CAIRO: table reader tail cursor can now go to bottom of data, which is useful when tailing code needs to access only additional data and not existing

GRIFFIN: fixed memory leak from 'sample by' factories, also fixed timestamp propagation via sub-queries

CUTLASS: started porting HTTP server components, removed "tuck" - previously attempted port. It is far easier to port Http server component by component rather than all at once. Started with HttpHeaderParser.

- 'select a.x from tab a' did not work because of redundant use of table alias
- 'select * from (SELECT' was sending parser into a stack overflow frenzy because of inconsistent case sensitivity

CAIRO: TableReader no longer relies on partition directories to be removed in order to remove logical data chunk. This is necessary for Windows where open files cannot be removed.

CAIRO: TableWriter maintains minTimestamp so that TableReader doesn't have to count directories to determine data interval

CAIRO: TableWriter.truncate() is unable to remove directories on Windows when TableReader has files in these directories open. truncate() requires major rethink. In this commit i just disabled relevant tests on Windows only. Linux and OSX are fine.

CAIRO: fixed transactional behaviour of TableWriter.truncate(). TableReader will now process this event correctly.