Revert PR #12250 - feat/non-superuser-dataverse-linking

[pdurbin]

What this PR does / why we need it:

Collection operations especially are crazy slow. See #12672.

This PR reverts the following PR:

• Dataverse collection linking/unlinking without superuser status #12250

Which issue(s) this PR closes:

• Closes 6.12 release candidate (d59202b) is super slow #12672

Special notes for your reviewer:

I wasn't able to use the revert button...

... so I asked Codex/ChatGPT to make a branch for me (I renamed the branch and rebased it):

Thank you to @youseihuayu-wonderful who suggested looking at this PR! This is what she had to say:

Suggestions on how to test this:

I have already deployed this PR/branch to the QA server. I would suggest testing the steps in #12672 there.

Does this PR introduce a user interface change? If mockups are available, please link/include them here:

No, the PR being reverted didn't change the UI so this PR doesn't either.

Is there a release notes update needed for this change?:

No, once this is merged, I'll fix the release notes to take this feature out.

Additional documentation:

None.

[coveralls]

Coverage is 25.412% — 12672-revert-pr12250 into develop. No base build found for develop.

[sonarqubecloud]

Quality Gate failed

Failed conditions
0.0% Coverage on New Code (required ≥ 80%)
3.6% Duplication on New Code (required ≤ 3%)

See analysis details on SonarQube Cloud

[github-actions]

📦 Pushed preview images as

ghcr.io/gdcc/dataverse:12672-revert-pr12250

ghcr.io/gdcc/configbaker:12672-revert-pr12250

🚢 See on GHCR. Use by referencing with full name as printed above, mind the registry name.

[github-actions]

Test Results

405 tests   390 ✅  33m 34s ⏱️
 55 suites   15 💤
 55 files      0 ❌

Results for commit d6598cc.

[landreev]

Some quick testing results:
Loading the top, root collection page for the superuser should take a few seconds and 700+ database queries. Confirming that this is the observed performance when testing this pr.
Testing the current develop branch, the same load takes ~126K queries. The number comes from:
~14K collections/"dataverses" x 3 additional 1:1 relation queries x the jsf rendering rule evaluated 3 times = 126K.