Multiplexer coding fixes

[andlaus]

This PR fixes the en- and decoding routines of multiplexers. The mistake was that we assumed that multiplexer.byte_position specifies the start position of the whole multiplexer object while in reality it only specifies the position of its content part.

Besides this, the PR contains a few smaller quality-of-life improvements:

• DiagLayer.decode() now re-raises encountered DecodeErrors if the binary data is also not decodable as a global negative response. The old behavior accidentally was to never raise DecodeError.
• The values of multiplexer parameters can now be specified using a 2-tuple or a list of length two in addition to a dictionary with a single entry
• The key of multiplexer parameters can now be specified by their raw integer value or the MultiplexerCase object in addition to the short name of the selected case
• The default case of multiplexers can now be selected by specifying an integer value which is not featured by any of the "ordinary" cases.

Andreas Lauser <andreas.lauser@mercedes-benz.com>, on behalf of MBition GmbH. Provider Information

[andlaus]

okay, I added a unit test for this functionality (I had to fix up a few mistakes in test_diag_data_dictionary_spec first) and rebased it on top of #319...

[andlaus]

@kayoub5: do you see any hurdles for merging this?

[kayoub5]

@andlaus could you publish a new version with this change?

[andlaus]

done