Dmitry Dygalo
Dmitry Dygalo
@caseyjhol First of all, I appreciate your effort! I'd be happy to assist with moving it forward. At this point, the difference between the branches is far too big for...
From my assessment, it looks like the issue is with the input schema. We can improve error reporting for missing references, so the problem is clear from the output.
The solution would be to copy all components to the root schema (prefixing them probably) + rewrite the references so they use the new location instead. Inlining loses the information...
Looks like `pytest-depends` is abandoned. The issue seems to happen due to `pytest-depends` not accounting for tests generated via `pytest_generate_tests` & others. Hence I am going to close the issue
Given that it is not clear how such cases should be serialized, we might just stick to some default and raise an error when we can't decide how to serialize...
Such values are filtered out during generation now
At this point, I don't think it is that beneficial, especially if #1158 is implemented
I tried to reproduce it, but it seems to be fixed by 1bd3f99b. That fix prevents overwriting regular parameters with security definitions, and therefore after that change, the original parameter...
Not something feasible atm
Now there is a way to control what characters are used for strings (globally, not for headers specifically). I also will consider adjusting the defaults for header generation in Schemathesis...