Irreal

Emacs Elements has an interesting video on something I didn’t really know about: 2-Column editing. I say “really know about” because I always assumed it was just another name for follow mode. But, of course, it’s an entirely different thing.

Even after watching the video and reading the Emacs manual entry, I still couldn’t give a clear explanation of its purpose. The TL;DR is that if you have a buffer that contains vaguely table like data where there’s some notion of columns, you can “split” the data between columns with each column being in its own buffer. These two buffers are displayed side by side and can be edited independently. When you’re finished editing, you can merge the two buffers back together.

I’m doubtlessly suffering from a lack of imagination but I’m having a hard time figuring out a use case for this. As far as I can see, you can do all the same editing almost as easily without learning the somewhat confusing 2-column commands.

You, of course, may disagree. Obviously some people do or the facility wouldn’t exist. Take a look at the video and see what you think. The video is 11 minutes, 42 seconds so it should be easy to fit in.

Mike Zamansky is back with the concluding video in his function header project. It’s the 9^th video in his Learning Elisp series. This video ties up the loose ends in the previous videos in the project and provides a serviceable function to generate header comments for C-like languages. The final function generates javadoc formatted headers but it’s trivial to change the form by altering the format strings to be what you need.

The new ideas in this video include:

1. Using string-split, string-join, and mapcar to parse and format the individual parameters and recombine them into a printable string.
2. Using save-excursion to return the point back to where it was when the function was invoked.

The next project will involve inserting emojis by name into Emacs text but that will have to wait until Zamansky gets back from a European trip. He says that will be in mid October but that he might fit in a short video before he leaves.

In any event, the current video is 24 minutes, 20 seconds so you’ll definitely need to schedule some time. This video presents the development of a real function that an Emacs user might want to write to aid in their day-to-day development. Indeed, the function is a recreation of an actual function that Zamansky wrote years ago when he was an active developer.

This should probably be a Red Meat Friday post but I really hate camel case. It’s hard to read, hard to type, and looks ugly. Among it’s many other sins, it has made abominations such as subword mode necessary.

Over at Emacs Elements there is a (very) short video on using subword mode. There’s not much to say. It’s a way of having forward-word skip to each capitalized segment of a camel case word. The video recommends enabling it via a hook function for whatever programming modes you use. You can, of course, toggle it manually by invoking Meta+x subword-mode.

Some people love this mode but—the awfulness of camel case notation aside—I’ve never seen the point of it. Yes, occasionally it’s convenient to have forward-word stop at each subword of a camel case identifier, but most often you want to treat the camel case word for what it is: a single symbol. In those rare cases (heh) where you want to put the cursor on one of the subwords, it’s easy enough to skip to the head of the composite symbol and move the point to the correct spot. Or you could use something easier like jump-char-forward. Regardless, it’s hard to see the utility of subword-mode.

Still, people differ in their needs and if this is something that’s useful for your workflow, Emacs, as usual, has you covered. The video is only two minutes, 25 seconds so you should be able to slip it in easily.

UPDATE [2023-09-01 Fri 12:39]: Mickey writes to tell me that I may have been too harsh on subword mode. What really happens when you invoke it is that the definition of a “word” is changed to encompass the subwords of SomethingLikeThis. That means that not just forward-word but any word-based command can stop and act on the subwords.

That makes it more useful than my original post suggested but I still think that usually you want to treat “SomethingLikeThis” as the single symbol that it is and subword mode is almost always overkill. As a real world, real time example, I misspelled SomethingLikeThis as SomethingLikrThis but rather than toggling on subword-mode and moving back to the error, I simply called jump-char-backward and went right to the error.

I ran across this Toot—or whatever Mastodon calls it—in Sacha’s weekly Emacs News:

Even though it’s sort of old news at this point, it’s worth calling attention to. The idea is to use occur to narrow down the files in a Dired listing and then invoke the occur edit function to change the names of those files.

You can do the same thing from within Dired only but I like how the two-step process narrows down the files you’re concerned with and avoids errors. It’s a nice tip and worth taking a look at.