Continuous Release 1.9.0

[Splines]

Warning

Ok, so I'm not sure why, but GitHub shows 4 commits up until "Add pgadmin..." that don't actually belong here. Instead, you can already find these commits on the main branch. This issue is exactly described here on Stack Overflow and has been viewed over more than 270k times. It seems like it is a limitation of GitHub. However, git log --oneline main..mampf-next also shows these commits in the diff and I'm not sure what we did wrong to end up in this weird situation.

1) I tried the rebasing solution in #579 and it was working perfectly fine and removed the 4 commits. The only downside of this approach is that due to the rebasing all the new commits of mampf-next would have the same timestamp: the time of performing the rebase. The will actually even get reassigned a new commit id.

2) If this first solution is not what we want, we could instead ignore the additional commits in this PR. The merge should work just fine. As we wanted to rename mampf-next to dev anyways, we could instead create this new branch dev from main after release 1.9.0 directly and ditch mampf-next. This way, we could be sure that main and dev match exactly and have the same commits on them.

No matter what option we choose, we will have a clean slate again starting with the next continuous release. _{Due to the nature of merge commits though, dev might be one commit behind main as it doesn't include the merge commit itself. See also this SO answer which explains this. That's not an issue though. One can avoid this by merging from main into dev directly after a continuous release such that dev can be fast-forwarded.}

Changes

Also see #565 for a background on the linting changes.

• Add aperiodic monotiles to landing page #543
• Track last sign in date #553
• Add pgadmin for easy local database access #558
• Improve bulk upload description #559
• Sensible configuration for rubocop #560
• Strip back pipeline to Rubocop linting only #562
• Redo the Rubocop config #564
• Enforce new RuboCop cops for all Ruby files #566
• Setup ESLint in order to lint JavaScript #568
• Init default settings/extensions for VSCode #569
• Enforce new ESLint rules for all .js and .js.erb files #575
• Reconfigure ESLint with flat config #577
• Upgrade Node.js to v20 (LTS) #578
• Fix inverse_of quiz certificate #582

For future releases?

• Fix unread comments status #574

---

• check schema.rb has correct verison

[codecov]

Codecov Report

Merging #561 (76cb7c0) into main (5b28b04) will increase coverage by 0.03%.
Report is 4 commits behind head on main.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #561      +/-   ##
==========================================
+ Coverage   66.49%   66.52%   +0.03%     
==========================================
  Files         311      311              
  Lines        9417     9426       +9     
==========================================
+ Hits         6262     6271       +9     
  Misses       3155     3155              

Files	Coverage Δ
app/models/user.rb	53.43% <100.00%> (+0.42%)	⬆️
app/uploaders/correction_uploader.rb	100.00% <100.00%> (ø)
app/uploaders/submission_uploader.rb	100.00% <100.00%> (ø)

[Splines]

Currently, all existing model tests pass except for quiz_certificate_spec.rb. Already the first test there ("has a valid factory") fails with this error:

ActiveRecord::InverseOfAssociationNotFoundError:
Could not find the inverse association for quiz (:quiz_certificate in Medium) Did you mean?  quiz_certificates

You can verify that by your own on #581 which has all the changes from here merged into it.

As a reference, here is the "mantra" I once wrote for the inverse_of. Might be useful.

[Splines]

For completeness: we've decided to go with option 2 of this comment. Also all tests were passing locally.