Tobias Megies

> What do you mean by `git rebase master`? It yields `fatal: invalid upstream 'master'` when I run it in my branch (this is the first time I contribute to...

> I did not enter a changelog entry because it was a minor change, but could do it Yes, please do so. It is a minor change but it *does*...

Two things left: - clean up flake8 (see `lint_code` CI action) - test the warning message being shown, see e.g. [here](https://github.com/obspy/obspy/blob/79558d2169e1f63b9f73d60d48f383d497c6ffb0/obspy/io/xseed/tests/test_core.py#L567-L574) using our helper context manager for that purpose

> What exactly do you mean with the second bullet point? Should I adjust the `test_inventory.py`? Ideally, you should add a test in `test_inventory.py` which does a `read_inventory` call that...

@jdassink the thing is, your test should hit the condition that a response is not available and then emit that warning. The context manager catches all warnings inside the block...

To check it you could also make the expected message match more via regex.. just to check ``` msg = ".*Could not find response tag.*" ```

Hmm.. so if I understand correctly what happens with this PR is that *a)* an empty figure will be opened or *b)* if an existing figure is used, it will...

> Therefore, I would propose that the behavior wrt parsing invalid datetime value is somewhat modified. I see the following possibilities: > - generate a warning or even an error...

I can confirm this, we (had?) similar misleading errors when trying lo load a non existent path with lxml (e.g. to read StationXML). This should ideally be fixed by centralizing...

See #2062 for a similar fix and we should reuse the same error message and exceptions as done there