Attendee:
Ivan, Heather, Scott, Roland, Giuseppe

Regrets:
Matthew

1 - Proposed project (https://github.com/IdentityPython/Governance/blob/master/idpy-projects.md)
a. https://github.com/peppelinux/pyMDOC-CBOR -> https://github.com/IdentityPython/pyMDOC-CBOR
See email to the idpy-discuss mailing list on 9 June 2023. We have strong support on the list.

Giuseppe is working on the 0.6.0 release; expects that to be done in a week or so. Goal is to make sure this is a tool that people can use without knowing Python.

There are still open questions re: mDoc vs VS and how that will play out in the EU Digital Wallet ARF.

This spec might be of interest (if only to make us all crazy) : https://datatracker.ietf.org/doc/draft-terbu-sd-jwt-vc/

See also: https://github.com/vcstuff

Note the Open Wallet Foundation has started holding meetings and has a task force focused on OIDC.

https://github.com/openwallet-foundation/OID4VC-due-diligence-tf

Italian specification of the wallet:

https://github.com/italia/eidas-it-wallet-docs/pulls

2 - Project review
   a. General

   b. OIDC - https://github.com/IdentityPython (idpy-oidc, JWTConnect-Python-CryptoJWT, etc)
Roland has been working on Satosa backend connecting to the OIDC library (a new backend using idpyoidc so Satosa can make client calls to OPs). When we get this module, we can start to look at logout flows. This will also relate to the browser changes and how cookies are effected.

Next step, wants to get SAML out of Satosa core (it has a dependency on pySAML). Also want to get rid of cryptodome (sp?) in favor of cryptoJWT. Concern that cryptoJWT may not handle certificates well. (https://cryptography.io/en/latest/x509/). Need to also determine where else cryptodome may still be a dependency. Would be nice to clean up that legacy.

Separating the core from the protocols is a good idea, so each protocol can be treated separately and libraries installed as relevant.

Suggest we move ahead with a PR for what Roland has worked on so far. Fully separating out pySAML will need to be treated as a separate effort that will require several steps.

There are older backends talking to specific services (facebook, github). It doesn't appear we need those specific connections any more. Should they be removed? We can remove the ORCID one; they are using standard protocol flows. Do need to check for the rest. They should be just a question of configuration not a specific module.

Roland has released 2.0 of idpy-oidc, which is what eduTEAMS is using. So, now eduTEAMS is (nearly) free to publish their frontend.

Note there is confusion regarding the older libraries. See https://github.com/IdentityPython/idpy-oidc/issues/53 . Ivan wrote a response, but if others have more to add, please do.

CZNIC is maintaining py-oidc. Roland will reach out and ask if they'd be interested in moving towards a more up to date implementation.

FYI: https://github.com/UniversitaDellaCalabria/SATOSA-oidcop/releases/tag/v2.0.0

   c. Satosa - https://github.com/IdentityPython/SATOSA - v8.3.0 and v8.4.0 released

nothing major; just pulled in a few PRs
Will be focusing next on https://github.com/IdentityPython/SATOSA/pull/405 to allow a change of paths
After that, will focus on pulling out the SAML dependencies

d. pySAML2 - https://github.com/IdentityPython/pysaml2 - also had a new release

main change keeps us compatible with xmlsec1
need to do some work around updating dependencies, possibly removing old dependencies such as cryptodome, and using as much as we can from cryptography
Giuseppe has made a PR for proper GitHub actions to test things as they are committed. Need to include that to make sure the CI is working for both pySAML2 and Satosa.

e. Any other project (pyFF, djangosaml2, etc)

djangosaml2 had two releases since the last idpy call. Now at v1.6.0. Now have a more granular control over samesite cookies.

3 - AOB
a. 2024 in-person meetings - https://github.com/fedidcg/meetings/wiki/2024-List-of-Identity-and-Related-Conferences

Note that the next idpy call is scheduled for July 4. Ivan is available, though probably the US participants will be offline that day.