Right now, attribute_map is using a regular dict, which means items are ordered starting CPython 3.6 (or Python 3.7), but before that they are not.
This is not a concern in JSON, since the order or node are not relevant per JSON spec, however in XML, attribute order does matter. This means in 2.7 or 3.5, serialization order might be random, and might create XML issue.
Exact impact is unclear yet, but worth creating an issue.
Right now, attribute_map is using a regular dict, which means items are ordered starting CPython 3.6 (or Python 3.7), but before that they are not.
This is not a concern in JSON, since the order or node are not relevant per JSON spec, however in XML, attribute order does matter. This means in 2.7 or 3.5, serialization order might be random, and might create XML issue.
Exact impact is unclear yet, but worth creating an issue.