You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have to admit seeing a PR like #518 without an accompanying test makes me a little nervous. It's so easy for stuff like this to regress in the future.
It would probably be straightforward to write headless browser-based tests, but my experience with such tests is that they're so slow, complex, and flaky to be almost not worth writing.
As an alternative, would it be possible to write a dummy DOM node implementation, and then run tests against that? For example, something like this:
I plan to investigate SSR soon, I think that may be combined (at least somewhat), with your idea of the TestNode.
Reason why I have not yet added tests, are that API changes are easier to make, but I guess we're consolidating slowly, so it probably makes sense to add them soon. I've also thought about headless browser-based tests, but as you say that may slow down CI significantly and requires a quite a little bit of setup.
I have to admit seeing a PR like #518 without an accompanying test makes me a little nervous. It's so easy for stuff like this to regress in the future.
It would probably be straightforward to write headless browser-based tests, but my experience with such tests is that they're so slow, complex, and flaky to be almost not worth writing.
As an alternative, would it be possible to write a dummy DOM node implementation, and then run tests against that? For example, something like this:
You could then write tests, and assert that dummy dom nodes were updated as expected.
The text was updated successfully, but these errors were encountered: