Community
Participate
Working Groups
If transaction was created at the time, when invalidation was running and this transactions calls waitForUpdate for already invalidating commit - it will wait for the next invalidation actions. If only single user is working, then simply deadlocks on waitForUpdate call.
Created attachment 198242 [details] Test and fix v1 Assigning transaction last update time the same value as session has
[NoMagic note: SVR-3225]
Changing to 4.1 to ensure that the fix will "last". Please clone this bugzilla to 4.0 if you want a maintenance fix, too.
With your patch I'm seeing an error in org.eclipse.emf.cdo.tests/META-INF/MANIFEST.MF: "Version must be increased to 4.0.100" An indication that you're not using our special version checker tool. Please export org.eclipse.emf.cdo.releng.version-feature to your host IDE's installation folder and do a clean build.
Your fix is good! The test however would, in case of a regression, hang (also on Hudson). I changed the logic to: freshTransaction.waitForUpdate(timestamp, DEFAULT_TIMEOUT);
Created attachment 198996 [details] Patch v2
Please commit to *trunk*
Fixed in trunk, revision 8575
Closing.