refactor(setup): extract DB user collision handling and normalize PostgreSQL suffixing

[joshtrichards]

Summary

This refactor extracts the database user/role collision handling into dedicated helpers in the MySQL and PostgreSQL setup classes, reduces hidden mutation, and makes DB-specific naming constraints explicit.

What changed

• extract userExists() helpers so existence checks no longer rely on implicit reads from $this->dbUser
• extract findAvailableUsername() helpers to isolate collision-avoidance logic
• make MySQL’s username-length limit explicit via a named constant
• add explicit PostgreSQL identifier-length handling
• move $this->dbUser assignment out of collision loops for better readability and less mutation
• normalize PostgreSQL suffixing to use base, base1, base2, ... when collisions occur

Why

This makes the setup flow easier to follow, reduces duplicated loop logic, and makes future refactors in the database setup classes safer and easier to reason about.

Notes

• MySQL/MariaDB username truncation and suffixing behavior is preserved
• PostgreSQL collision handling is now implemented in a dedicated helper
• PostgreSQL collision suffixing now starts at 1 instead of skipping to 2

Related: #59092

Checklist

• Code is properly formatted
• Sign-off message is added to all commits
• Tests (unit, integration, api and/or acceptance) are included
• Screenshots before/after for front-end changes
• Documentation (manuals or wiki) has been updated or is not required
• Backports requested where applicable (ex: critical bugfixes)
• Labels added where applicable (ex: bug/enhancement, 3. to review, feature component)
• Milestone added for target branch/version (ex: 32.x for stable32)

AI (if applicable)

• The content of this PR was partly or fully generated using AI

[MichaIng]

Shall we fix #59514 right here?

Suggested change

	$this->dbUser = $this->findAvailableUsername($connection, 'oc_' . $username);
	$this->dbUser = $this->findAvailableUsername($connection, $username);

setupDatabase calls this with hardcoded oc_admin, so the prefix is doubled, resulting in a oc_oc_admin database user. Compare with PostgreSQL, where no additional oc_ is prepended.