Peter Nerlich

> Does your last comment then mean, that this issue will be closed? I created this issue mostly because I need a valid issue ID in order to add a...

stress test to reproduce collisions: `tools/integreat-cms-cli shell` ```python from integreat_cms.cms.models import Page from integreat_cms.cms.utils.tree_mutex import tree_mutex @tree_mutex def forth(): alltag = Page.objects.get(id=21) sprachkurse = Page.objects.get(id=20) alltag.move(sprachkurse, 'last-child') @tree_mutex def back():...

mock `treebeard.models.Node._get_database_cursor()` within a custom `atomic` decorator (?) / context manager to always return the cursor of the connection of the currently wrapped function *EDIT: attempted here: https://github.com/digitalfabrik/integreat-cms/compare/7bc796c...bugfix/treebeard_transaction*

I had a crack at making `tree_mutex` be a transaction and inject its db cursor into treebeard, but it doesn't seem fully working yet: even though everything seems to work...

Maybe also risk a look at https://github.com/digitalfabrik/integreat-cms/compare/b40e283...bugfix/treebeard_transaction – I rebased it just now onto these latest changes. I didn't test it though. Have a look through the mutex at least,...

@timobrembeck I cannot figure out how to capture the command output in `tests/core/management/commands/test_repair_tree.py` *(btw. ignore `tests/cms/models/pages/tree_mutex.py`, it is a stub that we just copied, as you can see)* – the...

Then I guess I'll try that next then. Though if it works as I expect it to, it should be rather transparent. It just hands the strings to the desired...

Unfortunately, it's still the same

That was what I was forgetting. Forgive me, I was almost done, but then I discovered that the repair tree command would report the tree_id with its already fixed value,...

Okay, I had to re-run the CI pipeline a few times until it went through. At least twice I had a nonsensical failure (not finding the region Augsburg for example)...