0.1.0 - Dec 6, 2006
* Initial release

0.2.0 - Dec 7, 2006
* Added an XML parser for source parsing
* Added support for compound key constraints in destinations via the :unique => [] option
* Added ability to declare explicit columns in bulk import
* Added support for generators in destinations
* Added a SurrogateKeyGenerator for cases where the database doesn't support auto generated surrogate keys

0.3.0 - Dec 19, 2006
* Added support for calculated values in virtual fields with Proc

0.4.0 - Jan 11, 2006
* Added :skip_lines option to file source configurations, which can be used to skip the first n lines in the source data file
* Added better error handling in delimited parser - an error is now raised if the expected and actual field lengths do not match
* Added :truncate option for database destination. Set to true to truncate before importing data.
* Added support for :unique => [] option and virtual fields for the database destination

0.5.0 - Feb 17, 2007
* Changed require_gem to gem and added alias to allow for older versions of rubygems.
* Added support for Hash in the source configuration where :name => :parser_name defines the parser to use and
  :options => {} defines options to pass to the parser.
* Added support for passing a custom Parser class in the source configuration.
* Removed the need to include Enumerable in each parser implementation.
* Added new date_to_string and string_to_date transformers.
* Implemented foreign_key_lookup transform including an ActiveRecordResolver.
* Added real time activity logging which is called when the etl bin script is invoked.
* Improved error handling.
* Default logger level is now WARN.

0.5.1 - Feb 18, 2007
* Fixed up truncate processor.
* Updated HOW_TO_RELEASE doc.

0.5.2 - Feb 19, 2007
* Added error threshold.
* Fixed problem with transform error handling.

0.6.0 - Mar 8, 2007
* Fixed missing method problem in validate in Control class.
* Removed control validation for now (source could be code in the control file).
* Transform interface now defined as taking 3 arguments, the field name, field value and the row. This 
  is not backwards compatible.
* Added HierarchyLookupTransform.
* Added DefaultTransform which will return a specified value if the initial value is blank.
* Added row-level processing.
* Added HierarchyExploderProcessor which takes a single hierarchy row and explodes it to multiple rows
  as used in a hierarchy bridge.
* Added ApacheCombinedLogParser which parses Apache Combined Log format, including parsing of the 
  user agent string and the URI, returning a Hash.
* Fixed bug in SAX parser so that attributes are now set when the start_element event is received.
* Added an HttpTools module which provides some parsing methods (for user agent and URI).
* Database source now uses its own class for establishing an ActiveRecord connection.
* Log files are now timestamped.
* Source files are now archived automatically during the extraction process
* Added a :condition option to the destination configuration Hash that accepts a Proc with a single
  argument passed to it (the row).
* Added an :append_rows option to the destination configuration Hash that accepts either a Hash (to
  append a single row) or an Array of Hashes (to append multiple rows).
* Only print the read and written row counts if there is at least one source and one destination
  respectively.
* Added a depends_on directive that accepts a list of arguments of either strings or symbols. Each
  symbol is converted to a string and .ctl is appended; strings are passed through directly. The
  dependencies are executed in the order they are specified.
* The default field separator in the bulk loader is now a comma (was a tab).

0.6.1 - Mar 22, 2007
* Added support for absolute paths in file sources
* Added CopyFieldProcessor

0.7 - Apr 8, 2007
* Job execution is now tracked in a database. This means that ActiveRecord is required regardless
  of the sources being used in the ETL scripts. An example database configuration for the etl can
  be found in test/database.example.yml. This file is loaded from either a.) the current working
  directory or b.) the location specified using the -c command line argument when running the
  etl command.
* etl script now supports the following command line arguments:
** -h or --help: Prints the usage
** -l or --limit: Specifies a limit for the number of source rows to read, useful for testing
   your control files before executing a full ETL process
** -o or --offset: Specified a start offset for reading from the source, useful for testing your
   control files before executing a full ETL process
** -c or --config: Specify the database.yml file to configure the ETL execution data store
** -n or --newlog: Write to the logfile rather than appending to it
* Database source now supports specifying the select, join and order parts of the query.
* Database source understands the limit argument specified on the etl command line
* Added CheckExistProcessor
* Added CheckUniqueProcessor
* Added SurrogateKeyProcessor. The SurrogateKey processor should be used in conjunction with the 
  CheckExistProcessor and CheckUniqueProcessor to provide 
* Added SequenceProcessor
* Added OrdinalizeTransform
* Fixed a bug in the trim transform
* Sources now provide a trigger file which can be used to indicate that the original source
  data has been completely extracted to the local file system. This is useful if you need to
  recover from a failed ETL process.
* Updated README

0.7.1 - Apr 8, 2007
* Fixed source caching

0.7.2 - Apr 8, 2007
* Fixed quoting bug in CheckExistProcessor

0.8.0 - Apr 12, 2007
* Source now available through the current row source accessor.
* Added new_rows_only configuration option to DatabaseSource. A date field must be specified and
  only records that are greater than the date value in that field, relative to the last successful
  execution, will be returned from the source.
* Added an (untested) count feature which returns the number of rows for processing.
* If no natural key is defined then an empty array will now be used, resulting in the row being
  written to the output without going through change checks.
* Mapping argument in destination is now optional. An empty hash will be used if the mapping
  hash is not specified. If the mapping hash is not specified then the order will be determined
  using the originating source's order.
* ActiveRecord configurations loaded from database.yml by the etl tool will be merged with
  ActiveRecord::Base.configurations.
* Fixed several bugs in how record change detection was implemented.
* Fixed how the read_locally functionality was implemented so that it will find that last
  completed local source copy using the source's trigger file (untested).
  
0.8.1 - Apr 12, 2007
* Added EnumerableSource
* Added :type configuration option to the source directive, allowing the source type to be 
  explicitly specified. The source type can be a string or symbol (in which case the class will
  be constructed by appending Source to the type name), a class (which will be instantiated
  and passed the control, configuration and mapping) and finally an actual Source instance.
  
0.8.2 - April 15, 2007
* Fixed bug with premature destination closing.
* Added indexes to execution records table.
* Added a PrintRowProcessor.
* Added support for conditions and "group by" in the database source.
* Added after_initialize hook in Processor base class.
* Added examples directory

0.8.3 - May 13, 2007
* Added patches from Andy Triboletti