Kolmogorov-Smirnov test... a good overview

I was wondering if anyone here could recommend a review of the
Kolmogorov-Smirnov test, which goes slightly deeper than, for example,
the links on the wikipedia page or the material in Numerical recipes.

Publications available online would be more convenient, but anything
would help

I would be particularly interested in:

a) At what negative power of n (the geometric mean of the size of the
two samples) do corrections to the KS distribution typically appear?
b) What are the situations where the test is likely to fail, and why?

The level: I am a theoretical physicist (postdoc), but my knowledge
of statistics is "self-thought" with applications in mind.

Thanks, and best regards

GT
.