EmacsNews

This Week in Emacs

Packages & Tools: New Arrivals and Old Problems Solved

The new Typewriter package caught my attention this week — a writing-focused mode that gives Emacs a distraction-free, centred writing experience. It joins a growing set of packages aimed at writers and note-takers who want Emacs to feel less like a code editor and more like a focused composition environment. Worth checking out if you do any long-form writing in Emacs.

On the navigation front, a popular Reddit thread this week centred on repeat-mode finally making Emacs navigation "click" for one user. Repeat-mode — the built-in feature that lets you repeat transient key sequences without re-entering a prefix — is one of those features that feels invisible until it transforms how you move around. The discussion is worth reading for the concrete keybinding patterns people have built around it, particularly for window management and buffer navigation.

Also landing on MELPA this week: audacious, an Emacs interface for controlling the Audacious audio player; slothbar, a status bar for the Emacs X window manager; and e2ansi, which brings Emacs-powered syntax highlighting to less in the terminal — a clever idea for anyone who regularly reads code outside of Emacs. There's also nu-mode, a new keybinding scheme on MELPA, though it's early days and I'd treat it as experimental for now.

On the lighter side: betweenle.el brings the alphabetical word-guessing game Betweenle to Emacs. Procrastination, now with Elisp.

Performance & Practical Fixes: Terminal Speed and Long-Line Freezing

Two practical guides dropped this week covering common Emacs performance problems. The first — covered on Hacker News and noted by Irreal — explains why Emacs terminal emulators vary so dramatically in speed, comparing vterm, eat, ghostel, ansi-term, and term. If your terminal inside Emacs feels sluggish, this is the guide to read. The practical upshot is that the choice of terminal package matters significantly, with vterm and eat generally leading on performance due to their underlying architecture.

Separately, a Reddit post tackled how to prevent Emacs from freezing on files with very long lines — the classic minified JavaScript or SQL dump problem. The thread collects useful remedies including so-long-mode (built-in since Emacs 27) and configuration tweaks worth adding to your init file. A related Reddit

Deep Dive

The Performance Thread Running Through This Week

Looking across this week's content, a consistent theme stands out: Emacs performance, and the community's increasingly systematic approach to understanding and addressing it. Three distinct threads — terminal emulator speed, scrolling lag from line numbers, and long-line freezing — converged this week in ways that, taken together, paint a clear picture of where real-world Emacs friction still lives in 2026.

The terminal performance guide is the most immediately practical of the three. The key insight it surfaces — that architectural differences between terminal packages produce large, measurable differences in throughput — matters because many users simply pick a terminal emulator on recommendation and never revisit the choice. The guide explains why these differences exist, not just that they do: vterm's native module approach versus pure-Elisp implementations is a meaningful distinction, and understanding it makes the performance gap less surprising. That kind of mechanistic explanation tends to stick.

The display-line-numbers-mode lag investigation and Irreal's accompanying post are interesting for a different reason. Line numbers feel like a trivial feature, but they sit in a hot path of Emacs's redisplay engine. The community digging into why this costs what it costs — rather than simply advising people to turn it off — reflects a maturing understanding of Emacs internals that is increasingly filtering into everyday configuration advice.

The Org Ecosystem Keeps Expanding

The arrival of org-roam-timeline this week feels emblematic of a broader pattern: the Org-Roam ecosystem continues to attract focused satellite packages rather than expanding the core in all directions. A visual timeline for nodes is the kind of feature that a specific group of users — those doing research or journalling where temporal navigation matters — will find genuinely useful, while others will ignore it entirely. That modularity is healthy.

Combined with the community discussions around diagram tools and LaTeX engines, it's clear that Org-mode's place as a serious publishing and research platform is well-established. Users are now optimising within that space rather than questioning whether it's capable.

Distribution and Starting-Point Diversity

The Bedrock 2.0 update and the appearance of Nu-mode on MELPA reflect an ongoing tension in the Emacs community between minimal starting points and opinionated configurations. Bedrock represents one philosophy — provide the least possible to get started — while Nu-mode suggests there's still appetite for rethinking Emacs's keybinding model from scratch. Neither is wrong, but the continued proliferation of starting-point philosophies suggests the community hasn't settled on a single answer to "how should a new user begin?" That's probably fine; Emacs has always resisted that kind of convergence.

The Neo Emacs project is the most speculative item of the week, and I'm genuinely curious where it goes. "Emacs from the Future" is a bold framing, and it's too early to know whether it represents a meaningful direction for the ecosystem or an interesting experiment that stays niche. Worth watching, but not worth reconfiguring your setup around yet.