@ChrisRackauckas I have no idea why this test is failing. DiffEqSensitivity is in the TOML Extras and using DiffEqSensitivity is there. None of these tests or the TOML are touched by the commits and passed previously. Do you have any suggestions?

No issues locally when doing activate DiffEqUncertainty & test
Thanks.

https://travis-ci.org/github/SciML/DiffEqUncertainty.jl/jobs/721741699#L380

It's a versioning issue. Something is keeping it back in the stone age.

Why would it work locally though? I thought running Pkg> test ensured the same environment

hmmm... doing up on the dev environment knocked it back to the stoneage locally too

@ChrisRackauckas Can you rerun this test? This PR was held up by the [email protected] debacle.

I bumped the tolerances a bit to make the new AD tests pass but I think there's a real failure somewhere.

What's the blocker on this? My toy example fails (although it fails on both MonteCarlo and Koopman https://gist.github.com/sdwfrost/2c14346ced4d8279ef1769d1ef81d78d

I think the main blocker is that we will need to change the repo substantially in order to really get the performance in check. It needs a DSL. Without it, it cannot do codegen in a way that's type stable. So that's a pretty big project that will startup in a bit here.