new file:   AD2USB.pm

	modified:   AD2USB.pm

- Allows for use of all Generic_Item options including tie_events
- Remove states because this is not an "editable" object

Tabs are set at 3 spaces which is odd

Why are we hardcoding $Self?

Can't use a global variable, otherwise users would not be able to have multiple panels setup in MH.

Greatly simply DefineCmdMsg mess.

Make new() sub much easier to read, get rid of verbose coding.

Remove all changes to Config Hash, this should be read-only as it represents the contents of the ini file, this contents can be updated by MH

No need for this if regex includes start and end of line

No need for this

Also cleanup parser to make a little easier

…CheckCmd, Removed Switch Package

Returning a hash makes the code infinitely more readable, no longer need to remember the numbers for various message types.

Removed Switch to make compatible with Perl 5.8.8

Relying on substr is dangerous, particularly since a user can enable the !KPM: prefix.

Add more message parsing to GetStatusType, this routine can likely be converted to parse_message and save a lot of code

Get rid of all of the complicated tests.

Allow ChangeZones to loop around from a high zone to a low zone.

NOTE: May need to add a check to ChangeZones to see if it is reseting a mapped zone.

Anytime we get a keypad message we can check the status bits.

Stop using substr and use hash flags.

Wow, can't believe I missed that.

Using hash keys makes the code much more condensed and easier to read.

I have to admit I don't really understand all of the optional settings for wireless objects

Not clear to me why we don't reset mapped zones.  If it is necessary, we should be adding something to ChangeZones, not burying it all over the place.

Anytime it was set, it was being immediately cleared by ResetAdemco

These were always reset by ResetAdemco or not ever used

…e_Last_Num

First three served no purpose.

Zone_Last_Num is used only to keep track of state in fault loop.

Not sure how they work exactly, so everything defaults to 1

Change all calls to config_parms to use this method

No more hardcoding $Self!!

That was an awful idea.

Add an object for a partition.

Cleanup unused function sof door and motion items.  Some odd functions still remain.

This is only a start.

Mapped zones are faulted by direct messages received from the AD2.

This correction fixes the following oddity:

-------

Zones 1 and 3 are faulted.

The Alpha message describing the Zone 3 fault just arrived.

Zone 10, a mapped zone is then faulted.  The fault is immediately detected by an EXP! message.  MH updates the status of zone 10 to faulted.

AD2 then starts over with the Alpha fault list by sending the Alpha message that Zone 1 is faulted.

Since the last alpha fault received was zone 3, MH resets all zones between 3 -> 1 as ready, clearing out 10.

Two messages later, the Alpha message for zone 10 arrives, putting zone 10 back in a faulted state.

-------

This fixes the oddity by not reseting any mapped zones based on alpha messages.

The alphanumeric fault messages are more complicated than i realized.  The list of faulted zones restarts at the lowest faulted zone everytime a fault occurs.  So for example, if zones 1-5 are faulted, and the panel has just finished sending the zone 3 fault message, if zone 7 is then faulted, the next message will be a zone 1 fault again.

If we blindly perform wrap-around resets, it is very easy to create a "fluttering" zone when multiple zones are faulted.

To cut down on the potential for fluttering, this change only performs a wrap-around reset if the highest faulted zone that is reported remains constant for 2 cycles.  While not completly ensuring that all zones have been report, it is unlikely that the panel will be interrupted at exactly the same point in its cycle twice.

The major downside of this, is that there may be a slight delay in reseting zones from fault -> ready.  This is particularly true of the higher unmapped zones.

Try to optimize code as much as possible.

On a dual core machine, using AD2 code seems to add between .3-.7% CPU usage to one of the cores.  I think this is mostly a side effect of using a socket.

…eclaration

The old design for the ini declaration was a little confusing.  Plus, carrying it over directly into the Configation Hash did not allow for mht style declarations of a wireless object.

The new design seems more logical to me and adds the framework to support an mht style declaration.

The old design for ini declarations is stil supported for backwards compatibility.

…or MHT Declarations

They are essentially identical.  This gets rid of the duplicative code.

Added framework to support mht declaration of wireless, expander and relay items.

All references to AD2USB has been replaced with AD2.

This means all configuration settings and user code settings need to be updated as well.

The change was necessary because this module will work with at leaast AD2Serial, AD2USB, AD2Pi, and likely many more AD2 style products that will be released in the future.

…lity

Cannot use comma in mht definitions

Keyfob was not likely ever set to fault, since fault and ready were set in same loop.

…in Partition

Checking for faults using the Alphanumeric messages is really complicated.  The concept sounds simple, but building a test that doesn't get confused is difficult. I think this current system is more accurate, but a little slower to move from fault->ready.

I added a lot more comments to explain my logic since code alone is not really enough.

Also fixed a error in the partition child.