The risk was worth it, node was stuck (in 0.12), and it was the whole thing, not "one of the things that you use to run JS on the server". A fork of actix wouldn't enjoy the same conditions that existed with Node.
Does the fork care as much about TechEmpkwer benchmarks? What happens if it is slower because it forbids unsafe even where it makes sense?
There was a looming cliff in Async/await, which caused a split to some extent (some old libraries that no longer have maintainers, changes that make updates difficult). A fork with those changes looming would mean divergence when rewriting to support Async/await.
Maybe I'm being too cautious, but I doubt we'd have had a successful fork earlier. Let's see if it happens now ...
Does the fork care as much about TechEmpkwer benchmarks? What happens if it is slower because it forbids unsafe even where it makes sense?
There was a looming cliff in Async/await, which caused a split to some extent (some old libraries that no longer have maintainers, changes that make updates difficult). A fork with those changes looming would mean divergence when rewriting to support Async/await.
Maybe I'm being too cautious, but I doubt we'd have had a successful fork earlier. Let's see if it happens now ...