sdformat: naming conflicts in generated schema files (xsd)
The file world.sdf includes both model.sdf and state.sdf
state.sdf in turn includes model_state.sdf
Both model_state.sdf and model.sdf define a <model> tag:
which in itself shouldn’t be a problem, because they exist in different scopes, one inside <world> and the other inside <state>.
This include chain is reflected in the generated .xsd files; however, the includes are all placed at the top of their respective document, which makes both <xsd:element name='model'> tags appear on the same level inside world.xsd, i.e., the nesting of model_state’s model is not preserved. This - naturally - leads to a conflict and in the current version model_state’s <model> overwrites model’s <model>, which is (very) undesirable.
I’m not 100% sure how to fix this, since I am not familiar enough with the XMLSchema spec, but I assume the solution is something along the lines of renaming state tags, introducing namespaces, or ensuring that model_state’s model is included inside the state element (preferred, but not sure if feasible).
About this issue
- Original URL
- State: open
- Created 3 years ago
- Comments: 16 (9 by maintainers)
An update: I tried this and it seems to work. I (re-)wrote the ruby generation script in python - I’m more familiar with that language and hence faster - and I can now generate xsd which validates and which allows me to generate bindings from it. Are there internal unit tests for the generated xsd I can check against? It also addresses some of the other issues I raised.
If there is interest in this solution, I can contribute this generator upstream.