You mean that interactive is failing?

Strange, interactive-test is passing for me. You can open the devtools console on Electron and see what's wrong (I believe its ctrl-`shift`-`i`) or try to restart the compiler and then reload the electron app

Ha! Always room for confusion! By “interactive(?)” I meant can I run something non-interactive that would exit with 0 on success. I can see why you thought I was referring to interactive-test which frankly, I am not even aware of yet! I only ran the commands you shared with me in chat:

npm install
./script/watch
npm start

Right! So, one way is that after ./scripts/watch you run:

npx shadow-cljs clj-run repl-tooling.integration/run-tests-on-ci

That's what I use to run on CircleCI

Wait...

You don't even need to run ./scripts/watch 😄

Aha!

You can just run:

npm install
npx shadow-cljs clj-run repl-tooling.integration/run-tests-on-ci

This will fire up the watch process for shadow, start the "fixture app", run the tests, and will exit with code 0 if tests pass, or 1 if something fails. I believe it'll also retry tests that fail up to 5 times

(It's just an etaoin code to scrap the tests from the Electron app, really 😄. So I believe you'll need Chromedriver - if you're on Linux, you can get it from here: https://chromedriver.storage.googleapis.com/76.0.3809.126/chromedriver_linux64.zip)

cool, I’ll check repl-tooling for any pre-reqs needed.

I dev on macOS but CI on Linux.

I can do a brew install chromedriver so, maybe…

Ah no the brew installed chromedriver mismatches. But there is https://www.npmjs.com/package/electron-chromedriver, but no exact match for electron 6.0.2, but maybe 6.0.0 will work? Yeah, I can launch your tests.

But I’m getting failures…

I’ll try on a linux VM first.

Hmm.. I’m getting test failures on my linux VM as well. I shall probably put adding repl-tooling to rewrite-clj libs tests on on hold for now.

Neat what you did with Electron there though!

Thanks! The Electron version should not matter much, really. I'll try to see if there's something wrong on my side too

(I'll start some features in the near future, so I'll re-visit these tests 😄)