- From: Hayato Ito <web-platform-tests-notifications@w3.org>
- Date: Tue, 17 Jan 2017 11:45:08 GMT
- To: public-web-platform-tests-notifications@w3.org
Okay. We should judge this kind of *exhaustive*-ness case by case, however, in this particular case, I am not a fan of this kind of duplications. I got an impression that this duplication happened only because the usage of `assignedNodes()` were *spotted* in the original tests. Is my understanding correct? Is there any existing test which uses `assignedNodes()` API, but it was not *duplicated*, with a reason? Now we have doubled the number of tests. Then, later, if we add yet another new API, we would double the number again, having 4 x N tests? I think we want to avoid this kind of exponential-ness so that tests can be maintenance-friendly. Could we spend more efforts to reduce this kind of explosion by making tests more orthogonal, or having more maintenance-friendly way? For example, I think @rniwa parameterized some tests so that we can test both cases, open or closed, without *copy and pasting piece of code*. Can we do that? or is that too early optimization? Maybe I am worrying too much on this kind of duplication. :( @rniwa, WDYT? Is this okay for you? View on GitHub: https://github.com/w3c/web-platform-tests/pull/4541#issuecomment-273117828
Received on Tuesday, 17 January 2017 11:45:30 UTC