[
https://issues.apache.org/jira/browse/TUSCANY-2643?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Julien Bigot updated TUSCANY-2643:
----------------------------------
Attachment: second-solution.patch
After some discussions with slaws on IRC, I think this solution is better. The
behavior implemented is now the following : an implementation instance is
linked to a component.
i.e.
- if more than one component whose scope is composite share the same
componentType (even inside an unique composite), each one will have its own
implementation instance.
- if more than one composite instance (i.e. component whose implementation is
implementation.composite) are instances of the same composite (they share the
same type) and if this composite contains a component whose scope is composite,
these will share the same implementation instance.
> composite scope handled incorectly
> ----------------------------------
>
> Key: TUSCANY-2643
> URL: https://issues.apache.org/jira/browse/TUSCANY-2643
> Project: Tuscany
> Issue Type: Bug
> Components: C++ SCA
> Affects Versions: Cpp-Next
> Environment: Svn revision 704493
> Reporter: Julien Bigot
> Attachments: first-solution.patch, second-solution.patch
>
>
> Scope composite is not handled correctly.
> The actual behavior is to return the same implementation instance for all
> call to a service of a component whose scope is composite. In the case where
> there is more than one component with composite scope, this means that the
> first implementation instantiated is returned regardless of the actual
> component one is looking for.
> If the two components have a different ComponentType, this will likely lead
> to a segfault.
> I have implemented another behavior.
> With my patch, when using a service of a composite implementation, the name
> of the component is used as a unique ID, this is still not optimal as two
> components can have the same name inside two different composite or if more
> than one instance of a given composite are created.
> This does not scale well too as a map is used. The straightforward solution
> would be to have the implementation instance referenced in the Component
> class. but as this class is defined in the tuscany::sca::model namespace, I'm
> not sure it is Ok to modify it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
