Closed domanchi closed 3 years ago
slack-github integration test comment
In a quick examination of this issue, it looks like we need to make changes to current fundamental assumptions. Specifically, fuzzer.py
has no concept of which endpoints that it's currently fuzzing for, nor do the factories that are subscribed (https://github.com/Yelp/fuzz-lightyear/blob/master/fuzz_lightyear/fuzzer.py#L223).
Care should be taken when approaching this issue.
Issue
With some swagger specifications, there are "conflicting" input parameter names. That is, two endpoints may take in an
id
, but they refer to different IDs. With the current setup, we would not be able to support this case, because you can only register a factory that maps to a parameter name -- and the same factory would be used for both IDs.Solution
Provide a keyword-only interface to specify a list of test cases the factory applies to. e.g.
The precedence order would need to be adjusted as follows:
Furthermore, the
endpoints
parsing logic should support the same use case as factory registration. That is, it should allow for comma separated values, as well as a list of values to handle.