| |||||||||||||
KS2012: The future of kernel regression trackingKS2012: The future of kernel regression trackingPosted Aug 29, 2012 23:22 UTC (Wed) by raven667 (subscriber, #5198)In reply to: KS2012: The future of kernel regression tracking by dlang Parent article: KS2012: The future of kernel regression tracking
Maybe this is just a different usage of the term "regression". In kernel-land a regression as you describe is where a change negatively affects performance in some unique scenario, Linux kernel developers take this seriously. For obvious functionality type regressions it would seem a test suite could, over time, test every code path.
(Log in to post comments)
KS2012: The future of kernel regression tracking Posted Aug 29, 2012 23:27 UTC (Wed) by dlang (✭ supporter ✭, #313) [Link]
The problem is that many code paths will only be used with specific hardware, or in other specific conditions.
Plus, you need to remember that the kernel is multi-threaded, so timing of different things happening matters as well.
There are very few "obvious functionality" type regressions.
|
|||||||||||||