Daniel Hollas

Huh, seems like this is a known behaviour??? https://github.com/aiidateam/aiida-core/blob/212f6163b03b8762509ae2230c30172af8c02fed/src/aiida/orm/nodes/data/array/trajectory.py#L362

@sphuber @mbercx feel free to tag me for review once you converge, happy to take a look and do a bit of dogfooding. > am onboard with all suggestions, except...

@sphuber thanks for indulging me, and improving the tests! I left a couple of follow up comments. I have also tested this locally and it works great! 🎉

> Though, related to the most important open question: in a jiffy actually made me giggle when I was dogfooding now :D so I like it, but would ofc also...

> There are still some questions about desired behavior of the command I think the current behaviour is fine, after all the user can simple change a directory, right? Are...

I am also leaning towards local .aiida having preference, seems more intuitive. Another perhaps more general question, how does this interact with verdi daemon? Can there be more isolated daemons...

Hmm, tricky. I don't think erroring out is very user-friendly, rather we should have a predictable behaviour. > Leaves perhaps to ignore ~/.aiida when checking the CWD. So check CWD...

@mbercx some really insightful comments! The fact the this PR changes the semantics (not only behaviour) is a really great observation. But I would like to turn this around back...

I've checked the PR that implemented this feature in #5691, and indeed there are no testcases covering the interaction of variadic and keyword arguments. https://github.com/aiidateam/aiida-core/pull/5691/files CC @sphuber

This change makes a lot of sense, would be good to get into 2.5.0 if @sphuber agrees.