Updates inlcuded in this patch:

 * When processing FLEXlm debug log data (in filtrator)
   there was a problem when handing check out/in's that 
   lastet for 0 seconds causing over reporting of useage

 * Filtrator updated to makes processing of FLEXlm debug
   log ignore all lines in input that are not directly 
   used. Previously other lines was used to keep track of
   time, but for some reason these lines occationally 
   have wrong times (from #332)

 * Filtrator and filters updated to generate product
   and feature mapping in FLEXlm debug log based data
   (from #332)

 * Flab has been improved to handle input from multiple
   timezones correctly. 

 * Flab updated to ingore the same FLEXlm debug log lines
   as filtrator (and flexlogcollector)

 * config_license_filters updated to handle FLEXlm debug
   log data types (from #332)

This patch fixes correct product and feature mapping in
FLEXlm debug log based data types. 

The path also includes an update of filtrator that
makes processing of FLEXlm debug log ignore all lines
in input that are not directly used. Previously other
lines was used to keep track of time, but for some
reason these lines occationally have wrong times.

This updated archiver binary can be configured to prevent
generation of data (using filtrator) when incoming
data is old. 

By adding a process-threshold object to the "archiver"
object in

	/var/opt/openit/etc/daft.conf

The object has the following syntax:

	process-threshold()=""
	{
		max-age(Integer)="n"{}
		date-limit(Date)="YYYY-MM-DD"{}
	}

If the "max-age" object is present, only the archive files
that are less then "n" days old are run through filtator.

If the "date-limit" object is configured, only archive
files that are from the configured date or newer are run
through filtrator.

If both objects are present, only files satisfying both
thresholds are run through filtrator.

All data is however inserted into the archive, and 
for FLEXlm debug data, the "flab" program is still run
so that data later can be generated without additional
work.

The default setting for this object (e.g when not 
present in configuration file is to process all incoming
archive data through filtrator.

In addition to the issues in the superseeded patches, this
patch fixes a problem when calculating max available for
a product-mapped feature when two or more products mapped
to one super-product had the exact same max available at
any given time.

When licpoll has written data with a slightly too short wait for
offset (i.e if licpoll waits 149 instead of 150 seconds before
polling starts), the processing is disrupted and fails to give
data for full hour.

This patch fixes this behaviour.

This patch replaces the intra process locking mechanism
with thread mutex's instead of semaphores.

The patch also fixes a problem with correct handling 
of time zone for FLEXlm debug log data

The handling of license events type for flex log has also
been corrected (Problem with time zone).

This patch replaces the intra process locking mechanism
with thread mutex's instead of semaphores.

The patch also fixes a problem with correct handling 
of time zone for FLEXlm debug log data

This patch addes generation of correct prime time output for
FLEXlm debug log based data types.

The prime time information is added on the server unlike 
poll based license data where primetime information is added
at collection time. Any changes to the prime time file can
affect output when regenerating data.

This patch adds a lock to the archiver binary that prevents
more than one instance from running simultainously. It also
limits the number of open output files to 8.

In addition the patch allows archiver to accept that the last
field in some recorder data may be empty.

This patches update the archiver binary with a version
that is better at handling seriously corrupt input
files.

This patches update the archiver binary to be able to handle
strings that contain unprintable characters.