Trying to make hashcodes faster. Didn't achieve much on that front, so redirected into structural/consistency issues. The latter was lacking in terms of how/where "def seq" was defined. The documentation I can find doesn't give me much hint that the sequential form of my sequential collection might be a single-use iterator! (As in StringOps, ArrayOps.) If that's intentional it should be in huge letters. I'm assuming for now that it wasn't. Also, there was this: GenMapLike: def seq: Map[A, B] GenSetLike: def seq: Set[A] GenSeqLike: // nothing, returns Traversable So I added some def seqs where I needed the more specific types for my hashcode work. Hashcodewise, I broke the MurmurHash3 object into a reusable class and a collections-specific object, and I deprecated the methods which took GenTraversableOnce in favor of ones taking TraversableOnce, because there's no reason the hashcode library should have to know about things like "make sure to call seq before you traverse or you'll be sorry." Exclude things by their type and you can never make a mistake. End transmission.
This commit and the two subsequent commits were contributed by: Todd Vierling <email@example.com>. I combined some commits and mangled his commit messages, but all the credit is his. This pursues the same approach to classfile reduction seen in r19989 when AbstractFunctionN was introduced, but applies it to the collections. Thanks to -Xlint it's easy to verify that the private types don't escape. Design considerations as articulated by Todd: * Don't necessarily create concrete types for _everything_. Where a subtrait only provides a few additional methods, don't bother; instead, use the supertrait's concrete class and retain the "with". For example, "extends AbstractSeq[A] with LinearSeq[A]". * Examine all classes with .class file size greater than 10k. Named classes and class names ending in $$anon$<num> are candidates for analysis. * If a return type is currently inferred where an anon subclass would be returned, make the return type explicit. Don't allow the library-private abstract classes to leak into the public namespace [and scaladoc].
Addressing most of the warnings revealed by the patch to warn about unknown scaladoc variables. Updated and reran genprod. No review.
A new murmur3 hashcode implementation contributed by rex kerr. I'm really happy with it. There is benchmarking code included but you can use the pudding for proof: the compiler compiling itself is clearly faster. I deleted the existing murmur2 implementation which has never left trunk in a release. There remain some possible points of improvement with inlining and/or synthetic method generation, but this is a major improvement over the status quo. Closes #2537, review by prokopec.