Why don't test stable?

[ToBinio]

Just out of curiosity what is the reason that the test run on nighty, beta but not stable?

is it simple because when nighty and beta run stable will as well?

couldn't then the same be said about beta? when nightly works beta will as well

[Byron]

I actually checked when adding windows runners but couldn't find anything interesting in the respective commit messages.

With that said, I believe that lib.rs is built with nightly, which is why the maintainer was particularly interested in having this tested.
Using that as an argument, then we should probably add stable just for the sake of knowing it works there, not just assuming it by merit that nightly should never be less compatible than stable, nor should beta.

If all that is true, and the environment should be a concern, it should be possible to only test on stable and half CI time again, after it was doubled by adding windows runners.

My preference here is to only test stable and wait for issues to show up that indicate reasons for testing non-stable releases of Rust.

[Byron]

And I have made the change, let's see how that alters timings especially when caches finally kick in.