Fix SignalPdu unmarshal size issue

[Serubin]

SignalPdu unmarshal currently relies on using the data bucket size to calculate the unmarshal byte amount. This size is not consistent as the data bucket size may not represent any padded bytes included. Thus using the header size is safer and more consistent.

[leif81]

Thanks @Serubin

I'm a little confused now why we have getLength and getPduLength. Certainly room for confusion if nothing else.

[Serubin]

I absolutely agree. getLength calculates based on the size of the data coming in, whereas getPduLength retrieves the size from the PDU header. Getting the size from header is almost always more reliable.

There may be some use for getLength but it's not in the unmarshal process. Took me a while to determine why my bundled PDUs were consistently starting one byte off from their actual beginning.

[leif81]

Sorry I was away camping for a few days.

Lets merge this as is and we can think about refactoring getLength/getPduLength methods outside of this PR.