Irreal

The Emacs Command of the Day Twitter feed tells us something that I didn’t know:

Fun fact about the minibuffer: you can click it when it’s inactive! This opens the *Messages* buffer.

(Especially handy if you want to read messages that recently appeared in the minibuffer.)

— Ⓔ Command of the Day (@ecotd) April 18, 2022

That’s a really handy way of popping up the *Messages* buffer and it works even though the minibuffer is not active. Of course, you have to use the mouse to invoke it but that’s not too bad if you have a laptop with a trackpad—especially a MacBook where you can do the whole operation with your thumb while keeping your fingers on the home row.

There are many ways of popping up the *Messages* buffer—Ctrl+h e for example—but clicking on the minibuffer seems particularly easy. If you’ve taken a sacred pledge to abjure mouse usage from within Emacs, this method is obviously not for you but if you’re flexible about the mouse it’s a nice thing to know.

UPDATE [2022-04-26 Tue 13:24]: adjure → abjure

Here’s another post from Tony Aldon. Sacha already covered it in Emacs News but it’s interesting enough that it deserves another mention. The post is about catch/throw and examples of its use.

Aldon starts with some simple made up examples of using the catch/throw paradigm. He shows some of the ways it can be used. Most of the variation is in specifying the symbol that names the catch block but he also shows some examples of nested catch blocks.

After the simple introductory examples, Aldon presents some real examples from the Emacs source code. They’re more complicated, of course, and involve more code than just the catch/throw but they do show catch blocks in the wild.

The main takeaway about catch/throw is that it’s a mechanism for non-local exits—the throw can even be in a separate function—and is not an error mechanism like the try/catch sequence in Python, Javascript, C#, and many other languages. Its quintessential use is to break out of loops. A nice example is a dolist loop: there’s no way to specify an exit criterion so the catch/throw mechanism is perfect for exiting the loop when some criterion is satisfied.

Ever since I read Abo-abo’s post on using wgrep to edit multiple files, I’ve been a huge fan of using grep to search for some regular expression in multiple file, write enable the results buffer with wgrep, make changes to the results, and save those changes back to the original files. Although there are multiple workflows to accomplish this. they all involve using one of the grep commands to search for the regexp in multiple files, modifying the grep results, and then saving the results back to the files. This is an extraordinarily powerful technique that’s easy to generalize.

Tony Aldon has a post that offers another variation on the theme. It’s basically the same except that he uses rg.el to perform the grep duties. It’s very similar to what I do except that I use counsel-rg (part of the ivy swiper package) to do the searching.

Regardless of the details, the grep/wgrep paradigm is a powerful one and definitely worth knowing. It can make a complex refactoring utterly simple. The takeaway from all the posts on the matter is that you can use whatever flavor of grep you like and just write enable the grep results buffer to make changes that are reflected back to the original files.

Derek Taylor over at DistroTube has a nice video up on ERC, the builtin Emacs IRC chat client. IRC is, of course, invisible to most of the general public and maybe even some of the younger nerds. It’s a very old chat protocol that predates the Web and was popular back when Usenet was still an important way of sharing information. It’s sort of like the old Unix talk application but it’s run on a remote server and supports multiple people.

The #Emacs channel, for example is a place where Emacs users and developers can hang out, ask questions, and discuss matter important to the community. Sort of like Twitter but without the crazy people.

Taylor first gives a demonstration of vanilla ERC without any configuration. It’s surprisingly useful that way and probably more than adequate for occasional users. Then he goes over his, fairly minimal, configuration and shows the difference it makes in a session.

Most of his configuration just involves setting up some useful defaults—such as his nickname, IRC server, and the like—to make signing onto a session a bit simpler. It’s probably easier for new users to simply fire up ERC without worrying about configuration. The configuration can come later when they decide to become regular users.

The video is 14 minutes, 51 seconds so it should be relatively easy to fit in. If you’ve been curious about IRC and are an Emacs user, Taylor’s video is an excellent way to learn enough to try it out.

Update [2022-04-16 Sat 17:45]: pubic → public