GeoServer crunch time II

The GeoServer crunch time continues as all the last minuet changes trickle in. This beta represents a feature complete line in the sand before FOSS4G.

July 19th update: Okay we are in to over time (Canadian hockey reference). Lets see what new functionality arrived at the last moment, and why you do not have a GeoServer release being announced today.

  • Andrea has pulled together a GUI improvement allowing configuration (and reload) of band information. This is especially useful when working with scientific grid coverage such as NetCDF (i.e. where the bands do not simply map to GRAY_INDEX or RED_BAND, GREEN_BAND, BLUE_BAND).

    Customised band  complete with null value.

    Grayscale ban:

    RGB bands:
  • An enthusiastic new volunteer Brett Walker (go team Australia) has been doing a great job dusting off bug fixes and pull requests and applying them to the codebase. Brett has just signed his code contribution agreement and is already making his presence known - welcome to the team Brett!
    • Removing direct references to com.sub packages (hopefully allowing GeoTools to work smoothly on OpenJDK).
    • Upgrade to JAITools 1.3.1 which includes substantial changes to RangeLookupTable for thread safety.
      While we always try and use the latest and greatest (especially for a beta) this transition has not been smooth. While the GeoTools tests pass there has been some integration troubles downstream at the GeoServer level. We have tracked down Micheal Bedward (JAITools lead) for the details and Simone is going to take discussion over to geoserver-devel for resolution.
  • I was able to get a GeoTools 10-beta deployed to maven.
    • There has been a little bit of back and forth over an jdbc patch to allow gt-jdbc-oracle to support tables, views and synonyms. Thanks to Robin Pradel for setting up this patch (for both 10.x and 9.x) and to Ben for testing against Oracle.
    • Early testing showed problems building on JDK7 (the new NetCDF module), Brett Walker has stepped in with a patch.
      (Reminder that JDK7 is not yet a supported platform)
    • A couple groups are performing audits of the GeoTools codebase and have been pointing out the occasional GPL header. I managed to find a few on my own in the gt-swt and gt-jdbc-sqlserver modules. In most cases this stems from developers active on GPL projects leaving their Eclipse code template settings in place when creating a new file.
  • Kevin has deployed into Maven a 1.5-beta GeoWebCache - Thanks!
    Sadly I may have to ask him to run that again when we pass all the QA tests.
  • MapFish Community module is now functioning again. The original tile server basemap (from the Metacarta!) has been swapped for a GeoServer powered one at demo.opengeo.org.
    Yes the MapFish Community module is not strictly part of the release (community modules are not included until they pass QA checks). Thanks to my work colleague Brad Latus for this fix.

Why we are Stalled

The interesting thing is two changes are preventing me from going further. The upgrade to JAITools has resulted in the GeoServer nightly build failing, and that is the same release that we run CITE tests against (which is an important QA step before release).

While I could choose the revision on the 17th (which passes as noted above) that would leave out Andrea's great improvements to the GeoServer GUI.  As it is I am going to cool my jets until I hear back from the developer list.

Getting Involved

If you would like to help out please download a GeoServer nightly build and try it out in your organisation. Helping test during this critical time period is one of the best contribution you can make to the GeoServer community.

If you need a hand deploying GeoServer in your organisation check out our LISAsoft GeoServer Spatial Technologies page for support and training.