From a0a08598917c9d8523b39e25c9fcf667e57c1e2d Mon Sep 17 00:00:00 2001
From: Tamas Nagy <tamas@tamasnagy.com>
Date: Sun, 10 Jun 2018 15:57:18 -0700
Subject: [PATCH] use improved high numerical precision range from Base (#52)

KernelDensity.jl rolled its own range generator which suffers from
similar flaws as Base's linspace prior to being fixed in
JuliaLang/julia#18777. This closes #39.
---
 src/univariate.jl  | 3 +--
 test/univariate.jl | 2 ++
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/src/univariate.jl b/src/univariate.jl
index a82605cb..ae09eab6 100644
--- a/src/univariate.jl
+++ b/src/univariate.jl
@@ -66,8 +66,7 @@ function kde_range(boundary::Tuple{Real,Real}, npoints::Int)
     lo, hi = boundary
     lo < hi || error("boundary (a,b) must have a < b")
 
-    step = (hi - lo) / (npoints-1)
-    lo:step:hi
+    Compat.range(lo, stop=hi, length=npoints)
 end
 
 struct UniformWeights{N} end
diff --git a/test/univariate.jl b/test/univariate.jl
index dcefd08d..dd848950 100644
--- a/test/univariate.jl
+++ b/test/univariate.jl
@@ -13,6 +13,8 @@ end
 
 r = kde_range((-2.0,2.0), 128)
 @test step(r) > 0
+r2 = kde_range((0.12698109160784082, 0.9785547869337731), 256)
+@test length(r2) == 256
 
 for X in ([0.0], [0.0,0.0], [0.0,0.5], [-0.5:0.1:0.5;])
     w = default_bandwidth(X)