Use of obj.name in NocaseDict() init

[andy-maier]

Use of obj.name in NocaseDict() init

[andy-maier]

We agreed to determine from the pywbem code how often the functionality is used, grouped by (probably) pywbem client, mof compiler, mock support, test.

[andy-maier]

Here is the result of the investigation: The only place in pywbem where this capability used is in exactly 3 places in the unit test of the cim_obj.py module:

• tests/unittest/pywbem/test_cim_obj.py:6074: Initializing NocaseDict with iterable of CIMQualifier
  qualifiers=NocaseDict([
• tests/unittest/pywbem/test_cim_obj.py:6096: Initializing NocaseDict with iterable of CIMQualifier
  qualifiers=NocaseDict([
• tests/unittest/pywbem/test_cim_obj.py:6140: Initializing NocaseDict with iterable of CIMProperty
  properties=NocaseDict([

In the pywbem code itself, the capability is not used at all.

Still, it is a quite convenient capability for users and I think we should somehow retain it. One idea on that would be to provide the capability as a mixin class (say Keyable), so that users could easily construct a dict class with the capabilities they want:

class MyKeyableNocaseDict(NocaseDict, Keyable):
    pass

One could probably even do it such that the key attribute name can be specified:

class MyKeyableNocaseDict(NocaseDict, Keyable(attr='name')):
    pass

The same approach could be chosen for hashability:

class MyHashableNocaseDict(NocaseDict, Hashable):
    pass

And of course one could mix the capabilities.

DISCUSSION