Irreal

Yesterday I wrote that I had reconsidered my stance on Jinx and decided to give it a serious try. The only problem I had was that picking the correct spelling wasn’t working as advertised. I speculated that this was probably the result of some interaction between Jinx and Ivy and this, indeed, turned out to be the case.

Embarrassingly, mbork had already pointed this out in his post about Jinx that I linked to but since I wasn’t planning on installing Jinx, I hadn’t paid it much attention. His solution was simple: he merely disabled Ivy for Jinx by advising jinx-correct. That works fine but I’m not sure that I like that behavior either. You have to press Tab to get the list corrections and then it isn’t any easier choose the right one.

I’m sure there’s something simple that I’m missing so I’ll continue experimenting or perhaps one of Irreal’s more knowledgeable readers will tell me I’m just being stupid. In any event, I’m going back to the original behavior until I can figure out the correct way of fixing things.

In a stunning case of the Baader–Meinhof phenomenon, I’ve been running across the Jinx spelling corrector everywhere. I had never heard of it until last week when I watched and wrote about Prot’s video on writing in foreign languages with Emacs. Then Marcin Borkowski wrote about it and I posted about that. In all those instances, I concluded that Jinx was interesting but mainly of use for those who regularly wrote in more than one language. Since I seldom do that, I decided I was going to stick with flyspell.

Then Omar commented on my Borkowski post saying that Jinx was a better solution than flyspell even if you used it with only one language. I thought that perhaps I should reconsider my decision about sticking with flyspell. Finally, I saw this post by Bozhidar Batsov on why he has recently switched from flyspell to Jinx and became convinced that I should give it a try.

I installed it without difficulty and am writing this post with it as a sort of maiden voyage. So far, everything seems fine. By default, jinx-correct is bound to Meta+$, which I find a little clumsy so I changed that to Ctrl+;, which is where my muscle memory thinks it should be.

The only problem I’ve had is quickly choosing the correction. The documentation says that you can do that put typing the number of the correction but that simply exits the correction leaving the old spelling in place. The problem may be some interaction with Ivy, which I’ll investigate as soon as I get some time.

Marcin Borkowski (mbork) has a nice post on the Jinx spell corrector. Mbork has tried fly-spell but found it too laggy for his purposes. That was back in the days when he had a Pentium class machine running at 100 Mhz so he thought, in view of his more modern hardware, that it was worth trying out fly-spell again.

He still found it too slow and decided to try Jinx instead. I’ve always found fly-spell snappy and I hadn’t heard of Jinx until I read Prot’s post on writing (and spell correcting) foreign languages in Emacs but mbork wasn’t satisfied until he tried Jinx. He found that it was fast enough for his use cases and decided to embrace it as his spell corrector.

He discovered that Jinx interfered with some of his work flows but it turned out to be pretty easy to fix those. You can check out mbork’s post for the details.

If you are writing in multiple languages with Emacs, then Jinx may be an excellent solution as a spell checker. If, like me, you mostly write in your native tongue, you’ll probably find fly-spell is all you need. As I said above, it’s more than fast enough for me and provides a wonderfully interactive spell correction system.

Protesilaos Stavrou (prot) has an excellent video that addresses a problem I often see people asking about. The problem is how to globally set a key binding so that it won’t be overridden. He uses setting Meta+o to other-window as an example to explore the problem.

Prot begins by defining Meta+o in the global map in the usual way:

(define-key global-map (kbd "M-o") 'other-window )

That works well until he tries it in an HTML buffer. In the HTML buffer, Meta+o turns out to be a prefix key for inserting an HTML face command. The problem is that Emacs will use the most specific bindings first. Global bindings are the most general, followed by major mode bindings, and ending with minor mode bindings as the most specific.

Since HTML is a major mode, its keybinding take precedence over the global binding that we set. The answer is to get more specific yet by defining our binding in a minor mode so that it will always have precedence.

Prot shows how to define a minor where we can define our binding and activate it globally so that it’s always available. Of course, once we have this mechanism set up, we can add any number of bindings to our minor mode. Prot demonstrates this by adding a binding that Org uses.

Take a look at Prot’s video for all the details. He’s also included a copy of the file he was working with so you can see everything without worrying about copying it down from the video. The video is 12 minutes, 45 seconds so it should be easy to fit in.

My pal Watts Martin has a post that explores an artifact from the past: the insert key. Martin, who’s been a Mac head much longer than I have, remembers when Mac keyboards still had an Insert. They no longer do, of course, but my Unicomp Model M clone does as do several other PC keyboards. The key was used to toggle between inserting text and overwriting it. Emacs calls the latter overwrite-mode.

The idea of insert mode is that text in front of the cursor is pushed ahead as new text is added behind it. When insert mode is off—that is, when overwrite-mode is active—the text in front of cursor is overwritten. Overwriting text used to be more common but these days everyone uses insert mode and you rarely see it.

Still, as Martin says, overwriting can sometimes be useful and if you have an Insert key there’s no reason not to use it in Emacs. That’s a one liner and easy do in your init.el. On the Mac, the Insert key is usually mapped to Help but on other systems it probably has a different mapping so you may have to experiment to discover what it is.

Finally, Martin wanted a visual clue as to which mode he was in so he changed the shape of the cursor to reflect the mode. Again, that’s easy to do. He uses overwrite-mode-hook to toggle the cursor shape.

While I agree that overwrite-mode can occasionally be useful I don’t—and have never—used it enough to dedicate an actual key to it. I’m perfectly happy to simply call overwrite-mode to turn it on and off.