Open brooksdavis opened 8 years ago
I concur. This would be really nice (not just for CPU bound systems, but I/O bound and memory bound systems as well)
My preferred mechanism to fix this at the moment, which I've seen working reasonably well elsewhere, would be to implement what is described in issue #5. I've actually started working on that one.
Relying on test timeouts seems the wrong way to go, even with scaling, because the time a test takes to run can vastly vary due to the underlying platform. I think test timeouts make sense, but possibly only when the test itself is time-sensitive.
Timeouts are currently implemented in seconds. This can cause spurious failures on slow systems (for example qemu in full-emulation mode). It would be useful to be able to specify a global scaling factor for test timeouts across a run. I've been unable to find a way to do so.