client data values should be a list in the filter pipeline

[bnlawrence]

The filter pipeline is something that in principle is modifiable outside of the hdf library (and hence outside of pyfive). It's certainly something we need to get at for PyActiveStorage.

The spec says within the filter description:

Field Name	Description
Number of Client Data Values	Each filter can store integer values to control how the filter operates. The number of entries in the Client Data array is stored in this field.
Name	If the Name Length field is non-zero then it will contain the size of this field, not padded to a multiple of eight. This field contains a non-null-terminated, ASCII character string to serve as a comment/name for the filter.Filters that are defined in this format documentation (deflate, shuffle, etc.) do not store the Name Length or Name fields.
Client Data	This is an array of four-byte integers which will be passed to the filter function. The Client Data Number of Values determines the number of elements in the array.

Which clearly suggests we should have an array of these things.

Our current code has this:

    ...
    num_client_values = struct.unpack_from('<H', self.msg_data, offset)[0]
    offset += 2
    name = ""
    if name_length > 0:
          name = struct.unpack_from("{:d}s".format(name_length), self.msg_data, offset)[0]
          offset += name_length
    filter_info['name'] = name
    client_values = struct.unpack_from("<{:d}i".format(num_client_values), self.msg_data, offset)
    offset += (4 * num_client_values)
    filter_info['client_data'] = client_values
    filter_info['client_data_values'] = num_client_values
    ...

which produces a tupe in the filter_info['client_data'] evn if there is only one, but somewhere this is turned into an integer so we can expect things to break (and they do).

[bnlawrence]

Oh, for Pete's sake, this is now the consequence of the perfectly correct change we made associated with jjhelmus#66, so the problem is not here at all, it's in the things that made use of the wrong thing from pyfive because we had the wrong thing. Fixed downstream. No problems here.