Add a filtering capability to CCSDS parsers.

This feature was discussed in this forum thread.

The filtering capability could be used to:

1. parsing malformed CCSDS messages with some known small discrepancies that can be fixed
2. remove (ignore) unwanted data
3. add data that was not originally present in the files

The filtering capability would be very low level, at lexical analysis. At this stage, only one ParseToken is available and can be replaced by a list of replacement tokens. The list would contain only the original token when no filtering is needed on it, it could be one replacement token in case 1, it could be an empty list in case 2, or it can be a list with original token plus additional tokens in case 3.

This feature changes a public interface (MessageParser), hence it can be done only in a major release.