Hi, On Fri, Oct 15, 2010 at 6:48 AM, Tsirkin Evgeny <[email protected]> wrote: > On Fri, Oct 15, 2010 at 1:17 AM, Rakesh Vidyadharan <[email protected]>wrote: >> If all you are interested in is using the JCR API through JackRabbit, the >> easiest is to include the full jar file included in the standalone >> distribution in your classpath and use it. No need to worry about >> dependencies etc., since the full jar has it all. > > But that seems to have many unnecessary libs in it.
That's the tradeoff between the simplicity of just a single composite package and the complexity of proper dependency management. Jackrabbit supports both; the former through the large binaries available on the download page, and the latter through the POM-annotated artifacts on Maven Central. As mentioned in my other message, all modern Java dependency management tools support the Maven Central, so you don't really need Maven to access those jars. > Also if ,as other pointed out, the official releases are the source/maven > one. In that case using full standalone one will getting me into trouble > once i want a patch. Strictly speaking our only truly official release packages are the source jars available on the Jackrabbit download page. However, for convenience we also provide pre-built binaries in the form of the deployment packages you find on our download page and the component jars on Maven Central. For all practical purposes these are equivalent to binaries that you could build yourself from the release sources. > And there seems to be no way to do a custom configuration for the > standalone jars. You can certainly do that; see the --conf option on http://jackrabbit.apache.org/standalone-server.html. And if you use the standalone jar as a classpath dependency in your own application, you can directly access the RepositoryConf and RepositoryImpl classes from jackrabbit-core to set up a fully custom embedded repository. BR, Jukka Zitting
