Irreal

Pavel Korytov is a fairly new Emacs user. He’s been at it for about two and half years but as seasoned Emacs users know, the learning never ends and even those of us with 15 or more years in are still in the learning phase.

He’s got an interesting post, 916 days of Emacs, that looks at how he spent his time in Emacs. He uses ActivityWatch and WakaTime to track his tube time so he has some pretty accurate statistics on the matter.

It’s an interesting post: his usage patterns are very different from mine. It’s especially nice to see his initial learning phase times. None of this makes any real difference, of course—it’s just an engaging peek into someone else’s Emacs use, and the beginning of their Emacs journey. Take a look if you’re interested in such things.

I’ve written a couple of posts about Álvaro Ramírez’s implementation of Emacs shells for ChatGPT and DALL-E. After he wrote those, he noticed that they shared a lot of code so, of course, he abstracted that code out into a separate package of functions.

The result is a wrapper around comint, shell-maker, that makes it easy to build a specialized shell. He offers a trivial example of a shell, greeter-shell that accepts a name, X, and responds with Hello "/X/". That’s not terribly useful, of course, but it takes just a few lines of code—most of it boilerplate—to implement.

Right now shell-maker is bundled with chatgpt-shell but Ramírez says he will break it out to a separate package if there’s enough interest. Regardless, if you want to implement a specialized shell—whether or not it’s in service of the hot new AI apps—the shell-maker package is an easy way to do that without worrying about the intricacies of Comint. Take a look at Ramírez’s post for the details.

SnipingMan18 has posed an interesting questions on the Emacs sub-reddit: How would you describe the Emacs philosophy in a single line? If you’re an Emacser, take a moment to think how you would answer that question.

My answer popped into my head almost immediately so I read all the comments to see if anyone had the same take: no one did. Some of the answers were silly, but others made a lot of sense. For example, it’s hard to argue that “the extensible editor” captures (at least one) essence of Emacs. Another nice answer was “have it your way.” That certainly captures an important part of Emacs as I’ve often noted. Another good candidate is “the DIY editor”.

All of those answers provide a good, concise, précis of what Emacs is about but none of them nail down what makes it so special to me. Regular readers won’t be surprised that my answer is “Lisp machine lite” or something like that. The point is that Emacs is not merely an editor but an operating environment to handle all your routine chores. It is, in fact, an environment for hackers to solve problems in, even if those problems don’t involve writing code.

As I’ve said many times, the only app other than Emacs that I spend significant time in is Safari. When Emacs finally solves the problem of providing a usable Web interface—sorry, EWW is not there yet—probably 95% of my tube time will be spent in Emacs.

Your answer may differ but it’s hard to argue that Emacs provides an useful and powerful Lisp environment.

The other day, I wrote about Álvaro Ramírez’s post on his Emacs theme. In a nice coda to Ramírez’s post, Aimé Bertrand writes about how he configures his UI and the functions he uses to support it.

Bertrand has some nice ideas. He’s a Mac user so some of them are meant for macOS only but others work in any environment. Among the Mac specific functions are those to switch his theme based on whether the Mac is in light or dark mode.

As he points out, switching themes can be a problem because some aspects of the previous theme can remain active. He has a function that solves this problem by first deactivating the previous theme and then activating the new one.

Another nice trick is dealing with the “warning beep”. The conventional way for dealing with this is just to set ring-bell-function to 'ignore but then you get no notification at all. You can also set visible-bell to t to flash the screen instead but many find that annoying. Bertrand solves this problem with a function that flashes the mode line instead. That’s arguably better than no notice at all but less annoying than an audible beep or flashing the whole window.

Finally, he’s got a function to propertize a string in a way similar to set-face-attirbute. I don’t have a need for such a function but I can see how it’s just what you need if, say, you’re writing a package and want to specialize the appearance of a string.

As I say, Bertrand has some good ideas so it’s worth taking a look at his post. It’s fairly short so it won’t take much time to give it a read.

Marcin Borkowski (mbork) has an instructive post based on a question in the emacs-humanities mailing list. The question that Borkowski addresses is how to replace every occurrence of a word in a buffer with its downcased version without typing the word twice. Borkowski solves this problem by writing a bit of Elisp.

In a sense, his solution is silly—as even he admits—because there are pretty easy ways of doing this with built-in commands as a couple of the respondents replied. Still, Borkowski’s solution serves as a nice example of writing a bit of Elisp to solve a simple problem.

As usual, Borkowski shows us the scaffolding. By that I mean that he doesn’t just show us a solution but the steps—and missteps—he took arriving at it. That’s why it’s an instructive post. The main thing you learn from the post is how to search through a buffer for a given regex and replace it with something else. That’s basically the interactive query-replace-regexp but you’re not supposed to call that directly from Elisp for various reasons.

Borkowski gives us the template for this sort of operation and makes it obvious how to change the operation on the target or add other steps. It’s a nice post and worth reading even if you’re already conversant with Elisp.