I've separated legacy code away from CAIRO and its derivatives. The goal is to simplify top level package structure and create clean slate for future sub-projects

GRIFFIN: Put together a strawman SqlParser and run first SQL query through it! Also create LongTreeSet class to help with execution of key lookups with multiple key values. May be more!

Total refactoring. This is a re-born project in many ways. Existing storage engine is going to be replaced eventually with new, robust, well-tested, more flexible and yet faster engine - CAIRO. SQL implementation will remain, it is going to get integrated with CAIRO as well. HA package will die, it is going to be replaced with much more reliable and fast replication over UDP.

column type detector may decide that input file has not header if every value is "string". For this case I added "forceHeader" flag to have importer use first row as header.

fixed excessive garbage generation in concurrent map, also relaxed rate at which idle journals are removed.

JournalMetadata dual-purpose properties were causing confusion and failures on some use cases. This commit simplifies journal configuration.

bug: server does not multiplex journals correctly, when another copy of same journal subscribes, server does not respond sainly. This is not fixed yet, requires large server work.

bugfix: client gets into inconsistent state after subscribing incompatible journal

- OS-level remove and truncate file methods exposed [OSX so far]
- bugfix