remove short circuit for correctness

This optimization gives wrong result, if `NaN` and `missing` both exist.

add missing test

Author: N5N3

base/reduce.jl

@@ -637,11 +637,6 @@ function mapreduce_impl(f, op::Union{typeof(max), typeof(min)},
     start = first + 1
     simdstop  = start + chunk_len - 4
     while simdstop <= last - 3
-        # short circuit in case of NaN or missing
-        (v1 == v1) === true || return v1
-        (v2 == v2) === true || return v2
-        (v3 == v3) === true || return v3
-        (v4 == v4) === true || return v4
         @inbounds for i in start:4:simdstop
             v1 = _fast(op, v1, f(A[i+0]))
             v2 = _fast(op, v2, f(A[i+1]))

test/reduce.jl

@@ -395,6 +395,9 @@ A = circshift(reshape(1:24,2,3,4), (0,1,1))
         @test maximum(x) === minimum(x) === missing
         @test extrema(x) === (missing, missing)
     end
+    # inputs containing both missing and NaN
+    minimum([NaN;zeros(255);missing]) === missing
+    maximum([NaN;zeros(255);missing]) === missing
 end
 
 # findmin, findmax, argmin, argmax