- From: Ryosuke Niwa <notifications@github.com>
- Date: Tue, 09 Aug 2016 13:00:37 -0700
- To: w3c/webcomponents <webcomponents@noreply.github.com>
- Message-ID: <w3c/webcomponents/issues/78/238672734@github.com>
Given that we have a fundamental disagreement with the usefulness of `shadowRoot` in the first place, I don't find that argument convincing at all. What we need is a concrete list of use cases (there were a couple of good ones earlier in the thread), and we need to see if `/deep/` is the right API to provide for those use cases or not. For example, propagating Rect-render-like states to subcomponents is NOT an appropriate use case for `/deep/` since it won't be one off case to bypass encapsulation. This use case is better met by a new custom element callback designed to propagate states top-down. Writing [integration tests](https://en.wikipedia.org/wiki/Integration_testing) for a Web app consisting of hundreds of components is another use case that came up. I'd argue that such an integration test should not be relying on internal states of each component. Instead, they should be testing based on each component's public interface. Also, if authors really wanted to do white-box testing across components, they could simply write an equivalent of `/deep/` in JavaScript. And this would be fine for testing scenario because an extra few milliseconds of runtime wouldn't hurt end users. --- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/w3c/webcomponents/issues/78#issuecomment-238672734
Received on Tuesday, 9 August 2016 20:01:16 UTC