[ https://jira.jboss.org/jira/browse/DNA-194?page=com.atlassian.jira.plugin... ] Randall Hauch reassigned DNA-194: --------------------------------- Assignee: Randall Hauch -- 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-194?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-194: ----------------------------------- None of the nodes had a Node.getDefinition() implementation. To properly implement this, I refactored how JcrSession was creating and populating nodes, and I added some utility methods to the JcrNodeTypeManager (which is now a separate package-protected class rather than an inner class). This allows all of the Level 2 TCK unit tests to get further, but none of the other update functionality is implemented, so none of the Level 2 TCK tests run successfully. I also made a number of utility methods final, in the hopes that they can be inlined since they are called very frequently. -- 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-194?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-194: ----------------------------------- Applied the "...move_value_cast" patch. Also made the "createValueFormatException(...)" methods private to increase likelihood they'll be inlined. All tests pass. -- 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-194?page=com.atlassian.jira.plugin... ] Randall Hauch updated DNA-194: ------------------------------ Attachment: dna-194-session-cache.patch Uploaded a patch file that captures the refactoring to pull the management of the session's cached and updated state out of the individual Node and Property implementations and into a new "SessionCache" class. This class does a couple of different things. First, it centralizes all of the information about the nodes and properties that have been loaded from the repository source, making it easier to track the cached information as well as track/manage the changes made within the session. Remember that some activities, such as "refresh" or "save" can be performed on one node but could potentially affect many other objects. By centralizing all the cached and transient state, it's much easier to manage all this information. It also makes it much easier to alter the data structure used to manage the cache (for example to make it thread-safe), should that be required at a later time. Second, it manages the Node and Property implementation objects (which can be held onto by clients), ensuring that the same Node and Property instances are always returned for the same logical instance. These instances are managed by ReferenceMaps, allowing the garbage collector to remove entries when clients no longer hold onto the Node and Property objects (with normal references). The implementation classes of Node and Property (e.g., AbstractJcrNode and its concrete subclasses, and AbstractJcrProperty and its concrete subclasses) no longer directly hold onto the detailed node and property information being cached in SessionCache, but rather hold identifiers and dynamically look up the information when necessary. This ensures that a particular Node implementation object obtained for a node will always present the most up-to-date information about the node, even if that information is changed by the session or other objects. (Note that it would be very easy for each Node and Property implementation to locally cache the reference to the SessionCache's node information and/or property information, saving some repeated lookups using the SessionCache's methods, but since the SessionCache knows about the implementations it could invalidate these local references when needed. This could be done at a later time when we've proven that the current simple approach needs to be improved.) -- 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-194?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-194: ----------------------------------- Added most of the infrastructure to allow updates to the SessionCache. The cached representations of the nodes are changed and corresponding requests are accumulated in a Graph.Batch that are sent to the repository store when SessionCache.save() is called. This implementation hooks some of the infrastructure up to the JcrAbstractNode implementation, although more tests need to be written to verify the behavior. Currently, only deleting nodes has been tested and is thought to be complete. Changing properties, creating children, and moving nodes are partially implemented. (Some changes to the node type manager are required, but there is currently an ongoing refactoring of JcrNodeTypeManager. This is the primary reason for committing this partially-complete work.) -- 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-194?page=com.atlassian.jira.plugin... ] Randall Hauch commented on DNA-194: ----------------------------------- Implemented Node.addNode(...) methods and Node.save(). Some enhancements to the Request classes were required to allow the SessionCache to easily figure out whether a request affected a particular branch. Also enabled one TCK unit test. -- 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-194?page=com.atlassian.jira.plugin... ] Randall Hauch resolved DNA-194. ------------------------------- Resolution: Done I'm going to mark this as completed, since all of the work involving adding an implementation for update has been done. There may be small pieces of functionality (or bugs), but those should be addressed as specific issues. -- 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