[
https://issues.apache.org/jira/browse/JCR-1644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12618366#action_12618366
]
Jukka Zitting commented on JCR-1644:
------------------------------------
The stack overflows on NamespaceContext should really only be caused by
excessively nested XML elements (at least hundredths, probably thousands, of
levels of nesting). If this happens on documents with less than 100 levels of
nesting, then I suspect that there's something else wrong with the code; like
the NamespaceContext stack not being correctly rewound.
I tried reviewing the related code, but couldn't find any obvious errors. Could
you perhaps send me a troublesome XML document or come up with a cleaned up
example that illustrates the problem? It's not high priority as it seems like
the issue is already solved by the patch, but I'd rather verify that we
actually fixed the root cause instead of just working around it.
> make NamespaceContext#getPrefix(java.lang.String) iterative instead of
> recursive
> --------------------------------------------------------------------------------
>
> Key: JCR-1644
> URL: https://issues.apache.org/jira/browse/JCR-1644
> Project: Jackrabbit
> Issue Type: Improvement
> Components: jackrabbit-core, xml
> Affects Versions: core 1.4.5
> Reporter: Philippe Marschall
> Assignee: Jukka Zitting
> Priority: Minor
> Fix For: 1.5
>
> Attachments: JCR-1644-patch.diff, NamespaceContext.java,
> NamespaceContext.java.patch
>
>
> Currently the method
> org.apache.jackrabbit.core.xml.NamespaceContext#getPrefix(java.lang.String)
> uses recursion. For very large XML files (50 MB Magnolia website exports)
> this causes a stack overflow. The method can easily be rewritten using
> iteration.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
