Committers Call 2011-05-09


  • Naomi Dushay
  • Joe Gilbert
  • Rick Johnson
  • Jessie Keck
  • Simon Lamb
  • Julie Meloni (notetaker)
  • Molly Pickral
  • Bess Sadler
  • John Scofield
  • Eddie Shin
  • Matt Zumwalt


  • update on Hydra::Head plugin progress
  • update on hydra-jetty
  • update on Hudson
  • update on OM & ActiveFedora docs

Action Items

  • Eddie will follow up with Tom about Hi-Def line limit (note: this was fixed during the call)
  • Matt will make Hydra::Head a Rails 2 app that installs as plugin, push it, then Naomi will test that it is working app; then Matt will continue to make testable in rails2, then can move on to Rails 3 version
  • Bess will attempt to run Hudson on her OSX laptop, possibly also set up another Ubuntu-based instance at Amazon
  • Bess to look into port management plugin for Hudson
  • Matt will be finishing up OM & ActiveFedora docs this week; John Scofield will be finishing up solrizer docs this week

Next Week

  • Moderator: Matt Zumwalt
  • Notetaker: Jessie (w/ Julie backup if Jessie can't make it)


Naomi reporting on Hydra::Head: working with Matt, got Hydra::Head working all but 2 specs; those were failing because controllers were not working properly. It is tricky to get it working in the test environment, could not get it working in an app, doesn't even make it to test features...calling it an epic fail. Reported that Matt wants to take the tack of doing Hydra::Head w/ Rails 3; this was his suggested because he felt that making the spec pass was going to be too hacky, and we want to move to Rails 3 anyway, so why not just go for it now? Overall uncertainty as to the short-term future of Hydra::Head; Bess noted that no work can really move forward on Hydrus (and to some extent Hypatia) until this is settled.

The bulk of the call was devoted to finding solutions to this problem. Naomi rightly pointed out that code is only useful if it runs in context, and voiced concerns about leapfrogging directly to Rails 3. Matt suggested just working with ActiveFedora, Solrizer, Blacklight to build views, indexes, models while other developers worked on Rails 3 version of Hydra::Head. Rick noted that just starting with/working with ActiveFedora and Blacklight probably doesn't do much good because developers lose what Hydra offers above it; Eddie noted that specifically it doesn't make sense for Hydrus to go that route.

Matt described the situation as being at the point of diminishing returns: work going into Hydra::Head (Rails 2 version) just to make it testable, then proceeding forward to Rails 3 version. Bess asked for other options; Naomi suggested taking Hydra::Head + Blacklight, strip out testing workarounds, get working in app context, then try to get the tests to run. If it's all go, then move to Rails 3. Matt agreed this was a workable solution and a step forward.

There was discussion about similarities to Blacklight development and testing, as a sort of sanity check for the process that was eventually agreed-upon as an action item.

Bess reporting on hydra-jetty: upgraded to solr 1.4.1, single core, and also looked throgh the licensing files to ensure compliance. currently in a branch, called for some sanity checks (Molly did one, gave thumbs up). As soon as Hydra::Head is stabilized, will test with new hydra-jetty release, then move to master.

Eddie reporting on Hudson: Hull build on Hudson fails 1-2 specs because of a bug related to the Hudson build for ActiveFedora (an xpath query at the OM layer). Possible reasons include an old version of libxml2? Also of note, Eddie mention colliding ports for Hydra heads (e.g. not all can use 8983 when running tests). Had installed a port allocator plugin. Discussed having different apps decide on unique port numbers, or rake task that will take a port number as an argument. Bess suggested another plugin that will avoid port collisions.