Changes for Perl extension Xtract

0.10 to be released
	- If we are building indexes (because the consumer cares about speed)
	  then we now also run ANALYZE during the database cleanup.
	- Separated publish logic into a standalone module.

0.09 Sun  3 May 2009
	- Fleshing out some of the tests a bit more
	- Merged DBIx::Publish down into Xtract.pm
	- Added mysql table support

0.08 Sun  3 May 2009
	- Added LZMA compression support
	- Started migration to Moose

0.07 Sat  2 May 2009
	- The "DBI:" in --from is now optional
	- Rewrote the autotype code (faster and uses less memory)
	- Add support for autotype columns with only NULLs

0.06 Wed 29 Apr 2009
	- Adding timestamp on trace messages
	- Adding dependency on Time::Elapsed for human elapsed times
	- Adding experimental tuning option --sqlite_cache param to boost
	  SQLite memory use
	- Disabled the cleanup VACUUM, it takes an very long time to run
	  (for very little benefit)

0.05 Sun 26 Apr 2009
	- Factored out the indexing and filling code to seperate methods
	- Moved the old ->table to ->select
	- Made a new fast ->table for complete table extraction
	- Allow ->table to have a different implementation depending on
	  the type of the source database driver.
	- select, table and fill now track and return inserted rows
	- Automatically generated a compressed version of the database
	- Add a timer to the publishing run and report elapsed time
	- Use (faster?) UNIQUE indexes if column values are all different
	- Indexing is now a specific command line option --index
	- Added quiet mode --quiet

0.04 Sat 25 Apr 2009
	- Merged DBIx::Publish into this distribution for now

0.03 Sat 25 Apr 2009
	- More fixes
	- Updating to DBIx::Publish 0.03

0.02 Sat 25 Apr 2009
	- Forgot to add the script to the Makefile.PL

0.01 Sat 25 Apr 2009
	- Created the initial version