Fritz Meissner

The answer is that all the versions that failed are old versions of their respective Ruby families and the build includes an old `tools/config.sub` which does not know about `arm64-apple...`....

Update: it is possible to override the result of config.sub by using `CONFIGURE_OPTS="--build ..."`. Specifying the build argument made this error go away. I passed `CONFIGURE_OPTS="--build aarch64-apple-darwin20"` to ruby-build. In...

@noraj you seem to be implying that it's not worth getting `asdf` to install Ruby without problems because native dependencies will cause problems anyway. In fact all the native dependencies...

@airtonix what a bizarre comment. I'm not complaining about using M1, I'm ecstatic with it.

LSP logs show that the server responds to the request for autocompletion of methods on the association with many items from `Enumerable`, `ActiveRecord::Associations::CollectionProxy` etc, but they all have a `"return_type":...

Hmmm... I had this working in previous versions of solargraph-rails. With `has_many :things`, I specified the return type on `.things` as `CollectionProxy` and based on that, Solargraph was smart enough...

Your code at `lib/solargraph/rails/model.rb:69` looks right: ``` types: ["ActiveRecord::Associations::CollectionProxy"], ``` which is converted to a YARD comment: ``` comments

The mystery deepens. I rolled back to solargraph 0.3.1 and I still don't see this working. Here is what I definitely had in the past: In [the GIF here](https://github.com/iftheshoefritz/solargraph-rails#associations-experimental): you...

I tested with https://github.com/castwide/solargraph/pull/585, but still `first` and `last` have an undefined `return_type`. I'm wondering if there's something we could add to `types.yml` that would do the trick?

@grncdr I'll check the hacks out shortly