Fix for issue #151 - Warnings in message decoder test

[mstovenour]

Corrected bug exposed by warning when running MessageDecoder_test.pl. There are still two sets of messages that generate warnings. Those messages are intentionally used because they are not defined in the message definition hashes. It is reasonable to expect warnings given that those messages are not defined. Additional check logic could be added to prevent the warnings and write out error messages but that seems like a little bit overkill for a debug tool.

[jame]

I was curious as to how you resolved this and had a chance to look today. The only two issues I'm still seeing are as follows:

PLM Message: 025a15aabbcc
Use of uninitialized value in concatenation (.) or string at ../Insteon/MessageDecoder.pm line 520.
Use of uninitialized value in multiplication (*) at ../Insteon/MessageDecoder.pm line 521.
Use of uninitialized value in multiplication (*) at ../Insteon/MessageDecoder.pm line 524.
Use of uninitialized value in multiplication (*) at ../Insteon/MessageDecoder.pm line 524.
         PLM Command: (025a) 
  Undefined Cmd Data: (15aabbcc)

PLM Message: 0274000000
Use of uninitialized value in concatenation (.) or string at ../Insteon/MessageDecoder.pm line 520.
Use of uninitialized value in multiplication (*) at ../Insteon/MessageDecoder.pm line 521.
Use of uninitialized value in multiplication (*) at ../Insteon/MessageDecoder.pm line 524.
Use of uninitialized value in multiplication (*) at ../Insteon/MessageDecoder.pm line 524.
         PLM Command: (0274) 
  Undefined Cmd Data: (000000)

If those are the two you were referring to, should not the module still handle such PLM messages (i.e., undefined commands like those two) without errors? Or is that PLM messages like those two are not actually likely to be seen?

[mstovenour]

Those PLM messages will never be seen unless SmartLabs adds a new PLM command in the future (they haven’t added a “PLM” command in the past 8 years). I just added the tests to make sure it didn’t cause the code to abort. If you insist that the tests run clean one option is to remove those two tests J but I hardly think that is warranted here. Another option is to remove the tests all together so that the decoder will be like the remainder of MH (i.e. no tests at all) J

[krkeegan]

Merged as #209