-
Notifications
You must be signed in to change notification settings - Fork 10
Open
Description
The problem: we have too many things to test! For example:
- Stream into Node A, playback from Node B, for every node A + B — this is
O(nodes^2) - We want to test both long duration and short duration streams
- Recording enabled vs not
- Playback policy enabled vs not
- Multistream enabled vs not
Doubling the number of continuous automated tests every time we want to test a new feature is infeasible. Instead, we should implement some "fuzzing" — randomize the test parameters every time we want to boot up a new test. This will ensure we have coverage of all possible feature combinations without going over all of them exhaustively.
Metadata
Metadata
Assignees
Labels
No labels