cljs-dev

dnolen 2023-09-28T18:15:28.432279Z

@dominicm https://github.com/clojure/clojurescript/pull/215 if the tests pass it would be good for you to give this a try

👀 2
dnolen 2023-09-28T19:39:50.474479Z

also I think looking at this https://github.com/aranlucas/react-hook-form-mantine/blob/master/package.json - this PR may fix CLJS-3405 as well

dominicm 2023-09-28T19:45:43.294899Z

It does appear to work for both, yep!

dominicm 2023-09-28T19:51:34.122869Z

The problem I had with react-hook-form-mantine was twofold: 1. The use of .mjs extension should probably be handled the same way as .cjs us here: https://github.com/clojure/clojurescript/blob/0da9884836e9f8ba190f3b789adcb6b9b85a669c/src/main/clojure/cljs/foreign/node.clj#L127-L128 (because of https://github.com/aranlucas/react-hook-form-mantine/blob/master/package.json#L23) 2. The resolution only checks the first key that exists: https://github.com/clojure/clojurescript/blob/0da9884836e9f8ba190f3b789adcb6b9b85a669c/src/main/clojure/cljs/foreign/node.clj#L111 But if it's not able to use that key (e.g. because the .mjs tripped it up) then it doesn't try the next key. Either would solve use of react-hook-form-mantine with package-json resolution set to ["module"] .

dominicm 2023-09-28T20:01:38.615759Z

@dnolen Just checking, was this an intentional side-effect of your change? https://clojurians.slack.com/archives/CRRJBCX7S/p1695930813733249

dnolen 2023-09-28T20:41:27.509299Z

Intentional yes

lilactown 2023-09-28T23:38:37.506829Z

fixing this warning would mean helix would be broken for earlier versions of CLJS, right?