fix: skip re-attaching shadow roots during view transitions

[Arecsu]

Changes

• Fixes: Elements with declarative Shadow DOM in transition-persisted containers throw "Unable to re-attach to existing ShadowDOM" error during view transitions
• The existing attachShadowRoots() function attempts to call attachShadow() on all elements with <template shadowrootmode>, but doesn't check if the element already has a shadow root from a previous page
• This commonly occurs when using transition:persist on containers with web components that use declarative Shadow DOM (e.g., custom elements, third-party libraries)
• The fix adds a simple check: if a shadow root already exists, skip re-attachment and just remove the template
• Follow-up of a recent PR made by @delucis: Attach declarative Shadow DOM templates during view transition #14341

Testing

This happened when fixed hydration mismatches in number-flow/svelte, which is a wrapper around a web component that uses Declarative Shadow DOM. PR: barvian/number-flow#162

The component was inside a transition-persisted flyout in all page-routes.

• Before: Navigation throws "Unable to re-attach to existing ShadowDOM" error
• After: Navigation works smoothly, shadow roots preserved correctly

Docs

Not applicable — bug fix for existing view transitions feature.

[changeset-bot]

⚠️ No Changeset found

Latest commit: ac3a35a

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[codspeed-hq]

CodSpeed Performance Report

Merging #14737 will not alter performance

_{Comparing Arecsu:main (ac3a35a) with main (91780cf)}

Summary

✅ 6 untouched

[Arecsu]

If this needs to have a changeset, please let me know. I'm not sure I fully understood if this PR needs it 😅

[delucis]

Looks good — thanks for the fix @Arecsu!

Yes, a changeset would be great here. This is a bug fix so it should be a patch to the astro package.

We have more guidance on how to write a changeset here: https://contribute.docs.astro.build/docs-for-code-changes/changesets/