Alexander Plavin
Alexander Plavin
If @devmotion or others with a better understanding of weakdeps can resolve these warnings you see in CI, they are welcome!
Not sure from which side (Julia or packages), but those weakdeps issues are somehow getting resolved: new CI runs show no warnings or errors about them without any changes here....
Oh, it was just a wrong version! Now indeed we have no errors/warnings with this PR and can finally make the last dep weak. Anyone against this?
> PS: as mentioned before I am not opposed to this PR, it seems very reasonable for interval libraries to communicate. Though I am curious 🙂, is there any specific...
> I know nothing about IntervalSet.jl, can you perform any sort of arithmetic operations on IntervalSet.Interval? Not at all, they are "intervals as sets", not "intervals as numbers". Does it...
> A more perverse potential issue: does convert(::Type{
Thanks for looking into it! Updated the code.
> So accordingly, IA.Interval(1, Inf) should convert to IS.Interval{:closed, :open}(1, Inf). Hmm, I guess makes sense, even though it would be type-unstable... What's the best way to implement that? Just...
> Interval(2, 1). Is this supposed to be equivalent to Interval(1, 2)? No, it's an empty interval – no special casing, just following the regular interval definition: ```julia help?> Interval...
> isinf(lo) & (L == :closed) && return IA.nai(T) # eg IS.Interval(-Inf, 1) is invalid since closed > isinf(hi) & (R == :closed) && return IA.nai(T) # eg IS.Interval(1, Inf)...