Looking to improve the efficiency of the DataCache and QueryCache. Benchmarking is showing that the constant conversion from the POJO format to and from the PCData format is taking its toll. After talking with Patrick on the subject, he pointed out the PCDataGenerator which enhances the Entities for use in the Cache (vs using whole new PCData objects). The use of the PCDataGenerator is configured via the openjpa.DynamicDataStructs property.
When I dug into this a bit, I discovered that the testcases for this function have been disabled since they were introduced into our repository. As I continue to investigate the alternatives, I'm updating these testcases and re-enabling them. There is a whole set of testcases related to the datacache and querycache capabilities. I may be busy for a while...
This JIRA will be used to re-enable these testcases. And, maybe even start to improve the efficiency of this PCData object processing...
|1.||@DataCache and @Cacheable not extending to subclasses||Open||Unassigned|
|2.||Plural configuration property processing||Open||Unassigned|
|3.||Synching of DataCache with QueryCache outside of a transaction||Open||Unassigned|
|4.||PCData generation has performance issues||Open||Unassigned|