test: math: test floating-point behavior when mapping zero to zero

base/special/trig.jl

@@ -1112,6 +1112,7 @@ function _cosc(x::Number)
         # generic Taylor series: π ∑ (-1)^n (πx)^{2n-1}/a(n) where
         # a(n) = (1+2n)*(2n-1)! (= OEIS A174549)
         s = (term = -(π*x))/3
+        iszero(s) && return s  # preserve floating-point signed zero
         π²x² = term^2
         ε = eps(fastabs(term)) # error threshold to stop sum
         n = 1

test/math.jl

@@ -590,6 +590,39 @@ end
     @test ismissing(scdm[2])
 end
 
+@testset "behavior at signed zero of monotonic floating-point functions mapping zero to zero" begin
+    function rounder(rm::RoundingMode)
+        function closure(::Type{T}, x::AbstractFloat) where {T <: AbstractFloat}
+            round(T, x, rm)
+        end
+    end
+    rounding_modes = (
+        RoundNearest, RoundNearestTiesAway, RoundNearestTiesUp, RoundToZero, RoundFromZero, RoundUp, RoundDown,
+    )
+    rounders = map(rounder, rounding_modes)
+    @testset "typ: $typ" for typ in (Float16, Float32, Float64, BigFloat)
+        (n0, n1) = typ.(0:1)
+        rounders_typ = Base.Fix1.(rounders, typ)
+        @testset "f: $f" for f in (
+            # all strictly increasing
+            identity, deg2rad, rad2deg, cbrt, log1p, expm1, sinh, tanh, asinh, atanh,
+            sin, sind, sinpi, tan, tand, tanpi, asin, asind, atan, atand, Base.Fix2(atan, n1), Base.Fix2(atand, n1),
+            Base.Fix1(round, typ), Base.Fix1(trunc, typ), +, ∘(-, -), ∘(-, cosc),
+            rounders_typ...,
+            Base.Fix1(*, n1), Base.Fix2(*, n1), Base.Fix2(/, n1),
+        )
+            @testset "s: $s" for s in (-1, 1)
+                z = s * n0
+                z::typ
+                @test z == f(z)::typ
+                @test signbit(z) === signbit(f(z))
+                isbitstype(typ) &&
+                @test z === @inferred f(z)
+            end
+        end
+    end
+end
+
 @testset "Integer and Inf args for sinpi/cospi/tanpi/sinc/cosc" begin
     for (sinpi, cospi) in ((sinpi, cospi), (x->sincospi(x)[1], x->sincospi(x)[2]))
         @test sinpi(1) === 0.0