Apparent loop in the resolver

[Benson Margulies <be...@basistech.com>]

Using Karaf 4.0.1, with either felix 5.0.1 or 5.2.0, I have a
situation in which the Felix resolver seems to be capable of consuming
CPU time indefinitely.

I tried to isolate a test case, but I apparently missed some critical
bit of complexity, and failed. There's no telling what it will take to
get to a repro. So, instead, I ask, is there anything I can do to get
some insight into this that would either help me rearrange my bits and
pieces or help you fix a problem?


Here's a not very interesting jstack snapshot when using 5.2.0:

"pool-17-thread-8" #126 prio=5 os_prio=31 tid=0x00007febd2ac4800
nid=0x730b runnable [0x0000000127830000]
   java.lang.Thread.State: RUNNABLE
at org.apache.felix.resolver.util.OpenHashMap.get(OpenHashMap.java:686)
at org.apache.felix.resolver.ResolverImpl.getPackageSources(ResolverImpl.java:1720)
at org.apache.felix.resolver.ResolverImpl.mergeUses(ResolverImpl.java:1039)
at org.apache.felix.resolver.ResolverImpl.computeUses(ResolverImpl.java:873)
at org.apache.felix.resolver.ResolverImpl.access$500(ResolverImpl.java:60)
at org.apache.felix.resolver.ResolverImpl$6.run(ResolverImpl.java:1232)
at org.apache.felix.resolver.ResolverImpl$EnhancedExecutor$1.run(ResolverImpl.java:2419)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@felix.apache.org
For additional commands, e-mail: users-help@felix.apache.org