remove lgamma and use logabsgamma from Quadmath

Author: araujoms

Project.toml

@@ -18,7 +18,7 @@ SpecialFunctions = "276daf66-3868-5448-9aa4-cd146d93841b"
 
 [compat]
 Polynomials = "1, 2, 3, 4, 5, 6"
-Quadmath = "0.4, 0.5, 0.6, 0.7, 0.8"
+Quadmath = "0.5.10, 0.6, 0.7, 0.8"
 Requires = "1"
 SpecialFunctions = "1.0, 1.1, 2.0, 2.1, 2.2, 2.3, 2.4, 2.5"
 GenericLinearAlgebra = "0.2.5, 0.3, 0.4, 0.5, 0.6"

src/math/special/gamma_erf.jl

@@ -1,19 +1,16 @@
 erf(x::Double64) = Double64Float128(erf, x)
 erfc(x::Double64) = Double64Float128(erfc, x)
 gamma(x::Double64) = Double64Float128(gamma, x)
-lgamma(x::Double64) = Double64Float128(lgamma, x)
 
 erf(x::Double32) = Double32Float128(erf, x)
 erfc(x::Double32) = Double32Float128(erfc, x)
 gamma(x::Double32) = Double32Float128(gamma, x)
-lgamma(x::Double32) = Double32Float128(lgamma, x)
 
 erf(x::Double16) = Double16Float128(erf, x)
 erfc(x::Double16) = Double16Float128(erfc, x)
 gamma(x::Double16) = Double16Float128(gamma, x)
-lgamma(x::Double16) = Double16Float128(lgamma, x)
 
 function logabsgamma(x::DoubleFloat{T}) where {T<:IEEEFloat}
-    sign = x >= 0 ? 1 : 2*mod(ceil(Int64,x),2)-1
-    return lgamma(x), sign
+    result = logabsgamma(Float128(x))
+    return DoubleFloat{T}(result[1]), result[2]
 end

src/math/special/specialfunctions.jl

@@ -1,5 +1,5 @@
 import .SpecialFunctions
-import .SpecialFunctions: erf, erfc, gamma, lgamma, logabsgamma,
+import .SpecialFunctions: erf, erfc, gamma, logabsgamma,
     besselj0, besselj1, besselj, bessely0, bessely1, bessely,
     ellipk