[ https://jira.jboss.org/jira/browse/DNA-363?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-363: ----------------------------------- The first step is to create a CND parser. Using ANTLR 3 is probably by far the best approach, since that will require little of our code and will provide an excellent lexer/parser. However, that means that we'd have a dependency on 'antlr-runtime.jar'. This is probably acceptable (other libraries also have this, including Hibernate and Drools, to name just two). However, considering that we also want to have a CND sequencer, it doesn't make sense that the JCR implementation should be a dependency of this sequencer. Therefore, we should probably move the CND reading (and eventually generation) into a separate "dna-cnd" project that will depend on ANTLR (runtime) and "dna-graph". This CND reader could also behave as a graph importer, writing the node type content to a org.jboss.dna.graph.io.Destination instance and using the JCR-prescribed "nt:nodeType", "nt:propertyDefinition", and "nt:childNodeDefinition" structures. This would allow use to import the CND node types into a graph, whether that's a temporary in-memory repository (which is then processed by some other component), a real JCR repository (e.g., under the "/jcr:system"), or a configuration repository. It also would be very straightforward to wrap this in a sequencer. (We might even consider having a StreamSequencer subclass that wraps the SequencerOutput with a Destination, making it very easy to implement a sequencer from an importer.) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/DNA-363?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-363: ----------------------------------- Added a new 'dna-cnd' project containing a graph importer that reads JCR CND files and populates a repository with the content (based upon the JCR 'nt:nodeType', 'nt:propertyDefinition', and 'nt:childNodeDefinition'). This new project uses ANTLR 3.1.3, which is not yet in the JBoss Maven repository. I've already committed this newer version of ANTLR (with source) to the Maven repository using SVN, but it does not yet appear via HTTP. As soon as it is, we should change DNA's main POM.xml to include the new 'dna-cnd' project as a module. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/DNA-363?page=com.atlassian.jira.plugin... ] Brian Carothers updated DNA-363: -------------------------------- Attachment: DNA-363_preliminary.patch Added a rough patch that adds support for registering custom node types via CND files. The patch: - Cleans up a few places where the projection from CndImporter didn't match the JCR built-in type definitions for nt:nodeType, nt:childNodeDefinition, etc. - Adds a new abstract class GraphNodeTypeSource that returns the defined node types in the same graph format that CndImporter uses - Adds a new class CndNodeTypeSource that implements GraphNodeTypeSource and uses CndImporter to parse one or more CND files - Flipped RepositoryNodeTypeManager to expect a GraphNodeTypeSource as its input and internally projects the graph nodes into JcrNodeTypes for validation - Commented out TypeRegistrationTest since it depends on using (old) NodeTypeSources for custom type registration. This contributed the majority of the changed lines for the patch. - Modified JcrNodeDefinition to lazily load the required primary types and default primary type. This is important since nodes can define child nodes that have the parent node as a required or default primary type (q.v., nt:unstructured) - Added a test case to make sure that the CND registration was actually registering types - Flipped the TCK test type source from InMemoryRepositoryStub to register its types from a CND file Remaining work to make the patch ready for committing: - Add support for JSR-283-style typing (node type templates + finer-grained exceptions). This is the easiest way to re-enable the TypeRegistrationTest, as I have an original version based on NodeTypeTemplates that I could drop in. Converting 30-odd test cases to 30-odd CND files would be a mess. - Convert JCR and DNA built-in types to CND format - Delete old NodeTypeSource hierarchy and rename GraphNodeTypeSource to NodeTypeSource - Clean up/add comments - Review pass through the RNTM to make sure that there isn't any duplication of effort Future directions (separate bugs and not needed for 0.5, IMHO): - Add support for type unregistration (-283 optional feature) - Add support for type-reregistration (-283 optional feature) - Add support for type swapping (as discussed prev. on IRC). -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/DNA-363?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-363: ----------------------------------- When I apply the patch, I get several compile errors, and I believe that the latest patch is missing the new package "org.jboss.dna.nodetype" along with its classes: - InvalidNodeTypeDefinitionException - NodeTypeExistsException - NodeTypeDefinition - NodeTypeTemplate as well as the resource file: - jsr_170_builtins.cnd BTW, I really like how the node type template approach really simplifies the code that programmatically defines node types. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/DNA-363?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-363: ----------------------------------- Applied the "for_review" and "delta" patches. Made a few minor changes, including adding some JavaDoc, added JCIP concurrency annotations, and created a new ImmutableProblems wrapper so that some classes could be marked as @Immutable. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira