...
- Chris,
Notes:
See notes on IRC.
Dan's note: These are raw and I will edit them down later today.
<cwilper> https://wiki.duraspace.org/display/FCREPO/2012-02-09+-+Fedora+Committer+Meeting
<kompewter> [ 2012-02-09 - Fedora Committer Meeting - Fedora Repository Development - DuraSpace Wiki ] - https://wiki.duraspace.org/display/FCREPO/2012-02-09+-+Fedora+Committer+Meeting
Chris: Focus on the hot items today.
Adam: GSOC on agenda?
Chris: Yes. Adam got a high quality T-shirt from GSOC. DuraSpace will apply one for each major project.
Chris: Not a shoe-in, Need to apply, have good projects and mentors for each. Need mentor volunteers.
<ajs6f> https://github.com/fcrepo/fcrepo/tree/fcrepo-452
<kompewter> [ fcrepo/fcrepo at fcrepo-452 - GitHub ] - https://github.com/fcrepo/fcrepo/tree/fcrepo-452
Chris: Need to discuss SOAP faults and facilitate getting last years code into Fedora.
<ajs6f> https://jira.duraspace.org/browse/FCREPO-452
<kompewter> [ #FCREPO-452 Move away from Axis 1.3 - DuraSpace JIRA ] - https://jira.duraspace.org/browse/FCREPO-452
Chris: Problems in passing tests may be part of text in SOAP faults.
Chris: Adam: Human readible messages, code/test brittle, pulling message out of type.
Chris: Human version and "reason"
<ruckus> Hola guys, sry for being late
ruckus has changed nick to elschlomo
<elschlomo> <- frank
<ajs6f> http://stackoverflow.com/questions/3313328/how-to-get-a-soap-fault-message-to-be-ws-i-compliance
<Dan_Davis> Group intelligence doubled when Frank joined.
<kompewter> [ soapui - How to get a SOAP fault message to be WS-I compliance - Stack Overflow ] - http://stackoverflow.com/questions/3313328/how-to-get-a-soap-fault-message-to-be-ws-i-compliance
<ajs6f> http://www.xenta.nl/blog/2010/07/17/oracle-service-bus-11g-handling-soap-faults/
<kompewter> [ Oracle Service Bus 11g, handling SOAP Faults | Oracle .. Java .. OpenSource .. SOA ] - http://www.xenta.nl/blog/2010/07/17/oracle-service-bus-11g-handling-soap-faults/
<elschlomo> yes! GI * (-2)
Chris: admits he does not want to write more Java code, cannot get free labor from Mo.
Chris: What do we really need to pass as faults/exceptions.
Adam: Chris: Compatibility with client libraries.
Adam: set up some "fedora" specified fault codes.
29 == tdonohue ~tdonohue@c-67-177-108-221.hsd1.il.comcast.net has joined #duraspace
Chris: Code may be best kept in "human" readable return.
Dan: Possible include a correlation ID in SOAP fault and server log.
Chris: Do not go with fault subcode, Go with Oracle style with consistent code (detail), plus across rest in the HTTP header. And correlate with stack trace/addl details.
Adam: Java could serial/de-serial exception but cross client/language may need a better form. Takes it out of the web service to use.
Adam: Makes wind.
Chris: Summary, change exceptions to have code, language independent, Detail portion has code/message. Reason, stack trace. Adam: Check to see common approach but close.
42 <ajs6f> whttp://www.w3.org/TR/soap12-part1/#faultstringelement
42 <kompewter> [ SOAP Version 1.2 Part 1: Messaging Framework (Second Edition) ] - http://www.w3.org/TR/soap12-part1/#faultstringelement
Chris: Impose on API, identifiers
Chris: Not say exception to make less Java specific
Adam: Will come with some experience.
Adam: Not numbers but partially human readable.
Chris: Switch to newest Bamboo version.
Chris: Perhaps sometime this week.
Chris: Push snapshots to Sonatype, Fedora is hardest.
Chris: Maybe from an empty Maven repository.
Chris: Not a problem on build but on deploy
Adam: Cargo
<ajs6f> https://jira.duraspace.org/browse/FCREPO-1061
<kompewter> [ #FCREPO-1061 Create examples for Cargo-based integration tests - DuraSpace JIRA ] - https://jira.duraspace.org/browse/FCREPO-1061
<ajs6f> https://github.com/fcrepo/fcrepo/blob/fcrepo-1061/fcrepo-integrationtest/fcrepo-integrationtest-cargo/pom.xml
<kompewter> [ fcrepo-integrationtest/fcrepo-integrationtest-cargo/pom.xml at fcrepo-1061 from fcrepo/fcrepo - GitHub ] - https://github.com/fcrepo/fcrepo/blob/fcrepo-1061/fcrepo-integrationtest/fcrepo-integrationtest-cargo/pom.xml
Adam: Config A, B working, Installer not needed.
Adam: Cargo provides container.
Adam: IHO, installer not tested as part of ITs
Adam: Can keep running container. Trivial to switch containers.
Adam: Problems with Jetty, may never have been properly tested.
Adam: Use profiles, should helping us to change containers easily.
Adam: Getting a dissemination that does not exists, Tomcat 404 right, Jetty 500 wrong.
Chris: Ant script gone.
Chris: What tests run. Adam: Depends on profile, config specific.
Adam: Unfortunately, may not be best, Indicate when test inappropriate to run is better.
Chris: Adam: Can eliminate stuff reduces maintenance cost.
Chris: Would you suggest moving to Cargo, next step to get move done the refactor tests. Adam: Its orthoganal but likely move to Cargo first. Make refactoring faster. You don't lose config.
Adam: Check into Verify with different profiles.
Dan: How does this play into performance testing.
Adam: Add a Jmeter profile.
Adam: Or just a performance profile.
Aaron: Current approach defacto tests installer.
Adam: Separate the two. Black box plus testing installer by some other means.
Chris: Still need end-to-end.
Chris: War only goal would be helped.
Adam: (with much chagrin) brings up OSGI.
Chris: Installer goes away.
Adam: uses the word "quotidian"
Chris: Wrap up. Adam is still wrapping up Cargo stuff.
Chris: To make jump, Installer test as a unit. Maybe also end-to-end.
<ajs6f> I'll do one more config in Cargo, and that should be a good point at which to bring in more opinion and make a forward-looking decision.
Chris: Recent issues.
Chris: Some are major.
<cwilper> most important for now: https://jira.duraspace.org/browse/FCREPO-1062
<kompewter> [ #FCREPO-1062 invalid XML datastream checksums after migration from inline (X) to managed (M) via migration script - DuraSpace JIRA ] - https://jira.duraspace.org/browse/FCREPO-1062
Chris: Could add to API.
Dan: Adam: Maybe a utility, maybe part of rebuilder.
Adam: Utility running over Web services. Slow but easy.
Chris: Should this be critical. Consensus yes.
<cwilper> https://jira.duraspace.org/browse/FCREPO-190
<kompewter> [ #FCREPO-190 Inline XML Canonicalization (c14n) for SIZE and CHECKSUM - DuraSpace JIRA ] - https://jira.duraspace.org/browse/FCREPO-190
Adam: Checksumming in the future? Its a bytestream operation. For XML, maybe xml signature.
Chris: Promises (again) to make the meeting shorter.
...