Use correct n for calculating kstest pValue (#165)

* [#164] fixed variable error in cdf table argument

* [#164] corrected n to be size of ecdf output

inst/kstest.m

@@ -138,8 +138,8 @@
   endif
   ## Remove NaNs, get sample size and compute empirical cdf
   x(isnan (x)) = [];
-  n = length(x);
   [sampleCDF, x] = ecdf (x);
+  n = length(x);
   ## Remove 1st element
   x = x(2:end);
   ## Check the hypothesized CDF specified under the null hypothesis.
@@ -166,7 +166,7 @@
     endif
     ## Remove duplicates. Check for consistency
     rd = find (diff (xCDF) == 0);
-    if (! isempty (rm))
+    if (! isempty (rd))
       if (! all (ydiff(rd) == 0))
         error ("kstest: wrong duplicates in numericl CDF.");
       endif
@@ -335,4 +335,3 @@
 %! assert (p, 5.085438806199252e-05, 1e-14);
 %! assert (k, 0.2197, 1e-4);
 %! assert (c, 0.1207, 1e-4);
-