Remove performance regression when we're subscribing with a regex

[emasab]

and none of many topics in the cluster are matching it.

[confluent-cla-assistant]

🎉 All Contributor License Agreements have been signed. Ready to merge.
_{Please push an empty commit if you would like to re-run the checks to verify CLA status for all contributors.}

[Copilot]

Pull Request Overview

This PR addresses a performance regression when subscribing with regex patterns in scenarios where few topics in the cluster match the pattern. The main optimization replaces the inefficient approach of copying all cached topics to a list with direct iteration over the metadata cache, and introduces deduplication to handle multiple cache entries with the same topic name.

• Replaces topic list creation with direct cache iteration using TAILQ_FOREACH
• Adds a hash map for deduplication to avoid processing duplicate topic names
• Introduces early exit logic for already-matched topics to improve efficiency

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Line 1298 contains a grammatical error: 'add' should be 'added'.

Suggested change

	* Topic was already add to
	* Topic was already added to

[Copilot]

[nitpick] Extra space between 'mdt' and '=' creates inconsistent formatting. Should be single space like other assignments.

Suggested change

	mdt = &rkmce->rkmce_mtopic;
	mdt = &rkmce->rkmce_mtopic;