This is particularly a problem for webrtc's AppRTCDemoTest, which doesn't have any annotations, and so has 0 tests being run!
There are both whitelist and blacklist annotations.
Whitelist: @SmallTest, @MediumTest, @EnormousTest, etc
Blacklist: @Disabled, @FlakyTest
By default, test_runner.py runs tests that are annotated with any of the whitelisted annotation, but if it is lacking any of them, does not get run.
Possible solutions:
1. Raise an exception if any tests is found that doesn't have at least one annotation from the set of whitelist annotations
2. Always run tests that are lacking a whitelist annotation, but also show a warning for them
3. Change the default behaviour (where no explicit set of whitelist annotations is passed via the command-line), to run all tests.
1. Is probably the best outcome since it will ensure all tests are annotated
3. Is reasonable, but would result in tests that might get skipped on bots
2. Seems strictly better than right now, but #1 would be better longer term. Unsure if warnings would entice projects to annotate.
Comment 1 by jbudorick@chromium.org
, Apr 7 2016