adjoint test may work

test/rulesets/Base/mapreduce.jl

@@ -48,16 +48,10 @@
                 # Adjoint -- like PermutedDimsArray this may actually be used
                 xa = adjoint(rand(T,4,4))
                 test_rrule(prod, xa ⊢ rand(T,4,4))
-                @test_skip test_rrule(prod, xa ⊢ rand(T,4,4), fkwargs=(dims=2,)) # seems broken?
-#=
-xa = randn(3,3)'
-Zygote.gradient(x -> sum(prod(x,dims=1)), xa)[1]
-ForwardDiff.gradient(x -> sum(prod(x,dims=1)), xa)
-Zygote.gradient(x -> sum(prod(x,dims=1)), Matrix(xa))[1]
-# These all agree, so the fault is in the testing, somehow.
-=#
+                test_rrule(prod, xa ⊢ rand(T,4,4), fkwargs=(dims=2,))
                 @test unthunk(rrule(prod, adjoint(rand(T,3,3)))[2](1.0)[2]) isa Matrix
                 @test unthunk(rrule(prod, adjoint(rand(T,3,3)), dims=1)[2](ones(1,3))[2]) isa Matrix
+
                 # Diagonal -- a stupid thing to do, product of zeros! Shouldn't be an error though:
                 @test iszero(unthunk(rrule(prod, Diagonal(rand(T,3)))[2](1.0)[2]))
                 @test iszero(unthunk(rrule(prod, Diagonal(rand(T,3)), dims=1)[2](ones(1,3))[2]))