Irreal

Another great post from James Cherti. This one’s about speeding up Emacs terminal emulators. As Cherti says, Emacs terminals can be slow especially when they get pounded with a lot of output. He offers some suggestions to help.

Those changes are mostly what you’d expect. They turn off things like line wrapping that take extra cycles to calculate. Most of his post is his suggested code to remedy the slowness—although he does offer an explanations of what they’re doing—so rather than repeat what he’s already written, I’ll just send you to his post for the details.

If you’ve been bothered by slow Emacs terminal emulators, take at look at Cherti’s post.

If you’re a macOS user, you probably know about Homebrew, a repository of easy to install apps and libraries. I use it all the time because it’s so much easier than hunting up all the dependencies required for compiling them yourself.

If you’re a Linux user, you may or may not be familiar with Homebrew. It provides the same services for Linux but my—possible wrong—impression is that its Linux take-up is not as great as it is in macOS.

Regardless, if you’re on Linux or macOS, you should definitely check it out. The TL;DR of this post is that Homebrew has just released version 7.0.0, a major release. That probably wouldn’t be worth an Irreal post except for the “major release” part. On the macOS side, macOS 10.5 and earlier are no longer supported. Being a Mac-head, I didn’t check discontinued versions on Linux so if you’re using Homebrew on Linux you should check that out.

The other big change for Mac-heads is that Intel Macs are moved to Tier 3. You can check out the announcement for exactly what that means but basically it indicates that support is limited and end-of-life is approaching (in September of 2027).

Homebrew is an excellent repository and well worth investigating if you’re not familiar with it. I still have an Intel Mac so it’s a bit of a pain that support is diminishing but I need to upgrade anyway so this is probably the push I need.

James Cherti has been busy producing a lot of posts about optimizing various aspects of Emacs behavior. I’ve just come across his post that discusses optimizing buffer line numbering. It explains the different line numbering schemes and their relative costs.

I not a big user of line numbering. I usually turn it on for the Elfeed index but hardly every use it otherwise unless I have a specific need. In particular, I hardly ever use them in code buffers because I prefer to jump to a particular construct rather than a line number, which, irrationally, always seem to me to be ephemeral.

When I do use line numbers, I like absolute line numbers rather than the two types of relative numbers. That turns out to the right choice as far as efficiency is concerned but many people—including Cherti—prefer relative numbering. That can be expensive, especially for long lines, but there are some settings that can speed things up.

The descriptions of those settings don’t really capture what they do and why you might prefer one value over another so Cherti’s post is especially helpful. If you’re one of the people who like having your buffers numbered in one way or another, take a look at Cherti’s post to see how to get the best possible performance for your choice of representation.

Three years ago, I wrote about the Bedrock starter kit for Emacs by Ashton Wiersdorf. Now Wiersdorf has issued a new major release of the kit. It’s mainly an update for Emacs 31 and requires that version to run. I liked his announcement and description of Bedrock 2.0 and decided to write about it on Irreal. To be honest, I’d forgotten that I’d already written about Bedrock and didn’t discover that I had until I chose the obvious name for the post source file (bedrock.org) and found that such a file already existed.

You should read what I wrote before; everything from that post applies to the new version. The TL;DR is that although I don’t generally care for starter kits, I do like Wiersdorf’s because it mainly concerns itself with simple variable settings rather than configuring a bunch of packages that a user may or may not want¹. His init.el is well commented to explain what everything does and some items are commented out with an explanation of why you might want to enable them. You can see the file here.

If you’re looking for a simple, minimal configuration for Emacs as a starting point or you know a n00b who is, take a look at Bedrock.

… I’ve lost track. I’ve been using Jinx for my spelling correction for a while now and really like it. When I first started, my main problem was that the selection procedure described in the documentation didn’t work for me. I was almost certain that was because of an interaction with Ivy; my commenters agreed and offered some things to try.

After some experimentation, I still wasn’t able to get that aspect of Jinx working. I decided to wait for Emacs 31 to revisit the issue. Emacs 31 is here but I still haven’t addressed the problem. Most of the reason for that is that I’ve developed some efficient muscle memory for dealing with the problem. I have jinx-correct bound to Ctrl+; and when I call jinx-correct, I continue holding down Ctrl and can easily move down in the list of alternatives by pressing n. Usually the first choice is the one I want so it’s only a single press. Then I type Return to update the text.

That seems like a little more work than simply pressing a number but I don’t need to move my hands or reach to do it so it seems easy to me. I’ll probably do a bit more research and get the default method working when I get a chance but in the mean time, I’m pretty happy with how it’s working for me now.

I’m really enjoying how easy it is to put an “unknown spelling” in a variety of places—the current file, the current directory, or in a personal global dictionary—and am happy I switched. If you haven’t tried it yet, you should give it a spin.

Update [2026-09-09 Wed 11:55]: Fixed title typo.