Add plugin for AASM gem.

[connorshea]

This adds a few autogenerated methods that are created when a model has an AASM state machine.

Gem: https://github.com/aasm/aasm

I've tested this with my rails app vglist and am confident that it works correctly.

[codecov[bot]]

Codecov Report

Merging #427 (4640824) into master (b997f5b) will decrease coverage by 0.06%. The diff coverage is 0.00%.

@@            Coverage Diff             @@
##           master     #427      +/-   ##
==========================================
- Coverage   96.51%   96.44%   -0.07%     
==========================================
  Files         116      116              
  Lines        2894     2896       +2     
==========================================
  Hits         2793     2793              
- Misses        101      103       +2     

Impacted Files	Coverage Δ
lib/sorbet-rails/model_plugins/plugins.rb	66.12% <0.00%> (-2.21%)	:arrow_down:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update b997f5b...4640824. Read the comment docs.

[connorshea]

Looks like this is failing because Sorbet is using endless ranges, which aren't supported on Ruby 2.4 or 2.5: https://github.com/sorbet/sorbet/issues/4265 / https://github.com/sorbet/sorbet/pull/4281

We could either wait for the upstream fix (unclear if it'll be merged) or just drop Ruby 2.4 and 2.5 since they're EOL anyway.

CI is also failing for Rails 6.0 / Ruby 2.7, not entirely sure why that is.

[connorshea]

After trying to investigate why the signature for Wizard.all.in_batches changed (or, perhaps, why Sorbet only just recently started complaining about it not matching the asserted type in the test cases file), I have genuinely no idea. My best guess is that there was a change in Sorbet at some point in the last few weeks which made Sorbet understand the types better, or something? 🤷‍♂️

The change from T::Enumerable[Wizard::ActiveRecord_Relation] to ActiveRecord::Batches::BatchEnumerator for the return type of the in_batches method was made last May. No idea why this test was passing up until recently 🤷‍♂️

[connorshea]

Looks like this is failing because Sorbet is using endless ranges, which aren't supported on Ruby 2.4 or 2.5: sorbet/sorbet#4265 / sorbet/sorbet#4281

We could either wait for the upstream fix (unclear if it'll be merged) or just drop Ruby 2.4 and 2.5 since they're EOL anyway.

CI is also failing for Rails 6.0 / Ruby 2.7, not entirely sure why that is.

Looks like the latest Sorbet version restores compatibility with 2.4 and 2.5, so this should be fixed if we re-run travis.

[hdoan741]

@connorshea Thank you! I've fixed the tests (ActiveRecord::Batches::BatchEnumerator) and dropped ruby 2.4 and 2.5 from the test matrix, as well as Rails 5.0 and 5.1.

[connorshea]

Looks like CI passes now :)