[
https://issues.apache.org/jira/browse/AMQ-8299?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Piotr Klimczak updated AMQ-8299:
--------------------------------
Description:
The core principle of messaging system is that payload sent by producer is
received exactly same (unmodified) on consumer side.
However due to ActiveMQMapMessage using HashMap for map data, the order of
entries is not guaranteed effectively causing data corruption.
This is especially true when we are sending LinkedHashMap on producer side.
Then the consumer side will receive out of order data.
This is not newly introduced bug, looks like it is there for many years-
surprisingly.
>From what I have seen in client implementation, it is perfectly ready to work
>with LinkedHashMap. So it might be just one line change to fix + test coverage.
was:
The core principle of messaging system is that payload sent by producer is
received exactly same (unmodified) on consumer side.
However due to ActiveMQMapMessage using HashMap for map data, the order of
entries is not guaranteed effectively causing data corruption.
This is especially true when we are sending LinkedHashMap on producer side.
Then the consumer side will receive out of order data.
> ActiveMQMapMessage data corruption- should use LinkedHashMap to preserve
> original order
> ---------------------------------------------------------------------------------------
>
> Key: AMQ-8299
> URL: https://issues.apache.org/jira/browse/AMQ-8299
> Project: ActiveMQ
> Issue Type: Bug
> Components: JMS client
> Affects Versions: 5.16.2
> Reporter: Piotr Klimczak
> Priority: Major
>
> The core principle of messaging system is that payload sent by producer is
> received exactly same (unmodified) on consumer side.
> However due to ActiveMQMapMessage using HashMap for map data, the order of
> entries is not guaranteed effectively causing data corruption.
> This is especially true when we are sending LinkedHashMap on producer side.
> Then the consumer side will receive out of order data.
> This is not newly introduced bug, looks like it is there for many years-
> surprisingly.
> From what I have seen in client implementation, it is perfectly ready to work
> with LinkedHashMap. So it might be just one line change to fix + test
> coverage.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
