Ben Boeckel

See #1708 and #1709 for those two split out.

Thanks; I'll get around to it on my next HDF5 update cycle.

Oh, there's stuff to perserve here.

Would it be easy to just have two arguments, one for the build tree and another for the source tree?

Understood. There's no pressure for this to be implemented, just filing an issue. I agree that it is niche, but also wasn't sure how much work would be involved (maybe...

The lucky thing in our case is that the non-semver parts are just `struct` definitions with serde derivations; nothing goes across the boundary (or if it does, that's a problem)....

My attempt: https://gitlab.kitware.com/utils/rust-gitlab/-/merge_requests/423 (needs the feature published before we can actually use it for real).

This probably requires smarter logic in…something. Not sure if CPython, `pip`, or `setuptools` needs to get smarter or not.

A grep of my repackage of the relocatable install shows `Python.framework/Versions/3.11/lib/python3.11/_sysconfigdata__darwin_darwin.py` having hardcoded paths :/ . Other paths with hardcodings that may be relevant: - `Python.framework/Versions/3.11/lib/python3.11/config-3.11-darwin/python-config.py` - `Python.framework/Versions/3.11/lib/python3.11/config-3.11-darwin/Makefile` - `Python.framework/Versions/3.11/lib/pkgconfig/python-3.11.pc`...

I honestly don't know what triggers it (probably `logind`?), but X, dbus, etc. are all started by user units I've made. As soon as I login to a TTY on...