You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sling.apache.org by "Lukas Eder (JIRA)" <ji...@apache.org> on 2013/03/12 12:43:12 UTC

[jira] [Created] (SLING-2786) ResourceMetadata contains stale unmodifiableMap cache after clone()

Lukas Eder created SLING-2786:
---------------------------------

             Summary: ResourceMetadata contains stale unmodifiableMap cache after clone()
                 Key: SLING-2786
                 URL: https://issues.apache.org/jira/browse/SLING-2786
             Project: Sling
          Issue Type: Bug
          Components: API
    Affects Versions: API 2.3.0
            Reporter: Lukas Eder
            Priority: Minor


Super types of ResourceMetadata (java.util.HashMap and java.util.AbstractMap) correctly implement caching for members such as entrySet, keySet, values, etc. Two things should be considered:

1. The cache should be marked transient. It is not desireable to serialise / deserialise the "unmodifiableMap" cache. Only the isReadOnly flag should be serialised
2. clone() should be overridden in order to reset the cache on cloned instances

Consider the following code:

    	ResourceMetadata map1 = new ResourceMetadata();
    	map1.put("key1", "value1");
    	map1.lock();
    	map1.values(); // Enforce the creation of the cache
    	ResourceMetadata map2 = (ResourceMetadata) map1.clone();

Now, map2 contains an unmodifiable wrapper of map1. While this generally behaves correctly, there are two problems:

a) This is a potential memory leak as clones hold a reference to the original map.
b) If unlock() is ever added, this might lead to a subtle bug when (after the above code):
    - map1.unlock() is called
    - map1 is modified
    - map2 will reflect map1's modifications, even if map2 should still be locked

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira