search

jfmengels / node-elm-review

CLI for elm-review
https://package.elm-lang.org/packages/jfmengels/elm-review/latest/
BSD 3-Clause "New" or "Revised" License
48 stars 25 forks source link

Node 14, 16 installs error again #189

Closed lishaduck closed 4 months ago

lishaduck commented 4 months ago

Should still be fine for now, but glob (and the other issacs packages: rimraf, lru-cache, and jackspeak) just dropped 14 and 16 in a patch, and NPM warns on old Node versions again. Tests aren't catching it, and ls-engines doesn't resolve based on the ideal tree, just the lockfile. Though I think a flag should fix that, at least.

https://github.com/isaacs/node-glob/commit/8a69def3cad0de9ba26ca831065ffe448d153de3

CC: @jfmengels

lishaduck commented 4 months ago

Ok, yeah, npx ls-engine --mode=ideal fixes that. I guess I'll PR with the bump, and we can figure out how to handle it from there. Alternatively, I think we could explicitly express an upper bound.

lishaduck commented 4 months ago

Oh, and I just realized our engines is wrong, as well. I forgot to exclude 15 versions without some backports.

lishaduck commented 4 months ago

Ok, took another look. We can't express an upper-bound, as we don't depend on lru-cache or jackspeak. We could request that our transitives pin it, but then I think it'd be breaking for us to require a newer version for that. Oh, our transitions are glob.

See: https://github.com/isaacs/node-glob/issues/593#issuecomment-2211622255

lishaduck commented 4 months ago

To be clear, I personally agree that EOL node should be dropped, I'd just expect it in a major, and this is your project, and is depended on by many. I certainly don't mind if you'd rather not have (too major of a) a major (we are currently are in a major, given #157), I just don't know what we can do.