Shibboleth Developer's Meeting, June 16, 2011

Attendees: Scott Cantor, Daniel Fisher (Virgina Tech), Jim Fox (InCommon), Nate Klingenstein, Chad La Joie, Andrew Petro (Unicon), Brent Putman, Bill Thompson (Unicon), Rod Widdowson, Ian Young, Russell Yount (CMU), Tom Zeller

Code/SVN Fixups (Chad)

• license header on the source code files have been updated to use the new header; currently a typo in the Java header, will be fixed
• eol-style and mime-type SVN properties have been fixed on all SVN managed files
• developers need to be sure they have properly configured their SVN clients per the Configuring SVN Clients document

OpenSAML Update (Brent)

• Finished up the structural re-organization; breaking the xmltooling, openws, and opensaml v2 libraries into the single project, multi-module, maven project for v3. Also cleaned up some dependencies during this work.
• Finished up initial draft of v3 library initialization and configuration based on the Java Service mechanism. Brent will send an email to the dev list explaining this.
• Chad has been doing some more work in the opensaml-util module, mostly just pulling in other util classes from other Shib projects and consolidating them in one place

IdP Update (Chad)

• IdP v2.3.1 - a few minor bugs have been reported, one related to a bug in the vt-ldap library; will probably release v2.3.1 in a couple weeks.
• IdP v3.0 initial profile handler work:
  • so far work is progressing well; really like how WebFlow is allowing v2 profile handlers to be broken up in to "bite size" bits, making the code a lot cleaner and clear
  • currently a bit of API impedance between WebFlow and OpenSAML; WebFlow flow actions are singletons but various OpenSAML bits are prototypical so something is needed to bridge between the two. One option is just to write some singleton factories, another is to explore the Spring auto-proxy feature (which are essential singleton factories).
  • OpenSAML v2 decoder/encoder implementations sometimes contain things not specifically related to decoding/encoding messages (e.g., check that SAML message recipient URL matches IdP endpoint that received the message); these types of checks need to be extracted out of those classes

TestNG/XMLUnit Update (Brent)

• All JUnit tests are passing on OpenSAML v3
• XMLUnit has an assertion helper class which is all the JUnit test cases used
• XMLUnit offers two checks: equality vs similar. Default was similar, switch to equality and found a few bugs
• Next step, migration of all unit tests to TestNG

Git Experience (Scott)

• Eclipse Git tools doesn't have good support for git-svn
• Best things seems to be ability to stage local commit history across two machines; don't have to put work in SVN in order to get it to the other machine
• No elegant support for synchronizing across repository/machines
• More in-depth experience starting on the next release
• Good for working on multiple branches of the source code at once
• Number of git-svn GUIs, windows support better than expected

Code Reviews (Chad/Rod)

• MappedAttributeDefinition and TemplateAttributeDefinition Code Review, May 24th
• Do we need anything more formal? No, the email method appears to be good for now.
• If we do need something more formal in the future it may be worthwhile to look at Gerrit (example: OpenAFS Gerrit

Face to Face Meeting

• Do we need one before the I2 Member Meeting in October? No, just need a full day meeting at that time. Chad to check with I2 about getting a room.

Next Meeting: June 30, 1500UTC