Community
Participate
Working Groups
We should not be using the HelperContextMapKey class as the key to the helper context map. This is due the the fact that we cannot get a unique hashcode, as we could have an application name alone, or an application name and a classloader. We must support scenarios where the a given context is cached on application name and classloader, such that if a lookup is performed with the same loader in a non-WLS owned thread we can still return the correct helper context.
Created attachment 173700 [details] Contains proposed fix
Created attachment 173799 [details] Contains proposed fix
Created attachment 173870 [details] Contains proposed fix
Created attachment 174528 [details] Contains proposed fix for 2.2 stream
Fix checked into the 2.2 stream. Reviewed by: matt.macivor@oracle.com Tests: all unit tests pass as expected; sdo server tests all pass
Fix for bugs 320035 (SDO: JBoss redeploy scenario not handled correctly in SDOHelperContext) and 319171 (SDO: HelperContext cache should not be keyed on HelperContextMapKey) Reviewed by: matt.macivor@oracle.com Tests: all unit tests pass as expected
The Eclipselink project has moved to Github: https://github.com/eclipse-ee4j/eclipselink