David Widmann
David Widmann
It would be OK to add breaking changes here but IMO it's not good to be inconsistent with Julia base. In the sin(pi) PR triage made it clear that the...
> In that PR(JuliaLang/julia#42595), the discussion 1.0 vs 1 is not about accuracy, but we need to care about it in this PR. I don't see why we can't improve...
> Both sin(pi) and sin(quartπ) are Real, and there is no type-instability, so I think this is okay. Surely it is type stable but my point was that it is...
> In my local environment, the compilation time with this PR is much less than the current main branch. As @giordano mentioned, there are some problems with the implementation in...
> Julia v1.0.5 doesn't even support inv(π). That doesn't matter for us though as we don't define it? Since it's just a specific test that fails we could just not...
I thought one could just skip the tests that fail because some method in base is not in Julia 1.0. It's still tested properly in all other Julia versions so...
There's not much I can do here - I can't reproduce the issue (not the same specs but an M2 Mac). Did you try with a clean environment? And did...
> fresh installation of Julia After deleting `/Users/alexey/.julia`?
I still haven't been able to reproduce this issue, neither locally nor on CI. Therefore I'll close this issue for now.
`one` and `zero` for `AbstractIrrational`s was added in https://github.com/JuliaLang/julia/pull/34773, so this is the expected result in recent Julia versions.