@shaunlebron BTW my question about the demo site was more whether I could easily spin up a webpage to try the code right from the parinfer repo (without having to integrate it in Atom or whatever). If that’s not easy, don’t worry about it.

@shaunlebron Shouldn't this change bump the major version number? Seems like a backwards-incompatible change to the API.

published 3.0.0

@shaunlebron i can't wait to integrate it into all my projects! btw, do you find that streaming helps you work through problems? i plan on trying it for a new project i'm about to start. do you use OBS to record?

I apologize if this comes off as alarmist, but I am waiting to integrate 3.0.0 into atom-parinfer for the time being. I'm not sure all of the parinfer test cases are running correctly with this change.

The test case annotation code is somehow inside the parinfer library and the test harness isn't even calling public API methods. Maybe everything is working fine, but difficult to be sure. Also the test harness code still references cursorDx.

Everything is probably fine, just needs to be moved around some.

@sekao Are you still using parinfer-jvm, or are you only maintaining JS editors these days?

@sekao when he was cmd-tabbing around i saw the OBS icon

@cfleming just JS now. Once you get the changes into the jvm lib I'll definitely bring that into cross-parinfer though

@sekao Ok. Is anyone using cross-parinfer that you know of?

Just thinking about possible API changes - my version has diverged a bit from parinfer-jvm

@sekao: I expected streaming to slow me down a bit, but I think the focus/clarity of talking to an audience actually made me get a lot more done

it helped me get two major features done in two streaming sessions, so I’m definitely going to keep doing it

and yes, I was using OBS

@chrisoakman yeah, i understand your concern here. the tests call special test methods which in turn call the public API

so before, we trusted the external test parser to create reliable input and output, but now we are trusting the internal parser to do so. i don’t think it’s any different

if you’re still worried about it, I may not be able to address that this week

Been wondering if there's any benefit to updating nvim-parinfer to use a newer parinfer.js thinger?

I've been hoping for smarter indent/undent.

Oh, it's on 1.8.1

If I understand correctly from the chat log, 3.0.0 changed the interface in preparation for new features we don't have yet, right? So it's useful to update to 3.0.0, but mostly it is just fixes and interface changes?

3.0.0 changes the public API and adds significant new functionality

Ahhh. That's appealing.

I just bumped the nvim-parinfer refactor PR @snoe ... let me know if there's anything I need to do to make that acceptable for merge (or if it sucks :)

@cfleming i don't know of anyone else using it

@shaunlebron the "rubber duck" aspect of streaming sounds useful. i also expect i'll be less tempted to waste time browsing the web if i know my screen is being recorded 😄

@cfleming has already been playing around with it, but I updated the cursorDx proof-of-concept fork to use the new v3 API (and probably also introduced a ton of inefficiency, but hey): https://github.com/rgdelato/atom-parinfer/tree/cursor-dx-poc

I’ll be playing around with @rgdelato’s fork some more this afternoon, and creating test cases in the parinfer repo for problem cases that I find. I’ll try to fix them too.

Thinking about it last night, I suspect the change dx calculation might be a bit simplistic right now. For example, if someone pastes in some unbalanced code with a lot of extraneous close parens, they’ll be removed by the algorithm and should be taken into account for the dx.