This change catches the exception and returns false, invalidating the user session. This behavior is sensible, because if we are unable to validate the user we must assume the session is invalid.
Catching the error at this point reportedly helps with some scenarios and shouldn't cause any further harm.
* [JENKINS-43610] Split Trilead out from Core
* revert changes on ./core/src/main/resources/jenkins/split-plugin-cycles.txt
* remove JDK version on split-plugins.txt
* bump trilead-api version to 1.0.4
* fix: add trilead-api incremental version to detached plugins
* fix: add consume-incrementals setting to maven
* Update core/src/main/resources/jenkins/split-plugins.txt
Co-Authored-By: Jesse Glick <jglick@cloudbees.com>
* Avoid nodeprovisioner delay after consulting strategies
* Try another aproach - rate limit the review every second
* Throttle NodeProvisioner#update every second
* Re-indent to limit diff
* Update lastSuggestedReview
* Fix prior merge
* Logging messages
* Supplier
* Ignore additional calls to suggestReviewNow if we have a queued review
* Need to have only one copy of SystemCommandLanguage among detached plugins.
* Removing some cases from LoadDetachedPluginsTest which were sensitive to particular versions of detached plugins and thus hard to maintain.
Already had to do one such change in 84f939cebd and now upgradeFromJenkins2WithNewerDependency began failing after a war/pom.xml change.
* Display a warning in the Installed tab if the admin is about to disable or uninstall a detached plugin.
* @since
Co-Authored-By: Oleg Nenashev <o.v.nenashev@gmail.com>
* Proof of concept catching exceptions for java 11 removed classes
* Add ClassLoader to a lot of places. Backup
* Commenting out wip stuff, change packages, count hits findClass
* Improvements, sending and tests
* Removing the catcher from chain to make funcional tests work
* Removing ignored places and a bit of cleaning
* Cleaning unused code after removing the catcher from the chain
* Remove imports cleaning to avoid extra changes
* Cleaning and explaining comments
* Explaining comments
* fix javadoc
* [JENKINS-57223] Address feedback from Matt Sicker. Improve in concurrency
* [JENKINS-57223] Search on cause and suppressed exceptions and fix on name reported
* [JENKINS-57223] Add cycles control on exceptions and refactor reporting method
* [JENKINS-57223] Add trim to class name comparison
* [JENKINS-57223] Address Oleg's feedback and some improvements
* Avoid sending CNFE when ignored later in code
* Disable when running on Java 8 and a test
* Avoid sending empty events
* [JENKINS-57223] Polish assertion after printing all missing classes in logs
* [JENKINS-57223] Avoid extra if and add curly braces
* [JENKINS-57223] Better comment about time window check
This validates that the performance regression itself - redundant calls
to loadUserByUsername() - is also fixed besides simply the
implementation details of the user seed and session.
Signed-off-by: Matt Sicker <boards@gmail.com>
This fixes the remember me service to link the user seed to a
corresponding HTTP session. By doing so, this should significantly
reduce the number of calls to loadUserDetails().
Signed-off-by: Matt Sicker <boards@gmail.com>
* [JENKINS-57515] update jna library to 5.3.1
This solves problems with loading native libraries of jna on AIX
platform when using openJDK or other JDKs that assume .so is the
extension for shared libraries
* Pointer.SIZE -> Native.POINTER_SIZE
* remove unnecessary import
* adjust to changed api
(cherry picked from commit ffa9abf6e3)
Jeff Thompson