In order to implement fuzzing tests to the CLI, we could generate random config sets to catch potential errors that other tests could not catch.
Description
Description
Details
Details
- Difficulty level
- Unknown (require assessment)
- Version
- -
- Why the issue appeared?
- Will be filled on close
- Is it a breaking change?
- Perfectly compatible
- Issue type
- Internal change (not visible to end users)
Event Timeline
Comment Actions
The idea is feasible for parameters with constraints (like number ranges) defined in the XML, but there are many other cases where human input is necessary. We could give the tool a set of parameters to randomly generate, or a half-complete config with slots to fill in with random values. Worst case, we'd discover new constraints for more rigid templates; best case, we'd have a proper tool for generating corner cases for smoke tests and fuzzing.