Dr. Christian Tacke

I took a look at your db model about this. Instead of adding an orcid field to the `users` table, what about having orcid be represented using the `Identity` objects?...

Well, a first class citizen would also use an auth/oidc flow to verify the orcid, right? So I don't understand the difference, really? Except: * The "first class" implementation would...

> avoid any more special identity provider names I fully understand that! It just was a random thought while looking at the db models… > IMHO the difference is that...

Do I understand this change correctly in that "maintainer" would allow `Organization`s after this change would be implemented for 4.0? Would the 3.0 release at least allow a list of...

May I ask what is currently blocking this Pull Request from getting merged for the 4.0 release? Is there anything needing help?

> Nice idea! Can you open a PR for that? This is a very common case indeed, it's nice to optimize it. Yes, I could. But not before next week...

> Note that to be effective this change would also require the addition of a move constructor in `TString`. I think this exists already? https://root.cern.ch/doc/master/classTString.html#a97795e61556ec0ced81e15cdfa26a538 ```cpp TString::TString(TString&& s) ```

BTW, if we can't fix this for the 19.0 we should back out the geoloader stuff by ```cpp /// \bug Leaks GeoLoader and related resources, prevents memory issues (probably a...

For the time being #1518 has re-enabled the memory (including a proper marking). Still the problem of the unproper memory handling needs to be fixed.

Thanks for taking care and cleaning up! :-)