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-…
Note that the next idpy call is scheduled for July 4. Ivan is available, though probably the US participants will be offline that day.
Thanks! Heather
Hi,
I am attempting to set up a SAML environment to better understand how
Satosa works. Here is the model of my environment: SAML Service Provider
(SP, pysaml2) <-> Satosa Proxy <-> SAML Identity Provider (IdP, pysaml2).
During the single sign-on service, after entering the username and
password, when the SAML IdP posts the AssertionConsumerService to the
Satosa Proxy, an error(satosa log attached) occurs:
'saml2.sigver.SignatureError.', the following is my satosa proxy logs:
https://pastebin.com/YsRMVzKD
I am unsure about how to resolve this issue. Could anyone provide me with
some guidance?
Any help would be greatly appreciated.
Thanks,
Sam
Dear Friends,
more than a year passed from TNC22 at Trieste, an event during which we are
supposed to create basic components for issuing and verifying verifiable
credentials in both sd-jwt and mdoc cbor format (ISO 18013-5).
While sd-jwt was produced by Daniel Fett in parallel with writing the
draft, and written in python, for mdoc cbor we had nothing more than a
monolithic demo produced by Google and written in Java.
I confess to having meanly waited for someone to do something, sooner or
later. After a year I gathered my pieces and produced this
https://github.com/peppelinux/pyMDOC-CBOR
0.6.0 will be the first official release and I hereby express my
willingness to migrate this new library to identity python.
The Interop tests are going to be made in the GAIN-PoC WG with colleagues
of Authelete that produced something similar but in Java, and with any
other available participant.
Look forward for your kind answer,
best
--
____________________
Giuseppe De Marco
Centro ICT d'Ateneo
Università della Calabria
87036 Rende (CS) - Italy
Phone: +39 0984 496961
e-mail: giuseppe.demarco(a)unical.it
--
------------------------------------------------------------------------------------------------------------------
Il banner è generato automaticamente dal servizio di posta elettronica
dell'Università della Calabria
<https://www.unical.it/5x1000>
Hello all!
In case you still have calendar entries for idpy calls on the 23rd of May or 6th of June, please note those calls are canceled. Our next call is currently scheduled for 20 June 2023.
See you then!
Thanks! Heather
