Revision history for Perl extension POE::Component::SimpleDBI.

* 1.20

	Added the INSERT_ID to control $dbh->last_insert_id usage

* 1.19

	Added the PREPARE_CACHED argument to control caching

* 1.18

	Ignore the DBI error for last_insert_id and default to undef

* 1.17

	Added "INSERTID" to the result of DO

* 1.16

	Noticed a glaring documentation bug
		- SINGLE queries return mixedCaps rows ( not lowercase! )
		- MULTIPLE queries return lowercase rows

	This makes me *VERY* tempted to fix SINGLE to return lowercase, is this a good idea? Let me know!

	Fixed SimpleDBI failure on Win32 - thanks RT #23851

* 1.15

	Thanks to Fred Castellano, who stumbled on a DEADLOCK bug, fixed!
	Added sanity tests to CONNECT/DISCONNECT

* 1.14

	learned about the difference between ref $self and ref( $self )
	Kwalitee-related fixes

* 1.13

	Finally use a Changes file - thanks RT #18981
	Fixed a bug in SINGLE if returned_rows = 0 it will not return undef, but give us blank rows!
	Documentation tweaks

* 1.12

	In the SubProcess, added a binmode() to STDIN and STDERR, for the windows attempt
	Added code to make SimpleDBI work in Win32 boxes, thanks to the recent Wheel::Run patches!
	Documentation tweaks as usual

* 1.11

	Hannes had a problem:
		His IRC bot logs events to a database, and sometimes there is no events to log after
		hours and hours of inactivity ( must be a boring channel haha ), the db server disconnected!

	The solution was to do a $dbh->ping() before each query, if your DBI driver does it inefficiently, go yell at them!
	In the event that a reconnect is not possible, an error will be sent to the CONNECT event handler, look at the updated pod.

* 1.10

	Fixed a bug in the DO routine, thanks to Hannes!

* 1.09

	Removed the abstract LIMIT 1 to the SINGLE query

	Removed the silly 5.8.x requirement in Makefile.PL

	Made the SubProcess use less memory by exec()ing itself

	Added the new CONNECT/DISCONNECT commands

	Removed the db connection information from new()

	Minor tweaks here and there to not stupidly call() the queue checker when there is nothing to check :)

	Added the sysreaderr debugging output

	More intelligent SQL/PLACEHOLDERS/BAGGAGE handling

	Made the command arguments more stricter, it will only accept valid arguments, instead of just extracting what it needs

	Made sure all return data have ID/EVENT/SESSION/ACTION in them for easy debugging

	Added the SESSION parameter to all commands for easy redirection

	Updated the POD and generally made it better :)

	Added a new command -> Clear_Queue ( clears the queue )

* 1.08

	In the SubProcess, removed the select statement requirement

* 1.07

	In the SubProcess, fixed a silly mistake in DO's execution of placeholders

	Cleaned up a few error messages in the SubProcess

	Peppered the code with *more* DEBUG statements :)

	Replaced a croak() with a die() when it couldn't connect to the database

	Documented the _child events

* 1.06

	Fixed some typos in the POD

	Added the BAGGAGE option

* 1.05

	Fixed some typos in the POD

	Fixed the DEBUG + MAX_RETRIES "Subroutine redefined" foolishness

* 1.04

	Got rid of the EVENT_S and EVENT_E handlers, replaced with a single EVENT handler

	Internal changes to get rid of some stuff -> Send_Query / Send_Wheel

	Added the Delete_Query event -> Deletes an query via ID

	Changed the DO/MULTIPLE/SINGLE/QUOTE events to return an ID ( Only usable if call'ed )

	Made sure that the ACTION key is sent back to the EVENT handler every time

	Added some DEBUG stuff :)

	Added the CHANGES section

	Fixed some typos in the POD

* 1.03

	Increments refcount for querying sessions so they don't go away

	POD formatting

	Consolidated shutdown and shutdown_NOW into one single event

	General formatting in program

	DB connection error handling

	Renamed the result hash: RESULTS to RESULT for better readability

	SubProcess -> added DBI connect failure handling

* 1.02

	Initial release