[
https://issues.apache.org/jira/browse/SLING-10944?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stefan Seifert resolved SLING-10944.
------------------------------------
Resolution: Fixed
https://github.com/apache/sling-org-apache-sling-testing-sling-mock/commit/a223de695f7ce3b55e51767ac58874a0fac08adf
> sling-mock: Tear down MockSlingBindings on context tear down
> ------------------------------------------------------------
>
> Key: SLING-10944
> URL: https://issues.apache.org/jira/browse/SLING-10944
> Project: Sling
> Issue Type: Bug
> Components: Testing
> Reporter: Stefan Seifert
> Assignee: Stefan Seifert
> Priority: Major
> Fix For: Testing Sling Mock 3.1.4
>
>
> MockSlingBindings is integrated in a complex fashion:
> * it's created together with a mock sling http serlvet request (stored as
> request attributes)
> * it holds a reference to the current context to get access to all context
> objects (e.g. current resource) which may change during test execution
> * additionally it listens to OSGi events for registering or unregistering
> BindingsValuesProvider instances and allows to get values from them as value
> the asynchronous handling of registering/unregistering events and accessing
> the context may lead to conflicts if the context is in process of shutting
> down, or already was shut down and an event reaches the instance later in
> time (OSGi events are processed asynchronously). esp. i may happen that other
> services which are deactivated are sending such events late in the chain of
> deactivations.
> such a racing condition seems to be the root case of the problems described
> in [WTES-69|https://wcm-io.atlassian.net/browse/WTES-69].
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
