On Sat, 2014-04-26 at 18:15 +0200, Kurt Roeckx wrote: > So I can't see how the old code really sorted like it's supposed > to. We're probably lacking a test case that checks the result of > the coding is really what we produced. But this seems more likely > to actually produce the expected output.
Do you have some example of a SET OF structure that is correctly sorted? We can then have a test that decodes it and re-encodes it to verify the new code. regards, Nikos
