@polarathene any feedback on this?

Hey sorry I must have missed this (I get a lot of notifications).

I'm juggling quite a bit right now but at a quick glance this looks preferable to the other PR you referred to (I'm pretty sure I have feedback there where I expressed that PR would be fairly redundant for DMS to merge).

To review this I'll need to go over the prior discussions to refresh and be able to consider the changes (some which might require me to go over Rspamd docs on settings that have changed like try_fallback if those are missing context for their change).

One question I'd have from a quick look over is the docs advice change that referred to an RSA fallback, is that known to work with the selector map / path approach this PR switches over to? I know the existing one has a way to define multiple key types per domain (with the same or different selector IIRC).

Might need to verify if the RSA fallback works, so if you're able to look into that it'd be appreciated.

---

I should be able to make time for this PR sometime this month, and I think it looks like it will help progress towards the unified DKIM approach I'd like to get DMS to adopt, so thanks for taking the time to put this together ❤️

Presently when I can spare the time, I'll be first prioritizing a review on the Debian 13 Trixie upgrade PR as that's been quite a blocker for a while and needs my help/feedback, so I'll need to assist with getting that sorted.

I would still be interested in your PR being part of DMS v16 if all goes well after review 👍