[
https://issues.apache.org/jira/browse/AXIS2C-1701?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robert Lazarski resolved AXIS2C-1701.
-------------------------------------
Fix Version/s: 2.0.0
(was: 1.7.0)
Resolution: Fixed
The fix applies the same patch from the JIRA attachment - changing
axiom_element_find_namespace() to axiom_element_find_declared_namespace() so
duplicate namespace declarations from parent elements are redeclared locally
with proper ownership, preventing the memory leak.
> Memory leak on duplicate XML namespace declaration
> --------------------------------------------------
>
> Key: AXIS2C-1701
> URL: https://issues.apache.org/jira/browse/AXIS2C-1701
> Project: Axis2-C
> Issue Type: Bug
> Components: xml/om
> Reporter: Nicholas
> Priority: Minor
> Fix For: 2.0.0
>
> Attachments: axis.zip
>
>
> A memory leak exists in the axiom_element_declare_namespace function if the
> namespace exists in the parent hierarchy. For example when decoding this
> message:
> {code:java}
> <S:Envelope xmlns:S="http://schemas.xmlsoap.org/soap/envelope/";>
> <S:Header xmlns="http://www.w3.org/2005/08/addressing";>
> <To
> xmlns="http://www.w3.org/2005/08/addressing";>https://localhost/xxx</To>
> <Action xmlns="http://www.w3.org/2005/08/addressing";>ACTION</Action>
> <MessageID>XXXX</MessageID>"
> </S:Header>
> <S:Body />
> </S:Envelope>{code}
> The duplicate namespace declarations in the "To" and "Action" elements result
> in the leak of 2 namespace objects and their strings.
> Callers of the function expect that if it returns SUCCESS the namespace
> parameter is now owned by the element and then don't free it. But if the
> namespace has already been declared in a parent element it returns SUCCESS
> without changing the ref count of the parameter.
> A real fix would require changing the API so my quick fix is to only search
> the current element for a matching namespace and if not found it can be
> redeclared. Attached is a reproducer and a patch.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
