Jörg Hoh created SLING-3936:
-------------------------------

             Summary: Sling authenticator
                 Key: SLING-3936
                 URL: https://issues.apache.org/jira/browse/SLING-3936
             Project: Sling
          Issue Type: Improvement
          Components: Authentication
    Affects Versions: Auth Core 1.1.8
            Reporter: Jörg Hoh


When adding lot of paths, which need authentication, to the SlingAuthenticator, 
we encourage long times when the execution seems to starve at the 
PathBasedHolderCache.addHolder() method.

{code}
"FelixStartLevel" - Thread t@21
java.lang.Thread.State: RUNNABLE
at java.util.ComparableTimSort.sort(Unknown Source)
at java.util.ComparableTimSort.sort(Unknown Source)
at java.util.Arrays.sort(Unknown Source)
at java.util.Collections.sort(Unknown Source)
at 
org.apache.sling.auth.core.impl.PathBasedHolderCache.addHolder(PathBasedHolderCache.java:68)
at 
org.apache.sling.auth.core.impl.SlingAuthenticator$SlingAuthenticatorServiceListener.registerService(SlingAuthenticator.java:1511)
at 
org.apache.sling.auth.core.impl.SlingAuthenticator$SlingAuthenticatorServiceListener.addService(SlingAuthenticator.java:1528)
at 
org.apache.sling.auth.core.impl.SlingAuthenticator$SlingAuthenticatorServiceListener.serviceChanged(SlingAuthenticator.java:1494)

    locked <643be70> (a java.util.HashMap)
    at 
org.apache.felix.framework.util.EventDispatcher.invokeServiceListenerCallback(EventDispatcher.java:932)
    at 
org.apache.felix.framework.util.EventDispatcher.fireEventImmediately(EventDispatcher.java:793)
   ...
{code}

obviously the problem is the sorting of the ArrayList "byPathList" in the 
PathBasedCacheHolder.addHolder() method; this ArrayList is always taken from 
the already present cache, and then sorted again. We you add some tens of 
thousands items to this list, we might into a n^2 complexity.

Would it be possible to replace this implementation with something more 
performant, which benefits from the fact, that the byPathList is already 
sorted; or have a better datastructure (TreeSet?), which has the semantics we 
need?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to