Hi to all.

As you may noticed I'm developing a new driver for commercial Kenwoods, see details in issue 2999 here:  http://chirp.danplanet.com/issues/2999

I just have completed the last milestone (full bank support, a lot of bug fixes, a lot of validation, PEP8, and new settings) to send the patch to the daily tree, and the code it's stable enough in my testing bed here with 3 radios.

But the run_test.sh script fails, digging around this I found that the culprit is the run_test script it self in he section for testing banks with newly created channels exactly in two places (so far):

<code>
        {set memory empty}
        {assign some basic data to channel}
        mem_banks = model.get_memory_mappings(mem)
        if len(mem_banks) != 0:
            raise TestFailedError("Freshly-created memory has banks and " +
                                              "should not", "Banks: %s" % str(mem_banks))
{.....}

        model.remove_memory_from_mapping(mem, banks[0])
        reason = verify(banks[0])
        if reason is None:
            raise TestFailedError("Memory remains in bank after remove",
                                              reason)

</code>

As you can see in this code segments this assume that a channel may be assigned to a bank or not, but then here come this Kenwood commercial radios to break this:

The kenwood Serie 60G family has a enforcement to every channel: it hast belong to a bank (or groups in kenwood speaking)  

Then, should I patch the run_test script to reproduce the desired behavior?

It's the less traumatic solution from my point of view, but the less elegant.

My python is improving but it's not enough to recreate a bank model class with this enforcement link between, if such provision is made we need to change run_test.sh to reflect this changes.

As I'm new into this community I has to ask to the collective intelligence:

What is the correct solution to this?

Cheers.