On Feb 21, 2012, at 6:49 AM, Romain Manni-Bucau wrote: > the only reason i don't like it is because of its verbosity otherwise it is > a cool format.
I never understood the "verbosity argument" but then I don't use vi as an editor. My editor automatically adds end tags, quotes for attributes, and even automatically does auto-completion if it has the schema. What's not to like? :) > Well i pushed a version using the xml file. Here the several parts: > * > https://svn.apache.org/repos/asf/openejb/trunk/maven-plugins/spi-helper-maven-plugin/ > : > the maven plugin to scan an app > * rest-example pom ( > http://svn.apache.org/repos/asf/openejb/trunk/openejb/examples/webapps/rest-example/pom.xml) > : one usage of the plugin (even if here it is probably useless) Awesome! The power of this scanner is that it puts the scan report in a known fixed location. I think making it configurable weakens that feature. Why is JPA required? Do you intend to add the ability to look for certain annotations, classes, interfaces? > * https://svn.apache.org/repos/asf/openejb/trunk/maven-plugins/xbean-xml/ : > the module containing jaxb classes and a helper class to create the > annotation finder I can take a look at your plugin later tonight. Very excited! :) > * openejb-core/.../FinderFactory: updated to try to get the scan.xml file > before creating a standard finder I think this code belongs in xbean. Xbean can look for the file and use it. More specifically it needs to happen behind the scenes and not push the complexity on xbean users. > Note: the descriptor is in META-INF/org/apache/xbean/scan.xml excepted for > webapp where META-INF is replaced by WEB-INF to be consistent with other > descriptors. Makes sense. Regards, Alan
