Fix in the C++ DateTime class. The affected server files are included.

The following is taken from patch #537: After installation, if not
done before, the old flab files in the archive must be removed (as
openit user):

cd $ARCHIVE_DIR/license
find flexpoller -name "*FLAB*" | xargs rm

$ARCHIVE_DIR can be found in /etc/opt/openit/openit.cfg

The flab must be run for the affected period. Please proceed as
follows (as openit user):

/opt/openit/bin/flab -c /var/opt/openit/etc/daft.conf <start> <end>

<start> and <end> has format YYYY-MM-DD

Some bogus lines in the FLEXlm debug log combined with the 
wrong timestamps could cause errors in the output from 
flab, and thus the report data for the FLEXlm debug log
based data types.

Corrected package registration for features distributed under packages
without OPTIONS=SUITE.

Fixes a bug in flab. flab now correctly merges status lines from the
FLEXlm debug log and flab/flexpoller files regardless of wether
daemon/version/max available licenses information is available. flab
will no longer produce duplicate entires for same license use when
information level varies depending on source.

After installation, if not done before, the old flab files in the
archive must be removed (as openit user):

cd $ARCHIVE_DIR/license
find flexpoller -name "*FLAB*" | xargs rm

$ARCHIVE_DIR can be found in /etc/opt/openit/openit.cfg

The flab must be run for the affected period. Please proceed as
follows (as openit user):

/opt/openit/bin/flab -c /var/opt/openit/etc/daft.conf <start> <end>

<start> and <end> has format YYYY-MM-DD

NOTE: The following (server) platform does not yet support this
functionality:

space-sun-solaris2.5.1

Archiver does now set non-zero usage when splitting samcollector
records into multiple new records.

Update for handling server functionality of new LicenseOptimizer data
type (92), supporting actions: Terminate, Quit, Suspend, and Log.

NOTE: The following (server) platform does not yet support this
functionality:

space-sun-solaris2.5.1

NOTE-2: The functionality must be turned on with the console.

New rewritten flab missed support for UPGRADE lines in the FLEXlm
debug log. This patch includes update of binary that adds this.

After installation, if not done before, the old FLAB files in the
archive must be removed (as openit user):

cd $ARCHIVE_DIR/license
find flexpoller -name "*FLAB*" | xargs rm

$ARCHIVE_DIR can be found in /etc/opt/openit/openit.cfg

The flab must be run for the affected period. Please proceed as
follows (as openit user):

/opt/openit/bin/flab -c /var/opt/openit/etc/daft.conf <start> <end>

<start> and <end> has format YYYY-MM-DD

Because flab does not sort all events internally, there can be some
delayed data that should be appended to existing output files, but
files were sometimes overwritten instead of beeing appended to. This
has been fixed.

Shutdown events from vendor daemons were ignored. This has also been
fixed.

After installation, if not done before, the old FLAB files in the
archive must be removed (as openit user):

cd $ARCHIVE_DIR/license
find flexpoller -name "*FLAB*" | xargs rm

$ARCHIVE_DIR can be found in /etc/opt/openit/openit.cfg

The flab must be run for the affected period. Please proceed as
follows (as openit user):

/opt/openit/bin/flab -c /var/opt/openit/etc/daft.conf <start> <end>

<start> and <end> has format YYYY-MM-DD

flab and filtrator has been rewritten to share the processing of
FLEXlm debug log data (flexlog2) differently. This is done mainly for
simplifying maintenance.

After installation, the old FLAB files in the archive must be removed
and the flab must be run for the affected period. Please proceed as
follows (as openit user):

cd $ARCHIVE_DIR/license
find flexpoller -name "*FLAB*" | xargs rm
/opt/openit/bin/flab -c /var/opt/openit/etc/daft.conf <start> <end>

$ARCHIVE_DIR can be found in /etc/opt/openit/openit.cfg
<start> and <end> has format YYYY-MM-DD

Filtrator now correctly processes FLEXlm debug log data when the
case-sensitity options (username-forcelower and hostname-forcelower)
are set to "true".

Filtrator can now ignore flexpoller events that might conflict with
usage events from the FLEXlm debug log.

When to ignore, is set with:

ignore-flexpoller-usage(String)="<value>"{}

The <value> can be:

   <value>  = Never ignore
   period = Ignore when debug events occur within the poll period
   always = Always ignore

FLAB: A bug resulted in false detection of license server shutdown,
causing some license usage to be dropped.

flab now accepts the option "-d <prod-list>", where
<prod-list> is a comma separated list (without space) with names
of products. Old flab data for these products will be deleted.

filtrator will now reset all licenses in use to zero when a "SHUTDOWN
request" is read from the flexlog file.

Fixed bug causing events at 0:00:00 to be counted twice.
License clients, flexlogcollector.

Sometimes when filtrator processed OUT + IN usage across GMT midnight,
data could be missing. This has now been fixed for these platforms:

NOTE: The AIX4.2 binary will not be included.

When processing FLEXlm debug data, a weakness in the sorting algorithm
caused certain events to occur out of order, resulting in incorrect
usage reported. This has now been fixed.

NOTE: The AIX4.2 binary does not support FLEXnet debug log data.

Will now include debug events occurring after license server startup
but prior to first flexpoller/flab event. (These debug events would
previously be ignored.)

Corrected mapping to local time in stead of GMT.

Fixed bug in FlexLog Archive Builder which sometimes caused usage to
be lost when license daemon was shut down.

Fixes bug in prime time correction of licpoll data in filtrator.

Fixes bug in the flexlog-count-zero-as-one functionality, causing inaccurate
measures when multiple OUT and/or IN lines appeared in the FLEXlm debug log
with the same time stamp.

flab doesn't leave tmp files in TEMP_DIR.

filtrator has added logging of flexlog processing: To activate flexlog
process logging, edit /var/opt/openit/etc/daft.conf, and change the
value of daft.filtrator.log.processing.flexlog2 from 'false' to
'true'.

Fixed bug which in rare cases caused wildly high number of licenses
in use at daemon shutdown.

Corrected an error which in some special cases caused
license usage to be over-reported.

Added support for handling "REStarted" lines in FLEXlm output.
Better handling of simultaneous IN and OUT lines.
Better handling of partially completed Optimizer data logs.

Added support for license usage where a system service is
intermediating license distribution between the FLEXlm server and the
application.

After the patch is installed modify

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

and add line

  intermediate(String)=""{}

in object

  filtrator.corrections

To activate this functionality, edit /var/opt/openit/etc/daft.conf. Locate
the filtrator.corrections.intermediate object, and set it's value to reflect
the user + host intermediating the license usage.
Example:

        intermediate(String)="SYSTEM+myhost"{}

Where 'myhost' is the host, and 'SYSTEM' is the user.

Both user and host must be written as they appear in the lmstat output.

Fixed bug causing usage lasting longer than 4 hours to be dropped when
producing fewer than 6 FLAB-files per day.

This patch updates filtrator.

1: Added support for correcting usage loss caused by conflict
between flexlogcollector data and flexpoller data.

2: Removed deamon as a criterion for processing flexlogcollector
and flexpoller files.

After installing, use

  /opt/openit/bin/regenerate_data --archtype licpoll2 <start> <end>

to regenerate data (where <start> and <end> are dates on YYYY-MM-DD
format).

This update fixes and error when parsing hostnames from FLEXlm
debug log when generating the License Events data type. Hostnames
on the form

  HostA  [HostA1123]

will now correctly be parsed just as "HostA".

When events occured exactly at the same time in FLEXlm debug
log file and in the accompanying flexpoller file, a wrong
test (<= instead of <) would make filtrator ignore the 
event from the FLEXlm debug log file.

This error would be very rare in real data and was detected
in manually constructed data when testing other issues.

Only when using FLEXlm license data collection does the
installation "force" users to define product names. When installing
e.g IBM LUM and Elan/Abaqus the default name is most likely 
used (i.e IBM_LUM and ABAQUS). This is not a problem if there
are never any overlapping features on the licenses servers, but
this is often not the case.

This update allows filtrator to automatically rename the default
product names when they are encountered, from IBM_LUM to 
IBM_LUM-<host>, from ABAQUS to ABAQUS-<host> and from ClearCase
to ClearCase-<host>, to avoid trying to wrongly merge data from
different license servers.

To activate this edit

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

and in object

  filtrator.corrections

set

  default-product-adjustment(Bool)="true"{}

Please note that if there are redundant klients for e.g IBM_LUM
you can not use this options without having configured the LUM
product names to a shared name different from the default to get
the correct behaviour.

This patch adds functionallity that allows processing of FLEXlm
debug log data with check out and in lasting 0 seconds to be
accounted with a 1 second duration.

After the patch is installed modify

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

and add line

  flexlog-count-zero-as-one(Bool)="true"{}

in object

  filtrator.corrections

Back-end processing (flab) of FLEXlm debug log data would only read the
first section of data when multiple section where collected in the same
log file.

This patch fixes two problems with processing of license data


Multiple license agents with same polling offset
------------------------------------------------

If two or more license client poll the same product but do 
not have disting polling times (different offset) the
back-end has problems separating the data from the two 
correctly which will result in incorrect data.

After installing, use

  /opt/openit/bin/regenerate_data --archtype licpoll2 <start> <end>

to regenerate data (where <start> and <end> are dates on YYYY-MM-DD
format).


Improved parsing of data for License logfile events data type
-------------------------------------------------------------

Now ignoring "comments" when parsing FLEXlm debug for this data
type. They occationally contains wildly wrong dates which would
cause problems when parsing.

This patch updates the FLEXlm debug log back-end processing
program (flab) with a binary that can work with less frequent
status dumps than the current "one file per hour" scheme.

This improves both speed of this binary and also reduced disk
use (fewer files generated).

To get any benefit the following must be done after the
patch has been installed.

Edit

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

and change value of object

  flab.build.files-per-day

from "24" to "2".

You may also remove all FLAB files from the archive
and re-run flab over the whole periode where there
is flexlog2 data.

Note that newer releases will have the default value
changed from 24 to 2.

This patch update the back-end processing to correctly read
the "case-sensitivity" object entries. A bug in the code
caused the data reader to use the wrong config value for
flab/flexpoler data (it used the value of the 
"username-forcelower" instead of the "hostname-forcelower"
value).

This patch fixes the followings bugs found in the
license processing binary (filtrator):

USER_REMOVED lines
------------------

Filtrator now correctly handles USER_REMOVED lines
from FLEXlm debug log files.

License mapping files
---------------------

A bug in the parsing of the mapping file could cause
filtrator to crash when parsing 

  license-user-mapping

or

  license-host-mapping

LUM host names
--------------

A bug in license data collection (fixed by #371) caused LUM
node names to include the serial numbers, giving hostnames
like this:

  CAD-NODE-1.us(#194.19.77.134) Serial Number_ 72211XB1934

This patch updates the back-end processing so it filters
out that serial number for the hostnames in collected data.

After installing the patch regenerate data with

  /opt/openit/bin/regenerate_data

to get correct hostnames in the database.

This patch fixes the followings bugs found in the
license processing binary (filtrator):

USER_REMOVED lines
------------------

Filtrator now correctly handles USER_REMOVED lines
from FLEXlm debug log files.

License mapping files
---------------------

A bug in the parsing of the mapping file could cause
filtrator to crash when parsing 

  license-user-mapping

or

  license-host-mapping

LUM host names
--------------

A bug in license data collection (fixed by #371) caused LUM
node names to include the serial numbers, giving hostnames
like this:

  CAD-NODE-1.us(#194.19.77.134) Serial Number_ 72211XB1934

This patch updates the back-end processing so it filters
out that serial number for the hostnames in collected data.

After installing the patch regenerate data with

  /opt/openit/bin/regenerate_data

to get correct hostnames in the database.

A serious problem has been detected when processing
several debug log files simultaniously when some of the
logs has long gaps (periods where license servers has
been down).

In addition the processing is now able to handle FLEXlm
debug log data collected from host in different time
zones.

A bug in license data collection (fixed by #371) caused LUM
node names to include the serial numbers, giving hostnames
like this:

  CAD-NODE-1.us(#194.19.77.134) Serial Number_ 72211XB1934

This patch updates the back-end processing so it filters
out that serial number for the hostnames in collected data.

After installing the patch regenerate data with

  /opt/openit/bin/regenerate_data

to get correct hostnames in the database.

License data collected on windows clients sometime last
for longer than 1 hour (with overlapping collectors).

This patch make filtrator accept these files correctly.

A bug in license data collection (fixed by #371) caused LUM
node names to include the serial numbers, giving hostnames
like this:

  CAD-NODE-1.us(#194.19.77.134) Serial Number_ 72211XB1934

This patch updates the back-end processing so it filters
out that serial number for the hostnames in collected data.

After installing the patch regenerate data with

  /opt/openit/bin/regenerate_data

to get correct hostnames in the database.

Data from previous or next day may have to be processed for
completeness. This was sometimes omitted by archiver. This should now
be fixed.

A subtle error caused some archive files not to be recognized by
archiver, thereby being rejected. This patch fixes the file
recognition algorithm.

Improved automatic processing of recently archived flexlog data
(running FLAB with better parameter control).

New name for flexlog files in archive:

<logfile-time>-<archive-time>-<product>-<running-number>.in

Improves handling of files of test data type 88.

See also README of patch 455.

Allow for configuring what should happen to incoming archive files of
a format that is no longer supported.

In /var/opt/openit/etc/daft.conf, the following objects control this:

"archiver.unsupported" defines target directory

"archiver.unsupported-handling" defines action to be taken, one of
"delete", "move", "ignore"

This patch fixes archiver so that it properly moves corrupt
files from incoming to /var/opt/openit/temp/garbage/

It also fixed an issue which could cause archiver to skip handling
some files in database/incoming/

This patch fixes a problem with loading recorder data.
Archiver failed to accept that last field in line
could have an empty value.

This patch also include a recover option to help
with reprocessing previously failed recorder data.

To do this, manually run the following command
as the "openit" user:

  /opt/openit/bin/archiver -c /var/opt/openit/etc/daft.conf --recover

This will scan the files in the gargabe directory
(usually /var/opt/openit/temp/gargage), check if 
files "failed" with empty last field values and 
if so, insert the missing lines into the archive
and delete the processed file from gargage.

In addition to inserting the data into the archive
you will need to regenerate the recorder data
into the database. Run

  /opt/openit/bin/regenerate_data --archtype recorderevents <start> <end>

and

  /opt/openit/bin/regenerate_data --archtype recorderperiod <start> <end>

(where <start> and <end> are dates on YYYY-MM-DD format).

To void blocking normal archiver operations, it's possible to limit 
the running time of the recover operation by adding the parameter

  --timelimit <time>

where time is given using the P<period> format, e.g

	P2H		2 hours
	P90M		90 minutes 
	P900S		900 seconds

This patch fixes a problem with loading recorder data.
Archiver failed to accept that last field in line
could have an empty value.