getFlag() returns true when the property is equal, ignoring case, to the
string "true", not just "true"; "TrUe" also means true.
Closes gh-34295
Signed-off-by: Mengqi Xu <2663479778@qq.com>
This commit harmonizes how a candidate value is parsed to extract its
key and default, if any. Rather than returning {@code null} if no
default is available, `splitKeyAndValue` now consistently returns a
non-null array.
This prevents an escaped separator character to be mistakenly identified
as a placeholder in certain cases.
Closes gh-34289
For ComandLineArgs, supplying the same option multiple times with
different values is valid, and the values will be stored in a List.
This commit also updates the Javadoc for SimpleCommandLineArgsParser.
See gh-34282
Signed-off-by: puppy4c <puppy4c@foxmail.com>
This commit fixes a regression in property placeholder resolution where
the original key was no longer considered for an exact match before
processing the placeholder itself.
By default, property resolution uses ':' as the separator between the
key and the fallback value.
Consider a request to resolve ${prefix://service}. Previously,
placeholder resolution would first attempt to resolve the raw text, that
is 'prefix://service', before attempting to resolve the 'prefix' key and
then use '//service' if the key did not resolve.
This commit restores that behaviour purely for backward compatible
reason.
Closes gh-34124
Properly processes recursive types through always comparing generics via the top-level ResolvableType (rather than through nested TypeDescriptors with custom ResolvableType instances).
Closes gh-33932
This metadata information is required for supporting libraries using
`sun.misc.Unsafe#allocateInstance(Class<?>)`, even though Spring
Framework is not using this feature.
Closes gh-34055
This commit fixes a regression in PlaceHolderParser where it would no
longer resolve nested placeholders for a case where the fallback has a
placeholder itself.
This is due to the Part implementations and how they are structure, and
this commit makes sure that nested resolution happens consistently.
Closes gh-34020
Before this commit, in Spring Framework 6.2, Kotlin value class
unboxing was done at CoroutinesUtils level, which is a good fit
for InvocableHandlerMethod use case, but not for other ones like
AopUtils.
This commit moves such unboxing to InvocableHandlerMethod in
order to keep the HTTP response body support while fixing other
regressions.
Closes gh-33943
