Update ksdist.jl (#1581)

While the sqrt(n) factor is often part of the test statistic this is not what is implemented here. To witness, see
```julia
https://github.com/JuliaStats/Distributions.jl/blob/254f27f1b32a72d6529e2f78a3745b68bd38b31e/test/testutils.jl#L597-L607
```
where there is no sqrt(n) factor.

Author: jw3126

src/univariate/continuous/ksdist.jl

@@ -4,7 +4,7 @@
 Distribution of the (two-sided) Kolmogorov-Smirnoff statistic
 
 ```math
-D_n = \\sup_x | \\hat{F}_n(x) -F(x)| \\sqrt(n)
+D_n = \\sup_x | \\hat{F}_n(x) -F(x)|
 ```
 
 ``D_n`` converges a.s. to the Kolmogorov distribution.