Luke Boswell (Sep 10 2026 at 00:59): Luke Boswell (Sep 10 2026 at 00:59):What do people think about making this a warning? It's redundant and for some reason our Agentic friends love doing it.
import pf.Elem exposing [Elem]
Would import pf.Elem exposing [foo, bar] also expose Elem implicitly?
Luke Boswell (Sep 10 2026 at 02:18):Yes I think so
Richard Feldman (Sep 10 2026 at 02:24):seems reasonable to me! :thumbs_up:
Richard Feldman (Sep 10 2026 at 02:25):also could have the formatter autofix it
Richard Feldman (Sep 10 2026 at 02:25):maybe let's just do that and not give a warning
Richard Feldman (Sep 10 2026 at 02:25):ehh both is probably better actually, that way the warning gives an actual explanation
Luke Boswell (Sep 10 2026 at 03:02):The formatter doesn't have type information so it would have to be a heuristic "same name" thing I guess
Luke Boswell (Sep 10 2026 at 03:07):Made an issue for this https://github.com/roc-lang/roc/issues/11270
Richard Feldman (Sep 10 2026 at 03:18):I think same name is deterministic, right? :thinking:
Richard Feldman (Sep 10 2026 at 03:18):like I don't think type information would matter even if it had access to it :smile:
Luke Boswell (Sep 10 2026 at 03:21):Can I expose a nested type with the same name as the parent?
Luke Boswell (Sep 10 2026 at 03:21):I guess that should not be permitted
Luke Boswell (Sep 10 2026 at 03:22):Elem.Elem.Elem
Richard Feldman (Sep 10 2026 at 03:59):oh hm I guess you can
Richard Feldman (Sep 10 2026 at 04:00):I mean it should be permitted, it's just silly
Richard Feldman (Sep 10 2026 at 04:00):so is the status quo of exposing [Elem] actually a bug?
Richard Feldman (Sep 10 2026 at 04:01):like should it be an error bc that's like trying to expose Elem.Elem, which doesn't exist?
Last updated: Sep 24 2026 at 15:59 UTC