Prior to this commit, the "filename" parameter value for the
"Content-Disposition" header would contain non-printable characters,
causing parsing issues for HTTP clients.
This commit ensures that all non-printable characters are encoded.
Fixes gh-35034
Previously, when saving a WebSession, the system did not check whether
the session ID already existed. As a result, even if the session being
saved was an update to an existing one, it was incorrectly treated as a
new session, and a "maximum sessions exceeded" error was triggered.
This fix ensures that if a WebSession with the same ID already exists,
it will be updated rather than counted as a new session, thereby
preventing unnecessary session limit violations.
Closes gh-35013
Signed-off-by: Mohammad Saeed Nouri <msnsaeed71@gmail.com>
This commit adds `ProblemDetail` and `ErrorResponse` to the list of
supported return types for `@ExceptionHandler` methods.
Closes gh-34554
Signed-off-by: Vedran Pavic <vedran@vedranpavic.com>
AbstractKotlinSerializationHttpMessageConverter#getSupportedMediaTypes(Class<?>)
currently invokes transitively supports(Class<?>) which always return false
with generic types.
This commit adds an override that just invokes getSupportedMediaTypes().
Closes gh-34992
per https://www.rfc-editor.org/rfc/rfc7232#section-4.1
The server generating a 304 response MUST generate any of the
following header fields that would have been sent in a 200 (OK)
response to the same request: Cache-Control, Content-Location, Date,
ETag, Expires, and Vary.
Closes gh-34614
Signed-off-by: James Yuzawa <jtyuzawa@gmail.com>
Prior to this commit, combining the "/*" and "/x/y" path patterns
would result in a `StringIndexOutOfBoundsException`.
This commit fixes this problem and revisits the implementation for
better consistency:
* "/*" + "/x/y" is now "/x/y"
* "/x/*.html" + "/y/file.*" is now rejected because they don't share the
same prefix.
This change also adds the relevant Javadoc to the `PathPattern#combine`
method.
Fixes gh-34986
Add optimized DataBufferInputStream overrides for readNBytes, skip, and transferTo; all of them
allocate byte buffers which we can either avoid (in the case of skip) or size more precisely since
the number of remaining bytes is known.
Closes gh-34799
Signed-off-by: Patrick Strawderman <pstrawderman@netflix.com>
Prior to this commit, we would use the
`java.net.http.HttpRequest.Builder#method(String, BodyPublisher)` to
create HTTP requests for the JDK HttpClient. This method requires a
non-null body publisher; providing an empty publisher writes a
"Content-Length: 0" header to all requests.
As of Java 19, this behavior changes for `HttpRequest.Builder#GET` and
similar methods, where the body publisher is considered as null and no
"Content-Length" header is written.
This commit aligns with this behavior and favors dedicated HTTP methods
whenever available.`
Closes gh-34971
Prior to this commit, argument index handling in
AutowiredArgumentsCodeGenerator suffered from an off-by-one error when
generating code for an autowired inner class constructor.
Since the startIndex is already properly calculated for an inner class in
InstanceSupplierCodeGenerator.buildGetInstanceMethodForConstructor(...),
there is no need to adjust the argument indexes within
AutowiredArgumentsCodeGenerator.generateCode(...).
Closes gh-34974
Signed-off-by: Dmytro Nosan <dimanosan@gmail.com>
Prior to this commit, the verifyType() method in MessageHeaderAccessor
checked if REPLY_CHANNEL ended with the given header name which does
not make much sense and is inconsistent with the ERROR_CHANNEL check.
This commit therefore checks if the REPLY_CHANNEL is equal to the given
header name, analogous to the ERROR_CHANNEL check.
See gh-34881
Closes gh-34949
Signed-off-by: Mengqi Xu <2663479778@qq.com>
Brian Clozel
Johnny Lim
Sébastien Deleuze
Sam Brannen
Mohammad Saeed Nouri
秦利斌
rstoyanchev
Vedran Pavic
Jimmy Axenhus
Juergen Hoeller
James Yuzawa
DongNyoung Lee
addoDev
Tran Ngoc Nhan
Patrick Strawderman
Johannes Jank
Dmytro Nosan
Mengqi Xu
Aurh1l
Dhruv