Visual Test Shell for BACnet SVN

Allow arbirary number of detail lines (was 5000)

Moved V3.5.3 scripts under "old".
Replaced scripts with Buddy Lott's updates from SourceForge messages 1704394 and 1704507.
Futher updates are planned to generalize the updated scripts

Make script execution termination terminate ALL steps.
Make script execution terminate without crash when script edit or app is closed.
Rename "halt" to "suspend", and "kill" to "terminate" on menus to reduce ambiguity.
Revise toolbar butoons for suspend/terminate/step

Add a few more services and datatypes

Revise read-all-properties (from EPICS) to omit DER, invoke, BVLL etc. if destination is a Name
- makes for much shorter and more readable scripts
- makes the script reusable on different datalinks, via routers, etc.

Add "Survival Guide"
Small updates to other docs

Add EPICE cleanup code for new object types (fix memory leaks)
Revise and correct BACnetActionCommand parser
Revise property-and-index parser
Remove some cruft

Better handling of the string arrays, with #defines for the MAX_ of each type.
Then put the MAX_ size in the properties for each use, at ->PropET (the otherwise little-used table index).
This allows the statext: handling to be the same regardless of chosen size, per the object's properties.
Added handling for the subordinate-list so it could be parsed in as an array of DeviceObjectReferences instead of list, as it is for (most of?) the uses in other properties.

Rervisions to improve usability with non-local IUT (such as MS/TP device reached via IP-connected router)
- add routed name lookup to summary list: shows Name in source and destination column
- correct errors in NPCI building and router usage (scripting)
- revise dialog-based sending to eliminate inappropriate Names
- remove obsolete SummaryView from build

Declared the filename for EPICS errors as a #define, and moved it from the root of C: to the safer C:\temp folder.
Declared the number of MultiState state_text characterstring array elements as 128 (up from hard-coded 32) to accomodate lots of states.

I was seeing problems with the EPICS Error logfile, as though two separate threads/areas were both writing to it and invalidating the file handle (it was crashing on close).
Not convinced whether this was a debugging artifact or real problem; not seeing it at the end here.