Hi everyone, it’s been a while since I’ve used kakoune as I’ve discovered the excellent Dance for VSCode, which lets you use kakoune keys and editing in VSCode itself.
This day, I’ve decided to quit, however, as VSCode’s Intellisense is infuriating, and completely unconfigurable.
I thought “Now’s the time to go running back into the open arms of kakoune!”. But kak has the same problem. And I don’t know what to do.
I don’t have the scope to go into every problem that exists with it, but essentially, there are far too many suggestions, and it triggers for everything you type. The only real option available is to turn it off, which is hardly an option. In each project, the number of suggestions builds up until it becomes completely unsustainable. The best I can do is to give a couple of personal examples:
- Example 1: Dart / Flutter
I have a Flutter project, with a lot of classes, sub-libraries, libraries loaded, etc. Intellisense would be helpful, but I can’t concentrate on this complicated project with the fuzzy matching triggering for everything, every time, as the sheer amount of library classes summoned up is overwhelming.
Desired behaviour: Start Intellisense only after entering >=3 characters; bring in variables local to the file but not from other open files or from comments; use fuzzy matching for classes but strict for snippets as I have a lot of them.
Actual behaviour: Nothing, a total mess, just a completely unconfigurable, eye-rending spam of garbage after everything I enter.
- Example 2: BQN (an APL language)
I’m writing some small projects with this strange language (BQN), and the VSCode extension uses snippets to enter the peculiar glyphs, e.g.
∧∨¬⊣⊢↑↓«»⌾, used by this language. Snippets are in the form of
\ followed by an ASCII character; for example,
\q, ⍷ is
\$, and so on. It would be a perfect system, but it causes VSCode’s Intellisense to trigger for everything I write, because every single alphanumeric character is tied to one or two snippets. Better, also, would be if I could auto-enter snippets after completion (but for this project only)—it would save a huge amount of tab/enter keypresses.
Desired behaviour: Strictly complete variable names of >3 length, but only suggest snippets after entering the
\ character; auto-insert snippets after completion.
Actual behaviour: Zilch, nada, nothing, the Intellisense is a completely uncustomizable waste of time.
So, then, that’s my experience.
In short: I want to come back to
kak, but I’m afraid to because the fuzzy Intellisense / completions were always a frustration to me, and I’ve not seen any way to configure them. However, VSCode and Helix, for example, have the same problem. So I’m stuck, I don’t know what to do, and I thought we could share some fruitful debate on the matter?
It might seem like I’m being pernickety, but a lot of great things are started by somebody having a very specific vision, one that other people didn’t have. And I can’t be the only one. So let’s talk about it!