Revision f442e4230a41d0d40f81346c1707f946f743cbcf authored by Jeff Bezanson on 09 June 2021, 02:51:34 UTC, committed by GitHub on 09 June 2021, 02:51:34 UTC
This PR implements a way to keep tables of methods that are
not part of the internal method table, but still participate
in the special support we have for keeping tables of methods,
in particular unification through precompilation and efficient
lookup. The intended design use case is to allow for method overlay
tables for various non-CPU backends (e.g. GPU and TPU). These
backends would like to modify basic function like `sin` to
perform better on the device in question (or in the case of TPU
to define them over non-LLVM intrinsics).

It is worth noting that this PR only gives the ability to keep
these tables of methods. It assigns no particular meaning to them
and the runtime (and regular inference) do not look at them.
They are designed as an implementation detail for external
compilers and similar tools.

 # Demo

```julia
julia> using Base.Experimental: @overlay, @MethodTable

julia> @MethodTable(mt)
 # 0 methods:

julia> @overlay mt function sin(x::Float64)
           1
       end

julia> @overlay mt function cos(x::Float64)
           1
       end

julia> mt
 # 2 methods:
[1] cos(x::Float64) in Main at REPL[5]:1
[2] sin(x::Float64) in Main at REPL[4]:1

julia> Base._methods_by_ftype(Tuple{typeof(sin), Float64}, mt, 1, typemax(UInt))
1-element Vector{Any}:
 Core.MethodMatch(Tuple{typeof(sin), Float64}, svec(), sin(x::Float64) in Main at REPL[4]:1, true)

julia> Base._methods_by_ftype(Tuple{typeof(sin), Float64}, 1, typemax(UInt))
1-element Vector{Any}:
 Core.MethodMatch(Tuple{typeof(sin), Float64}, svec(Float64), sin(x::T) where T<:Union{Float32, Float64} in Base.Math at special/trig.jl:29, true)
```

Co-authored-by: Tim Besard <tim.besard@gmail.com>
Co-authored-by: Julian P Samaroo <jpsamaroo@jpsamaroo.me>
Co-authored-by: Keno Fischer <keno@juliacomputing.com>
1 parent 0e3276c
Raw File
math.jl
# This file is a part of Julia. License is MIT: https://julialang.org/license

using Random
using LinearAlgebra

function isnan_type(::Type{T}, x) where T
    isa(x, T) && isnan(x)
end

@testset "clamp" begin
    @test clamp(0, 1, 3) == 1
    @test clamp(1, 1, 3) == 1
    @test clamp(2, 1, 3) == 2
    @test clamp(3, 1, 3) == 3
    @test clamp(4, 1, 3) == 3

    @test clamp(0.0, 1, 3) == 1.0
    @test clamp(1.0, 1, 3) == 1.0
    @test clamp(2.0, 1, 3) == 2.0
    @test clamp(3.0, 1, 3) == 3.0
    @test clamp(4.0, 1, 3) == 3.0

    @test clamp.([0, 1, 2, 3, 4], 1.0, 3.0) == [1.0, 1.0, 2.0, 3.0, 3.0]
    @test clamp.([0 1; 2 3], 1.0, 3.0) == [1.0 1.0; 2.0 3.0]

    @test clamp(-200, Int8) === typemin(Int8)
    @test clamp(100, Int8) === Int8(100)
    @test clamp(200, Int8) === typemax(Int8)

    begin
        x = [0.0, 1.0, 2.0, 3.0, 4.0]
        clamp!(x, 1, 3)
        @test x == [1.0, 1.0, 2.0, 3.0, 3.0]
    end
end

@testset "constants" begin
    @test pi != ℯ
    @test ℯ != 1//2
    @test 1//2 <= ℯ
    @test ℯ <= 15//3
    @test big(1//2) < ℯ
    @test ℯ < big(20//6)
    @test ℯ^pi == exp(pi)
    @test ℯ^2 == exp(2)
    @test ℯ^2.4 == exp(2.4)
    @test ℯ^(2//3) == exp(2//3)

    @test Float16(3.0) < pi
    @test pi < Float16(4.0)
    @test widen(pi) === pi

    @test occursin("3.14159", sprint(show, MIME"text/plain"(), π))
    @test repr(Any[pi ℯ; ℯ pi]) == "Any[π ℯ; ℯ π]"
    @test string(pi) == "π"
end

@testset "frexp,ldexp,significand,exponent" begin
    @testset "$T" for T in (Float16,Float32,Float64)
        for z in (zero(T),-zero(T))
            frexp(z) === (z,0)
            significand(z) === z
            @test_throws DomainError exponent(z)
        end

        for (a,b) in [(T(12.8),T(0.8)),
                      (prevfloat(floatmin(T)), prevfloat(one(T), 2)),
                      (prevfloat(floatmin(T)), prevfloat(one(T), 2)),
                      (prevfloat(floatmin(T)), nextfloat(one(T), -2)),
                      (nextfloat(zero(T), 3), T(0.75)),
                      (prevfloat(zero(T), -3), T(0.75)),
                      (nextfloat(zero(T)), T(0.5))]

            n = Int(log2(a/b))
            @test frexp(a) == (b,n)
            @test ldexp(b,n) == a
            @test ldexp(a,-n) == b
            @test significand(a) == 2b
            @test exponent(a) == n-1

            @test frexp(-a) == (-b,n)
            @test ldexp(-b,n) == -a
            @test ldexp(-a,-n) == -b
            @test significand(-a) == -2b
            @test exponent(-a) == n-1
        end
        @test_throws DomainError exponent(convert(T,NaN))
        @test isnan_type(T, significand(convert(T,NaN)))
        x,y = frexp(convert(T,NaN))
        @test isnan_type(T, x)
        @test y == 0

        @testset "ldexp function" begin
            @test ldexp(T(0.0), 0) === T(0.0)
            @test ldexp(T(-0.0), 0) === T(-0.0)
            @test ldexp(T(Inf), 1) === T(Inf)
            @test ldexp(T(Inf), 10000) === T(Inf)
            @test ldexp(T(-Inf), 1) === T(-Inf)
            @test isnan_type(T, ldexp(T(NaN), 10))
            @test ldexp(T(1.0), 0) === T(1.0)
            @test ldexp(T(0.8), 4) === T(12.8)
            @test ldexp(T(-0.854375), 5) === T(-27.34)
            @test ldexp(T(1.0), typemax(Int)) === T(Inf)
            @test ldexp(T(1.0), typemin(Int)) === T(0.0)
            @test ldexp(prevfloat(floatmin(T)), typemax(Int)) === T(Inf)
            @test ldexp(prevfloat(floatmin(T)), typemin(Int)) === T(0.0)

            @test ldexp(T(0.0), Int128(0)) === T(0.0)
            @test ldexp(T(-0.0), Int128(0)) === T(-0.0)
            @test ldexp(T(1.0), Int128(0)) === T(1.0)
            @test ldexp(T(0.8), Int128(4)) === T(12.8)
            @test ldexp(T(-0.854375), Int128(5)) === T(-27.34)
            @test ldexp(T(1.0), typemax(Int128)) === T(Inf)
            @test ldexp(T(1.0), typemin(Int128)) === T(0.0)
            @test ldexp(prevfloat(floatmin(T)), typemax(Int128)) === T(Inf)
            @test ldexp(prevfloat(floatmin(T)), typemin(Int128)) === T(0.0)

            @test ldexp(T(0.0), BigInt(0)) === T(0.0)
            @test ldexp(T(-0.0), BigInt(0)) === T(-0.0)
            @test ldexp(T(1.0), BigInt(0)) === T(1.0)
            @test ldexp(T(0.8), BigInt(4)) === T(12.8)
            @test ldexp(T(-0.854375), BigInt(5)) === T(-27.34)
            @test ldexp(T(1.0), BigInt(typemax(Int128))) === T(Inf)
            @test ldexp(T(1.0), BigInt(typemin(Int128))) === T(0.0)
            @test ldexp(prevfloat(floatmin(T)), BigInt(typemax(Int128))) === T(Inf)
            @test ldexp(prevfloat(floatmin(T)), BigInt(typemin(Int128))) === T(0.0)

            # Test also against BigFloat reference. Needs to be exactly rounded.
            @test ldexp(floatmin(T), -1) == T(ldexp(big(floatmin(T)), -1))
            @test ldexp(floatmin(T), -2) == T(ldexp(big(floatmin(T)), -2))
            @test ldexp(floatmin(T)/2, 0) == T(ldexp(big(floatmin(T)/2), 0))
            @test ldexp(floatmin(T)/3, 0) == T(ldexp(big(floatmin(T)/3), 0))
            @test ldexp(floatmin(T)/3, -1) == T(ldexp(big(floatmin(T)/3), -1))
            @test ldexp(floatmin(T)/3, 11) == T(ldexp(big(floatmin(T)/3), 11))
            @test ldexp(floatmin(T)/11, -10) == T(ldexp(big(floatmin(T)/11), -10))
            @test ldexp(-floatmin(T)/11, -10) == T(ldexp(big(-floatmin(T)/11), -10))
        end
    end
end

# We compare to BigFloat instead of hard-coding
# values, assuming that BigFloat has an independently tested implementation.
@testset "basic math functions" begin
    @testset "$T" for T in (Float32, Float64)
        x = T(1//3)
        y = T(1//2)
        yi = 4
        @testset "Random values" begin
            @test x^y ≈ big(x)^big(y)
            @test x^1 === x
            @test x^yi ≈ big(x)^yi
            @test acos(x) ≈ acos(big(x))
            @test acosh(1+x) ≈ acosh(big(1+x))
            @test asin(x) ≈ asin(big(x))
            @test asinh(x) ≈ asinh(big(x))
            @test atan(x) ≈ atan(big(x))
            @test atan(x,y) ≈ atan(big(x),big(y))
            @test atanh(x) ≈ atanh(big(x))
            @test cbrt(x) ≈ cbrt(big(x))
            @test cos(x) ≈ cos(big(x))
            @test cosh(x) ≈ cosh(big(x))
            @test exp(x) ≈ exp(big(x))
            @test exp10(x) ≈ exp10(big(x))
            @test exp2(x) ≈ exp2(big(x))
            @test expm1(x) ≈ expm1(big(x))
            @test hypot(x,y) ≈ hypot(big(x),big(y))
            @test hypot(x,x,y) ≈ hypot(hypot(big(x),big(x)),big(y))
            @test hypot(x,x,y,y) ≈ hypot(hypot(big(x),big(x)),hypot(big(y),big(y)))
            @test log(x) ≈ log(big(x))
            @test log10(x) ≈ log10(big(x))
            @test log1p(x) ≈ log1p(big(x))
            @test log2(x) ≈ log2(big(x))
            @test sin(x) ≈ sin(big(x))
            @test sinh(x) ≈ sinh(big(x))
            @test sqrt(x) ≈ sqrt(big(x))
            @test tan(x) ≈ tan(big(x))
            @test tanh(x) ≈ tanh(big(x))
            @test sec(x) ≈ sec(big(x))
            @test csc(x) ≈ csc(big(x))
            @test secd(x) ≈ secd(big(x))
            @test cscd(x) ≈ cscd(big(x))
            @test sech(x) ≈ sech(big(x))
            @test csch(x) ≈ csch(big(x))
        end
        @testset "Special values" begin
            @test isequal(T(1//4)^T(1//2), T(1//2))
            @test isequal(T(1//4)^2, T(1//16))
            @test isequal(acos(T(1)), T(0))
            @test isequal(acosh(T(1)), T(0))
            @test asin(T(1)) ≈ T(pi)/2 atol=eps(T)
            @test atan(T(1)) ≈ T(pi)/4 atol=eps(T)
            @test atan(T(1),T(1)) ≈ T(pi)/4 atol=eps(T)
            @test isequal(cbrt(T(0)), T(0))
            @test isequal(cbrt(T(1)), T(1))
            @test isequal(cbrt(T(1000000000)), T(1000))
            @test isequal(cos(T(0)), T(1))
            @test cos(T(pi)/2) ≈ T(0) atol=eps(T)
            @test isequal(cos(T(pi)), T(-1))
            @test exp(T(1)) ≈ T(ℯ) atol=2*eps(T)
            @test isequal(exp10(T(1)), T(10))
            @test isequal(exp2(T(1)), T(2))
            @test isequal(expm1(T(0)), T(0))
            @test isequal(expm1(-floatmax(T)), -one(T))
            @test isequal(expm1(floatmax(T)), T(Inf))
            @test expm1(T(1)) ≈ T(ℯ)-1 atol=2*eps(T)
            @test isequal(hypot(T(3),T(4)), T(5))
            @test isequal(hypot(floatmax(T),T(1)),floatmax(T))
            @test isequal(hypot(floatmin(T)*sqrt(eps(T)),T(0)),floatmin(T)*sqrt(eps(T)))
            @test isequal(floatmin(T)*hypot(1.368423059742933,1.3510496552495361),hypot(floatmin(T)*1.368423059742933,floatmin(T)*1.3510496552495361))
            @test isequal(log(T(1)), T(0))
            @test isequal(log(ℯ,T(1)), T(0))
            @test log(T(ℯ)) ≈ T(1) atol=eps(T)
            @test isequal(log10(T(1)), T(0))
            @test isequal(log10(T(10)), T(1))
            @test isequal(log1p(T(0)), T(0))
            @test log1p(T(ℯ)-1) ≈ T(1) atol=eps(T)
            @test isequal(log2(T(1)), T(0))
            @test isequal(log2(T(2)), T(1))
            @test isequal(sin(T(0)), T(0))
            @test isequal(sin(T(pi)/2), T(1))
            @test sin(T(pi)) ≈ T(0) atol=eps(T)
            @test isequal(sqrt(T(0)), T(0))
            @test isequal(sqrt(T(1)), T(1))
            @test isequal(sqrt(T(100000000)), T(10000))
            @test isequal(tan(T(0)), T(0))
            @test tan(T(pi)/4) ≈ T(1) atol=eps(T)
            @test isequal(sec(T(pi)), -one(T))
            @test isequal(csc(T(pi)/2), one(T))
            @test isequal(secd(T(180)), -one(T))
            @test isequal(cscd(T(90)), one(T))
            @test isequal(sech(log(one(T))), one(T))
            @test isequal(csch(zero(T)), T(Inf))
        end
        @testset "Inverses" begin
            @test acos(cos(x)) ≈ x
            @test acosh(cosh(x)) ≈ x
            @test asin(sin(x)) ≈ x
            @test cbrt(x)^3 ≈ x
            @test cbrt(x^3) ≈ x
            @test asinh(sinh(x)) ≈ x
            @test atan(tan(x)) ≈ x
            @test atan(x,y) ≈ atan(x/y)
            @test atanh(tanh(x)) ≈ x
            @test cos(acos(x)) ≈ x
            @test cosh(acosh(1+x)) ≈ 1+x
            @test exp(log(x)) ≈ x
            @test exp10(log10(x)) ≈ x
            @test exp2(log2(x)) ≈ x
            @test expm1(log1p(x)) ≈ x
            @test log(exp(x)) ≈ x
            @test log10(exp10(x)) ≈ x
            @test log1p(expm1(x)) ≈ x
            @test log2(exp2(x)) ≈ x
            @test sin(asin(x)) ≈ x
            @test sinh(asinh(x)) ≈ x
            @test sqrt(x)^2 ≈ x
            @test sqrt(x^2) ≈ x
            @test tan(atan(x)) ≈ x
            @test tanh(atanh(x)) ≈ x
        end
        @testset "Relations between functions" begin
            @test cosh(x) ≈ (exp(x)+exp(-x))/2
            @test cosh(x)^2-sinh(x)^2 ≈ 1
            @test hypot(x,y) ≈ sqrt(x^2+y^2)
            @test sin(x)^2+cos(x)^2 ≈ 1
            @test sinh(x) ≈ (exp(x)-exp(-x))/2
            @test tan(x) ≈ sin(x)/cos(x)
            @test tanh(x) ≈ sinh(x)/cosh(x)
            @test sec(x) ≈ inv(cos(x))
            @test csc(x) ≈ inv(sin(x))
            @test secd(x) ≈ inv(cosd(x))
            @test cscd(x) ≈ inv(sind(x))
            @test sech(x) ≈ inv(cosh(x))
            @test csch(x) ≈ inv(sinh(x))
        end
        @testset "Edge cases" begin
            @test isinf(log(zero(T)))
            @test isnan_type(T, log(convert(T,NaN)))
            @test_throws DomainError log(-one(T))
            @test isinf(log1p(-one(T)))
            @test isnan_type(T, log1p(convert(T,NaN)))
            @test_throws DomainError log1p(convert(T,-2.0))
            @test hypot(T(0), T(0)) === T(0)
            @test hypot(T(Inf), T(Inf)) === T(Inf)
            @test hypot(T(Inf), T(x)) === T(Inf)
            @test hypot(T(Inf), T(NaN)) === T(Inf)
            @test isnan_type(T, hypot(T(x), T(NaN)))
            @test tanh(T(Inf)) === T(1)
        end
    end
    @testset "Float16 expm1" begin
        T=Float16
        @test isequal(expm1(T(0)), T(0))
        @test isequal(expm1(-floatmax(T)), -one(T))
        @test isequal(expm1(floatmax(T)), T(Inf))
        @test expm1(T(1)) ≈ T(ℯ)-1 atol=2*eps(T)
    end
end

@testset "exp function" for T in (Float64, Float32)
    @testset "$T accuracy" begin
        X = map(T, vcat(-10:0.0002:10, -80:0.001:80, 2.0^-27, 2.0^-28, 2.0^-14, 2.0^-13))
        for x in X
            y, yb = exp(x), exp(big(x))
            @test abs(y-yb) <= 1.0*eps(T(yb))
        end
    end
    @testset "$T edge cases" begin
        @test isnan_type(T, exp(T(NaN)))
        @test exp(T(-Inf)) === T(0.0)
        @test exp(T(Inf)) === T(Inf)
        @test exp(T(0.0)) === T(1.0) # exact
        @test exp(T(5000.0)) === T(Inf)
        @test exp(T(-5000.0)) === T(0.0)
    end
end

@testset "exp10 function" begin
    @testset "accuracy" begin
        X = map(Float64, vcat(-10:0.00021:10, -35:0.0023:100, -300:0.001:300))
        for x in X
            y, yb = exp10(x), exp10(big(x))
            @test abs(y-yb) <= 1.2*eps(Float64(yb))
        end
        X = map(Float32, vcat(-10:0.00021:10, -35:0.0023:35, -35:0.001:35))
        for x in X
            y, yb = exp10(x), exp10(big(x))
            @test abs(y-yb) <= 1.2*eps(Float32(yb))
        end
    end
    @testset "$T edge cases" for T in (Float64, Float32)
        @test isnan_type(T, exp10(T(NaN)))
        @test exp10(T(-Inf)) === T(0.0)
        @test exp10(T(Inf)) === T(Inf)
        @test exp10(T(0.0)) === T(1.0) # exact
        @test exp10(T(1.0)) === T(10.0)
        @test exp10(T(3.0)) === T(1000.0)
        @test exp10(T(5000.0)) === T(Inf)
        @test exp10(T(-5000.0)) === T(0.0)
    end
end

@testset "test abstractarray trig functions" begin
    TAA = rand(2,2)
    TAA = (TAA + TAA')/2.
    STAA = Symmetric(TAA)
    @test Array(atanh.(STAA)) == atanh.(TAA)
    @test Array(asinh.(STAA)) == asinh.(TAA)
    TAA .+= 1
    @test Array(acosh.(STAA)) == acosh.(TAA)
    @test Array(acsch.(STAA)) == acsch.(TAA)
    @test Array(acoth.(STAA)) == acoth.(TAA)
    @test sind(TAA) == sin(deg2rad.(TAA))
    @test cosd(TAA) == cos(deg2rad.(TAA))
    @test tand(TAA) == tan(deg2rad.(TAA))
    @test asind(TAA) == rad2deg.(asin(TAA))
    @test acosd(TAA) == rad2deg.(acos(TAA))
    @test atand(TAA) == rad2deg.(atan(TAA))
    @test asecd(TAA) == rad2deg.(asec(TAA))
    @test acscd(TAA) == rad2deg.(acsc(TAA))
    @test acotd(TAA) == rad2deg.(acot(TAA))

    m = rand(3,2) # not square matrix
    ex = @test_throws DimensionMismatch sind(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch cosd(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch tand(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch asind(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch acosd(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch atand(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch asecd(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch acscd(m)
    @test startswith(ex.value.msg, "matrix is not square")
    ex = @test_throws DimensionMismatch acotd(m)
    @test startswith(ex.value.msg, "matrix is not square")
end

@testset "check exp2(::Integer) matches exp2(::Float)" begin
    for ii in -2048:2048
        expected = exp2(float(ii))
        @test exp2(Int16(ii)) == expected
        @test exp2(Int32(ii)) == expected
        @test exp2(Int64(ii)) == expected
        @test exp2(Int128(ii)) == expected
        if ii >= 0
            @test exp2(UInt16(ii)) == expected
            @test exp2(UInt32(ii)) == expected
            @test exp2(UInt64(ii)) == expected
            @test exp2(UInt128(ii)) == expected
        end
    end
end

@testset "deg2rad/rad2deg" begin
    @testset "$T" for T in (Int, Float64, BigFloat)
        @test deg2rad(T(180)) ≈ 1pi
        @test deg2rad.(T[45, 60]) ≈ [pi/T(4), pi/T(3)]
        @test rad2deg.([pi/T(4), pi/T(3)]) ≈ [45, 60]
        @test rad2deg(T(1)*pi) ≈ 180
        @test rad2deg(T(1)) ≈ rad2deg(true)
        @test deg2rad(T(1)) ≈ deg2rad(true)
    end
    @test deg2rad(180 + 60im) ≈ pi + (pi/3)*im
    @test rad2deg(pi + (pi/3)*im) ≈ 180 + 60im
end

@testset "degree-based trig functions" begin
    @testset "$T" for T = (Float32,Float64,Rational{Int})
        fT = typeof(float(one(T)))
        fTsc = typeof( (float(one(T)), float(one(T))) )
        for x = -400:40:400
            @test sind(convert(T,x))::fT ≈ convert(fT,sin(pi/180*x)) atol=eps(deg2rad(convert(fT,x)))
            @test cosd(convert(T,x))::fT ≈ convert(fT,cos(pi/180*x)) atol=eps(deg2rad(convert(fT,x)))

            s,c = sincosd(convert(T,x))
            @test s::fT ≈ convert(fT,sin(pi/180*x)) atol=eps(deg2rad(convert(fT,x)))
            @test c::fT ≈ convert(fT,cos(pi/180*x)) atol=eps(deg2rad(convert(fT,x)))
        end
        @testset "sind" begin
            @test sind(convert(T,0.0))::fT === zero(fT)
            @test sind(convert(T,180.0))::fT === zero(fT)
            @test sind(convert(T,360.0))::fT === zero(fT)
            T != Rational{Int} && @test sind(convert(T,-0.0))::fT === -zero(fT)
            @test sind(convert(T,-180.0))::fT === -zero(fT)
            @test sind(convert(T,-360.0))::fT === -zero(fT)
            if T <: AbstractFloat
                @test isnan(sind(T(NaN)))
            end
        end
        @testset "cosd" begin
            @test cosd(convert(T,90))::fT === zero(fT)
            @test cosd(convert(T,270))::fT === zero(fT)
            @test cosd(convert(T,-90))::fT === zero(fT)
            @test cosd(convert(T,-270))::fT === zero(fT)
            if T <: AbstractFloat
                @test isnan(cosd(T(NaN)))
            end
        end
        @testset "sincosd" begin
            @test sincosd(convert(T,-360))::fTsc === ( -zero(fT),  one(fT) )
            @test sincosd(convert(T,-270))::fTsc === (   one(fT), zero(fT) )
            @test sincosd(convert(T,-180))::fTsc === ( -zero(fT), -one(fT) )
            @test sincosd(convert(T, -90))::fTsc === (  -one(fT), zero(fT) )
            @test sincosd(convert(T,   0))::fTsc === (  zero(fT),  one(fT) )
            @test sincosd(convert(T,  90))::fTsc === (   one(fT), zero(fT) )
            @test sincosd(convert(T, 180))::fTsc === (  zero(fT), -one(fT) )
            @test sincosd(convert(T, 270))::fTsc === (  -one(fT), zero(fT) )
            if T <: AbstractFloat
                @test_throws DomainError sincosd(T(Inf))
                @test all(isnan.(sincosd(T(NaN))))
            end
        end

        @testset "$name" for (name, (sinpi, cospi)) in (
            "sinpi and cospi" => (sinpi, cospi),
            "sincospi" => (x->sincospi(x)[1], x->sincospi(x)[2])
        )
            @testset "pi * $x" for x = -3:0.3:3
                @test sinpi(convert(T,x))::fT ≈ convert(fT,sin(pi*x)) atol=eps(pi*convert(fT,x))
                @test cospi(convert(T,x))::fT ≈ convert(fT,cos(pi*x)) atol=eps(pi*convert(fT,x))
            end

            @test sinpi(convert(T,0.0))::fT === zero(fT)
            @test sinpi(convert(T,1.0))::fT === zero(fT)
            @test sinpi(convert(T,2.0))::fT === zero(fT)
            T != Rational{Int} && @test sinpi(convert(T,-0.0))::fT === -zero(fT)
            @test sinpi(convert(T,-1.0))::fT === -zero(fT)
            @test sinpi(convert(T,-2.0))::fT === -zero(fT)
            @test_throws DomainError sinpi(convert(T,Inf))

            @test cospi(convert(T,0.5))::fT === zero(fT)
            @test cospi(convert(T,1.5))::fT === zero(fT)
            @test cospi(convert(T,-0.5))::fT === zero(fT)
            @test cospi(convert(T,-1.5))::fT === zero(fT)
            @test_throws DomainError cospi(convert(T,Inf))
        end
        @testset "Check exact values" begin
            @test sind(convert(T,30)) == 0.5
            @test cosd(convert(T,60)) == 0.5
            @test sind(convert(T,150)) == 0.5
            @test sinpi(one(T)/convert(T,6)) == 0.5
            @test sincospi(one(T)/convert(T,6))[1] == 0.5
            @test_throws DomainError sind(convert(T,Inf))
            @test_throws DomainError cosd(convert(T,Inf))
            T != Float32 && @test cospi(one(T)/convert(T,3)) == 0.5
            T != Float32 && @test sincospi(one(T)/convert(T,3))[2] == 0.5
            T == Rational{Int} && @test sinpi(5//6) == 0.5
            T == Rational{Int} && @test sincospi(5//6)[1] == 0.5
        end
    end
    scdm = sincosd(missing)
    @test ismissing(scdm[1])
    @test ismissing(scdm[2])
end

@testset "Integer and Inf args for sinpi/cospi/sinc/cosc" begin
    for (sinpi, cospi) in ((sinpi, cospi), (x->sincospi(x)[1], x->sincospi(x)[2]))
        @test sinpi(1) == 0
        @test sinpi(-1) == -0
        @test cospi(1) == -1
        @test cospi(2) == 1
    end

    @test sinc(1) == 0
    @test sinc(complex(1,0)) == 0
    @test sinc(0) == 1
    @test sinc(Inf) == 0
    @test cosc(1) == -1
    @test cosc(0) == 0
    @test cosc(complex(1,0)) == -1
    @test cosc(Inf) == 0

    @test sinc(Inf + 3im) == 0
    @test cosc(Inf + 3im) == 0

    @test isequal(sinc(Inf + Inf*im), NaN + NaN*im)
    @test isequal(cosc(Inf + Inf*im), NaN + NaN*im)
end

# issue #37227
@testset "sinc/cosc accuracy" begin
    setprecision(256) do
        for R in (BigFloat, Float16, Float32, Float64)
            for T in (R, Complex{R})
                for x in (0, 1e-5, 1e-20, 1e-30, 1e-40, 1e-50, 1e-60, 1e-70, 5.07138898934e-313)
                    if x < eps(R)
                        @test sinc(T(x)) == 1
                    end
                    @test cosc(T(x)) ≈ pi*(-R(x)*pi)/3 rtol=max(eps(R)*100, (pi*R(x))^2)
                end
            end
        end
    end
    @test @inferred(sinc(0//1)) === 1.0
    @test @inferred(cosc(0//1)) === -0.0

    # test right before/after thresholds of Taylor series
    @test sinc(0.001) ≈ 0.999998355066745 rtol=1e-15
    @test sinc(0.00099) ≈ 0.9999983878009009 rtol=1e-15
    @test sinc(0.05f0) ≈ 0.9958927352435614 rtol=1e-7
    @test sinc(0.0499f0) ≈ 0.9959091277049384 rtol=1e-7
    @test cosc(0.14) ≈ -0.4517331883801308 rtol=1e-15
    @test cosc(0.1399) ≈ -0.45142306168781854 rtol=1e-14
    @test cosc(0.26f0) ≈ -0.7996401373462212 rtol=5e-7
    @test cosc(0.2599f0) ≈ -0.7993744054401625 rtol=5e-7
    setprecision(256) do
        @test cosc(big"0.5") ≈ big"-1.273239544735162686151070106980114896275677165923651589981338752471174381073817" rtol=1e-76
        @test cosc(big"0.499") ≈ big"-1.272045747741181369948389133250213864178198918667041860771078493955590574971317" rtol=1e-76
    end
end

@testset "Irrational args to sinpi/cospi/sinc/cosc" begin
    for x in (pi, ℯ, Base.MathConstants.golden)
        for (sinpi, cospi) in ((sinpi, cospi), (x->sincospi(x)[1], x->sincospi(x)[2]))
            @test sinpi(x) ≈ Float64(sinpi(big(x)))
            @test cospi(x) ≈ Float64(cospi(big(x)))
            @test sinpi(complex(x, x)) ≈ ComplexF64(sinpi(complex(big(x), big(x))))
            @test cospi(complex(x, x)) ≈ ComplexF64(cospi(complex(big(x), big(x))))
        end
        @test sinc(x)  ≈ Float64(sinc(big(x)))
        @test cosc(x)  ≈ Float64(cosc(big(x)))
        @test sinc(complex(x, x))  ≈ ComplexF64(sinc(complex(big(x),  big(x))))
        @test cosc(complex(x, x))  ≈ ComplexF64(cosc(complex(big(x),  big(x))))
    end
end

@testset "half-integer and nan/infs for sincospi,sinpi,cospi" begin
    @testset for T in (ComplexF32, ComplexF64)
        @test sincospi(T(0.5, 0.0)) == (T(1.0,0.0), T(0.0, -0.0))
        @test sincospi(T(1.5, 0.0)) == (T(-1.0,0.0), T(0.0, 0.0))
        @test sinpi(T(1.5, 1.5)) ≈ T(-cosh(3*π/2), 0.0)
        @test cospi(T(0.5, 0.5)) ≈ T(0.0, -sinh(π/2))
        s, c = sincospi(T(Inf64, 0.0))
        @test isnan(real(s)) && imag(s) == zero(real(T))
        @test isnan(real(c)) && imag(c) == -zero(real(T))
        s, c = sincospi(T(NaN, 0.0))
        @test isnan(real(s)) && imag(s) == zero(real(T))
        @test isnan(real(c)) && imag(c) == zero(real(T))
        s, c = sincospi(T(NaN, Inf64))
        @test isnan(real(s)) && isinf(imag(s))
        @test isinf(real(c)) && isnan(imag(c))
        s, c = sincospi(T(NaN, 2))
        @test isnan(real(s)) && isnan(imag(s))
        @test isnan(real(c)) && isnan(imag(c))
    end
end

@testset "trig function type stability" begin
    @testset "$T $f" for T = (Float32,Float64,BigFloat,Rational{Int16},Complex{Int32},ComplexF16), f = (sind,cosd,sinpi,cospi)
        @test Base.return_types(f,Tuple{T}) == [float(T)]
    end
    @testset "$T sincospi" for T = (Float32,Float64,BigFloat,Rational{Int16},Complex{Int32},ComplexF16)
        @test Base.return_types(sincospi,Tuple{T}) == [Tuple{float(T),float(T)}]
    end
end

# useful test functions for relative error, which differ from isapprox (≈)
# in that relerrc separately looks at the real and imaginary parts
relerr(z, x) = z == x ? 0.0 : abs(z - x) / abs(x)
relerrc(z, x) = max(relerr(real(z),real(x)), relerr(imag(z),imag(x)))
≅(a,b) = relerrc(a,b) ≤ 1e-13

@testset "subnormal flags" begin
    # Ensure subnormal flags functions don't segfault
    @test any(set_zero_subnormals(true) .== [false,true])
    @test any(get_zero_subnormals() .== [false,true])
    @test set_zero_subnormals(false)
    @test !get_zero_subnormals()
end

@testset "evalpoly" begin
    @test @evalpoly(2,3,4,5,6) == 3+2*(4+2*(5+2*6)) == @evalpoly(2+0im,3,4,5,6)
    a0 = 1
    a1 = 2
    c = 3
    @test @evalpoly(c, a0, a1) == 7
    @test @evalpoly(1, 2) == 2
end

@testset "evalpoly real" begin
    for x in -1.0:2.0, p1 in -3.0:3.0, p2 in -3.0:3.0, p3 in -3.0:3.0
        evpm = @evalpoly(x, p1, p2, p3)
        @test evalpoly(x, (p1, p2, p3)) == evpm
        @test evalpoly(x, [p1, p2, p3]) == evpm
    end
end

@testset "evalpoly complex" begin
    for x in -1.0:2.0, y in -1.0:2.0, p1 in -3.0:3.0, p2 in -3.0:3.0, p3 in -3.0:3.0
        z = x + im * y
        evpm = @evalpoly(z, p1, p2, p3)
        @test evalpoly(z, (p1, p2, p3)) == evpm
        @test evalpoly(z, [p1, p2, p3]) == evpm
    end
    @test evalpoly(1+im, (2,)) == 2
    @test evalpoly(1+im, [2,]) == 2
end

@testset "cis" begin
    for z in (1.234, 1.234 + 5.678im)
        @test cis(z) ≈ exp(im*z)
    end
    let z = [1.234, 5.678]
        @test cis.(z) ≈ exp.(im*z)
    end
end

@testset "modf" begin
    @testset "$elty" for elty in (Float16, Float32, Float64)
        @test modf( convert(elty,1.2) )[1] ≈ convert(elty,0.2)
        @test modf( convert(elty,1.2) )[2] ≈ convert(elty,1.0)
        @test modf( convert(elty,1.0) )[1] ≈ convert(elty,0.0)
        @test modf( convert(elty,1.0) )[2] ≈ convert(elty,1.0)
        @test isequal(modf( convert(elty,-Inf) ), (-0.0, -Inf))
        @test isequal(modf( convert(elty,Inf) ), (0.0, Inf))
        @test isequal(modf( convert(elty,NaN) ), (NaN, NaN))
    end
end

@testset "frexp" begin
    @testset "$elty" for elty in (Float16, Float32, Float64)
        @test frexp( convert(elty,0.5) ) == (0.5, 0)
        @test frexp( convert(elty,4.0) ) == (0.5, 3)
        @test frexp( convert(elty,10.5) ) == (0.65625, 4)
    end
end

@testset "log/log1p" begin
    # using Tang's algorithm, should be accurate to within 0.56 ulps
    X = rand(100)
    for x in X
        for n = -5:5
            xn = ldexp(x,n)

            for T in (Float32,Float64)
                xt = T(x)

                y = log(xt)
                yb = log(big(xt))
                @test abs(y-yb) <= 0.56*eps(T(yb))

                y = log1p(xt)
                yb = log1p(big(xt))
                @test abs(y-yb) <= 0.56*eps(T(yb))

                if n <= 0
                    y = log1p(-xt)
                    yb = log1p(big(-xt))
                    @test abs(y-yb) <= 0.56*eps(T(yb))
                end
            end
        end
    end

    for n = 0:28
        @test log(2,2^n) == n
    end
    setprecision(10_000) do
        @test log(2,big(2)^100) == 100
        @test log(2,big(2)^200) == 200
        @test log(2,big(2)^300) == 300
        @test log(2,big(2)^400) == 400
    end

    for T in (Float32,Float64)
        @test log(zero(T)) == -Inf
        @test isnan_type(T, log(T(NaN)))
        @test_throws DomainError log(-one(T))
        @test log1p(-one(T)) == -Inf
        @test isnan_type(T, log1p(T(NaN)))
        @test_throws DomainError log1p(-2*one(T))
    end
    @testset "log of subnormals" begin
        # checked results with WolframAlpha
        for (T, lr) in ((Float32, LinRange(2.f0^(-129), 2.f0^(-128), 1000)),
                        (Float64, LinRange(2.0^(-1025), 2.0^(-1024), 1000)))
            for x in lr
                @test log(x)   ≈ T(log(widen(x))) rtol=2eps(T)
                @test log2(x)  ≈ T(log2(widen(x))) rtol=2eps(T)
                @test log10(x) ≈ T(log10(widen(x))) rtol=2eps(T)
            end
        end
    end
end

@testset "vectorization of 2-arg functions" begin
    binary_math_functions = [
        copysign, flipsign, log, atan, hypot, max, min,
    ]
    @testset "$f" for f in binary_math_functions
        x = y = 2
        v = [f(x,y)]
        @test f.([x],y) == v
        @test f.(x,[y]) == v
        @test f.([x],[y]) == v
    end
end

@testset "issues #3024, #12822, #24240" begin
    p2 = -2
    p3 = -3
    @test_throws DomainError 2 ^ p2
    @test 2 ^ -2 == 0.25 == (2^-1)^2
    @test_throws DomainError (-2)^(2.2)
    @test_throws DomainError (-2.0)^(2.2)
    @test_throws DomainError false ^ p2
    @test false ^ -2 == Inf
    @test 1 ^ -2 === (-1) ^ -2 == 1 ^ p2 === (-1) ^ p2 === 1
    @test (-1) ^ -1 === (-1) ^ -3 == (-1) ^ p3 === -1
    @test true ^ -2 == true ^ p2 === true
end

@testset "issue #13748" begin
    let A = [1 2; 3 4]; B = [5 6; 7 8]; C = [9 10; 11 12]
        @test muladd(A,B,C) == A*B + C
    end
end

@testset "issue #19872" begin
    f19872a(x) = x ^ 5
    f19872b(x) = x ^ (-1024)
    @test 0 < f19872b(2.0) < 1e-300
    @test issubnormal(2.0 ^ (-1024))
    @test issubnormal(f19872b(2.0))
    @test !issubnormal(f19872b(0.0))
    @test f19872a(2.0) === 32.0
    @test !issubnormal(f19872a(2.0))
    @test !issubnormal(0.0)
end

# no domain error is thrown for negative values
@test invoke(cbrt, Tuple{AbstractFloat}, -1.0) == -1.0

@testset "promote Float16 irrational #15359" begin
    @test typeof(Float16(.5) * pi) == Float16
end

@testset "sincos" begin
    @test sincos(1.0) === (sin(1.0), cos(1.0))
    @test sincos(1f0) === (sin(1f0), cos(1f0))
    @test sincos(Float16(1)) === (sin(Float16(1)), cos(Float16(1)))
    @test sincos(1) === (sin(1), cos(1))
    @test sincos(big(1)) == (sin(big(1)), cos(big(1)))
    @test sincos(big(1.0)) == (sin(big(1.0)), cos(big(1.0)))
    @test sincos(NaN) === (NaN, NaN)
    @test sincos(NaN32) === (NaN32, NaN32)
    @test_throws DomainError sincos(Inf32)
    @test_throws DomainError sincos(Inf64)
end

@testset "test fallback definitions" begin
    @test exp10(5) ≈ exp10(5.0)
    @test exp10(50//10) ≈ exp10(5.0)
    @test log10(exp10(ℯ)) ≈ ℯ
    @test log(ℯ) === 1
    @test exp2(Float16(2.0)) ≈ exp2(2.0)
    @test exp2(Float16(1.0)) === Float16(exp2(1.0))
    @test exp10(Float16(1.0)) === Float16(exp10(1.0))
end

@testset "isapprox" begin
  # #22742: updated isapprox semantics
  @test !isapprox(1.0, 1.0+1e-12, atol=1e-14)
  @test isapprox(1.0, 1.0+0.5*sqrt(eps(1.0)))
  @test !isapprox(1.0, 1.0+1.5*sqrt(eps(1.0)), atol=sqrt(eps(1.0)))

  # #13132: Use of `norm` kwarg for scalar arguments
  @test isapprox(1, 1+1.0e-12, norm=abs)
  @test !isapprox(1, 1+1.0e-12, norm=x->1)
end

# test AbstractFloat fallback pr22716
struct Float22716{T<:AbstractFloat} <: AbstractFloat
    x::T
end
Base.:^(x::Number, y::Float22716) = x^(y.x)
let x = 2.0
    @test exp2(Float22716(x)) === 2^x
    @test exp10(Float22716(x)) === 10^x
end

@testset "asin #23088" begin
    for T in (Float32, Float64)
        @test asin(zero(T)) === zero(T)
        @test asin(-zero(T)) === -zero(T)
        @test asin(nextfloat(zero(T))) === nextfloat(zero(T))
        @test asin(prevfloat(zero(T))) === prevfloat(zero(T))
        @test asin(one(T)) === T(pi)/2
        @test asin(-one(T)) === -T(pi)/2
        for x in (0.45, 0.6, 0.98)
            by = asin(big(T(x)))
            @test T(abs(asin(T(x)) - by))/eps(T(abs(by))) <= 1
            bym = asin(big(T(-x)))
            @test T(abs(asin(T(-x)) - bym))/eps(T(abs(bym))) <= 1
        end
        @test_throws DomainError asin(-T(Inf))
        @test_throws DomainError asin(T(Inf))
        @test isnan_type(T, asin(T(NaN)))
    end
end

@testset "sin, cos, sincos, tan #23088" begin
    for T in (Float32, Float64)
        @test sin(zero(T)) === zero(T)
        @test sin(-zero(T)) === -zero(T)
        @test cos(zero(T)) === T(1.0)
        @test cos(-zero(T)) === T(1.0)
        @test sin(nextfloat(zero(T))) === nextfloat(zero(T))
        @test sin(prevfloat(zero(T))) === prevfloat(zero(T))
        @test cos(nextfloat(zero(T))) === T(1.0)
        @test cos(prevfloat(zero(T))) === T(1.0)
        for x in (0.1, 0.45, 0.6, 0.75, 0.79, 0.98)
            for op in (sin, cos, tan)
                by = T(op(big(x)))
                @test abs(op(T(x)) - by)/eps(by) <= one(T)
                bym = T(op(big(-x)))
                @test abs(op(T(-x)) - bym)/eps(bym) <= one(T)
            end
        end
        @test_throws DomainError sin(-T(Inf))
        @test_throws DomainError sin(T(Inf))
        @test_throws DomainError cos(-T(Inf))
        @test_throws DomainError cos(T(Inf))
        @test_throws DomainError tan(-T(Inf))
        @test_throws DomainError tan(T(Inf))
        @test sin(T(NaN)) === T(NaN)
        @test cos(T(NaN)) === T(NaN)
        @test tan(T(NaN)) === T(NaN)
    end
end

@testset "rem_pio2 #23088" begin
    vals = (2.356194490192345f0, 3.9269908169872414f0, 7.0685834705770345f0,
              5.497787143782138f0, 4.216574282663131f8, 4.216574282663131f12)
    for (i, x) in enumerate(vals)
        for op in (prevfloat, nextfloat)
            Ty = Float32(Base.Math.rem_pio2_kernel(op(vals[i]))[2].hi)
            By = Float32(rem(big(op(x)), pi/2))
            @test Ty ≈ By || Ty ≈ By-Float32(pi)/2
        end
    end
end

@testset "atan #23383" begin
    for T in (Float32, Float64)
        @test atan(T(NaN)) === T(NaN)
        @test atan(-T(Inf)) === -T(pi)/2
        @test atan(T(Inf)) === T(pi)/2
        # no reduction needed |x| < 7/16
        @test atan(zero(T)) === zero(T)
        @test atan(prevfloat(zero(T))) === prevfloat(zero(T))
        @test atan(nextfloat(zero(T))) === nextfloat(zero(T))
        for x in (T(7/16), (T(7/16)+T(11/16))/2, T(11/16),
                  (T(11/16)+T(19/16))/2, T(19/16),
                  (T(19/16)+T(39/16))/2, T(39/16),
                  (T(39/16)+T(2)^23)/2, T(2)^23)
            x = T(7/16)
            by = T(atan(big(x)))
            @test abs(atan(x) - by)/eps(by) <= one(T)
            x = prevfloat(T(7/16))
            by = T(atan(big(x)))
            @test abs(atan(x) - by)/eps(by) <= one(T)
            x = nextfloat(T(7/16))
            by = T(atan(big(x)))
            @test abs(atan(x) - by)/eps(by) <= one(T)
        end
        # This case was used to find a bug, but it isn't special in itself
        @test atan(1.7581305072934137) ≈ 1.053644580517088
    end
end
@testset "atan" begin
    for T in (Float32, Float64)
        @test isnan_type(T, atan(T(NaN), T(NaN)))
        @test isnan_type(T, atan(T(NaN), T(0.1)))
        @test isnan_type(T, atan(T(0.1), T(NaN)))
        r = T(randn())
        absr = abs(r)
        # y zero
        @test atan(T(r), one(T)) === atan(T(r))
        @test atan(zero(T), absr) === zero(T)
        @test atan(-zero(T), absr) === -zero(T)
        @test atan(zero(T), -absr) === T(pi)
        @test atan(-zero(T), -absr) === -T(pi)
        # x zero and y not zero
        @test atan(one(T), zero(T)) === T(pi)/2
        @test atan(-one(T), zero(T)) === -T(pi)/2
        # isinf(x) == true && isinf(y) == true
        @test atan(T(Inf), T(Inf)) === T(pi)/4 # m == 0 (see atan code)
        @test atan(-T(Inf), T(Inf)) === -T(pi)/4 # m == 1
        @test atan(T(Inf), -T(Inf)) === 3*T(pi)/4 # m == 2
        @test atan(-T(Inf), -T(Inf)) === -3*T(pi)/4 # m == 3
        # isinf(x) == true && isinf(y) == false
        @test atan(absr, T(Inf)) === zero(T) # m == 0
        @test atan(-absr, T(Inf)) === -zero(T) # m == 1
        @test atan(absr, -T(Inf)) === T(pi) # m == 2
        @test atan(-absr, -T(Inf)) === -T(pi) # m == 3
        # isinf(y) == true && isinf(x) == false
        @test atan(T(Inf), absr) === T(pi)/2
        @test atan(-T(Inf), absr) === -T(pi)/2
        @test atan(T(Inf), -absr) === T(pi)/2
        @test atan(-T(Inf), -absr) === -T(pi)/2
        # |y/x| above high threshold
        atanpi = T(1.5707963267948966)
        @test atan(T(2.0^61), T(1.0)) === atanpi # m==0
        @test atan(-T(2.0^61), T(1.0)) === -atanpi # m==1
        @test atan(T(2.0^61), -T(1.0)) === atanpi # m==2
        @test atan(-T(2.0^61), -T(1.0)) === -atanpi # m==3
        @test atan(-T(Inf), -absr) === -T(pi)/2
        # |y|/x between 0 and low threshold
        @test atan(T(2.0^-61), -T(1.0)) === T(pi) # m==2
        @test atan(-T(2.0^-61), -T(1.0)) === -T(pi) # m==3
        # y/x is "safe" ("arbitrary values", just need to hit the branch)
        _ATAN_PI_LO(::Type{Float32}) = -8.7422776573f-08
        _ATAN_PI_LO(::Type{Float64}) = 1.2246467991473531772E-16
        @test atan(T(5.0), T(2.5)) === atan(abs(T(5.0)/T(2.5)))
        @test atan(-T(5.0), T(2.5)) === -atan(abs(-T(5.0)/T(2.5)))
        @test atan(T(5.0), -T(2.5)) === T(pi)-(atan(abs(T(5.0)/-T(2.5)))-_ATAN_PI_LO(T))
        @test atan(-T(5.0), -T(2.5)) === -(T(pi)-atan(abs(-T(5.0)/-T(2.5)))-_ATAN_PI_LO(T))
        @test atan(T(1235.2341234), T(2.5)) === atan(abs(T(1235.2341234)/T(2.5)))
        @test atan(-T(1235.2341234), T(2.5)) === -atan(abs(-T(1235.2341234)/T(2.5)))
        @test atan(T(1235.2341234), -T(2.5)) === T(pi)-(atan(abs(T(1235.2341234)/-T(2.5)))-_ATAN_PI_LO(T))
        @test atan(-T(1235.2341234), -T(2.5)) === -(T(pi)-(atan(abs(-T(1235.2341234)/T(2.5)))-_ATAN_PI_LO(T)))
    end
end

@testset "atand" begin
    for T in (Float32, Float64)
        r = T(randn())
        absr = abs(r)

        # Tests related to the 1-argument version of `atan`.
        # ==================================================

        @test atand(T(Inf))  === T(90.0)
        @test atand(-T(Inf)) === -T(90.0)
        @test atand(zero(T)) === T(0.0)
        @test atand(one(T))  === T(45.0)
        @test atand(-one(T)) === -T(45.0)

        # Tests related to the 2-argument version of `atan`.
        # ==================================================

        # If `x` is one, then `atand(y,x)` must be equal to `atand(y)`.
        @test atand(T(r), one(T))    === atand(T(r))

        # `y` zero.
        @test atand(zero(T), absr)   === zero(T)
        @test atand(-zero(T), absr)  === -zero(T)
        @test atand(zero(T), -absr)  === T(180.0)
        @test atand(-zero(T), -absr) === -T(180.0)

        # `x` zero and `y` not zero.
        @test atand(one(T), zero(T))  === T(90.0)
        @test atand(-one(T), zero(T)) === -T(90.0)

        # `x` and `y` equal for each quadrant.
        @test atand(+absr, +absr) === T(45.0)
        @test atand(-absr, +absr) === -T(45.0)
        @test atand(+absr, -absr) === T(135.0)
        @test atand(-absr, -absr) === -T(135.0)
    end
end

@testset "acos #23283" begin
    for T in (Float32, Float64)
        @test acos(zero(T)) === T(pi)/2
        @test acos(-zero(T)) === T(pi)/2
        @test acos(nextfloat(zero(T))) === T(pi)/2
        @test acos(prevfloat(zero(T))) === T(pi)/2
        @test acos(one(T)) === T(0.0)
        @test acos(-one(T)) === T(pi)
        for x in (0.45, 0.6, 0.98)
            by = acos(big(T(x)))
            @test T((acos(T(x)) - by))/eps(abs(T(by))) <= 1
            bym = acos(big(T(-x)))
            @test T(abs(acos(T(-x)) - bym))/eps(abs(T(bym))) <= 1
        end
        @test_throws DomainError acos(-T(Inf))
        @test_throws DomainError acos(T(Inf))
        @test isnan_type(T, acos(T(NaN)))
    end
end

#prev, current, next float
pcnfloat(x) = prevfloat(x), x, nextfloat(x)
import Base.Math: COSH_SMALL_X, H_SMALL_X, H_MEDIUM_X, H_LARGE_X

@testset "sinh" begin
    for T in (Float32, Float64)
        @test sinh(zero(T)) === zero(T)
        @test sinh(-zero(T)) === -zero(T)
        @test sinh(nextfloat(zero(T))) === nextfloat(zero(T))
        @test sinh(prevfloat(zero(T))) === prevfloat(zero(T))
        @test sinh(T(1000)) === T(Inf)
        @test sinh(-T(1000)) === -T(Inf)
        @test isnan_type(T, sinh(T(NaN)))
        for x in Iterators.flatten(pcnfloat.([H_SMALL_X(T), H_MEDIUM_X(T), H_LARGE_X(T)]))
            @test sinh(x) ≈ sinh(big(x)) rtol=eps(T)
            @test sinh(-x) ≈ sinh(big(-x)) rtol=eps(T)
        end
    end
end

@testset "cosh" begin
    for T in (Float32, Float64)
        @test cosh(zero(T)) === one(T)
        @test cosh(-zero(T)) === one(T)
        @test cosh(nextfloat(zero(T))) === one(T)
        @test cosh(prevfloat(zero(T))) === one(T)
        @test cosh(T(1000)) === T(Inf)
        @test cosh(-T(1000)) === T(Inf)
        @test isnan_type(T, cosh(T(NaN)))
        for x in Iterators.flatten(pcnfloat.([COSH_SMALL_X(T), H_MEDIUM_X(T), H_LARGE_X(T)]))
            @test cosh(x) ≈ cosh(big(x)) rtol=eps(T)
            @test cosh(-x) ≈ cosh(big(-x)) rtol=eps(T)
        end
    end
end

@testset "tanh" begin
    for T in (Float32, Float64)
        @test tanh(zero(T)) === zero(T)
        @test tanh(-zero(T)) === -zero(T)
        @test tanh(nextfloat(zero(T))) === nextfloat(zero(T))
        @test tanh(prevfloat(zero(T))) === prevfloat(zero(T))
        @test tanh(T(1000)) === one(T)
        @test tanh(-T(1000)) === -one(T)
        @test isnan_type(T, tanh(T(NaN)))
        for x in Iterators.flatten(pcnfloat.([H_SMALL_X(T), T(1.0), H_MEDIUM_X(T)]))
            @test tanh(x) ≈ tanh(big(x)) rtol=eps(T)
            @test tanh(-x) ≈ -tanh(big(x)) rtol=eps(T)
        end
    end
    @test tanh(18.0) ≈ tanh(big(18.0)) rtol=eps(Float64)
    @test tanh(8.0) ≈ tanh(big(8.0)) rtol=eps(Float32)
end

@testset "asinh" begin
    for T in (Float32, Float64)
        @test asinh(zero(T)) === zero(T)
        @test asinh(-zero(T)) === -zero(T)
        @test asinh(nextfloat(zero(T))) === nextfloat(zero(T))
        @test asinh(prevfloat(zero(T))) === prevfloat(zero(T))
        @test isnan_type(T, asinh(T(NaN)))
        for x in Iterators.flatten(pcnfloat.([T(2)^-28,T(2),T(2)^28]))
            @test asinh(x) ≈ asinh(big(x)) rtol=eps(T)
            @test asinh(-x) ≈ asinh(big(-x)) rtol=eps(T)
        end
    end
end

@testset "acosh" begin
    for T in (Float32, Float64)
        @test_throws DomainError acosh(T(0.1))
        @test acosh(one(T)) === zero(T)
        @test isnan_type(T, acosh(T(NaN)))
        for x in Iterators.flatten(pcnfloat.([nextfloat(T(1.0)), T(2), T(2)^28]))
            @test acosh(x) ≈ acosh(big(x)) rtol=eps(T)
        end
    end
end

@testset "atanh" begin
    for T in (Float32, Float64)
        @test_throws DomainError atanh(T(1.1))
        @test atanh(zero(T)) === zero(T)
        @test atanh(-zero(T)) === -zero(T)
        @test atanh(one(T)) === T(Inf)
        @test atanh(-one(T)) === -T(Inf)
        @test atanh(nextfloat(zero(T))) === nextfloat(zero(T))
        @test atanh(prevfloat(zero(T))) === prevfloat(zero(T))
        @test isnan_type(T, atanh(T(NaN)))
        for x in Iterators.flatten(pcnfloat.([T(2.0)^-28, T(0.5)]))
            @test atanh(x) ≈ atanh(big(x)) rtol=eps(T)
            @test atanh(-x) ≈ atanh(big(-x)) rtol=eps(T)
        end
    end
end

# Define simple wrapper of a Float type:
struct FloatWrapper <: Real
    x::Float64
end

import Base: +, -, *, /, ^, sin, cos, exp, sinh, cosh, convert, isfinite, float, promote_rule

for op in (:+, :-, :*, :/, :^)
    @eval $op(x::FloatWrapper, y::FloatWrapper) = FloatWrapper($op(x.x, y.x))
end

for op in (:sin, :cos, :exp, :sinh, :cosh, :-)
    @eval $op(x::FloatWrapper) = FloatWrapper($op(x.x))
end

for op in (:isfinite,)
    @eval $op(x::FloatWrapper) = $op(x.x)
end

convert(::Type{FloatWrapper}, x::Int) = FloatWrapper(float(x))
promote_rule(::Type{FloatWrapper}, ::Type{Int}) = FloatWrapper

float(x::FloatWrapper) = x

@testset "exp(Complex(a, b)) for a and b of non-standard real type #25292" begin

    x = FloatWrapper(3.1)
    y = FloatWrapper(4.1)

    @test sincos(x) == (sin(x), cos(x))

    z = Complex(x, y)

    @test isa(exp(z), Complex)
    @test isa(sin(z), Complex)
    @test isa(cos(z), Complex)
end

# Define simple wrapper of a Float type:
struct FloatWrapper2 <: Real
    x::Float64
end

float(x::FloatWrapper2) = x.x
@testset "inverse hyperbolic trig functions of non-standard float" begin
    x = FloatWrapper2(3.1)
    @test asinh(sinh(x)) == asinh(sinh(3.1))
    @test acosh(cosh(x)) == acosh(cosh(3.1))
    @test atanh(tanh(x)) == atanh(tanh(3.1))
end

@testset "cbrt" begin
    for T in (Float32, Float64)
        @test cbrt(zero(T)) === zero(T)
        @test cbrt(-zero(T)) === -zero(T)
        @test cbrt(one(T)) === one(T)
        @test cbrt(-one(T)) === -one(T)
        @test cbrt(T(Inf)) === T(Inf)
        @test cbrt(-T(Inf)) === -T(Inf)
        @test isnan_type(T, cbrt(T(NaN)))
        for x in (pcnfloat(nextfloat(nextfloat(zero(T))))...,
                  pcnfloat(prevfloat(prevfloat(zero(T))))...,
                  0.45, 0.6, 0.98,
                  map(x->x^3, 1.0:1.0:1024.0)...,
                  nextfloat(-T(Inf)), prevfloat(T(Inf)))
            by = cbrt(big(T(x)))
            @test cbrt(T(x)) ≈ by rtol=eps(T)
            bym = cbrt(big(T(-x)))
            @test cbrt(T(-x)) ≈ bym rtol=eps(T)
        end
    end
end

@testset "hypot" begin
    @test hypot(0, 0) == 0.0
    @test hypot(3, 4) == 5.0
    @test hypot(NaN, Inf) == Inf
    @test hypot(Inf, NaN) == Inf
    @test hypot(Inf, Inf) == Inf

    isdefined(Main, :Furlongs) || @eval Main include("testhelpers/Furlongs.jl")
    using .Main.Furlongs
    @test (@inferred hypot(Furlong(0), Furlong(0))) == Furlong(0.0)
    @test (@inferred hypot(Furlong(3), Furlong(4))) == Furlong(5.0)
    @test (@inferred hypot(Furlong(NaN), Furlong(Inf))) == Furlong(Inf)
    @test (@inferred hypot(Furlong(Inf), Furlong(NaN))) == Furlong(Inf)
    @test (@inferred hypot(Furlong(0), Furlong(0), Furlong(0))) == Furlong(0.0)
    @test (@inferred hypot(Furlong(Inf), Furlong(Inf))) == Furlong(Inf)
    @test (@inferred hypot(Furlong(1), Furlong(1), Furlong(1))) == Furlong(sqrt(3))
    @test (@inferred hypot(Furlong(Inf), Furlong(NaN), Furlong(0))) == Furlong(Inf)
    @test (@inferred hypot(Furlong(Inf), Furlong(Inf), Furlong(Inf))) == Furlong(Inf)
    @test isnan(hypot(Furlong(NaN), Furlong(0), Furlong(1)))
    ex = @test_throws ErrorException hypot(Furlong(1), 1)
    @test startswith(ex.value.msg, "promotion of types ")

    @test_throws MethodError hypot()
    @test (@inferred hypot(floatmax())) == floatmax()
    @test (@inferred hypot(floatmax(), floatmax())) == Inf
    @test (@inferred hypot(floatmin(), floatmin())) == √2floatmin()
    @test (@inferred hypot(floatmin(), floatmin(), floatmin())) == √3floatmin()
    @test (@inferred hypot(1e-162)) ≈ 1e-162
    @test (@inferred hypot(2e-162, 1e-162, 1e-162)) ≈ hypot(2, 1, 1)*1e-162
    @test (@inferred hypot(1e162)) ≈ 1e162
    @test hypot(-2) === 2.0
    @test hypot(-2, 0) === 2.0
    let i = typemax(Int)
        @test (@inferred hypot(i, i)) ≈ i * √2
        @test (@inferred hypot(i, i, i)) ≈ i * √3
        @test (@inferred hypot(i, i, i, i)) ≈ 2.0i
        @test (@inferred hypot(i//1, 1//i, 1//i)) ≈ i
    end
    let i = typemin(Int)
        @test (@inferred hypot(i, i)) ≈ -√2i
        @test (@inferred hypot(i, i, i)) ≈ -√3i
        @test (@inferred hypot(i, i, i, i)) ≈ -2.0i
    end
    @testset "$T" for T in (Float32, Float64)
        @test (@inferred hypot(T(Inf), T(NaN))) == T(Inf) # IEEE754 says so
        @test (@inferred hypot(T(Inf), T(3//2), T(NaN))) == T(Inf)
        @test (@inferred hypot(T(1e10), T(1e10), T(1e10), T(1e10))) ≈ 2e10
        @test isnan_type(T, hypot(T(3), T(3//4), T(NaN)))
        @test hypot(T(1), T(0)) === T(1)
        @test hypot(T(1), T(0), T(0)) === T(1)
        @test (@inferred hypot(T(Inf), T(Inf), T(Inf))) == T(Inf)
        for s in (zero(T), floatmin(T)*1e3, floatmax(T)*1e-3, T(Inf))
            @test hypot(1s, 2s)     ≈ s * hypot(1, 2)   rtol=8eps(T)
            @test hypot(1s, 2s, 3s) ≈ s * hypot(1, 2, 3) rtol=8eps(T)
        end
    end
    @testset "$T" for T in (Float16, Float32, Float64, BigFloat)
        let x = 1.1sqrt(floatmin(T))
            @test (@inferred hypot(x, x/4)) ≈ x * sqrt(17/BigFloat(16))
            @test (@inferred hypot(x, x/4, x/4)) ≈ x * sqrt(9/BigFloat(8))
        end
        let x = 2sqrt(nextfloat(zero(T)))
            @test (@inferred hypot(x, x/4)) ≈ x * sqrt(17/BigFloat(16))
            @test (@inferred hypot(x, x/4, x/4)) ≈ x * sqrt(9/BigFloat(8))
        end
        let x = sqrt(nextfloat(zero(T))/eps(T))/8, f = sqrt(4eps(T))
            @test hypot(x, x*f) ≈ x * hypot(one(f), f) rtol=eps(T)
            @test hypot(x, x*f, x*f) ≈ x * hypot(one(f), f, f) rtol=eps(T)
        end
        let x = floatmax(T)/2
            @test (@inferred hypot(x, x/4)) ≈ x * sqrt(17/BigFloat(16))
            @test (@inferred hypot(x, x/4, x/4)) ≈ x * sqrt(9/BigFloat(8))
        end
    end
    # hypot on Complex returns Real
    @test (@inferred hypot(3, 4im)) === 5.0
    @test (@inferred hypot(3, 4im, 12)) === 13.0
end

struct BadFloatWrapper <: AbstractFloat
    x::Float64
end

@testset "not impelemented errors" begin
    x = BadFloatWrapper(1.9)
    for f in (sin, cos, tan, sinh, cosh, tanh, atan, acos, asin, asinh, acosh, atanh, exp, log1p, expm1, log) #exp2, exp10 broken for now
        @test_throws MethodError f(x)
    end
end
back to top