~singpolyma/em_fiberscheduler

Add contributing/getting help to readme
Allow disabling/changing autostart/autostop behaviour
Run rubocop in CI
Import io_read and io_write from fiber_scheduler
Work around race condition

Only seems to come up on Ruby 3.0, but safe to do in any case
Use specs from git
Do not resume dead Fibers

They may be dead due to a Fiber#raise, especially due to Timeout
Use EM.defer for ProcessWait fallback
Add CI for ruby 3.0.2
Run when there is any EM work left
Refactor process_wait
Add address_resolve
Try timeout_after