Fix ecdf bugs

Corrects output of  ecdf in the case where there are no values in X
between two values in v and the case where all values in X were less
than multiple values in v.

src/others.jl

@@ -81,14 +81,17 @@ function ecdf{T}(X::AbstractVector{T})
         r0 = 0
         i = 1
         for x in Xs
-            if x > v[ord[i]]
+            while i <= m && x > v[ord[i]]
                 r[ord[i]] = r0
                 i += 1
             end
             r0 += 1
             if i > m break end
         end
-        if i == m r[ord[i]] = n end
+        while i <= m
+            r[ord[i]] = n
+            i += 1
+        end
         return r / n
     end
     return e

test/01.jl

@@ -28,3 +28,6 @@ fnecdf = ecdf(randn(10000000))
 @test_approx_eq_eps fnecdf([-1.96, -1.644854, -1.281552, -0.6744898, 0, 0.6744898, 1.281552, 1.644854, 1.96]) [0.025, 0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95, 0.975] 1e-3
 @test_approx_eq_eps fnecdf(1.96) 0.975 1e-3
 @test_approx_eq fnecdf([-1.96, -1.644854, -1.281552, -0.6744898, 0, 0.6744898, 1.281552, 1.644854, 1.96]) map(fnecdf, [-1.96, -1.644854, -1.281552, -0.6744898, 0, 0.6744898, 1.281552, 1.644854, 1.96])
+
+fnecdf = ecdf([0.5])
+@test fnecdf([zeros(5000), ones(5000)]) == [zeros(5000), ones(5000)]