[
https://issues.apache.org/jira/browse/CAMEL-3983?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13036103#comment-13036103
]
Claus Straube commented on CAMEL-3983:
--------------------------------------
Claus, you're absolutely right if we talk about the body. Everybody is free to
choose one of the marshalers inside his routes. This alone would be no issue at
all. But we have to serialize the headers, too. I would await if I put any
object into my header and throw the message into the queue, that the object is
still in the header as I stored it there - without any effort. The
serialization part is absolute transparent for the user. He put's in a not
serializable Foo object and gets that back (not the same instance, but the same
object with equal values). So a serialization framework like Jaxb, protobuf or
jackson, where you have to define the class you want to serialize explicitly is
an overhead with no win in my eyes. I've choosen xstream because it's already
used in Camel and because it can serialize / deserialize any object without
extra effort - just for transportation...
> Added Support for Serialization and Message Headers to Hazelcast SEDA
> functionality
> -----------------------------------------------------------------------------------
>
> Key: CAMEL-3983
> URL: https://issues.apache.org/jira/browse/CAMEL-3983
> Project: Camel
> Issue Type: New Feature
> Components: camel-hazelcast
> Affects Versions: 2.8.0
> Reporter: Claus Straube
> Fix For: Future
>
> Attachments: hazelcast_seda_serialization_and_headers_01.diff
>
>
> The current implementation looses headers that are given to a
> 'hazelcast:seda:foo' route and is has problems serializing complex objects
> inside body that are not serializable.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
