Fixed selection attributes handling after softBreak.

[arkflpc]

🚀 Summary

Improved document selection attribute inheritance around <softBreak> so returning the caret after a soft break restores expected formatting.

When selection attributes are recalculated across <softBreak>, only attributes marked with copyOnEnter are inherited. Other inline non-object elements still act as hard boundaries.

---

📌 Related issues

• Closes Double Shift+Enter should preserve selection attributes. #19853.

---

💡 Additional information

Optional: Notes on decisions, edge cases, or anything helpful for reviewers.

---

🧾 Checklists

Use the following checklists to ensure important areas were not overlooked.
This does not apply to feature-branch merges.
If an item is not relevant to this type of change, simply leave it unchecked.

Author checklist

• Is the changelog entry intentionally omitted?
• Is the change backward-compatible?
• Have you considered the impact on different editor setups and core interactions? (e.g., classic/inline/multi-root/many editors, typing, selection, paste, tables, lists, images, collaboration, pagination)
• Has the change been manually verified in the relevant setups?
• Does this change affect any of the above?
• Is performance impacted?
• Is accessibility affected?
• Have tests been added that fail without this change (against regression)?
• Have the API documentation, guides, feature digest, and related feature sections been updated where needed?
• Have metadata files (ckeditor5-metadata.json) been updated if needed?
• Are there any changes the team should be informed about (e.g. architectural, difficult to revert in future versions or having impact on other features)?
• Were these changes documented (in Logbook)?

Reviewer checklist

• PR description explains the changes and the chosen approach (especially when performance, API, or UX is affected).
• The changelog entry is clear, user‑ or integrator-facing, and it describes any breaking changes.
• All new external dependencies have been approved and mentioned in LICENSE.md (if any).
• All human-readable, translateable strings in this PR been introduced using t() (if any).
• I manually verified the change (e.g., in manual tests or documentation).
• The target branch is correct.

[Mati365]

Looks ok. I left one comment. Make sure manual QA testing is performed, and E2E testing was executed on this PR.

[Mati365]

@arkflpc Please merge recent master, or cherry-pick 1d846d2 in order to fix failing manual tests.

[Mati365]

I'm not sure whether the engine should be aware of the existence of the softBreak element. It's registered in an external plugin. Would a schema flag be better? What do you think?

[Mati365]

We discussed with @niegowski, and it might be good idea to store some info in schema / attribute properties. However, there might be another workaround related to storing attributes on soft-breaks itself. We need to discuss it a little bit more.