The way innative currently does tests does not work well in systems that expect a result to be the distributed end-result. The tests must be factored out and made optional, and exposed to cmake so other tools can hook into the testing process.
This is complicated by the known test failures from the float NaNs that either need to be fixed or skipped.
The way innative currently does tests does not work well in systems that expect a result to be the distributed end-result. The tests must be factored out and made optional, and exposed to cmake so other tools can hook into the testing process.
This is complicated by the known test failures from the float NaNs that either need to be fixed or skipped.