Issues gh-25038 and gh-25618 collectively introduced a regression for
thread-scoped and transaction-scoped beans.
For example, given a thread-scoped bean X that depends on another
thread-scoped bean Y, if the names of the beans (when used as map keys)
end up in the same bucket within a ConcurrentHashMap AND an attempt is
made to retrieve bean X from the ApplicationContext prior to retrieving
bean Y, then the use of Map::computeIfAbsent in SimpleThreadScope
results in recursive access to the same internal bucket in the map.
On Java 8, that scenario simply hangs. On Java 9 and higher,
ConcurrentHashMap throws an IllegalStateException pointing out that a
"Recursive update" was attempted.
In light of these findings, we are reverting the changes made to
SimpleThreadScope and SimpleTransactionScope in commits 50a4fdac6e and
148dc95eb1.
Closes gh-25801
PR gh-25038 introduced regressions in SimpleThreadScope and
SimpleTransactionScope in Spring Framework 5.2.7. Specifically, if a
thread-scoped or transaction-scoped bean has a dependency on another
thread-scoped or transaction-scoped bean, respectively, a
ConcurrentModificationException will be thrown on Java 11 or higher.
The reason is that Java 11 introduced a check for concurrent
modification in java.util.HashMap's computeIfAbsent() implementation,
and such a modification can occur when a thread-scoped bean is being
created in order to satisfy a dependency of another thread-scoped bean
that is currently being created.
This commit fixes these regressions by switching from HashMap to
ConcurrentHashMap for the instance maps in SimpleThreadScope and
SimpleTransactionScope.
Closes gh-25618
Includes consistent constructor-level storage of derived resource in ScannedGenericBeanDefinition and ConfigurationClassBeanDefinition.
Closes gh-24978
Prior to this commit, the testTestLazyConnectionToRemote() method in
MBeanClientInterceptorTests expected an exception to be thrown while
attempting to access the state of a proxied MBean after the MBeanServer
had been shutdown; however, the test occasionally failed if the server
had not been properly shutdown.
Since an attempt to wait on the server to shutdown proved not to be
consistently helpful in this scenario, we are entirely removing this
check from the test.