[
https://issues.jboss.org/browse/ISPN-1212?page=com.atlassian.jira.plugin....
]
Galder Zamarreño commented on ISPN-1212:
----------------------------------------
Another thing that could be useful for the testsuite is around those timing related
tests.
- Instead of relying on timing, we could rely on notifications. So, if you have a key=k1
that has to be live for 10 seconds, wait for a notification that the entry has been
expired, probably via a listener. That way we'd avoid completely thread scheduling
issues. However, Sanne's code that checks that a key is present until just before
those 10 seconds and then waits for a much bigger time for the data to disappear works for
the moment.
Improved PING and FD protocols for the testsuite
------------------------------------------------
Key: ISPN-1212
URL:
https://issues.jboss.org/browse/ISPN-1212
Project: Infinispan
Issue Type: Task
Components: Test Suite
Reporter: Galder Zamarreño
Assignee: Galder Zamarreño
Fix For: 5.0.0.CR7
Create:
- INFINISPAN_TEST_PING: use thread locals to maintain cluster membership since all tests
run in same JVM. This would make discovery rock solid and would make the testsuite more
robust.
- INFINISPAN_TEST_FD: instead of relying on pings to check whether the node is there,
attach a cache or cache manager listener to detect when a cache stops or fails and
propagate that information back to the membership.
--
This message is automatically generated by JIRA.
For more information on JIRA, see:
http://www.atlassian.com/software/jira