@lsenta: It’s a bit more complex than that unfortunately. I could probably plug into autocomplete, but that’s far from a complete solution. You also need navigation, doc, view source etc - it’s all doable from the REPL but it’s a totally different implementation from what’s there. Many things are more difficult and having two implementations would be a source of many bugs.

I’ll have some improvements for this soon using the current model.

I just had the weirdest bug, Cursive moved around a bunch of ] and ) all over the place

@danielcompton: Are you in parinfer mode?

I think I toggled through it

Hmm, I hadn’t considered that.

Why would it change anything?

So when you enter parinfer mode, it has to run paren mode to ensure that the indentation is within the limits required by indent mode.

I had assumed that anyone entering parinfer mode had explicitly chosen it, but of course toggling through it will trigger that.

Damn.

but why would it mess up my ] and )?

It shouldn’t have actually broken anything.

It created invalid code

Can you show a diff of an example?

I’ll try to reproduce it

I have a fancy diff tool which doesn’t show the +/- so a screenshot is better

So in that case, the indentation causes the parens to move to invalid locations. I’ll have to check if parinfer handles that case (multi-arity fns) - IIRC there was something in their issue tracker about that.

In any case, sorry 😞

I’m starting to think that indentation is too fragile in Clojure code for a global mode that assumes it.

It looks like the parinfer demo editor handles that case, I must have a bug in the integration.

There were quite a few things broken

I think it’s more general than just multi arity functions

Right, but in that case it hasn’t broken code, right? It’s just pulled the parens over the comments.

(which I agree is bad, but that is a known issue with parinfer)

ah yeah, that’s not a good example