Skip to content

support for MSC4439: PGP key URIs in .well-known#5090

Closed
thetayloredman wants to merge 1 commit into
spantaleev:masterfrom
thetayloredman:msc4439
Closed

support for MSC4439: PGP key URIs in .well-known#5090
thetayloredman wants to merge 1 commit into
spantaleev:masterfrom
thetayloredman:msc4439

Conversation

@thetayloredman
Copy link
Copy Markdown

for details on the proposed (not yet accepted) MSC, see msc4439 (rendered).

I haven't messed with Ansible/Jinja2 much, so I'm not sure if this could be improved on. also a pointer as to documentation changes seen as fit would be appreciated!

@spantaleev
Copy link
Copy Markdown
Owner

Thanks, but I'd rather not land this.

MSC4439 isn't accepted yet, and baking the dev.zirco. unstable prefix into the playbook would just have to be torn out once the MSC lands (likely with a different prefix or key name).

Also, this isn't needed. The existing template serializes each contact as-is via to_json, so any custom key in a contact entry passes through to the output. You can emulate MSC4439 today without any playbook change:

matrix_static_files_file_matrix_support_property_m_contacts:
  - role: m.role.admin
    email_address: admin@example.com
    dev.zirco.msc4439.pgp_key: https://example.com/key.asc

On the implementation side: the key-rewrite logic (namespace + dict2items/rejectattr/items2dict) turns a one-line template into something fairly hard to follow, for a rename that users can do themselves in config.

Happy to revisit once the MSC is accepted and has a stable name. Closing.

@spantaleev spantaleev closed this Apr 24, 2026
@thetayloredman
Copy link
Copy Markdown
Author

thetayloredman commented Apr 24, 2026

@spantaleev All good, totally understand it. FWIW, the MSC process itself is built to flow in that way (implementing the unstable prefix, then once it merges dropping it). I'm maybe not familiar with how others use the playbook, but I thought adding the key that way so once it gets merged the user's vars didn't need changing?

The rewriting process was definitely not great looking though.

But anyway, no issue here, just needed to have an extra something as an implementation for the MSC process. Doesn't need to be merged, just thought I'd have it at least looked at :)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants