The JDKs 9 and 10 are no longer supported by Oracle and the
CAs are not up to date, so removing from the build.
18:13:02 * What went wrong:
18:13:02 Execution failed for task ':buildSrc:compileJava'.
18:13:02 > Could not resolve all files for configuration ':buildSrc:compileClasspath'.
18:13:02 > Could not resolve com.thaiopensource:trang:20091111.
18:13:02 Required by:
18:13:02 project :buildSrc
18:13:02 > Could not resolve com.thaiopensource:trang:20091111.
18:13:02 > Could not get resource 'https://repo.maven.apache.org/maven2/com/thaiopensource/trang/20091111/trang-20091111.pom'.
18:13:02 > Could not GET 'https://repo.maven.apache.org/maven2/com/thaiopensource/trang/20091111/trang-20091111.pom'.
18:13:02 > sun.security.validator.ValidatorException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target
18:13:02 > Could not resolve net.sourceforge.saxon:saxon:9.1.0.8.
18:13:02 Required by:
18:13:02 project :buildSrc
18:13:02 > Could not resolve net.sourceforge.saxon:saxon:9.1.0.8.
18:13:02 > Could not get resource 'https://repo.maven.apache.org/maven2/net/sourceforge/saxon/saxon/9.1.0.8/saxon-9.1.0.8.pom'.
18:13:02 > Could not GET 'https://repo.maven.apache.org/maven2/net/sourceforge/saxon/saxon/9.1.0.8/saxon-9.1.0.8.pom'.
18:13:02 > sun.security.validator.ValidatorException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to request
Allow this to be gathered by Gradle enterprise since if build is up to
date there will be no tests ran which causes failure. Additionally,
Gradle Enterprise displays the tests better than Jenkins.
Prior to this commit ClaimTypeConverter returned the claims with the
original value for all the claims with a null converted value.
The changes allows ClaimTypeConverter to overwrite and return claims
with converted value of null.
Closes gh-10135
Flux#skip discards its corresponding elements, meaning that they
aren't intended for reuse. When using RSocket's ByteBufPayloads,
this means that the bytes are releaseed back into RSocket's pool.
Since the downstream request may still need the skipped payload,
we should construct the publisher in a different way so as to
avoid the preemptive release.
Deferring Spring JavaFormat to clarify what changed.
Closes gh-9345
