3 Replies Latest reply on Jul 21, 2010 3:00 AM by galder.zamarreno

Is this a bug in cache.replace?

pengyan Jul 16, 2010 3:00 AM

When replacing a entry with the following class as a Key, I found the replace method doesn't work as the API said.

I've attached my test case here. It seems the root cause is it pass an instance of org.infinispan.marshall.MarshalledValue into the equals method.

1. Re: Is this a bug in cache.replace?

galder.zamarreno Jul 20, 2010 4:55 AM (in response to pengyan)

Hi Chun, thanks for attaching the test case. I managed to replicate the failure and I'm looking into it.
Actions
2. Re: Is this a bug in cache.replace?

galder.zamarreno Jul 20, 2010 5:48 AM (in response to galder.zamarreno)

This is a bug with replace() calls and lazy deserialization: https://jira.jboss.org/browse/ISPN-546
Actions
3. Re: Is this a bug in cache.replace?

galder.zamarreno Jul 21, 2010 3:00 AM (in response to galder.zamarreno)

I've fixed the bug but unfortunately it didn't make it to Infinispan 4.1.0.CR2. It will however be included in 4.1.0.Final.

If you want to test the fix, you can checkout the 4.1.x branch from http://anonsvn.jboss.org/repos/infinispan/branches/4.1.x/ and build it yourself. Make sure you follow recommendations in http://community.jboss.org/docs/DOC-15189
Actions