Restored original singleton-only adaptInnerBeanName behavior, avoiding endless unique name calculation for every single prototype instance. Non-named inner BeanDefinition objects now suffixed with an identity hex code analogous to non-named XML bean definitions, largely avoiding naming collisions to begin with. After SPR-11246, post-processors can deal with unstable classes per bean name, so occasional collisions aren't a hard problem anymore.
Issue: SPR-11545
(cherry picked from commit 5308b3e)
Also uses addAll instead of iteration over untyped collection now, supporting optimized addAll in target collection type, and avoids repeated getElementTypeDescriptor calls.
Issue: SPR-11479
(cherry picked from commit bea94d5)
When a node of an SPeL expression was a call to a bean referenced
in a method argument, the expression was resolved twice.
The resolved arguments are now specified to MethodValueRef instead
of resolving the arguments again in the constructor
Issue: SPR-11445
(cherry picked from commit 519799e)
This change removes the use of a CountDownLatch to wait for the
asynchronously computed controller method return value. Instead we
check in a loop every 200 milliseconds if the result has been set.
If the result is not set within the specified amount of time to wait
an IllegalStateException is raised.
Additional changes:
- Use AtomicReference to hold the async result
- Remove @Ignore annotations on AsyncTests methods
- Remove checks for the presence of Servlet 3
Issue: SPR-11516
Update ConcurrentReferenceHashMap to protect against references that
have been garbage collected but for some reason do not appear as a
`pollForPurge` result.
Also added purgeUnreferencedEntries() method to allow for programmatic
cleanup.
Issue: SPR-11440
(cherry picked from commit 2b4c81e6)
Replaced static import of org.junit.Assert.* with implicit imports in
order to avoid ambiguity between JUnit's static Assert methods and those
in org.custommonkey.xmlunit.XMLAssert.
This commit introduces a test in XmlBeanFactoryTests that verifies that
CGLIB generated subclasses for method injected beans are reused across
bean factories for identical bean definitions. In other words, by
verifying that the same CGLIB generated class is reused for identical
bean definitions, we can be certain that Spring is no longer generating
identical, duplicate classes that consume memory in the VM.
Issue: SPR-10785, SPR-11420
(cherry picked from commit f2a4537)
Prior to this commit, the inclusion of the 'overloaded' flag in the
implementations of equals() and hashCode() in MethodOverride could lead
to adverse effects in the outcome of equals() in AbstractBeanDefinition.
For example, given two bean definitions A and B that represent the
exact same bean definition metadata for a bean that relies on method
injection, if A has been validated and B has not, then A.equals(B) will
potentially return false, which is not acceptable behavior.
This commit addresses this issue by removing the 'overloaded' flag from
the implementations of equals() and hashCode() for MethodOverride.
Issue: SPR-11420
(cherry picked from commit 9534245)
MarshallingView also explicitly skips BindingResult when searching for a model object now, implementing common custom subclass behavior out-of-the-box.
Issue: SPR-11411
Issue: SPR-11417
Before this change the getPathWithinServletMapping method of
UrlPathHelper could not work properly when a default servlet mapping
(i.e. "/") was used in combination with urlDecode=false. The fact that
the getServletPath() method of HttpServletRequest always returns a
decoded path was getting in the way.
Although there is no way to check Servlet mappings through the Servlet
API, this change aims to detect the given scenario and returns the full
path following the context path thus avoiding URL decoding.
Note that the same can be achieved by setting urlDecode=false and
alwaysUseFullPath=true. However this change ensures that urlDecode
works properly without having to know that.
Issue: SPR-11101
(cherry picked from commit 12598f8)