Overhauled the HuskyCoder interface

[rchillyard]

HuskySortIntegrationTest: added Rule for timing; PureHuskySortTest: split up into two segments: onne in test and one in it; HuskyCoder: instead of imperfect(length), we now have perfect(); Extended HuskyCoder: HuskySequenceCoder: defines perfect(x) and perfectForLength(n); IntroHuskySort: if interim inversions not supporeted, return NaN; HuskySortHelper:

• added longCoder;
• fixed bigIntegerCoder;
• fixed doubleCoder;
• gave names to the various magic numbers;
• renamed printableAsciiCoder to englishCoder; HuskySortHelperTests: add some more unit tests; ProcessorDependentTimeout: generalized time units to allow for microseconds, etc.

[rchillyard]

I've made some fairly drastic changes to the HuskyCoder implementations. I don't really understand the UTF8 stuff and would appreciate more testing on that. I'd also like to do more testing with the special IntroHuskySort implementation that keeps track of interim inversions -- for unicode and maybe UTF8 sources. It seems to me that the number of inversions after the first pass when the coder is the unicode coder could be quite high.

[rchillyard]

There are some failures. Two are integration tests timing out. We need to figure the correct default processor for the Java CI. Maybe it's bit slower than a 1.0

[rchillyard]

It appears that the failures in testUnicodeCoder are Java-version dependent. This issue is now fixed.

[rchillyard]

Everything is looking good. Am merging.