supportive of it[1] and the wider Linux community is pretty positive[2] (see all the quotes at the end of the slides). 1: https://lore.kernel.org/rust-for-linux/2025021954-flaccid-pucker-f7d9@gregkh/ 2:

Do you have a favourite way of highlighting text in a CLI error message? Currently I'm using backticks, but I'm wondering what's the most readable.

I hear people say that Go is often hard to search online (hence sometimes "Golang"), but the vast majority of language names are common words. Names with punctuation (C++, C#) are hard too. Is this a big problem in practice? "Perl" isn't a dictionary word, but it's an exception.

Copilot doesn't offer a way of disabling completion inside comments. People have discovered a comical workaround: swearing in the comment!

In LSP, a position is represented as a line number and a column offset (in Unicode code units): https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#position This is pretty elegant. You'll get the correct line regardless of encoding bugs, and the editor already knows the line number so it's cheap to compute.

I've released difftastic 0.63! In this release: * Better parsing of Elixir, LaTeX, Make, Nix, Rust and YAML * Better detecting of text encoding, especially on Windows * Prebuilt musl binaries, so you can run released binaries on older systems!

I made some changes to a node express project that I haven't touched in almost five years. I was pleasantly surprised that I only needed to update one dependency to get it working again! (It was sqlite3, which is a native dependency using node-gyp.)

I'm experimenting with jj this weekend. It seems pretty nice so far, but the mental model is pretty different from git. For example, there's no `checkout` command. You do `jj new ABC` to switch to a commit, which creates a new empty commit on top of ABC.

I've just squeezed another 5% of performance out of difftastic by finding a few HashSet values that weren't FxHashSet. I do wonder whether hash DoS resistance is a good default. Sure, Rust programs are often pretty fast anyway, but it feels like a different threat model to the rest of Rust.

I'm experimenting with live-evaluating tests in my programming language project. It's relatively fiddly to hook up a UI for this, but it saves a precious keystroke to run the tests! I'm hoping that it results in more, better tests due to the convenience.

I'm experimenting with Atuin for searching my terminal history: https://atuin.sh/ Previously I'd just used fzf to find items, which does work nicely. Occasionally I *really* want to search "commands which were run in this directory" though, which Atuin offers.

I have a bunch of open tasks on my GH repos, so I tried asking ChatGPT and Claude to write a card-based web UI that showed some random open issues. ChatGPT gave me something that worked, but the Claude mock-ups look better (and render inline!).

"Example Driven Development" using Glamorous and Pharo Smalltalk: https://medium.com/feenk/an-example-of-example-driven-development-4dea0d995920 Tests returning values and composing is a really interesting model. It establishes structure and shows which test failure is the most 'fundamental'.