https://github.com/web-platform-tests/wpt/pull/16080 has tests and https://github.com/whatwg/html/pull/4447 has a refactoring of the specification that makes it no longer rely on descendant change notifications.
Have we pulled that test in yet? I’d like to look at it; might tackle this bug.
(In reply to Darin Adler from comment #2)
> Have we pulled that test in yet? I’d like to look at it; might tackle this
Yes. It's at LayoutTests/imported/w3c/web-platform-tests/html/semantics/forms/the-output-element/mutations.window.html
Wow, this is easy to fix. I will.
Thank you for making this not be defined in terms of change notifications!
The test seems to not match the specification. The test has a test case where the output element's value is set to the string "some". The specification says that should do a string replace all operation with that value. One result of that is that all the children are replaced. The test then sets the textContent of a child element to "y". But after the replace all operation, that child element is no longer one of the children, so that has no effect. The test expects that output.value will be "y" after that operation.
Added a comment about this in GitHub: https://github.com/web-platform-tests/wpt/pull/16080/commits/610c0a602dc302c9927ab976f0e75a60c505ab33#r632250067
Created attachment 428597 [details]
I opened <https://github.com/web-platform-tests/wpt/issues/28994>. Anne, would love to hear your take on this.
Created attachment 428600 [details]
Created attachment 428603 [details]
All those EWS bubbles green; hoping for review soon.
Created attachment 428665 [details]
Good news, the bug in the test was fixed back in March.
This patch modifies the imported WPT tests. Please ensure that any changes on the tests (not coming from a WPT import) are exported to WPT. Please see https://trac.webkit.org/wiki/WPTExportProcess
Committed r277527 (237755@main): <https://commits.webkit.org/237755@main>
All reviewed patches have been landed. Closing bug and clearing flags on attachment 428665 [details].