Nejc Habjan

That works for me if it helps developers! But keep in mind there are many alternatives, like podman, Rancher, [`nerdctl compose`](https://github.com/containerd/nerdctl/blob/master/docs/compose.md) and so on. I thought people create aliases for...

Thanks for the report @lhausermann. You're right this needs to be implemented still. We can probably reuse the project upload and make an `UploadMixin` or so.

In the meantime you can make use of the raw `http_post()` methods as a workaround.

Exactly - we'll soon have this documented a bit better on RTD but you can already take a peek at https://github.com/python-gitlab/python-gitlab/pull/2230/files.

Seems like we've been doing this for a long time, I can reproduce it with 3.0.0 at least, but probably goes way back. GitLab converts attribute names of arrays when...

@Mogztter thanks. I should say we found this issue via plantuml-markdown, which does some pre-processing of its own, and manually calls the `/map/` endpoint: https://github.com/mikitex70/plantuml-markdown/blob/master/plantuml_markdown.py#L228-L240. So maybe my wording above...

> @nejch your opinion would be useful here, too :-) @lgatellier keep in mind that python-gitlab 2.0.0 dropped support for python < 3.6. Otherwise I agree :) some things like...

I think this could also be implemented as a top-level fixture, to keep things in one language and less code. `tests/conftest.py`: ```python import os from pathlib import Path @pytest.fixture(autouse=True, scope="session")...

I'm going to close this as I think it's a GitLab issue (or intended behavior if re-confirmation is required), especially if other attributes can be updated normally. The client in...

Just to follow up here, there are different scenarios: Sometimes we _don't want_ to persist attributes because they are stale compared to the real state on the server after certain...