Shibboleth Developer's Meeting, June 14, 2013
Brent, Ian, Marvin, Rod, Scott, Tom
Next call is next Friday.
60 to 90 minute call window.
Work on message decoders and other messaging code. Protocol message signing and other security code refactoring.
Got permission to put UK federation "private" code (e.g. aggregator components) on GitHub, some of which is there now, more to come. Experiencing pain of svn-git conversion tools.
Attribute filters are functionally complete. I need to do at leas one pass over to check logging, annotations and javadoc.
Attribute mapping (needed for the AttributeInMetadataFilter) is WIP.
Dealing with administrivia regarding SP patch
Checking in storage API additions and tweaks and moved into dedicated maven modules
Rebased transientId definition on top of storage API
- Open issue around re-running resolver within a transaction
- Structured vs. "hacky" storage of multiple data fields (SP uses XML, thinking we probably want JSON here)
- Current def produces a string attribute value, I'd rather generate a NameID directly
- No principal connector thing yet?
Code review for idp-attribute-filter next week Friday June 21 ?
Spent most of week considering issues about Webflow, most of which discussed on list.
Some work on velocity templates.
We agreed not to meet on June 28.
Agreed to rev java-support module with some refactoring discussed on shib-dev.
Scott and Rod discussed the need to flesh out some issues around name identifiers and the dependent concern of name canonicalization.
Rod proposed another face-to-face meeting and mentioned the high value of the Columbus meeting. Scott suggested travel expenses from the OH meeting might provide input to a budget for the proposed face-to-face.
Scott briefly summarized Shib Consortium Board meeting where consortium membership fees were discussed. Scott advocated for scaling back membership fees, which he felt would exclude most US institutions due to excessive cost. He was successful in reducing costs and some changes in institutional size categories. Developer contracts were also discussed; draft version preserves developer intellectual property rights.
Scott has been reviewing and updating the project roadmap. He noted the schedule and estimates are fairly accurate in most cases.
Scott mentioned some shortcomings in using Jira as a project tracking tool for a project of this size. He noted that it's important for devs to use Jira time tracking for project reporting purposes.
Scott discussed project risk briefly; the most risk is on the SP side and it could benefit most from an additional developer. An additional IdP developer has little to no value due to high communication and collaboration costs.