Merge pull request #2 from JuliaIntervals/update_0.5

Updates for Julia 0.5
JuliaIntervals · Apr 24, 2017 · 4f19b92 · 4f19b92
2 parents 7f3734e + 659a41c
commit 4f19b92
Show file tree

Hide file tree

Showing 4 changed files with 65 additions and 33 deletions.
diff --git a/.travis.yml b/.travis.yml
@@ -0,0 +1,30 @@
+language: julia
+
+sudo: required
+
+services:
+  - docker
+
+os:
+  - linux
+  - osx
+
+before_install:
+  - sudo apt-get -qq update
+  - sudo apt-get install -y libmpfi-dev
+
+julia:
+  - 0.5
+  - 0.6
+  - nightly
+
+# matrix:
+#   allow_failures:
+#     - julia: 0.6
+#     - julia: nightly
+
+notifications:
+  email: false
+
+after_success:
+- julia -e 'cd(Pkg.dir("MPFI")); Pkg.add("Coverage"); using Coverage; Coveralls.submit(Coveralls.process_folder()); Codecov.submit(process_folder())'
diff --git a/REQUIRE b/REQUIRE
@@ -1 +1 @@
-julia 0.2-
+julia 0.5
diff --git a/src/MPFI.jl b/src/MPFI.jl
@@ -12,15 +12,16 @@ export
     mid,
     mig,
     isbounded,
-    right
+    right,
+    square
 
-import
-    Base: precision, string, print, show, showcompact, promote_rule,
-        promote, convert, +, *, -, /, exp, isinf, isnan, nan, inf, sqrt,
-        square, exp, exp2, expm1, cosh, sinh, tanh, sech, csch, coth, inv,
-        sqrt, cbrt, abs, log, log2, log10, log1p, sin, cos, tan, sec,
-        csc, acos, asin, atan, acosh, asinh, atanh, isempty, union,
-        intersect, in, cmp, ldexp, rand, rand!
+import Base:
+    precision, string, print, show, showcompact, promote_rule,
+    promote, convert, +, *, -, /, exp, isinf, isnan, sqrt,
+    exp, exp2, expm1, cosh, sinh, tanh, sech, csch, coth, inv,
+    sqrt, cbrt, abs, log, log2, log10, log1p, sin, cos, tan, sec,
+    csc, acos, asin, atan, acosh, asinh, atanh, isempty, union,
+    intersect, in, cmp, ldexp, rand, rand!
 
 type Interval <: Number
     left_prec::Clong
@@ -31,8 +32,9 @@ type Interval <: Number
     right_sign::Cint
     right_exp::Clong
     right_d::Ptr{Void}
+
     function Interval()
-        N = get_bigfloat_precision()
+        N = precision(BigFloat)
         z = new(zero(Clong), zero(Cint), zero(Clong), C_NULL,
                 zero(Clong), zero(Cint), zero(Clong), C_NULL)
         ccall((:mpfi_init2,:libmpfi), Void, (Ptr{Interval}, Clong), &z, N)
@@ -72,18 +74,18 @@ end
 
 function Interval(x::String, base::Int)
     z = Interval()
-    err = ccall((:mpfi_set_str, :libmpfi), Int32, (Ptr{Interval}, Ptr{Uint8}, Int32), &z, x, base)
+    err = ccall((:mpfi_set_str, :libmpfi), Int32, (Ptr{Interval}, Ptr{UInt8}, Int32), &z, x, base)
     if err != 0; error("Invalid input"); end
     return z
 end
 Interval(x::String) = Interval(x, 10)
 
 Interval(x::Integer) = Interval(BigInt(x))
 
-Interval(x::Union(Bool,Int8,Int16,Int32)) = Interval(convert(Clong,x))
-Interval(x::Union(Uint8,Uint16,Uint32)) = Interval(convert(Culong,x))
+Interval(x::Union{Bool,Int8,Int16,Int32}) = Interval(convert(Clong,x))
+Interval(x::Union{UInt8,UInt16,UInt32}) = Interval(convert(Culong,x))
 
-Interval(x::Float32) = Interval(float64(x))
+Interval(x::Float32) = Interval(Float64(x))
 Interval(x::Rational) = Interval(num(x)) / Interval(den(x))
 
 # Dyadic constructors
@@ -110,26 +112,26 @@ function Interval(x::BigFloat, y::BigFloat)
     return z
 end
 
-Interval(x::MathConst) = convert(Interval, x)
+Interval(x::Irrational) = convert(Interval, x)
 
 # Promotion of constants
-promote_rule{s}(::Type{MathConst{s}}, ::Type{Interval}) = Interval
+promote_rule{s}(::Type{Irrational{s}}, ::Type{Interval}) = Interval
 
 # Conversions to Interval
 convert(::Type{Interval}, x::Rational) = Interval(x)
 convert(::Type{Interval}, x::Real) = Interval(x)
-convert(::Type{Interval}, x::MathConst) = Interval(big(x))
-function convert(::Type{Interval}, ::MathConst{:π})
+convert(::Type{Interval}, x::Irrational) = Interval(big(x))
+function convert(::Type{Interval}, ::Irrational{:π})
     z = Interval()
     ccall((:mpfi_const_pi,:libmpfi), Cint, (Ptr{Interval},), &z)
     return z
 end
-function convert(::Type{Interval}, ::MathConst{:γ})
+function convert(::Type{Interval}, ::Irrational{:γ})
     z = Interval()
     ccall((:mpfi_const_euler,:libmpfi), Cint, (Ptr{Interval},), &z)
     return z
 end
-function convert(::Type{Interval}, ::MathConst{:catalan})
+function convert(::Type{Interval}, ::Irrational{:catalan})
     z = Interval()
     ccall((:mpfi_const_catalan,:libmpfi), Cint, (Ptr{Interval},), &z)
     return z
@@ -144,15 +146,15 @@ end
 convert(::Type{Float64}, x::Interval) =
     ccall((:mpfi_get_d,:libmpfi), Float64, (Ptr{Interval},), &x)
 
-for to in (Int8, Int16, Int32, Int64, Uint8, Uint16, Uint32, Uint64, BigInt, Float32)
+for to in (Int8, Int16, Int32, Int64, UInt8, UInt16, UInt32, UInt64, BigInt, Float32)
     @eval begin
         function convert(::Type{$to}, x::Interval)
             convert($to, convert(BigFloat, x))
         end
     end
 end
 convert(::Type{Integer}, x::Interval) = convert(BigInt, x)
-convert(::Type{FloatingPoint}, x::Interval) = convert(BigFloat, x)
+convert(::Type{AbstractFloat}, x::Interval) = convert(BigFloat, x)
 
 # Interval functions with floating-point results
 for f in (:diam_abs, :diam_rel, :diam, :mag, :mig, :mid)
@@ -165,7 +167,7 @@ end
 
 # Basic operations between intervals
 for (fJ, fC) in ((:+,:add), (:-,:sub), (:*,:mul), (:/,:div))
-    @eval begin 
+    @eval begin
         function ($fJ)(x::Interval, y::Interval)
             z = Interval()
             ccall(($(string(:mpfi_,fC)),:libmpfi), Int32, (Ptr{Interval}, Ptr{Interval}, Ptr{Interval}), &z, &x, &y)

diff --git a/test/MPFI_test.jl b/test/MPFI_test.jl
@@ -1,8 +1,8 @@
 # Check conversion and output
 for i = 1:10000
-    a = {rand(Int8), rand(Int16), rand(Int32), rand(Int64), rand(Int128), 
-         rand(Uint8), rand(Uint16), rand(Uint32), rand(Uint64), rand(Uint128),
-         rand(Float32), rand(Float64)}
+    a = [rand(Int8), rand(Int16), rand(Int32), rand(Int64), rand(Int128),
+         rand(UInt8), rand(UInt16), rand(UInt32), rand(UInt64), rand(UInt128),
+         rand(Float32), rand(Float64)]
     b = map(string, a)
     for j in a
         if j != 0
@@ -15,7 +15,7 @@ end
 
 # Check precision
 a = Interval(1)
-@test precision(a) == get_bigfloat_precision()
+@test precision(a) == precision(BigFloat)
 
 # Check squaring
 a = Interval(-1, 1)
@@ -27,9 +27,9 @@ a = Interval(0, 1)
 @test (left(-a), right(-a)) == (-1, 0)
 @test signbit(right(-a)) == 1
 
-# Check rand
-a = Interval(-1, 1)
-@test -1 <= rand(a) <= 1
-r = rand(a, 1, 3)
-@test size(r) == (1, 3)
-@test typeof(r) == Matrix{BigFloat}
+# # Check rand
+# a = Interval(-1, 1)
+# @test -1 <= rand(a) <= 1
+# r = rand(a, 1, 3)
+# @test size(r) == (1, 3)
+# @test typeof(r) == Matrix{BigFloat}