EmacsNews

This Week in Emacs

Packages & Tools: graph-fa2 0.2.3 has landed on MELPA, offering a pure-elisp ForceAtlas2 graph layout engine that runs in the background. It's a niche tool, but an impressive one for anyone doing graph visualisation inside Emacs. More immediately practical is runbox.el, which tackles a long-standing pain point: running shell commands, compilation, and eglot over TRAMP in local buffers. If you use TRAMP for remote or homelab workflows — and there's clear community interest in exactly that — runbox.el addresses the friction that comes from language servers and compile commands not behaving reliably across remote connections.

The eshell corner had some activity too, with separate discussions on shared history across eshell sessions and integrating zoxide for fast directory jumping within eshell. Both are small quality-of-life improvements that add up for people who live in eshell. The zoxide theme extended to Dired as well, with dired-z bringing frecency-based directory jumping to Dired — a useful complement for users who prefer keyboard-driven file management.

On the supply-chain security front, a thought-provoking piece explored using an LLM to review package upgrades managed via straight.el. Whether or not you adopt this workflow, it's a timely reminder that package management hygiene matters, and it's interesting to see Emacs-specific thinking applied to a problem the broader software world is still working through.

Packages & UI: TextUI, which adds responsive Flex and Grid layout for widget.el interfaces, went public this week. It's an ambitious effort to bring modern layout primitives to Emacs UI building. The author also published a detailed write-up of what they learned about widget.el while building it — well worth reading if you've ever wanted to build a polished Emacs interface. widget.el is underused and poorly documented, and this kind of hands-on exploration fills a real gap.

Org Mode: Irreal covered markup editing in Org Mode this week, highlighting finer-grained controls for inline formatting — useful for writers who work heavily in Org and want more precise markup behaviour. Separately, a Reddit thread explored reading .docx files in Emacs with tracked changes and comments preserved as Org Mode — which speaks to a real collaborative writing problem: staying in Org while working with colleagues who use Word. If this workflow can be made reliable, it's a meaningful bridge.

Irreal also covered a customised Org agenda view, a recurring topic that remains useful because Org agenda customisation is both powerful and genuinely confusing. Anyone building their own agenda setup will likely find something worth borrowing there.

Development & Elisp: Irreal's piece on ongoing Emacs development is a useful corrective to occasional claims that Emacs has stagnated — the argument being that development is steady and purposeful rather than headline-grabbing. On a more concrete note, Emacs 31 brings markdown-ts-mode and there's already community discussion around its keybindings, which suggests people are actively testing the pre-release. The speedbar also got attention in an Irreal post on Emacs 31 and the Speedbar, covering changes or renewed interest in that long-standing tool.

Prot published a new video on conditionally expanding abbrev-mode definitions — the kind of

Deep Dive

The Emacs ecosystem right now: steady, growing, and absorbing newcomers

Looking at this week's activity, I notice a dual rhythm in the Emacs community. On one side, there's the steady, experienced-user layer: Prot releasing a nuanced video on abbrev-mode conditionals, Irreal reassuring readers that Emacs development is ongoing and healthy, developers shipping tools like runbox.el to solve specific, gnarly problems around TRAMP and language servers. On the other side, there's a clear influx of newcomers — Neovim and NvChad users particularly — arriving with genuine curiosity and finding the community receptive.

These two groups want different things, and it's encouraging to see both being served. The experienced-user work this week is characteristically focused: runbox.el is not a flashy tool, but it solves a real problem for anyone developing against remote machines or homelabs. The eglot-over-TRAMP problem has frustrated people for years, and a clean solution deserves attention. Similarly, dired-z and the zoxide-in-eshell thread reflect a broader trend: popular shell tooling getting first-class Emacs integration rather than just being invoked through a terminal emulator.

The TextUI announcement is one I'm watching with interest. widget.el is one of those Emacs subsystems that works but has historically been impenetrable — the author's accompanying write-up is almost as valuable as the package itself, because it maps territory that has gone unmapped for too long. If TextUI gains traction, it could raise the ceiling on what polished Emacs interfaces look like without requiring a move to a heavier framework.

The LLM-assisted package review workflow is harder to categorise. Using an LLM to summarise diffs in package upgrades before applying them is a practical security habit that maps well onto straight.el's lockfile-based approach. I'm not convinced it catches truly sophisticated supply-chain attacks, but as a supplementary habit for security-conscious users, it's actionable today and worth knowing about.

The newcomer wave brings its own energy. Guides like Emacs for NvChads matter not because veteran users need them, but because they lower the barrier for people who might otherwise give up after a frustrating first weekend. The Hacker News thread on how you'd relearn Emacs today generated thoughtful discussion about learning paths and tooling choices that serves both newcomers orienting themselves and veterans reflecting on how the ecosystem has changed. The fact that this question still attracts genuine engagement — rather than a tired list of the same links — says something healthy about the community.