Jim Cornette wrote:
Is this similar to what the Fedora Automated Test Suite [...] aims to accomplish?
Both share the goal of making software suck less. :-) The key difference is that the Fedora Automated Test Suite uses an explicit, fixed, human-designed battery of tests. CBI treats regular daily usage as the "test suite", where any run that ends in failure (e.g. crash) is taken as an example of a failed test.
These are complementary approaches, of course. A well-designed automated test suite is extremely valuable but can be hard to create. CBI offers a different view of things, with less developer steering and more focus on ordinary day-to-day usage.
Having automated programs to diagnose problems does sound like it will aid in getting more information back to improve Linux program versions.
Agreed!
