Matt Bauman

Ah, that's unfortunate. Perhaps this should be an upstream issue? Seems like a 1-element menu should make sense and be easy to support. cf. https://github.com/JuliaLang/julia/pull/35915#issuecomment-638316817

This would just be all kinds of breaking. Just skim across all the usages of [`isdigit` throughout the ecosystem](https://juliahub.com/ui/Search?q=isdigit&type=code) — so many packages do things that are effectively `isdigit(c) &&...

I took the liberty to add a few tests here. I would be surprised if the Windows CI passes, but I figured it was worth a shot. We can disable...

Argh, I didn't expect that changing a test file would break the build.

Well I was gonna complain that this didn't add tests, but then I realized that `getpass` didn't have _any_ tests in the first place... and as someone who has touched...

`abs` would also be a similar candidate — in fact its behavior is derived from unary `-`. Is loosing precision more compelling? ```julia julia> x = nextfloat(BigFloat(-1, precision=512)) -0.999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999925 julia>...

Might be worth asking triage for a second look? :) We have a bit more context now — this _would be_ a minor change and it's not necessarily the bug...

Ah, `copysign` is another `abs`-like generic function that relies on unary minus: ```Julia julia> copysign(prevfloat(BigFloat(1.0; precision=512)), 1) 0.999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999999925 julia> copysign(prevfloat(BigFloat(1.0; precision=512)), -1) -1.0 ```

Oh, yeah, public exprs print like this should work: ```Julia julia> Expr(:public, :foo) :(public foo) julia> @eval Base begin foo = 1 $(Expr(:public, :foo)) end julia> Base.ispublic(Base, :foo) true ```

Rudimentary integration tests added in #21. It'd still be nice to ~use deterministic seeding and~ add some stronger bounds on the results.