NanoSparqlServer does not locate "html" resources when run as an embedded server from the jar

[thompsonbry]

The NanoSparqlServer? does not locate "html" resources when run as an embedded server from the jar. The problem is on the one hand that the bigdata-war/src directory was not being included within the jar. However, even with that directory copied into the jar the NanoSparqlServer? is still not locating these resources.

Most people will run the NanoSparqlServer? from the WAR, which does not have this problem. This issue will mainly be encountered by people using the NanoSparqlServer? as a benchmarking utility.

[thompsonbry]

Modified build.xml to bundle the web app resources in the bigdata jar.

[thompsonbry]

Martyn suggests an incantation along the following lines to locate the web resources in the class path:

String webDir = this.class.getClassLoader().getResource("com/company/project/mywebdir").toExternalForm();

ServletContextHandler context = new ServletContextHandler(); context.setContextPath("/"); context.setResourceBase(webDir);

[thompsonbry]

We have three different evaluation environments which need to be handled. The IDE, the JAR, and the WAR. The IDE and JAR are both "embedded" deployments and are run through NanoSparqlServer#main?(). The web resources for the NanoSparqlServer? are deployed to bigdata-war/src/html in the JAR, which is their path in the top-level project. However, we can not use that path in the IDE as it would require the project root to be on the build path. Therefore, I have modified NanoSparqlServer? to first look for the "bigdata-war/src/html" resource (JAR) and then look for the "html" resource (IDE). I have also added "bigdata-war/src" to the build path for Eclipse.

This fixes the deployment under the IDE (the index.html page is now located).