[
https://issues.apache.org/jira/browse/ISIS-3223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17611880#comment-17611880
]
ASF subversion and git services commented on ISIS-3223:
-------------------------------------------------------
Commit b4aa542410d1176845c76c13d3daf0cd71189d09 in isis's branch
refs/heads/master from Andi Huber
[ https://gitbox.apache.org/repos/asf?p=isis.git;h=b4aa542410 ]
ISIS-3223: improve AsyncTask, so we can wait for result on on calling
thread
> WrapperFactory with async executor service throws exception
> -----------------------------------------------------------
>
> Key: ISIS-3223
> URL: https://issues.apache.org/jira/browse/ISIS-3223
> Project: Isis
> Issue Type: Bug
> Affects Versions: 2.0.0-M8
> Reporter: Daniel Keir Haywood
> Assignee: Andi Huber
> Priority: Major
> Fix For: 2.0.0-M8
>
>
> To reproduce:
> * checkout ISIS-3223 branch
> * locate the `WrapperFactory_async_IntegTest` in new
> `regressiontests/stable-core-wrapperfactory` module
> * (enhance, using sh enhance.sh -w)
> * run the test; it will fail because the async executor code (as result of
> line 60) hits an exception when calling back to WrapperFactoryDefault, line
> 649, which in turn is the RepositoryServiceDefault failing an invariant
> assert on line 244.
> I don't understand enough to know why that invariant is checked. The method
> used in DN is "detachCopy", so it would seem that there _are_ legitimate
> reasons when the entity we get back is different to the one being detached
> (the DN code has a entityClass.newInstance(...) in one of its code paths,
> that is what we hit).
> Perhaps the fact that this is multi-threaded code is the issue - though I
> don't believe we are sharing the reference to the Counter bean across
> threads, each thread has its own reference.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
