Re: [udig-dev] lowagie.itext dependency for GeoScript

[Frank Gasdorf <fgdrf@xxxxxxxxxxxxxxxxxxxxx>]

All right, updated my side, tested pdf export from GeoScript and it works, with some minor changes in the omsbox bundle:

* OMS bundle checks hard coded bundle names but doesn't check whether the version so correct (2.1.7 for lowagie.itext/lowagie.text)

* I would suggest to use a dynamic way to resolve dependencies

   * import directive for com.lowagie.text packages (maybe marked as optional) for omsbo bundle or

   * use PackageAdmin to find a bundle that exports lowagie.text packages and checks the version too

However, because we changed to com.lowagie.text bundle resolved from ORBIT the bundle name has changed from self-made bundle com.lowagie.itext 

I committed a change that fixes the resulting NullPointerException in getPath() if getBundle() returns null. (see https://github.com/uDig/udig-platform/commit/e9e3d2664e920b71007626ee2b65f72bbdb87b43).

This issue lead into "disabled" functionality for GeoScript and everything else build on OMS.

HTH,

Frank

2013/8/16 Frank Gasdorf <fgdrf@xxxxxxxxxxxxxxxxxxxxx>

I reviewed the dependency-tree for the libs bundle and have seen, that we have a dependency from geoscript to lowagie.itext.

[INFO] +- org.geoscript:geoscript-groovy:jar:1.1-SNAPSHOT:compile

[INFO] |  +- org.geotools:gt-geojson:jar:9.0-M0:compile

[INFO] |  |  \- com.googlecode.json-simple:json-simple:jar:1.1:compile

[INFO] |  +- org.geotools:gt-gtopo30:jar:9.0-M0:compile

[INFO] |  +- org.geotools:gt-imagepyramid:jar:9.0-M0:compile

[INFO] |  +- org.jaitools:jt-jiffleop:jar:0.2.0:compile

[INFO] |  +- org.json:json:jar:20090211:compile

[INFO] |  \- com.lowagie:itext:jar:2.1.7:compile

Because of already approved CQ for itext 2.1.7 (https://dev.eclipse.org/ipzilla/show_bug.cgi?id=7477) we can and included this bundle already from orbit repository.

I like to suggest to exclude com.lowagie.itext from pom-libs, update the manifest and classpath files (in addition to build.properties) and import packages in version 2.1.7 (resolved from bundle from CQ 3475)

I assume the geoscript would work like expected. 

Otherwise I guess we have to create an other CQ for transitive dependencies for GeoScript. 

@Jody: Any hints from your side after the call with Sharon? Whats your feeling about it?

- Frank