fabric8io/kubernetes-client v5.4.0

5.4.0 (2021-05-19)

on GitHub

Bugs

• Fix #3040: Consistently order printer columns by JSON path to prevent undue changes in generated CRDs
• Fix #3041: Properly output additionalProperties field for Maps, output warning for unsupported complex maps
• Fix #3036: Fix file descriptor leak when loading cacerts file
• Fix #3038: Upgrade TLS versions in mock servers to 1.2.
• Fix #3037: Account for JsonProperty annotations when computing properties' name
• Fix #3014: Resync Future is canceled and resync executor is shutdown on informer stop
• Fix #2529: SelfSubjectAccessReview not working with OpenShiftClient
• Fix #2978: Fix SharedInformer NPE on initial requests while syncing
• Fix #2989: serialization will generate valid yaml when using subtypes
• Fix #2991: reduced the level of ReflectWatcher event received log
• Fix #2992: allowing Watch auto-reconnect for shared informers
• Fix #2994: updating the SharedIndexInformer indexer state for a delete event generated by resync
• Fix #2996: Generating CRDs from the API should now properly work
• Fix #3000: Set no_proxy in the okhttp builder in case the proxy_url is null
• Fix #3011: properly handle enum types for additional printer columns
• Fix #3020: annotations should now properly have their associated values when processing CRDs from the API
• Fix #3016: Use scheduleWithFixedDelay for resync task execution
• Fix #2991: reduced the level of ReflectWatcher event received log
• Fix #3027: fix NPE when sorting events in KubernetesResourceUtil
• Fix #3054: missing entry for Trigger in TektonTriggersResourceMappingProvider
• Fix #3047: NPE when getting version when there is no build date
• Fix #3024: stopAllRegisteredInformers will not call startWatcher
• Fix #3067: Added a patch(PatchContext, item) operation to be more explicit about patching and diffing behavior
• Fix #3097: refresh token with autoconfigure even if authprovider is null

Improvements

• Fix #2788: Support FIPS mode in kubernetes-client with BouncyCastleFipsProvider
• Fix #2910: Move crd-generator tests from kubernetes-itests to kubernetes-tests
• Fix #3005: Make it possible to select which CRD version is generated / improve output
• Fix #3015: Thread interruption in a nominal case (like closing the client) are now logged in debug
• Fix #3057: Removed debug calls for CustomResource during deserialization
• Fix #3050: More enforcement of the informer lifecycle
• Fix #3061: Removed the deltafifo from the informer logic
• Fix #3081: Use apiGroupName in generated package for OpenShiftClient Handler/OperationsImpl classes
• Fix #3089: Allowing patch/edit to infer context from the item
• Fix #3066: Added replaceStatus (PUT), editStatus (JSON PATCH), and patchStatus (JSON MERGE PATCH) methods to support non-locking status updates

Dependency Upgrade

• Fix #2979: Update Kubernetes Model to v1.21.0
• Fix #3099: Update Tekton Triggers Model to v0.13.0
• Fix #3118: Update to sundrio 0.40.1

New Features

• Fix #2984: Add support for flowcontrol.apiserver.k8s.io/v1beta1 FlowSchema and PriorityLevelConfiguration
• Fix #2980: Add DSL Support for apps/v1#ControllerRevision resource
• Fix #2981: Add DSL support for storage.k8s.io/v1beta1 CSIDriver, CSINode and VolumeAttachment
• Fix #2912: Add DSL support for storage.k8s.io/v1beta1 CSIStorageCapacity
• Fix #2701: Better support for patching in KuberntesClient
• Fix #3034: Added a SharedInformer.isRunning method
• Fix #3088: mock server will assume /status is a subresource, and other refinements to match kube behavior
• Fix #3111: Add DSL Support for config.openshift.io/v1 resources in OpenShiftClient

Note: Breaking changes in the API

DSL Changes:

• client.batch().jobs() deprecated, suggestion to move to client.batch().v1().jobs()
• client.batch().cronjobs() deprecated, suggestion to move to client.batch().v1().cronjobs() or client.batch().v1beta1().cronjobs()
• client.policy().podSecurityPolicies() deprecated, suggestion to move to client.policy().v1beta1().podSecurityPolicies()
• client.policy().podDisruptionBudget() deprecated, suggestion to move to client.policy().v1().podDisruptionBudget() or client.policy().v1beta1().podDisruptionBudget()

Model Changes:

• Classes in io.fabric8.kubernetes.api.model.batch have been moved to io.fabric8.kubernetes.api.model.batch.v1 and io.fabric8.kubernetes.api.model.batch.v1beta1
• Classes in io.fabric8.kubernetes.api.model.policy have been moved to io.fabric8.kubernetes.api.model.policy.v1 and io.fabric8.kubernetes.api.model.policy.v1beta1