2022-11-18
Shibboleth Developer's Meeting, 2022-11-18
Call Administrivia
09:00 Central US / 10:00Â Eastern US /Â 15:00Â UK / 17:00 FI
Calls are normally the 1st and 3rd Fridays of each month. Next call would be Friday 2022-12-02. Any reason to deviate from this?
60 to 90 minute call window.
Call Details
This week's call will use the OSU Zoom system, per message sent to dev list (https://marc.info/?l=shibboleth-dev&m=166869233232713&w=2)
AGENDA
Plan for migration off HttpServletXXX proxies (largely mapped out, just summing up cause, solution)
I believe the Supplier hook was added post-4.2? We may want to switch to a custom non-null interface…
JSSH-5: ServiceableComponent should implement AutoCloseClosed
See net.shibboleth.idp.attribute.filter.impl.AttributeFilterImplTest for test weirdness around this
Rod is reworking this change
(PS) Pin release plugin versions?
Phil will look at pinning the initial cases.
(RDW) Whither
SAML1NameIdentifierAttributeDefinition
andSAML2NameIDAttributeDefinition
?Still TBD, would like to know what the use cases are at this point. Willing to change warning message if we need to.
Attendees:
Brent
Absent, updated via Slack that HttpClient 5.2 refactor is moving along
Daniel
Â
Henri
Absent today
Relying party bean-wiring problem described in a mail
Ian
Â
John
Â
Marvin
Â
Phil
DuoOIDC plugin improvements and testing
DuoOIDC plugin 1.3.0 release
Back to RP next week
Hopefully a new release before TechEx.
Improve code
Add key set document endpoint (hopefully before TechEx)
Â
Rod
SSPCPP-961: Shibboleth SP - Windows - possible Privilege EscalationClosed
IDP-1793: Use Suppliers for HttpRequest/ResponseClosed I need to document the new beans (but things are still moving in this space)
IDP-2037: Computed, Pairwise and StoredId DataConnectors to not implement all the common attributesClosed, IDP-2036: Mapped Attribute Definition does not implements all the common AttributesClosed
Somebody from Red Hat Product Security claims they didn't put any hibernate jars into Maven Central. Doesn’t that make you feel comfortable.
Scott
Walking the libraries with the Null analyzer enabled, finding a few bugs here and there
We probably want to take the V5 opportunity to revisit our method contracts and reduce use of nulls, especially around collections
Using asserts to some extent, per Slack to convince analyzer, these are on in the tests automatically
Removed what seemed to be an unused method, but @Rod Widdowson may want to review
Tom
testbed
Other
Â