Enable a few tests #511
Enable a few tests #511
Conversation
Signed-off-by: Jafar Al-Gharaibeh <[email protected]>
Signed-off-by: Jafar Al-Gharaibeh <[email protected]>
There was a problem hiding this comment.
This PR looks ok to me but I have a comment on the tests that are enabled (which were recently committed in e43d66be ). They include the running time as part of the test results. Given we have a system that expects actual test results to be identical to a "Gold Standard" output, how is this not almost guaranteed to fail when run on machines with a different performance to the machine that produced the .std file?
I suggest not enabling lib/stand/uniittest whilst we decide what to do.
|
I had the same thought. This is a new testing paradigm that I would like to make use of. I'd like to split this into a new directory and add more tests using the framework. In the short term, I at least wanted to make sure there is some test coverage for it. We already have support for ignoring differences in output, which I used in this case to allow the tests to pass if you noticed. By having it run, we can catch issues with visual inspection of the test output at least, even on the CI side. |
|
@Don-Ward in summary, the test passed the CI on all platforms as part of this PR. So this is not an issue. |
No description provided.