Skip to content

Adopting children of an intentional null shadow root goes wrong #1429

@annevk

Description

@annevk

@annevk @sorvell
One thing I was thinking today about the adoption steps:
From the above example, based on the new spec, it's clear that shadowroot should be getting effective global registry or null depending on "keep custom element registry null" and the original factors (null or is using another global registry) during adoption.
But what about x-foo? I think the intended behavior is that x-foo' s registry will remain to be null. However, based on the new spec, the element has null registry originally, so during adoption in above case, it will gain the effective global registry of document, which does not align with the expected behavior.
When we set registry for null registry element during insertion in previous spec, we assign the registry based off the inclusiveDescendant's parent's registry. Do we need to do something similar when we're setting registry for null registry element during adoption? Perhaps something like "set inclusiveDescendant's parent's (or inclusiveDescendant's tree scope's) registry if it is an effective global registry in the new document, otherwise set null"

Originally posted by @ja-y-son in #1423 (comment)

Metadata

Metadata

Assignees

No one assigned

    Labels

    needs testsMoving the issue forward requires someone to write teststopic: nodes

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions