Replace the existing ASM based readers with new implementations that
also support MergedAnnotations. The meta-data classes themselves are
now immutable, and constructed via separate reader classes.

The `SimpleMetadataReader` class has been updated to return the new
classes, however the old ones remain since some of them are public
and might be being used directly.

Closes gh-22884

Add `AnnotatedTypeMetaData.getAnnotations()` that can be used to access
annotation details using the `MergedAnnotations` interface.

Where possible, the existing annotation methods have been migrated to
call `getAnnotation()`, rather than needing their own implementation.

The existing ASM based meta-data implementations have not been updated
since they will be deprecated and replaced in a subsequent commit.

See gh-22884

Add some additional tests to provide more coverage of class and method
meta-data support.

See gh-22884

Update `StandardAnnotationMetadata` and `AnnotationMetadataReadingVisitor`
so that `java.lang.annotation` annotations are consistently skipped.

Closes gh-22885

Update ASM based metadata readers so that only RetentionPolicy.RUNTIME
annotations are exposed. This aligned behavior with the reflection based
implementation.

Closes gh-22886

Add a factory method to `MergedAnnotation` that allows an instance to
be created for an explicit collection of root annotations. This method
will allow ASM based readers to expose a `MergedAnnotation` instance
that has root annotations loaded from bytecode, and meta-annotations
loaded using reflection.

See gh-22884

Update TypeMappedAnnotation so that Strings can be used to represent
Class attribute values. This will allow ASM annotation readers to
present a `MergedAnnotation` instance without necessarily having the
actual class values on the classpath.

When the underlying value is a String, any calls to
`getValue(name, String.class)` or `asMap(Adapt.CLASS_TO_STRING)` will
simply return the original String. Calls that need the actual Class
result (such as `getClass`) will use `Class.forName` and may throw
a `ClassNotFoundException` at that point.

This commit also allows an empty Object[] to be used to represent
any empty primitive array.

See gh-22884

Rename `from` to `of` for the `MergedAnnotation` factory methods that
work with Maps. The previous name was a little confusing, especially
when an annotation source parameter was specified. The new method name
helps to make it clearer when the user is explicitly defining the
attributes of the annotation, as opposed to picking them up from the
source.

Add a `getTypeHierarchy()` method to `MergedAnnotation` that can be used
to return the full type hierarchy information. This method is
specifically designed to be used in combination with
`MergedAnnotationPredicates.unique`.

This update also allows us to delete the `parentAndType` method
from `AnnotatedElementUtils`.

Closes gh-22908

Closes gh-22912

Added two methods to DataBufferUtils:

* matcher(byte[]), which returns a Matcher object that can be used to
  find a delimiter in a data buffer.
* split(Publisher<DataBuffer>, byte[] delimiter), which splits a given
  stream of data buffers around a given delimiter.

Add method retainedSlice to the DataBuffer, defaulting to using
DataBufferUtils for retain, but allowing for ByteBuf specific override.

org.junit.Assert.assertThat() is deprecated as of JUnit 4.13.

See gh-22894

Ensure that results returned from keySet, entrySet & values are tracked
to remove case insensitive keys from the source map.

Closes gh-22821

Update `MergedAnnotation` with a `getRoot()` method that allows the
root direct annotation to be retrieved easily.

Closes gh-22818

See gh-22818

Following on 3ebbfa21 where the local
refCount was removed in favor of using the internal refCount of the
native data buffer, this commit ensures that LeakAwareDataBufferFactory
uses a PooledDataBufferFactory delegate by default.

There are also fixes for test issues with eager allocation uncovered by
these changes in StringDecoder and ResourceDecoder.

LeakAwareDataBuffer was keeping its own refCount rather than checking
through the delegate. This leads to false leak reports in a sequence
where an allocated buffer is retained and then sliced since it is not
aware of the changes to the refCount through the slice.

Due to the changes in a7425c81, we no
longer need to execute tests using the JMXMP protocol in PERFORMANCE
builds.

This commit removes the JMXMP constant from the TestGroup enum and
updates affected tests, thereby effectively including such tests in
the standard build from now on.

See gh-22757

Closes gh-22766

Add a convenience method that allows a `MergedAnnotation` to be
converted into an `AnnotationAttributes` instance. Also rename
the `MapValues` enum to `Adapt` which generally seems to read
better.

Closes gh-22738

Add tests for `MissingMergedAnnotation`

See gh-21697

Delete `AnnotationAttributeExtractor` which is package private and
no longer used.

Closes gh-22753

See gh-22726

LogMessage is an abstract class now: with internal subclasses for Supplier bindings as well as printf-style format strings with a variable number of arguments (some fixed for efficiency, varargs array as fallback), created through corresponding static factory methods.

Closes gh-22726

See gh-21697

For consistency within the framework, this commit renames the SUPER_CLASS enum constant
to SUPERCLASS.

See gh-21697

Closes gh-22726

Update `AnnotationUtils` so that `RepeatableContainers.none()` is used
when performing an exhaustive search for merged annotations. These
parameters were accidentally removed in commit 210b1789 and weren't
caught earlier because we were missing a test.

Closes gh-22702

Update `addConventionAnnotationValues` so that `value` attributes
do not override existing annotation values. The aligns the new
`MergedAnnotations` API with the previous `AnnotatedElementUtils`
logic.

Closes gh-22703

Following a response on
https://github.com/reactor/reactor-core/issues/1634 apparently
FluxSink is respecting doOnDiscard but there is a race condition
in DataBufferUtils with file reading.

This commit makes two changes:
1) Add more checks for onDispose to detect cancellation signals
as well as to deal with potentially concurrent such signal.
2) Do not close the channel through the Flux.using callback but
rather allow the current I/O callback to take place and only then
close the channel or else the buffer is left hanging.

Despite this tests still can fail due to a suspected issue in Reactor
itself with the doOnDiscard callback for FluxSink. That's tracked under
the same issue https://github.com/reactor/reactor-core/issues/1634
and for now the use of DefaultDataBufferFactory is enforced as a
workaround until the issue is resolved.

See gh-22107