Astro Cactus

• Source: chrismwilliams/astro-theme-cactus (mine, 2026-08-25) — demo
• Verdict: keep — as sentiment, not strategy
• Why: I want to support it and possibly port it into boris and rotkeeper. It is a bad fit for both systems and for me, I wasn't even that excited about it when I was using Astro, which I don't anymore. The desire persists anyway. Rich in irony, as advertised.

The facts

Astro Cactus is Chris Williams's opinionated blog starter: 1.7k stars, MIT, Astro v6 + Tailwind v4, MD/MDX posts and notes, dark/light mode, Pagefind static search, Satori-generated OG images, webmentions, automatic RSS, sitemap, robots.txt, web manifest, Expressive Code blocks. 939 commits of maintenance. It is, genuinely, one of the good ones — the theme you point people at when they say "I want a blog."

It was also born as a Hexo theme. Cactus has now survived two JavaScript ecosystems and is eyeing a third that isn't JavaScript at all.

The irony, examined

The port urge is sentiment wearing a strategy costume. Boris already does most of Cactus's feature list natively — RSS, sitemap, drafts excluded from feeds, static search, dark mode — with no Node, no pnpm lockfile, and no postbuild script. Porting Cactus into boris means translating Tailwind classes into a Zig compiler's template slots to reproduce a design I was lukewarm on, for a site that already has a theme I actively like. Rotkeeper's spooky-dark crypt would fit it even less.

The honest version of the impulse: star it, sponsor Chris if the sponsor button's warm, and if the affection needs an outlet, steal the design language — the type scale, the spacing, the quiet — into the lab theme, without importing a single dependency. Themes migrate; frameworks don't get to come along.

Disposition

Held as draft until the owner decides: publish the irony as-is, or promote to Technology Shit with a full "porting a theme across three frameworks" writeup. The middle path (design-language transplant into lab/) is real work with no deadline.