remove GenericSVD and GenericSchur as dependencies

Due to type piracy, these packages cause problems with precompilation and don't play nice with GenericLinearAlgebra (see discussion in JuliaLinearAlgebra/GenericLinearAlgebra.jl#71). Additionally, GenericSVD has now been deprecated (JuliaLinearAlgebra/GenericSVD.jl#30).

Author: simonbyrne

Project.toml

@@ -12,8 +12,6 @@ Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 Libdl = "8f399da3-3557-5675-b5ff-fb832c97cbdb"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 BinaryProvider = "b99e7846-7c00-51b0-8f62-c81ae34c0232"
-GenericSVD = "01680d73-4ee2-5a08-a1aa-533608c188bb"
-GenericSchur = "c145ed77-6b09-5dd9-b285-bf645a82121e"
 SpecialFunctions = "276daf66-3868-5448-9aa4-cd146d93841b"
 Readables = "0d4725de-cd7c-5e44-8a85-a48caeef9fa5"
 GMP_jll = "781609d7-10c4-51f6-84f2-b8444358ff6d"
@@ -30,7 +28,5 @@ test = ["Test"]
 [compat]
 SpecialFunctions = "0.7, 0.8, 0.9, 0.10, 1"
 BinaryProvider = "0.4, 0.5, 0.6, 0.7, 0.9, 1"
-GenericSVD = "0.3, 0.4, 0.5"
-GenericSchur = "0.4, 0.5, 0.6"
 Readables = "0.3"
 julia = "1"

docs/exemplars/expmatrix.jl

@@ -18,7 +18,7 @@ with a   500x500 matrix of Float64s, working with 350 bits used  25 seconds
 with a   750x750 matrix of Float64s, working with 175 bits used  25 seconds
 =#
 
-using ArbNumerics # GenericSVD, LinearAlgebra, Readables
+using ArbNumerics # LinearAlgebra, Readables
 
 #=
 ArbNumerics' interactive users see the "viewable precision"; computations utilize "working precision".

docs/src/matrixfunctions.md

@@ -7,7 +7,7 @@
 
 ## Matrix Functions
 
-#### `using ArbNumerics, GenericSchur, GenericSVD, LinearAlgebra`
+#### `using ArbNumerics, LinearAlgebra`
 
 - `det`, `tr`
 - `transpose`, `adjoint`, `inv`

src/ArbNumerics.jl

@@ -127,8 +127,6 @@ import SpecialFunctions: gamma, lgamma, lfact, digamma, invdigamma, polygamma, t
      besselj, besselj0, besselj1, bessely, bessely0, bessely1, besseli, besselk,
      eta, zeta
 
-using  GenericSVD
-
 using  LinearAlgebra
 import LinearAlgebra: tr, det, transpose, transpose!, norm, lu, ldlt,
                       cholesky, tril, triu, eigvals, svdvals, floatmin2,