Skellam (special case) is broken with Zygote and SpecialFunctions >= 1

[devmotion]

Currently, these distributions are broken due to FluxML/Zygote.jl#873. As discussed in #155, the corresponding tests are marked as broken. The issue will be fixed upstream.

See also: TuringLang/DistributionsAD.jl#145

[devmotion]

Copy from TuringLang/DistributionsAD.jl#145 (comment):

The main issue was fixed upstream, only some interesting issue due to Zygote._zero with ChainRulesCore.NotImplemented for matrices of Skellam variates is left: TuringLang/DistributionsAD.jl#172

[torfjelde]

So I guess this can be closed?

[devmotion]

No, Skellam is still broken with Zygote in one test.

[torfjelde]

Do you know which one? I just had to remove a bunch of broken = (:zygote, ) to make tests pass in #177 so jus curious which one still remains.

[devmotion]

Are you sure? I fixed these in #182, I thought.

[devmotion]

Ah, it seems you updated the tests before merging the master branch. I updated this issue here when I fixed the tests in #182, so the remaining error is still not fixed.

[torfjelde]

Aaaah, my bad; thanks!