More detail about the implementation. Currently, namespaces are accessed directly with C functions which expect to get an exact dict type: these functions are designed for performance, not to handle dict subclasses, on purpose.

Supporting dict subclasses first require to check if the namespace type is dict: that's the first performance overhead. Then new code should be added to handle subclasses: that's the additional maintenance burden. This code path should be tested, maintained, reviewed, etc.

The risk is also to forget to support subclasses in one function which would break the feature. Same problem is new code is added which doesn't support dict subclasses.

This feature is far from being free to implement. It's very expensive on multiple aspects.